William R. Jarnagin (Auth.) - Blumgart's Surgery of The Liver, Biliary Tract and Pancreas, 2-Volume Set - Elsevier (2016) .Compressed

Prometheus, chained to the rocky Mount Caucasus, has his liver eaten by the eagle of Zeus.
(Engraving, 1566, possibly after a work by Titian.

Reproduced by permission of the Hulton Getty Picture Collection, London.)
Cảm ơn bạn đã tải sách từ Doctor Plus Club.
Tất cả ebook được Doctor Plus Club sưu tầm & tổng hợp từ nhiều
nguồn trên internet, mạng xã hội. Tất cả sách Doctor Plus Club chia sẽ vì
đích duy nhất là để đọc, tham khảo, giúp sinh viên, bác sĩ Việt Nam tiếp cận,
hiểu biết nhiều hơn về y học.
Chúng tôi không bán hay in ấn, sao chép, không thương mại hóa những
ebook này (nghĩa là quy đổi ra giá và mua bán những ebook này).
Chúng tôi sẵn sàng gỡ bỏ sách ra khỏi website, fanpage khi nhận được
yêu cầu từ tác giả hay những người đang nắm giữ bản quyền những sách
này.
Chúng tôi không khuyến khích các cá nhân hay tổ chức in ấn, phát hành
lại và thương mại hóa các ebook này nếu chưa được sự cho phép của tác
giả.
Nếu có điều kiện các bạn hãy mua sách gốc từ nhà sản xuất để ủng
hộ tác giả.
Mọi thắc mắc hay khiếu nại xin vui lòng liên hệ chúng tôi qua email:
support@doctorplus.club
Website của chúng tôi: https://doctorplus.club

Fanpage của chúng tôi: https://www.facebook.com/doctorplus.club/
Like, share là động lực để chúng tôi tiếp tục phát triển hơn nữa
Chân thành cảm ơn. Chúc bạn học tốt!
BLUMGART’S
Surgery of the Liver,
Biliary Tract, and Pancreas
6th EDITION
EDITOR-IN-CHIEF
William R. Jarnagin, MD, FACS
ASSOCIATE EDITORS
Peter J. Allen, MD
William C. Chapman, MD, FACS
Michael I. D’Angelica, MD, FACS
Ronald P. DeMatteo, MD, FACS
Richard Kinh Gian Do, MD, PhD
Jean-Nicolas Vauthey, MD, FACS
EDITOR EMERITUS
Leslie H. Blumgart, BDS, MD, DSc(Hon), FACS, FRCS(Eng, Edin), FRCPS(Glas)
VOLUME 1
1600 John F. Kennedy Blvd.
Ste 1800
Philadelphia, PA 19103-2899
Blumgart’s Surgery of the Liver, Biliary Tract, and Pancreas ISBN: 978-0-323-34062-5
Copyright © 2017 by Elsevier, Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage and retrieval
system, without permission in writing from the publisher. Details on how to seek permission, further
information about the Publisher’s permissions policies and our arrangements with organizations such
as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website:
www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the
Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical
treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described herein. In
using such information or methods they should be mindful of their own safety and the safety of
others, including parties for whom they have a professional responsibility.
With respect to any drug or pharmaceutical products identified, readers are advised to check
the most current information provided (i) on procedures featured or (ii) by the manufacturer of
each product to be administered, to verify the recommended dose or formula, the method and
duration of administration, and contraindications. It is the responsibility of practitioners, relying
on their own experience and knowledge of their patients, to make diagnoses, to determine dosages
and the best treatment for each individual patient, and to take all appropriate safety precautions.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of products
liability, negligence or otherwise, or from any use or operation of any methods, products,
instructions, or ideas contained in the material herein.
Previous editions copyrighted 2012, 2007, 2000, 1994, and 1988 by Saunders, an imprint of
Elsevier Inc.
Library of Congress Cataloging-in-Publication Data
Names: Jarnagin, William R. (Surgeon), editor.

Title: Blumgart’s surgery of the liver, biliary tract, and pancreas / editor-in-chief, William R. Jarnagin ;
associate editors, Peter J.Allen, William C. Chapman, Michael I. D’Angelica, Ronald P. DeMatteo,
Richard Kinh Gian Do, Jean-Nicolas Vauthey ; editor emeritus, Leslie H. Blumgart.
Other titles: Surgery of the liver, biliary tract, and pancreas
Description: Sixth edition. | Philadelphia, PA : Elsevier, [2017] | Includes bibliographical references
and index.
Identifiers: LCCN 2016013975 | ISBN 9780323340625 (hardcover)
Subjects: | MESH: Liver–surgery | Biliary Tract Surgical Procedures | Pancreas–surgery
Classification: LCC RD669 | NLM WI 770 | DDC 617.5/56–dc23 LC record available at http://
lccn.loc.gov/2016013975
Executive Content Strategist: Michael Houston

Senior Content Development Specialist: Dee Simpson
Publishing Services Manager: Patricia Tannian
Project Manager: Stephanie Turza
Design Direction: Maggie Reid
Printed in China
Last digit is the print number: 9 8 7 6 5 4 3 2 1

This book is dedicated to
the development of surgery of the liver, biliary tract,
and pancreas as a specialty.
EDITORS
EDITOR-IN-CHIEF Ronald P. DeMatteo, MD, FACS

Vice Chair, Department of Surgery
William R. Jarnagin, MD, FACS Chief, Division of General Surgical Oncology
Chief, Hepatopancreatobiliary Surgery Leslie H. Blumgart Chair in Surgery
Benno C. Schmidt Professor of Surgical Oncology Memorial Sloan Kettering Cancer Center
Memorial Sloan Kettering Cancer Center; New York, New York
Professor of Surgery
Weill Medical College of Cornell University Richard Kinh Gian Do, MD, PhD
New York, New York Associate Professor of Radiology
Weill Medical College of Cornell University;
ASSOCIATE EDITORS Assistant Attending Physician
Department of Radiology
Peter J. Allen, MD Memorial Sloan Kettering Cancer Center
Professor of Surgery New York, New York
Department of Surgery
Memorial Sloan Kettering Cancer Center Jean-Nicolas Vauthey, MD, FACS
New York, New York Professor of Surgical Oncology
Chief, Hepato-Pancreato-Biliary Section
William C. Chapman, MD, FACS Bessie McGoldrick Professor in Clinical Cancer Research
Professor Department of Surgical Oncology
Chief, Division of General Surgery University of Texas MD Anderson Cancer Center
Chief, Abdominal Transplantation Section Houston, Texas
Washington University School of Medicine
St. Louis, Missouri
EDITOR EMERITUS
Michael I. D’Angelica, MD, FACS Leslie H. Blumgart, BDS, MD, DSc(Hon), FACS,
Attending Surgeon FRCS(Eng, Edin), FRCPS(Glas)
Hepatopancreatobiliary Surgery Member
Enid A. Haupt Chair in Surgery Professor of Surgery and Attending Surgeon
Memorial Sloan Kettering Cancer Center; Memorial Sloan Kettering Cancer Center;
Associate Professor Professor of Surgery
Department of Surgery Weill Medical College of Cornell University
Weill Medical College of Cornell University New York, New York
New York, New York
vii
CONTRIBUTORS
Ghassan K. Abou-Alfa, MD Jesper B. Andersen, MD

Assistant Attending Physician Biotech Research and Innovation Centre (BRIC)
Memorial Sloan Kettering Cancer Center; Department of Health and Medical Sciences
Assistant Professor University of Copenhagen
Weill Medical College of Cornell University Copenhagen, Denmark
New York, New York
Christopher D. Anderson, MD
Jad Abou Khalil, MD, CM James D. Hardy Professor and Chair
Chief Resident Department of Surgery
McGill University Health Centre University of Mississippi Medical Center
Montreal, Quebec, Canada Jackson, Mississippi
Pietro Addeo, MD Vittoria Arslan-Carlon, MD

Attending Surgeon Assistant Anesthesiologist
Hepato-Pancreato-Biliary Surgery and Liver Transplantation Department of Anesthesiology and Critical Care
University of Strasbourg Memorial Sloan Kettering Cancer Center
Strasbourg, France New York, New York
N. Volkan Adsay, MD Horacio J. Asbun, MD, FACS

Professor and Vice-Chair Professor of Surgery
Director of Anatomic Pathology Department of Surgery
Emory University Mayo Clinic
Atlanta, Georgia Jacksonville, Florida
Anil Kumar Agarwal, MCh, FRCS, FACS Béatrice Aussilhou, MD

Professor Department of Hepato-Pancreatic-Biliary Surgery and Liver
Director Transplantation
Department of GI Surgery and Liver Transplant Beaujon Hospital
Govind Ballabh Pant institute of Postgraduate Medical Clichy, France
Education & Research
Maulana Azad Medical College Joseph Awad, MD
New Delhi, India Professor of Medicine
Department of Gastroenterology and Hepatology
Farzad Alemi, MD Vanderbilt University;
Assistant Professor and Section Chief Chief
Department of Hepatopancreatobiliary Surgery Transplant Center
University of Missouri–Kansas City Tennessee Valley Healthcare System
Kansas City, Missouri Nashville, Tennessee
Peter J. Allen, MD Daniel Azoulay, MD, PhD

Professor of Surgery Professor of Surgery
Department of Surgery Department of Digestive, HPB and Liver Transplant Surgery
Memorial Sloan Kettering Cancer Center Hôpital Henri Mondo
New York, New York Assistance Publique-Hôpitaux de Paris Faculté de Médecine
Université Paris-Est-Créteil
Ahmed Al-Mukhtar, MD Créteil, France
Consultant
Hepatobiliary Surgeon Philippe Bachellier, MD, PhD
Sheffield Teaching Hospitals Professor and Chairman
Sheffield, England Hepato-Pancreato-Biliary Surgery and Liver Transplantation
Pôle des Pathologies Digestives
Thomas A. Aloia, MD Hépatiques et de la Transplantation
Associate Professor of Surgical Oncology Hôpital de Hautepierre-Hôpitaux
Department of Surgical Oncology Universitaires de Strasbourg
University of Texas MD Anderson Cancer Center Strasbourg, France
Houston, Texas
viii
CONTRIBUTORS ix
Talia B. Baker, MD Marc G. H. Besselink, MD, PhD

Associate Professor of Surgery Hepato-Pancreato-Biliary Surgeon
Division of Transplantation Academic Medical Center
Department of Surgery Amsterdam, The Netherlands
Northwestern University Feinberg School of Medicine
Chicago, Illinois Anton J. Bilchik, MD, PhD
Zubin M. Bamboat, MD Chief of Medicine
Department of Surgery Chief of Gastrointestinal Research Program
Memorial Sloan Kettering Cancer Center John Wayne Cancer Institute
New York, New York Providence Saint John’s Health Center
Santa Monica, California
Jeffrey Stewart Barkun, MD, FRSC(C)
Professor of Surgery Leslie H. Blumgart, BDS, MD, DSc(Hon), FACS,
Department of Hepatobiliary & Transplant Surgery FRCS(Eng, Edin), FRCPS(Glas)
McGill University Health Centre Member
Montreal, Quebec, Canada Professor of Surgery and Attending Surgeon
Memorial Sloan Kettering Cancer Center;
Claudio Bassi, FRCS, FACS, FEBS Professor of Surgery
Professor of Surgery Weill Medical College of Cornell University
Pancreas Institute New York, New York
Verona University Hospital Trust
Verona, Italy Franz Edward Boas, MD, PhD
Assistant Attending Physician
Olca Basturk, MD Department of Radiology
Assistant Attending Physician Memorial Sloan Kettering Cancer Center
Department of Pathology New York, New York
Memorial Sloan Kettering Cancer Center;
Assistant Professor Lynn A. Brody, MD
Department of Pathology and Laboratory Medicine Attending Interventional Radiologist
Weill Medical College of Cornell University Department of Diagnostic Radiology
New York, New York Memorial Sloan Kettering Cancer Center
New York, New York
Rachel E. Beard, MD
Resident Physician Karen T. Brown, MD, FSIR
Department of Surgery Attending Radiologist
Beth Israel Deaconess Medical Center Department of Radiology
Boston, Massachusetts Memorial Sloan Kettering Cancer Center;
Professor of Clinical Radiology
Pierre Bedossa, MD, PhD Department of Radiology
Professor Weill Medical College of Cornell University
Department of Pathology New York, New York
Beaujon Hospital
Paris, France Jordi Bruix, MD, PhD
Senior Consultant
Jacques Belghiti, MD Liver Unit, BCLC Group
Professor Hospital Clinic
Physician University of Barcelona
Department of Hepato-Pancreatic-Biliary Surgery and Liver Centro de Investigación Biomédica en Red de Enfermedades
Transplantation Hepáticas y Digestivas (CIBERehd)
Beaujon Hospital Barcelona, Spain
Clichy, France
David A. Bruno, MD
Omar Bellorin-Marin, MD Assistant Professor of Surgery
Administrative Chief Resident Transplant Division
Department of Surgery University of Maryland School of Medicine
New York Presbyterian Queens/Weill Medical College of Baltimore, Maryland
Cornell University
Flushing, New York Elizabeth M. Brunt, MD
Professor
Pathology and Immunology
Washington University School of Medicine
St. Louis, Missouri
x CONTRIBUTORS
Justin M. Burns, MD See Ching Chan, MD, PhD

Assistant Professor of Surgery Clinical Professor
Department of Transplantation Department of Surgery
Mayo Clinic The University of Hong Kong
Jacksonville, Florida Hong Kong, China
Giovanni Butturini, MD, PhD William C. Chapman, MD, FACS

Department of Surgery Professor
The Pancreas Institute Chief, Division of General Surgery
Verona University Hospital Trust Chief, Abdominal Transplantation Section
Verona, Italy Washington University School of Medicine
St. Louis, Missouri
Juan Carlos Caicedo, MD
Adult and Pediatric Transplant Surgeon Daniel Cherqui, MD
Associate Professor of Surgery Professor
Department of Surgery Hepatobiliary Surgery and Liver Transplantation
Northwestern Memorial Hospital Paul Brousse Hospital
Northwestern University; Villejuif, France
Pediatric Transplant Surgeon
Department of Surgery Clifford S. Cho, MD
Lurie Children’s Hospital Chief, Division of Surgical Oncology
Chicago, Illinois University of Wisconsin School of Medicine and Public
Health
Mark P. Callery, MD Madison, Wisconsin
Harvard Medical School; Jin Wook Chung, MD, PhD
Chief, Division of General Surgery Professor
Beth Israel Deaconess Medical Center College of Medicine
Boston, Massachusetts Seoul National University
Seoul, Korea
Abdul Saied Calvino, MD
Assistant Professor of Surgery Jesse Clanton, MD
Boston University School of Medicine/Roger Williams Hepatopancreatobiliary Surgery Fellow
Medical Center Section of General, Thoracic and Vascular Surgery
Providence, Rhode Island Virginia Mason Medical Center
Seattle, Washington
Danielle H. Carpenter, MD
Assistant Professor Bryan Marshall Clary, MD
Pathology and Immunology Department of Surgery
Washington University School of Medicine University of California, San Diego
St. Louis, Missouri San Diego, California
C. Ross Carter, MD, FRCS Sean Patrick Cleary, MD, FRCSC

Consultant Pancreatic Surgeon Associate Professor
Glasgow Royal Infirmary Department of Surgery
Glasgow, Scotland University of Toronto
Toronto, Ontario, Canada
François Cauchy, MD
Physician Kelly M. Collins, MD
Hepatobiliary Surgery and Liver Transplantation Unit Senior Staff Surgeon, Transplant and Hepatobiliary Surgery
Beaujon Hospital Henry Ford Hospital
Clichy, France Surgical Director, Liver Transplant
Children’s Hospital of Michigan
Chung Yip Chan, MBBS, MMed(Surgery), Detroit, Michigan
MD, FRCSEd
Senior Consultant John Barry Conneely, MCh, FRCSI
Department of Hepatopancreatobiliary and Transplant Consultant
Surgery Hepatopancreatobiliary Surgeon
Singapore General Hospital Department of Surgery
Singapore Mater Misericordiae Hospital
Dublin, Ireland
CONTRIBUTORS xi
Louise C. Connell, MD Michael D. Darcy, MD

Fellow Professor of Radiology and Surgery
Memorial Sloan Kettering Cancer Center Washington University in St Louis;
New York, New York Chief of Interventional Radiology
Mallinckrodt Institute of Radiology
Carlos U. Corvera, MD, FACS St. Louis, Missouri
Chief, Liver, Biliary and Pancreatic Surgery Jeremy L. Davis, MD
Department of Gastrointestinal Surgical Oncology Assistant Research Physician
Maurice Galante Distinguished Professorship in Center for Cancer Research
Hepatobiliary Surgery National Cancer Institute, NIH
UCSF Helen Diller Family Comprehensive Cancer Center Bethesda, Maryland
San Francisco, California
Jeroen de Jonge, MD, PhD
Guido Costa, MD Assistant Professor
Resident Department of Hepatobiliary and Transplant Surgery
Division of Hepatobiliary and General Surgery Erasmus MC Rotterdam
Humanitas Research Hospital Rotterdam, The Netherlands
Rozzano-Milan, Italy
Ronald P. DeMatteo, MD, FACS
Anne M. Covey, MD Vice Chair, Department of Surgery
Attending Interventional Radiologist Chief, Division of General Surgical Oncology
Department of Diagnostic Radiology Leslie H. Blumgart Chair in Surgery
Memorial Sloan Kettering Cancer Center; Memorial Sloan Kettering Cancer Center
Professor of Radiology New York, New York
Department of Diagnostic Radiology
Weill Medical College of Cornell University Danielle K. DePeralta, MD
New York, New York Surgical Resident
Department of General Surgery
Jeffrey S. Crippin, MD Massachusetts General Hospital
Marilyn Bornefeld Chair in Gastrointestinal Research and Boston, Massachusetts
Treatment
Department of Internal Medicine Niraj M. Desai, MD
Washington University School of Medicine Assistant Professor
St. Louis, Missouri Department of Surgery
Johns Hopkins University School of Medicine
Kristopher P. Croome, MD Baltimore, Maryland
Assistant Professor
Department of Transplant Surgery Eduardo de Santibañes, MD, PhD
Mayo Clinic Chairman
Jacksonville, Florida General Surgery and Liver Transplantation
Professor
Hany Dabbous, MD Department of General Surgery
Professor of Tropical Medicine and Liver Diseases Hospital Italiano
Ain Shams University Buenos Aires, Argentina
Cairo, Egypt
Martin de Santibañes, MD
Michael I. D’Angelica, MD, FACS Associate Professor of Surgery
Attending Surgeon Hepato-Biliary-Pancreatic Unit
Hepatopancreatobiliary Surgery Liver Transplantation Unit
Enid A. Haupt Chair in Surgery Hospital Italiano
Memorial Sloan Kettering Cancer Center; Buenos Aires, Argentina
Associate Professor
Department of Surgery Euan J. Dickson, MD, FRCS
Weill Medical College of Cornell University Consultant Pancreatic Surgeon
New York, New York University of Glasgow
Glasgow, Scotland
xii CONTRIBUTORS
Christopher John DiMaio, MD Imane El Dika, MD

Director of Therapeutic Endoscopy Fellow
Division of Gastroenterology Memorial Sloan Kettering Cancer Center
Icahn School of Medicine at Mount Sinai New York, New York
New York, New York
Yousef El-Gohary, MA, MD, MRCS (Glasg)
Richard Kinh Gian Do, MD, PhD Physician
Associate Professor of Radiology Department of Surgery
Weill Medical College of Cornell University; Stony Brook University Medical Center
Assistant Attending Physician Stony Brook, New York
Memorial Sloan Kettering Cancer Center Itaru Endo, MD, PhD
New York, New York Professor and Chairman
Department of Gastroenterological Surgery
Safi Dokmak, MD Yokohama City University Graduate School of Medicine
Physician Yokohama, Japan
Hepatobiliary Surgery and Liver Transplantation Unit
Beaujon Hospital C. Kristian Enestvedt, MD, FACS
Clichy, France Assistant Professor of Surgery
Division of Abdominal Organ Transplantation/Hepatobiliary
Marcello Donati, MD, PhD Surgery
Assistant Professor of Surgery Oregon Health & Science University
General and Oncologic Surgery Unit Portland, Oregon
Department of Surgery and Medical-Surgical Specialties
University of Catania N. Joseph Espat, MD, FACS
Catania, Italy Professor of Surgery
M. B. Majella Doyle, MD, FACS Roger Williams Medical Center
Director, Liver Transplant Boston University School of Medicine
Director, Transplant HPB Fellowship Program Providence, Rhode Island
Section of Abdominal Transplantation
Washington University School of Medicine Cecilia G. Ethun, MD
St. Louis, Missouri Research Fellow
Division of Surgical Oncology
Vikas Dudeja, MBBS Department of Surgery
Assistant Professor Winship Cancer Institute of Emory University
Department of Surgical Oncology Atlanta, Georgia
Miller School of Medicine
University of Miami Sheung Tat Fan, MD, PhD, DSc
Miami, Florida Director
Liver Surgery Centre
Mark Dunphy, DO Hong Kong Sanatorium and Hospital;
Assistant Attending Physician Honorary Clinical Professor of Surgery
Department of Radiology Department of Surgery
Memorial Sloan Kettering Cancer Center The University of Hong Kong
New York, New York Hong Kong, China
Truman M. Earl, MD Paul T. Fanta, MD

Associate Professor Associate Clinical Professor
Department of Surgery Division of Hematology and Oncology
Division of Transplant and Hepatobiliary Surgery Department of Medicine
University of Mississippi Medical Center University of California, San Diego
Jackson, Mississippi San Diego, California
Tomoki Ebata, MD Olivier Farges, MD, PhD

Associate Professor Department of Hepato-Pancreatic-Biliary Surgery and Liver
Division of Surgical Oncology Transplantation
Department of Surgery Beaujon Hospital
Nagoya University Graduate School of Medicine Clichy, France
Nagoya, Japan
CONTRIBUTORS xiii
Cristina R. Ferrone, MD David A. Geller, MD

Associate Professor of Surgery Richard L. Simmons Professor of Surgery
Massachusetts General Hospital Chief, Division of Hepatobiliary and Pancreatic Surgery
Boston, Massachusetts Department of Surgery
University of Pittsburgh
Ryan C. Fields, MD Pittsburgh, Pennsylvania
Assistant Professor of Surgery
Section of Hepatopancreatobiliary, Gastrointestinal, and Hans Gerdes, MD
Oncologic Surgery Attending Physician
Department of Surgery; Department of Medicine
Associate Program Director Memorial Hospital for Cancer and Allied Diseases;
General Surgery Residency Program; Professor of Clinical Medicine
Director Weill Medical College of Cornell University
Resident Research; New York, New York
Barnes-Jewish Hospital
Washington University School of Medicine Scott R. Gerst, MD
St. Louis, Missouri Associate Attending Radiologist
Mary Fischer, MD Memorial Sloan Kettering Cancer Center
Anesthesiologist New York, New York
Department of Anesthesiology and Critical Care
Memorial Sloan Kettering Cancer Center; George K. Gittes, MD
Associate Professor Professor of Surgery and Surgeon-in-Chief
Department of Anesthesiology Department of Surgery
Weill Medical College of Cornell University Children’s Hospital of Pittsburgh
New York, New York University of Pittsburgh School of Medicine
Pittsburgh, Pennsylvania
Sarah B. Fisher, MD
Department of Surgery Jaime Glorioso, MD
Division of Surgical Oncology Resident
Emory University Department of General Surgery
Atlanta, Georgia Mayo Clinic
Rochester, Minnesota
Devin C. Flaherty, DO, PhD
Fellow Jill S. Gluskin, MD
Department of Surgical Oncology Assistant Attending Radiologist
John Wayne Cancer Institute Department of Radiology
Providence Saint John’s Health Center Memorial Sloan Kettering Cancer Center
Santa Monica, California New York, New York
Yuman Fong, MD Brian K. P. Goh, MBBS, MMed(Surgery), MSc,

Chairman FRCSEd
Department of Surgery Senior Consultant
City of Hope National Medical Center Department of Hepatopancreatobiliary and Transplant
Duarte, California Surgery
Singapore General Hospital
Scott L. Friedman, MD Singapore
Fishberg Professor of Medicine
Division of Liver Diseases Stevan A. Gonzalez, MD
Icahn School of Medicine at Mount Sinai Medical Director of Liver Transplantation
New York, New York Department of Hepatology
Baylor All Saints Medical Center
Ahmed Gabr, MD Fort Worth, Texas
Clinical Research Fellow
Department of Radiology Karyn A. Goodman, MD
Northwestern University Feinberg School of Medicine Professor
Chicago, Illinois Department of Radiation Oncology
University of Colorado
John R. Galloway, MD Denver, Colorado
Emory University School of Medicine
Atlanta, Georgia
xiv CONTRIBUTORS
Gregory J. Gores, MD James J. Harding, MD

Professor of Medicine Assistant Attending Physician
Transplant Center Memorial Sloan Kettering Cancer Center;
Mayo Clinic Instructor
Rochester, Minnesota Weill Medical College at Cornell University
New York, New York
Eduardo H. Gotuzzo, MD, FACP, FIDSA
Director Ewen M. Harrison, MB ChB, PhD, FRCS
Instituto de Medicina Tropical Alexander von Humboldt Senior Lecturer
Universidad Peruana Cayetano Heredia; Department of Clinical Surgery
Head, Enfermedades Ifecciosas y Tropicales University of Edinburgh;
Hospital Nacional Cayetano Heredia Consultant
Lima, Peru Hepatopancreatobiliary Surgeon
Department of Clinical Surgery
Dirk J. Gouma, MD Royal Infirmary of Edinburgh
Professor Edinburgh, Scotland
Academic Medical Center Werner Hartwig, MD
Amsterdam, The Netherlands Associate Professor
Deputy Medical Director
Paul D. Greig, MD, FRCSC Head, Division of Pancreatic Surgery and LMU Munich
Professor Pancreatic Center;
Department of Surgery Department of General, Visceral, and Transplantation Surgery
University of Toronto; LMU University Hospital
Staff Surgeon Munich, Germany
Toronto General Hospital Kiyoshi Hasegawa, MD, PhD
Toronto, Ontario, Canada Associate Professor
Division of Hepato-Biliary-Pancreatic Surgery
James F. Griffin, MD Graduate School of Medicine
Assistant Resident University of Tokyo
Department of Surgery Tokyo, Japan
Johns Hopkins Hospital
Baltimore, Maryland Jaclyn F. Hechtman, MD
Assistant Member
Christopher M. Halloran, MD, FRCS, FAcadTM Department of Pathology
Clinical Senior Lecturer/Consultant Memorial Sloan Kettering Cancer Center
Pancreato-Biliary Surgeon New York, New York
Department of Molecular and Clinical Cancer Medicine
University of Liverpool Julie K. Heimbach, MD
Liverpool, England Professor of Surgery
Division of Transplantation Surgery
Neil A. Halpern, MD Mayo Clinic
Chief, Critical Care Medicine Rochester, Minnesota
Department of Anesthesiology and Critical Care Medicine
Memorial Sloan Kettering Cancer Center; William S. Helton, MD
Professor of Clinical Anesthesiology and Medicine Director, Liver, Biliary and Pancreas Surgery Center
Department of Anesthesiology Department of General, Vascular and Thoracic Surgery
Weill Medical College of Cornell University Virginia Mason Medical Center
New York, New York Seattle, Washington
Chet W. Hammill, MD, MCR, FACS Alan W. Hemming, MD

Department of Liver and Pancreas Surgery Professor and Chief
The Oregon Clinic Division of Transplantation and Hepatobiliary Surgery
Portland, Oregon University of California, San Diego
San Diego, California
Paul D. Hansen, MD, FACS
Medical Director J. Michael Henderson, MB, ChB, FRCE(Ed), FACS
Department of Surgical Oncology Chief Medical Officer
Providence Portland Cancer Center; University of Mississippi Medical Center
Department of Liver and Pancreas Surgery Jackson, Mississippi
The Oregon Clinic
Portland, Oregon
CONTRIBUTORS xv
Asher Hirshberg, MD, FACS Joseph Ralph Kallini, MD

Director of Emergency Vascular Surgery Clinical Research Fellow
Kings County Hospital Center Department of Radiology
Brooklyn, New York Northwestern University Feinberg School of Medicine
Chicago, Illinois
James R. Howe V, MD
Director Ivan Kangrga, MD, PhD
Division of Surgical Oncology and Endocrine Surgery Professor and Chief of Clinical Anesthesiology
Department of Surgery Department of Anesthesiology
Carver College of Medicine Washington University School of Medicine
University of Iowa St. Louis, Missouri
Iowa City, Iowa
Paul J. Karanicolas, MD, PhD
Christopher B. Hughes, MD Assistant Professor
Associate Professor of Surgery Department of Surgery
Surgical Director University of Toronto
Department of Liver Transplantation Toronto, Ontario, Canada
Starzl Transplantation Institute
University of Pittsburgh Medical Center Seth S. Katz, MD, PhD
Pittsburgh, Pennsylvania Assistant Clinical Member
Christine Iacobuzio-Donahue, MD, PhD Memorial Sloan Kettering Cancer Center
Attending Pathologist New York, New York
Department of Pathology
Affiliate Member Steven C. Katz, MD, FACS
Human Oncology and Pathogenesis Program Associate Professor of Surgery
Associate Director for Translational Research Boston University School of Medicine
David M. Rubenstein Center for Pancreatic Cancer Research Director, Complex Surgical Oncology
Memorial Sloan Kettering Cancer Center Fellowship Director, Surgical Immunotherapy
New York, New York Roger Williams Medical Center
Providence, Rhode Island
William R. Jarnagin, MD, FACS
Chief, Hepatopancreatobiliary Surgery Kaitlyn J. Kelly, MD
Benno C. Schmidt Professor of Surgical Oncology Assistant Professor of Surgery
Memorial Sloan Kettering Cancer Center; Division of Surgical Oncology
Professor of Surgery University of California, San Diego
Weill Medical College of Cornell University; San Diego, California
New York, New York
Nancy E. Kemeny, MD
Roger L. Jenkins, MD Professor of Medicine
Professor of Surgery Weill Medical College of Cornell University;
Tufts Medical School Attending Physician
Boston, Massachusetts; Solid Tumor–GI Division
Chief of Surgery Memorial Sloan Kettering Cancer Center
Division of Hepatobiliary Surgery New York, New York
Department of Transplantation
Lahey Hospital and Medical Center Eugene P. Kennedy, MD
Burlington, Massachusetts Associate Professor
Zeljka Jutric, MD Sidney Kimmel Medical College
Fellow, Liver and Pancreas Surgery Thomas Jefferson University
Providence Portland Cancer Center Philadelphia, Pennsylvania
Portland, Oregon
Korosh Khalili, MD, FRCPC
Christoph Kahlert, MD Associate Professor
Universitätsklinikum Carl Gustav Carus Dresden Department of Medical Imaging
Klinik und Poliklinik für Viszeral-, Thorax- und University of Toronto
Gefäßchirurgie Toronto, Ontario, Canada
Dresden, Germany
xvi CONTRIBUTORS
Adeel S. Khan, MD, FACS Norihiro Kokudo, MD, PhD

Instructor of Transplant Surgery Professor
Department of Surgery Division of Hepato-Biliary-Pancreatic Surgery
Washington University School of Medicine Graduate School of Medicine
St. Louis, Missouri University of Tokyo
Tokyo, Japan
Saboor Khan, PhD, FRCS, FACS
Consultant Hepatobiliary Pancreatic and General Surgeon Dionysios Koliogiannis, MD
University Hospitals Coventry and Warwickshire NHS Trust; Resident Surgeon
Associate Professor of Surgery (Hon.) Department of General, Visceral, and Transplantation Surgery
Department of Surgery LMU University Hospital
Warwick Medical School Munich, Germany
Coventry, England
David A. Kooby, MD
Heung Bae Kim, MD Professor of Surgery
Weitzman Family Chair in Surgical Innovation Directory of Surgical Oncology
Department of Surgery Emory St. Joseph’s Hospital;
Director Associate Professor of Surgery
Pediatric Transplant Center Director of Minimally Invasive Gastrointestinal Oncologic
Boston Children’s Hospital; Surgery
Associate Professor of Surgery Emory University School of Medicine
Harvard Medical School Atlanta, Georgia
Boston, Massachusetts
Kevin Korenblat, MD
T. Peter Kingham, MD Professor of Medicine
Assistant Professor Department of Medicine
Department of Surgery Washington University School of Medicine
Memorial Sloan Kettering Cancer Center St. Louis, Missouri
New York, New York
Simone Krebs, MD
Allan D. Kirk, MD, PhD, FACS Department of Radiology
David C. Sabiston, Jr. Professor and Chairman Molecular Imaging and Therapy Service
Department of Surgery Memorial Sloan Kettering Cancer Center
Duke University New York, New York
Durham, North Carolina
Michael J. LaQuaglia, MD
David S. Klimstra, MD General Surgery Resident
Chairman and James Ewing Alumni Chair in Pathology Department of Surgery
Department of Pathology Albert Einstein College of Medicine and Montefiore Medical
Memorial Sloan Kettering Cancer Center; Center
Professor Bronx, New York
Department of Pathology and Laboratory Medicine
Weill Medical College of Cornell University Michael P. LaQuaglia, MD
New York, New York Chief, Pediatric Service
Michael Kluger, MD Memorial Sloan Kettering Cancer Center;
Assistant Professor of Surgery Professor of Surgery
Division of GI and Endocrine Surgery Department of Surgery
Columbia University College of Physicians and Surgeons Weill Medical College of Cornell University
New York-Presbyterian Hospital New York, New York
New York, New York
Nicholas F. LaRusso, MD
Stuart J. Knechtle, MD Medical Director
Professor of Surgery Center for Connected Care
Department of Surgery Mayo Clinic;
Duke University School of Medicine Charles H. Weinman Professor of Medicine
Durham, North Carolina Biochemistry and Molecular Biology
Mayo Clinic College of Medicine;
Jonathan B. Koea, MD, FACS, FRACS Distinguished Investigator
Hepatobiliary Surgeon Mayo Foundation
Upper Gastrointestinal Unit Rochester, Minnesota
North Shore Hospital
Auckland, New Zealand
CONTRIBUTORS xvii
Alexis Laurent, MD, PhD Chung-Wei Lin, MD

Professor of Surgery Attending Physician
Department of Digestive, HPB and Liver Transplant Surgery Department of Surgery
Hôpital Henri Mondor Koo Foundation Sun Yat-Sen Cancer Center
Assistance Publique-Hôpitaux de Paris Faculté de Médecine Taipei, Taiwan
Université Paris-Est-Créteil
Créteil, France David C. Linehan, MD
Seymour I. Schwartz Professor and Chairman
Konstantinos N. Lazaridis, MD Department of Surgery
Professor of Medicine University of Rochester
Division of Gastroenterology & Hepatology Rochester, New York
Center for Basic Research in Digestive Diseases
Mayo Clinic College of Medicine Roberto Carlos Lopez-Solis, MD, FACS
Rochester, Minnesota Assistant Professor of Surgery
Director of Organ Procurement
Julie N. Leal, MD, FRCSC Department of Transplant Surgery
Fellow University of Pittsburgh Medical Center;
Division of Hepatopancreatobiliary Surgery Department of General Surgery
Memorial Sloan Kettering Cancer Center McGowan Center of Regenerative Medicine
New York, New York University of Pittsburgh
Pittsburgh, Pennsylvania
Eliza J. Lee, MD
Resident Jeffrey A. Lowell, MD, FACS
Department of Surgery Professor of Surgery and Pediatrics
Beth Israel Deaconess Medical Center Department of Surgery
Boston, Massachusetts Washington University School of Medicine
St. Louis, Missouri
Major Kenneth Lee IV, MD, PhD
Assistant Professor of Surgery David C. Madoff, MD
Department of Surgery Professor of Radiology
University of Pennsylvania Perelman School of Medicine Chief
Philadelphia, Pennsylvania Division of Interventional Radiology
Ser Yee Lee, MBBS, MMed(Surgery), MSc, Weill Medical College of Cornell University
FAMS, FRCSEd New York, New York
Consultant
Department of Hepatopancreatobiliary and Transplant Jason Maggi, MD
Surgery General Surgeon
Singapore General Hospital Department of General Surgery
Singapore Naval Hospital Camp Pendleton
United States Navy
Riccardo Lencioni, MD Oceanside, California
Professor of Radiology
Vice-Chair for Clinical and Translational Research Shishir K. Maithel, MD, FACS
Department of Interventional Radiology Associate Professor of Surgery
University of Miami Miller School of Medicine Division of Surgical Oncology
Sylvester Comprehensive Cancer Center Department of Surgery
Miami, Florida Winship Cancer Institute of Emory University
Atlanta, Georgia
Alexandre Liccioni, MD, PhD
Gastroenterologist Ali W. Majeed, MD, FRCS(Edin), FRCS(Gen)
Barcelona Clinic Liver Cancer (BCLC) Group, Liver Unit Consultant
Hospital Clinic Barcelona, IDIBAPS Hepatobiliary Surgeon
University of Barcelona Department of Hepatobiliary Surgery
Barcelona, Spain Northern General Hospital
Sheffield, England
Michael E. Lidsky, MD
Resident Peter Malfertheiner, MD
Department of Surgery Department of Gastroenterology
Duke University School of Medicine University of Magdeburg
Durham, North Carolina Magdeburg, Germany
xviii CONTRIBUTORS
Giuseppe Malleo, MD, PhD Jessica E. Maxwell, MD, MBA

Department of Surgery Surgery Resident
The Pancreas Institute Department of General Surgery
Verona University Hospital Trust Carver College of Medicine
Verona, Italy University of Iowa
Iowa City, Iowa
Shennen A. Mao, MD
Resident Oscar M. Mazza, MD
Department of General Surgery Professor of Surgery
Mayo Clinic Staff Surgeon
Rochester, Minnesota Hepato-Biliary-Pancreatic Unit
Hospital Italiano
Giovanni Marchegiani, MD Buenos Aires, Argentina
Pancreas Institute
University of Verona Ian D. McGilvray, MDCM, PhD
Verona, Italy Associate Professor
Luis A. Marcos, MD Staff Surgeon
Associate Professor of Clinical Medicine, Molecular Genetics University Health Network
and Microbiology University of Toronto
Division of Infectious Diseases Toronto, Ontario, Canada
Stony Brook University (State University of New York)
Stony Brook, New York Colin J. McKay, MD, FRCS
Consultant Pancreatic Surgeon
James F. Markmann, MD, PhD Glasgow Royal Infirmary;
Chief Hon. Clinical Associate Professor
Division of Transplant Surgery University of Glasgow
Claude E. Welch Professor of Surgery Glasgow, Scotland
Harvard Medical School Doireann M. McWeeney, MB, FFRRCSI
Massachusetts General Hospital Joint Department of Medical Imaging
Boston, Massachusetts University Health Network
Toronto, Ontario, Canada
J. Wallis Marsh, MD, MBA
Raizman-Haney Professor of Surgery Jose Melendez, MD
Starzl Transplantation Institute Associate Chair of Anesthesiology
University of Pittsburgh Medical Center Associate Professor of Anesthesiology
Pittsburgh, Pennsylvania University of Colorado School of Medicine
Denver, Colorado
Robert C. G. Martin II, MD, PhD, FACS
Professor of Surgery Robin B. Mendelsohn, MD
Sam and Loita Weakley Endowed Chair of Surgical Oncology Assistant Attending Physician
Director, Division of Surgical Oncology Department of Medicine
University of Louisville Gastroenterology and Nutrition Service
Louisville, Kentucky Memorial Sloan Kettering Cancer Center
New York, New York
Ryusei Matsuyama, MD, PhD
Assistant Professor George Miller, MD
Department of Gastroenterological Surgery Assistant Professor
Yokohama City University Graduate School of Medicine Department of Surgery and Cell Biology
Yokohama, Japan New York University School of Medicine
New York, New York
Matthias S. Matter, MD
Institute of Pathology Klaus E. Mönkemüller, MD, PhD, FASGE
Molecular Pathology Division Professor
University Hospital of Basel Director, Division of Gastroenterology and Hepatology
Basel, Switzerland Hirschowitz I. Endoscopy Center of Excellence
University of Alabama
Francisco Juan Mattera, MD Birmingham, Alabama
Surgeon
Liver Transplantation Unit
Hospital Italiano
Buenos Aires, Argentina
CONTRIBUTORS xix
Ryutaro Mori, MD, PhD Rachel O’Connor

Assistant Professor Division of Surgical Oncology
Department of Gastroenterological Surgery University of Louisville School of Medicine
Yokohama City University Graduate School of Medicine Louisville, Kentucky
Yokohama, Japan
John G. O’Grady, MD, FRCPI
Vitor Moutinho, MD Professor
Surgical Oncology Fellow Institute of Liver Studies
Department of Surgery King’s College Hospital
Memorial Sloan Kettering Cancer Center London, England
New York, New York
Frances E. Oldfield, MBChB (Hons), MRCS
Masato Nagino, MD, PhD Clinical Research Fellow
Professor and Chairman University of Liverpool
Division of Surgical Oncology Liverpool, England
Nagoya University Graduate School of Medicine Karl J. Oldhafer, MD, PhD
Nagoya, Japan Professor
David M. Nagorney, MD Asklepios Hospital Barmbek
Professor of Surgery Semmelweis University Budapest
Department of Surgery Asklepios Campus Hamburg
Mayo Clinic Hamburg, Germany
Rochester, Minnesota
Kim M. Olthoff, MD
Satish Nagula, MD Donald Guthrie Professor of Surgery
Associate Professor of Medicine Division of Transplantation
Division of Gastroenterology Department of Surgery
Department of Medicine University of Pennsylvania
Icahn School of Medicine at Mount Sinai Philadelphia, Pennsylvania
New York, New York
Susan L. Orloff, MD, FACS, FAASLD
Attila Nakeeb, MD Professor of Surgery
Professor Chief
Department of Surgery Division of Abdominal Organ Transplantation/Hepatobiliary
Indiana University of School of Medicine Surgery
Indianapolis, Indiana Department of Surgery;
Adjunct Professor
Geir I. Nedredal, MD, PhD Department of Microbiology and Immunology
Assistant Adjunct Professor Oregon Health & Science University;
Department of Surgery Chief
University of California, San Francisco Transplant Program
San Francisco, California Portland VA Medical Center
Portland, Oregon
John P. Neoptolemos, MD, FRCS, FMedSci
Chair of Surgery Alessandro Paniccia, MD
Department of Molecular and Clinical Cancer Medicine General Surgery Resident
University of Liverpool Department of Surgery
Liverpool, England University of Colorado Anschutz Medical Campus
Aurora, Colorado
James Neuberger, DM, FRCP
The Liver Unit Valérie Paradis, MD, PhD
Queen Elizabeth Hospital Professor
Birmingham, England Department of Pathlogy
Beaujon Hospital
Scott L. Nyberg, MD, PhD Paris, France
Professor
Department of Surgery Rowan W. Parks, MD, FRCSI, FRCSEd
Mayo Clinic Professor of Surgical Sciences
Rochester, Minnesota Department of Clinical Surgery
University of Edinburgh
Edinburgh, Scotland
xx CONTRIBUTORS
Gérard Pascal, MD Fabio Procopio, MD

Department of Digestive, HPB and Liver Transplant Surgery Staff Surgeon
Hôpital Henri Mondo Division of Hepatobiliary and General Surgery
Assistance Publique-Hôpitaux de Paris Faculté de Médecine Humanitas Research Hospital
Université Paris-Est-Créteil Rozzano-Milan, Italy
Créteil, France
Michael J. Pucci, MD
Stephen M. Pastores, MD Assistant Professor of Surgery
Program Director, Critical Care Medicine Department of Surgery
Department of Anesthesiology and Critical Care Medicine Sidney Kimmel Medical College
Memorial Sloan Kettering Cancer Center Thomas Jefferson University
New York, New York; Philadelphia, Pennsylvania
Professor of Clinical Anesthesiology and Medicine
Department of Anesthesiology Motaz Qadan, MD, PhD
Weill Medical College of Cornell University Surgical Oncology Fellow
New York, New York Department of Surgery
Memorial Sloan Kettering Cancer Center
Timothy M. Pawlik, MD, PhD New York, New York
Professor of Surgery and Oncology
Chair, Department of Surgery Kheman Rajkomar, FRACS
The Urban Meyer III and Shelley Meyer Cancer Research Hepatopancreatobiliary Fellow
Chair Upper Gastrointestinal Unit
The Ohio State Wexner Medical Center Department of Surgery
Columbus, Ohio North Shore Hospital
Auckland, New Zealand
Venu G. Pillarisetty, MD
Assistant Professor Srinevas K. Reddy, MD
Department of Surgery Associate Professor of Oncology
University of Washington; Department of Surgical Oncology
Attending Surgeon Roswell Park Cancer Institute
University of Washington Medical Center Buffalo, New York
Seattle, Washington
Maria E. Reig, MD, PhD
James Francis Pingpank Jr., MD Hepatologist
Associate Professor of Surgery Barcelona Clinic Liver Cancer (BCLC) Group, Liver Unit
Department of Surgery Hospital Clinic Barcelona, IDIBAPS
University of Pittsburgh University of Barcelona
Pittsburgh, Pennsylvania Centro de Investigación Biomédica en Red de Enfermedades
Hepáticas y Digestivas (CIBERehd)
C. Wright Pinson, MD, MBA Barcelona, Spain
Deputy Vice Chancellor for Health Affairs
Vanderbilt University Medical Center; Joseph Arturo Reza, MD
Chief Executive Officer Resident in General Surgery
Vanderbilt Health System Department of Surgery
Nashville, Tennessee University of California San Francisco
San Francisco, California
Henry Anthony Pitt, MD
Chief Quality Officer John Paul Roberts, MD
Temple University Health System; Chief
Associate Vice Dean for Clinical Affairs Division of Transplantation
Department of Surgery Department of Surgery
Lewis Katz School of Medicine at Temple University University of California San Francisco
Philadelphia, Pennsylvania San Francisco, California
James J. Pomposelli, MD Piera Marie Cote Robson, MSN

Professor of Surgery Clinical Nurse Specialist
University of Colorado Departments of Nursing and Radiology
Denver, Colorado Memorial Sloan Kettering Cancer Center
New York, New York
CONTRIBUTORS xxi
Flavio G. Rocha, MD Charbel Sandroussi, MMSc, FRACS

Staff Surgeon, Hepatopancreatobiliary Service Clinical Associate Professor
Section of General, Thoracic and Vascular Surgery Department of Hepatobiliary and Upper Gastrointestinal
Virginia Mason Medical Center; Surgery
Clinical Assistant Professor of Surgery Royal Prince Alfred Hospital
University of Washington Sydney, Australia
Seattle, Washington
Dominic E. Sanford, MD
Garrett Richard Roll, MD Department of Surgery
Liver Transplant and Multiorgan Retrieval Surgeon Washington University School of Medicine
Liver Unit St. Louis, Missouri
Queen Elizabeth Hospital
Birmingham, England Olivier Scatton, MD, PhD
Professor
Sean M. Ronnekleiv-Kelly, MD Physician
Surgical Oncology Fellow Hepatobiliary Surgery and Liver Transplantation Unit
Department of Surgical Oncology Pitié-Salpêtrière Hospital
Johns Hopkins Hospital Paris, France
Baltimore, Maryland
Mark Andrew Schattner, MD
Alexander S. Rosemurgy II, MD Associate Clinical Member
Director of Hepatopancreaticobiliary Surgery Department of Medicine
Florida Hospital Tampa Memorial Sloan Kettering Cancer Center;
Tampa, Florida Associate Professor of Clinical Medicine
Department of Medicine
Charles B. Rosen, MD Weill Medical College of Cornell University
Professor of Surgery New York, New York
Chair, Division of Transplantation Surgery
Mayo Clinic William Palmer Schecter, MD
Rochester, Minnesota Professor of Clinical Surgery Emeritus
Pierre F. Saldinger, MD University of California, San Francisco
Chairman and Surgeon-in-Chief San Francisco, California
New York Hospital Queens Hans Francis Schoellhammer, MD
Flushing, New York Assistant Clinical Professor
Professor of Clinical Surgery Department of Surgery
Weill Medical College of Cornell University City of Hope National Medical Center
New York, New York Duarte, California
Riad Salem, MD, MBA Richard D. Schulick, MD, PhD

Professor Professor and Chair
Departments of Radiology, Medicine (Hematology- Department of Surgery
Oncology), and Surgery University of Colorado School of Medicine
Northwestern University Feinberg School of Medicine; Aurora, Colorado
Director, Interventional Oncology
Robert H. Lurie Comprehensive Cancer Center Lawrence H. Schwartz, MD
Northwestern Memorial Hospital Professor of Radiology
Chicago, Illinois Columbia University College of Physicians and Surgeons
New York Presbyterian Hospital
Suhail Bakr Salem, MD New York, New York
Advanced Endoscopy Fellow
Department of Gastroenterology Kevin N. Shah, MD
Memorial Sloan Kettering Cancer Center Department of Surgery
New York, New York Duke University Medical Center
Durham, North Carolina
Roberto Salvia, MD
Pancreas Institute
University of Verona
Verona, Italy
xxii CONTRIBUTORS
Ross W. Shepherd, MD, FRACP, FRCP Stephen B. Solomon, MD

Adjunct Professor Chief, Interventional Radiology Service
Pediatric Gastroenterology and Hepatology Director, Center for Image-Guided Intervention
Baylor College of Medicine Memorial Sloan Kettering Cancer Center
Houston, Texas; New York, New York
Honorary Professor
Queensland Institute for Medical Research Olivier Soubrane, MD
University of Queensland School of Medicine Professor
Brisbane, Australia Physician
Hepatobiliary Surgery and Liver Transplantation Unit
Hiroshi Shimada, MD, PhD Beaujon Hospital
Professor Clichy, France
Department of Gastroenterological Surgery
Yokohama City University Graduate School of Medicine Nicholas Spinelli, MD
Yokohama, Japan General Surgery
Hepatobiliary and Pancreatic Surgery
Masafumi Shimoda, MD, PhD Sentara Martha Jefferson Hospital
Assistant Professor Charlottesville, Virginia
Osaka University Graduate School of Medicine John A. Stauffer, MD, FACS
Osaka, Japan Associate Professor of Surgery
Junichi Shindoh, MD, PhD Mayo Clinic
Attending Surgeon Jacksonville, Florida
Division of Hepatobiliary-Pancreatic Surgery
Toranomon Hospital Lygia Stewart, MD
Tokyo, Japan Professor of Clinical Surgery
Hosein Shokouh-Amiri, MD, FACS University of California, San Francisco;
Clinical Professor of Surgery Chief, General Surgery
Louisiana State University Health Sciences Center; San Francisco VA Medical Center
Surgical Director, Liver Transplantation San Francisco, California
John C. McDonald Regional Transplant Center
Willis-Knighton Health System Matthew S. Strand, MD
Shreveport, Louisiana Resident Physician in General Surgery
Surgical Oncology Research Fellow
Jason K. Sicklick, MD, FACS Department of Surgery
Assistant Professor of Surgery Washington University School of Medicine
Division of Surgical Oncology St. Louis, Missouri
Moores UCSD Cancer Center
University of California, San Diego James H. Tabibian, MD, PhD
UC San Diego Health System Instructor of Medicine
San Diego, California Division of Gastroenterology and Hepatology
Mayo Clinic
Robert H. Siegelbaum, MD Rochester, Minnesota;
Assistant Attending Radiologist Instructor of Medicine
Department of Radiology, Interventional Radiology Service Division of Gastroenterology
Memorial Sloan Kettering Cancer Center University of Pennsylvania
New York, New York Philadelphia, Pennsylvania;
Assistant Professor of Medicine
Gagandeep Singh, MD, FACS Division of Gastroenterology and Hepatology
Chief, Division of Surgical Oncology UC Davis Medical Center
Head, Hepatobiliary and Pancreatic Surgery Sacramento, California
City of Hope National Medical Center Guido Torzilli, MD, PhD, FACS
Duarte, California Professor of Surgery
School of Medicine
Rory L. Smoot, MD Humanitas University;
Assistant Professor of Surgery Chairman of the Department of Surgery
Department of Surgery Director of the Division of Hepatobiliary and General
Mayo Clinic Surgery
Rochester, Minnesota Humanitas Research Hospital, IRCCS
Rozzano-Milan, Italy
CONTRIBUTORS xxiii
James F. Trotter, MD Alejandra Maria Villamil, MD

Medical Director of Liver Transplantation Universidad de Buenos Aires;
Department of Hepatology Liver Transplantation Unit
Baylor University Medical Center Hospital Italiano de Buenos Aires;
Dallas, Texas Director,
Argentine School of Hepatology
Simon Turcotte, MD, FRCSC Argentine Association for the Study of Liver Disease
Assistant Professor of Surgery Buenos Aires, Argentina
Université de Montréal;
Hepatopancreatobiliary and Liver Transplantation Service Louis P. Voigt, MD
Centre Hospitalier de l’Université de Montréal; Associate Attending Physician
Scientist Department of Anesthesiology and Critical Care Medicine
Centre de Recherche du Centre Hospitalier de l’Université Memorial Sloan Kettering Cancer Center;
de Montréal Assistant Professor of Clinical Anesthesiology and Medicine
Montreal, Quebec, Canada Department of Anesthesiology
Weill Medical College of Cornell University
Yumirle P. Turmelle, MD New York, New York
Associate Professor
Department of Pediatrics Charles M. Vollmer Jr., MD
Washington University School of Medicine Professor of Surgery
St. Louis, Missouri Department of Surgery
University of Pennsylvania Perelman School of Medicine
Demetrios J. Tzimas, MD Philadelphia, Pennsylvania
Assistant Professor of Medicine
Division of Gastroenterology and Hepatology Jack R. Wands, MD
Department of Medicine Jeffrey and Kimberly Greenberg-Artemis and Martha
Stony Brook University Hospital Joukowsky Professor in Gastroenterology
Stony Brook, New York Professor of Medical Science
Director, Division of Gastroenterology and Liver Research
Thomas Van Gulik, MD, PhD Center
Professor Rhode Island Hospital
Department of Surgery Warren Alpert Medical School of Brown University
Academic Medical Center Providence, Rhode Island
Amsterdam, The Netherlands
Julia Wattacheril, MD
Andrea Vannucci, MD Assistant Professor of Medicine
Associate Professor of Anesthesiology Division of Digestive and Liver Diseases
Department of Anesthesiology Center for Liver Disease and Transplantation
Washington University School of Medicine Columbia University College of Physicians and Surgeons
St. Louis, Missouri New York, New York
Jean-Nicolas Vauthey, MD, FACS Sharon Marie Weber, MD

Professor of Surgical Oncology Tim and MaryAnn McKenzie Chair of Surgical Oncology
Chief, Hepato-Pancreato-Biliary Section Department of General Surgery
Bessie McGoldrick Professor in Clinical Cancer Research Vice Chair, General Surgery
Department of Surgical Oncology University of Wisconsin;
University of Texas MD Anderson Cancer Center Director for Surgical Oncology
Houston, Texas University of Wisconsin Carbone Cancer Center
Madison, Wisconsin
Diana Vetter, MD
Physician Matthew J. Weiss, MD
Department of Abdominal Surgery Assistant Professor of Surgery and Oncology
University Hospital Zurich Department of Surgery
Zurich, Switzerland Johns Hopkins Hospital
Baltimore, Maryland
Valérie Vilgrain, MD
Professor of Radiology Jürgen Weitz, MD
Beaujon Hospital Chair, Department of Gastrointestinal, Thoracic and Vascular
Paris Diderot University Surgery
Paris, France Medizinische Fakultät Carl Gustav Carus
Technische Universität Dresden
Dresden, Germany
xxiv CONTRIBUTORS
Jens Werner, MD Chang Jin Yoon, MD, PhD

Director and Chairman Professor
Professor of Surgery College of Medicine
Department of General, Visceral, and Transplantation Surgery Seoul National University
LMU University Hospital Seoul, Korea;
Munich, Germany Professor
Megan Winner, MD Seoul National University Bundang Hospital
Attending Surgeon Gyeonggi, Korea
Winthrop University Hospital
Mineola, New York Adam Yopp, MD
Assistant Professor of Surgery
John Wong, MD, PhD Department of Surgery
Honorary Clinical Professor University of Texas Southwestern Medical Center
Department of Surgery Dallas, Texas
The University of Hong Kong
Hong Kong, China D. Owen Young, MD
General Surgery Resident
Dennis Yang, MD Department of Graduate Medical Education
Assistant Professor of Medicine Virginia Mason Medical Center
Division of Gastroenterology Seattle, Washington
University of Florida College of Medicine
Gainesville, Florida Kai Zhao, MD
Hooman Yarmohammadi, MD Stony Brook University
Assistant Professor of Radiology Stony Brook, New York
Memorial Sloan Kettering Cancer Center Gazi B. Zibari, MD, FACS
New York, New York Director of Transplantation Services
Director of Advanced Surgery Center
Charles J. Yeo, MD John C. McDonald Regional Transplant Center
Samuel D. Gross Professor and Chair Willis-Knighton Health System;
Department of Surgery Clinical Professor of Surgery
Sidney Kimmel Medical College Malcolm Feist Chair in Transplant Surgery
Thomas Jefferson University Louisiana State University Health Sciences Center
Philadelphia, Pennsylvania Shreveport, Louisiana
Theresa Pluth Yeo, PhD, MPH, ACNP-BC George Zogopoulos, MD, PhD, FRCS(C), FACS
Co-Director, Jefferson Pancreas Tumor Registry Assistant Professor of Surgery, McGill University
Department of Surgery Hepato-Pancreato-Biliary and Abdominal Organ Transplant
Thomas Jefferson University Hospital; Surgery
Adjunct Associate Professor McGill University Health Centre
Jefferson College of Nursing Montreal, Quebec, Canada
Philadelphia, Pennsylvania
PREFACE
The sixth edition of Blumgart’s Surgery of the Liver, Biliary Tract, to discuss specific topics based not only on the published litera-
and Pancreas is the largest and most ambitious effort in the long ture but on their own views. To that effect, overlap between
history of this textbook and maintains the tradition of embrac- chapters and discussion of controversy was encouraged in order
ing change in order to keep the book relevant to what Dr. Leslie to allow for conflicting points of view.
H. Blumgart established long ago. The initial section of the book remains dedicated to HPB
The sixth edition continues the tradition of using associate anatomy and physiology, with Chapter 2, “Surgical and Radio-
editors to comprehensively cover the extraordinary growth of logic Anatomy of the Liver, Biliary Tract, and Pancreas,” forming
knowledge and advances over the past four years.The associate the backbone of this section. Chapter 2 is one of the most
editors are all world-class experts in the field and bring great important chapters in the entire book and forms the foundation
insight to the book based on personal experience. Dr. Jean- for understanding much of the discussion in the subsequent
Nicolas Vauthey of University of Texas MD Anderson Cancer sections on physiology, molecular biology and immunology,
Center joins Dr. William Chapman of Washington University imaging, and perioperative management. Technical advances
in St. Louis in taking primary oversight of sections dealing in liver, biliary and pancreatic resection, transplantation, and
largely with hepatic resection and transplantation, reflecting the minimally invasive surgery are covered in detail, particularly in
substantial contributions they have made in these areas. My the new chapters “Advances in the Molecular Characterization
colleagues at Memorial Sloan Kettering Cancer are owed a of Liver Tumors” and “Stones in the Bile Duct: Minimally
great debt of gratitude for their great efforts. Drs. Ronald Invasive Surgical Approaches.” Recent advances in the imaging
DeMatteo and Michael D’Angelica once again brought their field are highlighted in Chapter 14, “Emerging Techniques in
expertise to bear in the sections on basic science/physiology and Diagnostic Imaging,” which should be an invaluable resource
biliary tract disease, respectively. Drs. Peter Allen and Richard to those in the field of radiology. We have again included a great
Kinh Gian Do joined the editorship with this edition and made deal of technical detail, which can be viewed with the expanded
substantive improvements in the sections on pancreatic disease list of videos on the Expert Consult website.
and radiology, respectively. In summary, the sixth edition attempts to include all aspects
The current edition reflects the ongoing and major changes of the anatomy, pathology diagnosis, and surgical and non-
in the field of hepatopancreatobiliary (HPB) surgery, including surgical treatments related to HPB disorders. We hope the work
minimally invasive resection techniques, molecular biology of is of value to a wide range of readers, from seasoned HPB
HPB malignancy, and advances in systemic and ablative thera- practitioners to surgical trainees and physicians in related dis-
pies. The organization of the book remains similar to the fifth ciplines. We have expanded our list of contributors in order to
edition; several new chapters have been added while others have ensure the broadest and most contemporary viewpoints possible.
been expanded. We have maintained the general format by I would like to again express sincere thanks to the co-editors
covering all surgical aspects of the management of HPB disor- who have collaborated with me in this project. We hope that the
ders, while the radiologic, endoscopic and other nonsurgical readers find this text to be a valuable resource for many years
approaches are presented in detail and highlighted when they to come.
represent the preferred therapy. As with past editions, contribu- W.R. Jarnagin, MD
tors were chosen largely based on their expertise and were asked New York, 2016
xxv
ACKNOWLEDGMENTS
The Editors are very grateful to our colleagues in surgery and other disciplines who have
contributed to the current edition. Without their enthusiastic support and insightful contri-
butions, often highlighting areas of controversy and differing opinion, this project would
never have been possible. Special thanks to our respective staffs in New York, St. Louis, and
Houston who have assisted in the preparation of this work. Finally, thanks and appreciation
are due to Dee Simpson, Michael Houston, and all of the staff of our esteemed publisher,
Elsevier, for their great support throughout the project.
xxvi
VIDEO CONTENTS
T. Peter Kingham, MD/Video Editor
CHAPTER 2 Surgical and Radiologic Anatomy of the Liver, Biliary • Pedicle Ligation, Right Pedicle
Tract, and Pancreas • Pedicle Ligation, Left Pedicle
• Basic Hepatic Anatomy: Three-Dimensional Model Outflow Control
• Right Hepatic Vein
CHAPTER 6 Liver Regeneration: Mechanisms and Clinical • Left and Middle Hepatic Veins
Relevance • Left Hepatic Vein
• Atrophy/Hyperthrophy Complex of the Liver Parenchymal Transection
• Clamp Crushing
CHAPTER 31 Bile Duct Exploration and Biliary-Enteric • Stapler
Anastomosis • Water Jet
• Exploration of Common Bile Duct Major Hepatic Resections
• Biliary Entreric Anastomosis—Running, Interrupted • Right Hemihepatectomy
• Extended Right Hepatectomy
CHAPTER 33 Cholecystitis • Left Hemihepatectomy
• Mirizzi Syndrome • Extended Left Hepatectomy
Hepatic Tumors
CHAPTER 49 Tumors of the Gallbladder • Enucleation of Benign Hepatic Tumors
• General Principles for Gallbladder and Hilar Cholangiocarcinoma Resection for Biliary Malignancy
• Biliary and Liver Resection for Gallbladder and Hilar Cancer • Left Hepatectomy with En Bloc Resection of the Bile Duct and
Caudate Lobe
CHAPTER 66 Techniques of Pancreatic Resection: • Right Hepatectomy with En Bloc Resection of the Bile Duct
Pancreaticoduodenectomy, Distal Pancreatectomy, Segmental Gallbladder Cancer
Pancreatectomy, Total Pancreatectomy, and Transduodenal • Porta Hepatitis Lymphadenectomy with Bile Duct Preservation for
Resection of the Papilla of Vater Incidental Gallbladder Cancer
• Pylorus-Preserving Pancreaticoduodenectomy for Adenocarinoma • Segmentectomy IVb/V for Gallbladder Cancer
of the Pancreas CHAPTER 104 Resection Technique for Live-Donor Transplantation
• Central Pancreatectomy and Pancreatiocogastrostomy for
Endocrine Tumor • Resection for Living Donor Transplantation: Donor Right
• Distal Pancreatectomy and Splenectomy Hepatectomy Including the MHV
• Ampullectomy for Ampullary Tumors CHAPTER 105 Minimally Invasive Techniques in Hepatic
CHAPTER 67 Minimally Invasive Pancreatic Resectional Resection
Techniques • Laparoscopic Right Hepatectomy
• Laparoscopic Distal Pancreatectomy • Laparoscopic Left Lateral Sectionectomy
• Laparoscopic Pancreaticoduodenectomy CHAPTER 107 Vascular Reconstruction Techniques
CHAPTER 74 Hydatid Disease of the Liver • Reconstruction During Hepatectomy
• Right Hepatectomy and Pericystectomy for Hydatid Disease CHAPTER 108B Segment-Oriented Anatomic Liver Resections
CHAPTER 99 Regional Chemotherapy for Liver Tumors Segmental Resections
• Robotic Hepatic Artery Pump Placement • Left Lateral Sectionectomy, Segments 2 and 3
• Right Posterior Sectionectomy, Segments 6 and 7
CHAPTER 103B Hepatic Resection for Benign Disease and for • Central Hepatectomy, Segments 4, 5, and 8
Liver and Biliary Tumors • Resection of Segment 3
TECHNIQUES OF HEPATIC RESECTION • Caudate Hepatectomy, Segment 1
Inflow Control CHAPTER 116 Orthotopic Liver Transplantation
• Extrahepatic Approach, Right Hepatic Artery and Portal Vein
• Extrahepatic Approach, Left Hepatic Artery and Portal Vein • Orthotopic Liver Transplantation
xxxii
INTRODUCTION
Hepatobiliary and pancreatic surgery:
historical perspective
Leslie H. Blumgart
A cursory examination of the Table of Contents of this book (Celsus, 1935). Celsus lived in the first century and described
reveals how some of the extraordinary achievements of modern symptoms attributable to liver disease. Gallstones were recog-
medical science and clinical practice have been applied to the nized in the embalming of mummies in ancient Egypt. In 1909
management of liver, biliary, and pancreatic disease. The near- a mummy with a preserved liver and a gallbladder containing
miraculous developments in our understanding of the molecu- 30 gallstones was presented to the Museum of the Royal College
lar nature of disease, combined with the wonders of modern of Surgeons in London. This mummy came from Deir-el-Bahn
imaging technology, advances in anesthesiology, and refine- at Thebes and was that of a priestess of the 21st dynasty around
ments in surgical technique, can be so absorbing as to blind a 1500 BCE. Elliot-Smith, an outstanding English anatomist and
contemporary physician to the rich history that lies behind Egyptologist, described the gallbladder as being large and con-
today’s practice. This introductory essay examines the key taining “many spherical calculi.” The specimen was destroyed
achievements of the past, on which the modern practice of liver, by German bombs during World War II, but the description of
biliary, and pancreatic surgery depends. it was accepted as evidence that the gallstones contained therein
were the earliest specimen of such calculi to have survived from
antiquity (1971). Gordon-Taylor, a noted surgical historian in
ANCIENT HISTORY UNTIL THE England, had likewise called attention to the terminal illness of
EIGHTEENTH CENTURY Alexander the Great at the age of 34 (323 BCE) as an example
of fatal biliary tract disease. The description of Alexander’s
“For the King of Babylon stood at the parting of the way, at illness, as recounted by Weigall (1933), suggests that Alexander
the head of the two ways, to use divination: he made his died of complications culminating in peritonitis.
arrows bright, he consulted with images, he looked in the Rhazes (850–923) and Avicenna (980–1037), two Persians,
liver.” wrote on general surgical topics and the nature of disease and
appreciated the gallbladder but lacked knowledge of the
EZEKIEL 21:21
common bile duct. Biliary fistulae were known to have formed
In ancient societies, the priest had the simultaneous duties of after the drainage of an abdominal wall abscess, and it was
divination, protection, and treatment of the ill. For individuals known that individuals with fistulae had a better prognosis than
with the responsibility of divining the future, the liver was of those who had an external communication with the intestines
central importance. As exposed by wounds sustained in combat, (Glenn, 1971).
at sacrificial offerings, and at incisions after death, the liver is Greek academic achievements surpassed that of all other
the most obvious organ in the abdomen and was observed to civilizations until the fifth century BCE. Hippocrates, widely
contain the most blood; because life and blood were perceived acclaimed in medicine then and since that time, recognized the
to be synonymous, the liver was considered the seat of the soul. seriousness of biliary tract disease as evident in the following
A tablet believed to be from the time of Hammurabi (about passage from The Genuine Works of Hippocrates (translated by
2000 BCE), now in the British Museum, names the various parts Adams in 1939): “[I]n a bilious fever, jaundice coming on with
of the liver and indicates the prognostic significance of each rigor before the seventh day carries off the fever, but if it occurs
(Jastrow, 1908). The liver was divided into about 50 portions without the fever, and not at the proper time, it is a fatal
for individual inspection in an effort to overlook as little as symptom.” He also noted that in the case of jaundice, “[I]f the
possible. Such hepatoscopy was widely in vogue over the fol- liver hardens it is a bad sign.”
lowing centuries and was practiced by the Etruscans, as evi- A few centuries after Aristotle recognized jaundice as an
denced by a bronze tablet in a museum at Piacenza depicting element of disease, Galen viewed biliary tract disease as a rec-
the liver, this being strikingly similar to the Babylonian clay ognized clinical entity to be treated successfully in part by diet.
tablet in the British Museum. Although Celsus had drawn attention to the gallbladder and
The Roman Celsus in his text De Medicina, translated by liver in the first century CE, in the second century, Galen was
W.G. Spencer in 1935, mentioned the liver and described its the most noted author of the Greco-Roman age. Galen named
anatomic location: “[T]he liver, which starts from the actual three main organs that govern the body: the heart, the source
partition under the praecordia on the right side, is concave of heat and the principal organ; the brain, the source of sensitiv-
within (that is on the inferior surface) and convex without; its ity for all parts; and the liver, as a principal part of the nutritive
projecting part rests lightly on the stomach and it is divided into organs (Green, 1951). Galen considered jaundice to be due to
four lobes. Outside its lower part, the gallbladder adheres to it” yellow bile flowing into the skin and recognized that although
1
2 Introduction Hepatobiliary and pancreatic surgery: historical perspective
jaundice could be the result of hepatic disease, it also could

arise when the liver was not involved at all. Galen’s teaching
persisted for centuries, and until the middle of the 17th century,
many disorders were described as alterations in the balance of
the main humors of the body brought about by hepatic dysfunc-
tion. As recorded by Rosner (1992), ancient Hebrew physicians
used pigeons to treat jaundice by placing the pigeon on the
patient’s umbilicus and believed the pigeon would cure the
patient: “The pigeon will draw all yellowness out.”
Many subsequent advances came from Italy about the time
of the Renaissance. Benivieni (1506) in Florence described a
series of autopsies in his patients, which was the first record of
special reference to biliary tract disease and its clinical mani-
festations. Two such examples are cited in the translation by
Singer and Long (1954) of Benivieni’s book, The Hidden Causes
of Disease:
“A case of stones found in the coat of the liver—a woman of

noble birth had been for long tormented by pain in the region
of the liver. She had consulted many physicians but could not
drive out the evil by any remedy … a few days afterward, the
disease took stronger hold and she departed this life. … I then
had her dead body cut open. There were found in the lower
part of the membrane around the liver a collection of small
stones varying in shape and color. Some were round, some
pointed, some square, according as position and chance had
determined, and they were also marked with reddish, blue and
white spots. These stones by their weight had caused the mem-
brane to hang down in a bag a palm’s length and two-fingers
wide. This we judge the cause of her death.”
The second quotation also suggests death from gallstones:

FIGURE 0.1 Andreas Vesalius, anatomist. (From De Humani Corporis
“Stone in the middle of the gallbladder–there died recently a Fabrica, 1543.)
noble lady, Diamantes, laid low by the pain due to stones. As
she had felt no harm from it before, the physicians decided to
cut open her body. A great many stones were found but none in 1543, and that Fernell, in his De Morbis Universilibus et Par-
in the urinary bladder as they had expected, but save for one, ticularibus (1588), proposed that the predisposing cause of gall-
which was in the gallbladder, was black and as big as a dried stone formation was stasis and observed that in jaundice the
chestnut. [A]ll were in the covering of the liver and made its feces become white and the urine dark, and that stones may be
coat hang down in a little bag. This we decided was the cause passed per anum. Falloppio, Vesalius, and Fernell all were
of her death and declared that a wise man would be well active in the first half of the 16th century and probably dis-
advised to make no definite pronouncements on obscure and cussed with one another their theories of the cause of gallstones
uncertain diseases.” and the changes in the liver with which gallstones were
associated.
Much of the next phase of development of knowledge was William Harvey (1578–1657), who also worked in Padua, is
centered around northeast Italy, particularly Padua, and many held by many to be the greatest of the contributors to the study
of the eponyms that we now use in surgery evolved from that of anatomy and physiology, in addition to his having established
period. The publication of Vesalius’s De Humani Corporis a clear understanding of the circulation. There is little doubt
Fabrica in 1543 (Fig. 0.1), and Harvey’s De Mortu Cordis 100 that Harvey also gave thought to the liver and its relation to the
years later, marked the emergence of a new scientific spirit in circulation and to the heart in particular. Harvey’s student and
anatomy and physiology. As was the custom of the time, Vesa- younger contemporary, Francis Glisson, investigated the struc-
lius often recorded descriptions of findings of dissections on ture and function of the liver extensively. His book Anatomia
individuals who had recently died, and expressed an opinion Hepatis was published in 1654 and is the first major modern
on the cause of death. Rains (1964), in his treatise Gallstones— work on hepatology. Glisson gave a clear description of hepatic
Causes and Treatment, stated: “Vesalius found [that he had] a anatomy, especially of the hepatic capsule and of the investment
hemoperitoneum coming from an abscess, which had eroded of the hepatic artery, portal vein, and bile duct. He described
the portal vein. The gallbladder was yellow and contained 18 the fibrous framework of the liver and illustrated hepatic vas-
calculi. Very light, of a triangular shape with even edges and cular and biliary anatomy on the basis of cast and injection
surfaces everywhere, green by color somewhat blackish. The studies (Fig. 0.2). Glisson was the first to mention a sphincteric
spleen was very large.” Similarly, Rains recounted that Fallop- mechanism around the orifice of the common bile duct (Boyden,
pio described stones in the gallbladder and common bile duct 1936). He also deduced the flow of blood through the portal
Introduction Hepatobiliary and pancreatic surgery: historical perspective 3
FIGURE 0.2 Illustration of the vasculature of the liver. (From Glisson F,

1654, Anatomia Hepatis.)
veins traversing the capillaries into the vena cava at a time when
no microscopic studies of the liver had been done (Walker,
1966). Some of his illustrations are remarkably similar to those
of Couinaud (1954, 1957, 1981) and to the images displayed
today in three-dimensionally constructed computed tomogra-
phy (CT) scans. Glisson was one of the great clinical physicians
of all time, yet his name is related to an anatomic structure of
FIGURE 0.3 Giovanni Battista Morgagni (1682–1771), physician.
limited importance.
In the century that followed the time of Harvey, great activ-
ity in publication among individuals engaged in medicine tumor, having arisen in the epigastrium, and being opened,
occurred. As today, senior professors often held opposing opin- either by art or spontaneously discharged cystic calculi at its
ions and expressed their feelings based on fact or fancy with aperture … the first was cured; the second had a fistula left,
equal fervor. In an attempt to select the factual, Morgagni (Fig. by which a thin and chylous kind of liquor, but of a yellow
0.3), senior professor of anatomy and president of the Univer- color, distilled; and the third had an ulcer remain, which with
sity of Padua, published in 1761 an analysis of disease (transits sanies discharged bilious calculi at times.
lated by Benjamin Alexander in 1960) under the title Seats and
Causes of Disease, among which are those of the liver and biliary Morgagni, as is evident in his text, presented the matter of
tract. In referring to gallstones, Morgagni analyzed the distribu- biliary tract disease in logical sequence. He considered the
tion of stones in male and female patients, including age, inci- incidence and the population affected in the different decades
dence, and treatment: of the life span. His descriptions of the possible mechanisms by
which calculi might be formed remind one of much of the writ-
[A]s to the lithotomy, which has also lately been thought of, ings today. Finally, he considered therapy under conservative
in the gallbladder, do not be surprised that I made no mention medical management and warned about attempting to treat the
of it above. For in the first place, the pains which were excited condition by operation (Morgagni, 1960).
by gallstones, that were endeavoring to discharge themselves, As outlined by Wood (1979), numerous eponyms currently
are not only brought on by those which come from the cysts used pertain to the names of these early physicians. Johann
but also by those that come from the hepatic duct. In the second Wirsung (1600–1643) also studied in Padua and was the first
place, those cystic stones which are the largest, and on account to dissect the human pancreatic duct and to describe it in a
of which lithotomy seems, to some persons, to be chiefly desir- 1642 letter to Riola, professor of anatomy and botany at the
able, neither endeavor to disengage themselves nor create any University of Paris. Wirsung was subsequently murdered by a
great uneasiness; or at least for the most part. Dalmatian physician, in a dispute probably related to who had
described the duct first (Major, 1954; Morgenstern, 1965).
This is a remarkable analysis recognizing the presence of stones Abraham Vater (1684–1751) was the first to describe the
in the common bile duct and the relatively innocuous nature papilla of the duodenum. In 1720 he wrote that “those double
of the large stone. A further quotation is as follows: ducts (bile and pancreatic ducts) … come together in single
combination” (Boyden, 1936). He described not an ampulla
Last of all, although there were no danger in cutting, can you but an elevation of the mucosa of the duodenum and described
suppose there would be of no great difficulty in healing the the first reported case of an ampulla with two orifices. Likewise,
wound? We have, before our eyes, examples of three women, the duct of Santorini takes its name from the Venetian Giovanni
one of Bologna, of Frankfurt, and of Göttingen, in whom a Domenico Santorini (1681–1737), a brilliant anatomist and
one of the most exact and careful dissectors of his day. While
Vater was describing the tubercle at the confluence of the pan-
creatic and bile ducts, Santorini was relating the first detailed
observation of the orifice of the two ducts (Boyden, 1936). In
observations printed posthumously, Santorini noted a second
pancreatic duct of normal occurrence and named the upper one
the superior pancreatic duct and the lower one the main pan-
creatic duct. It was not until Oddi in 1864, working in Bologna,
rediscovered “Glisson’s sphincter” and did studies in dogs that
the holding quality of the sphincter of the outlet of the choledo-
chus was recognized (Boyden, 1936).
The work of Morgagni was seminal in the understanding of
liver and biliary disease. It started the movement toward ana-
tomic and pathologic appreciation. Belief in the humoral causa-
tion of disease became untenable, and scientific medicine was
launched. The end of the 17th century and the beginning of
the 18th could be said to mark the initiation of modern medical
science.
EIGHTEENTH CENTURY TO MODERN TIMES

The period from 1700 to the present day can be considered in
two phases: the era before the discovery of general anesthesia,
antisepsis, and x-rays in the middle of the 19th century and the
period that followed, bringing important discoveries in blood
transfusion, antibiotics, and immunology. These major devel-
opments have been followed by the birth of the electronic age,
the flowering of radiologic imaging, and the increasing power
of genetics and gene manipulation. The passage of surgery
through these milestones is fascinating to observe, and the
enormous strides in the maturation of hepatobiliary and pan-
creatic surgery to a recognized specialty are an important part
of this story. FIGURE 0.4 W.K. Roentgen (1845–1923). He discovered x-rays on
November 8, 1895.
IMAGING IN HEPATOBILIARY AND

PANCREATIC SURGERY biliary and pancreatic radiology and approaches to common
bile duct stones. The 1970s saw not only the introduction of
“Where there is no vision, the people perish. …” endoscopic approaches to the biliary tract but also the develop-
ment of good methods for percutaneous transhepatic cholangi-
Proverbs 29:18
ography (Ohto & Tsuchiya, 1969; Okuda et al, 1974), CT of
After the discovery of x-rays by Roentgen (Fig. 0.4) in 1895, the liver (Grossman et al, 1977), and the use of ultrasound
there have been continual important and extraordinary devel- (Bryan et al, 1977) in liver and biliary surgery. Magnetic reso-
opments in radiology. A contemporary surgeon who examines nance (MR) axial imaging (Damadian, 1971; Lauterbur, 1973)
images to determine the cause of disease, sometimes before was conceived and has led to the development of magnetic
taking a history or examining the patient, can hardly contem- resonance imaging cholangiopancreatography (MRCP).
plate a time when precise imaging was not available. A land- Not only were endoscopic and transhepatic approaches to
mark development in hepatobiliary imaging occurred when, stones now possible, endoscopic and percutaneous transhepatic
after experimenting with various iodine compounds, Graham, intubation of the biliary tree for the relief of jaundice and for
a North American surgeon, developed oral cholecystography dilation of strictures became a reality. Arteriography developed
(Graham & Cole, 1924). Although biliary calculi had been to a fine degree, so that good hepatic arteriography and por-
observed using x-rays alone (Buxbaum, 1898), the problem of tography were developed (Hemingway & Allison, 1988), which
detecting radiotransparent calculi was evident, and the develop- inevitably led to the development of techniques for hepatic
ment of oral cholecystography marked an important turning artery embolization in the management of liver tumors and of
point. Postoperative cholangiography was soon developed by hemobilia (Allison et al, 1977).
Mirizzi (1932) in Argentina. Intraoperative cholangiography Surgeons today have experienced their world transformed
(Mirizzi, 1937) and choledochoscopy (Bakes, 1923; McIver, by imaging; it may be said that the extraordinary developments
1941) also were developed. in the radiology of the liver, biliary tract, and pancreas have
In the 1970s, endoscopic cholangiopancreatography enabled the rapid progress in the management of the diseases
(Blumgart et al, 1972; Cotton et al, 1972a, 1972b; Demling & of these organs. The extraordinary advances in image-guidance
Classen, 1970; McCune et al, 1968; Oi et al, 1970) and endo- technology have progressed to a level that CT, ultrasound, and
scopic papillotomy (Classen & Demling, 1974) revolutionized angiography now allow percutaneous image-guided methods
for tumor ablation, arterial embolization, alcohol injection, and

injection of other substances directly into tumors. A host of
other techniques are currently in development. It is of note that
advances in imaging have enabled the development of stereo-
tactic approaches in the management of liver tumors.
SURGERY OF THE BILIARY TRACT

AND PANCREAS
Biliary Tract
The early part of the 18th century saw rapid development in
the understanding of biliary disease. The ideas of Morgagni
rapidly spread and were a firm basis for the changes to come.
Studies regarding the structure, origin, and functions of bile
were initiated in the latter half of the 18th century and culmi-
nated in Wieland’s clarification of the structure of the various
bile acids and Windaus’s demonstration of the relationship of
bile acids to steroids. For his contribution, Windaus was
awarded a Nobel prize in 1928. Boerhaave (1668–1738) estab-
lished a wide surgical reputation at the University of Leiden in
Holland and worked extensively on diseases of the liver and
biliary tract. In 1743 Jean Louis Petit presented to the Paris
Academy of Surgery a paper entitled “Considerations Concerning
Tumors Produced by Retained Bile in the Gallbladder of the Liver,
Sometimes Taken for Liver Abscesses.” He introduced the term
biliary colic (colique hépatique) and established the clinical treat-
ment of it and of obstructive cholecystitis. In cases in which the
gallbladder adhered firmly to the abdominal wall, Petit advised
that it be punctured and opened to remove calculi, and he
applied this procedure successfully in the case of one patient;
this preceded Morgagni’s publication of 1761 referred to earlier.
Biliary surgery began in Indiana on July 15, 1867, when
John Bobbs operated on a woman who had a large tumor that
he believed to be an ovarian cyst. To his amazement, when the
abdomen was opened, he found an enormous gallbladder filled
with stones. He opened it and extracted the calculi, sutured it
carefully, and placed it back in the abdomen (Bobbs, 1868). It FIGURE 0.5 Theodore Kocher (1841–1917), surgeon.
was nearly a decade before a similar cholecystostomy was per-
formed. As is happenstance, this was accomplished almost
simultaneously by a Swiss, Theodore Kocher (Fig. 0.5); a duodenum as described by the great master (Kocher, 1903).
North American, Marion Sims; and an Englishman, Lawson Kocher originally described the maneuver for use in gastric
Tait. All three surgeons planned operations in which the gall- surgery, but similar to so many other firsts in surgery at that
bladder was affixed to the abdominal wall to allow extraction time, the maneuver was first described by Jourdan (1895) and
of stones and pus and to leave it open to the exterior, so that was first performed in biliary surgery by Vautrin (1896). Perhaps
peritonitis as a result of maneuvers within the abdomen could the maneuver should be referred to as “Vautrin’s maneuver”
be avoided. rather than Kocher’s. Tait, the great British surgeon, is given
Sims worked on both sides of the Atlantic, and in Paris he credit for performing the first cholecystostomy for gallbladder
operated on a patient with long-standing jaundice and a tumor lithiasis in one stage. The patient, a 40-year-old woman, sur-
in the right hypochondrium. With antiseptic technique, he vived, and by 1884, he had performed the procedure in 14 cases
opened the gallbladder and extracted 60 calculi, and then with only one death (Tait, 1885).
sutured the gallbladder to the abdominal wall; this was perhaps The first elective cholecystectomy was done by Langenbuch
the first elective surgical procedure for obstructive jaundice (Fig. 0.6). While others were pursuing the construction of
(Sims, 1878). In the same year, in Bern, Switzerland, Kocher biliary fistulae and direct removal of gallstones, Langenbuch
performed a cholecystostomy in two stages (Glenn, 1971). In observed that because stones were known to recur, others had
the first stage, he packed the wound with gauze to the bottom “busied themselves with the product of the disease, not the
of the gallbladder, and 8 days later, he emptied the residual disease itself” (Langenbuch, 1882). As he was later to recount,
stones from the gallbladder. Incidentally, Kocher, who won the two thoughts kept occurring to him: first, that in animal experi-
Nobel prize in medicine in 1909 for his work on the physiology ments, Zambeccari in 1630 and Teckoff in 1667 had performed
and surgery of the thyroid gland, also described sphincterotomy, cholecystostomy and cholecystectomy in dogs and had shown
or internal choledochoduodenostomy. His name is remembered that the gallbladder was not essential to life (Langenbuch,
by every biliary surgeon who performs mobilization of the 1896); second, that his medical colleagues believed that the
video laparoscopic cholecystectomy. He did not publish his

experience, but the news spread rapidly, and Dubois in Paris
published the first series of laparoscopic cholecystectomies
(Dubois et al, 1989, 1990). Perissat, working in Bordeaux,
further developed the laparoscopic approach and introduced it
to the United States in 1989 (Perissat et al, 1990). The proce-
dure has since been extended, so that laparoscopic exploration
of the common bile duct is now possible and is carried out
routinely. Today, most cholecystectomies are done using lapa-
roscopic techniques, such that the modern surgical trainee
often has little or no experience in open cholecystectomy. More
recently, cholecystectomy has been done via minimally invasive
techniques using robotics (Marescaux et al, 1998; Satava,
1999) to manipulate the instruments.
Not long after the performance of the first cholecystectomy,
attempts were made to remove stones from the bile ducts. In
1898 Thornton performed the first removal of a stone from the
common bile duct. A year later Courvoisier operated success-
fully on another case of choledocholithiasis and published his
well-known monograph on the pathology and surgery of the
biliary ducts. He also enunciated the law that was to bear his
name, which established that in patients with jaundice, if the
gallbladder is not distended, the case is more likely to be one
of stones. The operative procedure for exploration of the
common bile duct for choledocholithiasis was not popularized,
however, because of the risk of peritoneal infiltration by bile.
In 1897 Kehr (Fig. 0.7) performed exploration of the
common bile duct and placed a rubber tube in the common
bile duct through the cystic duct; this was the first systematic
use of biliary intubation. Kehr’s name is properly associated
with the development of biliary intubation; Kehr (1912) and
Quenu and Duval (1908) were able to extract stones along the
tunnels created by the drainage tubes. These were the precur-
sors for techniques later developed by Mondet (1962) and
Mazzariello (1966, 1974). Kehr had developed numerous com-
binations of drainage in patients with biliary stones, and many
FIGURE 0.6 Karl Langenbuch (1846–1901), surgeon.
other surgeons rapidly developed choledochotomy without
suture and using tube drainage.
gallbladder itself gave rise to stones. He developed the tech- Surgery of the bile ducts rapidly disseminated across Europe,
nique for cholecystectomy over several years of cadaver dissec- England, and the United States, and in 1912, Kehr developed
tion and performed the operation at the Lazarus hospital in what came to be known as the T-tube. Not only was choledo-
Berlin on July 15, 1882. The patient had experienced biliary chotomy simplified, but biliary tract repair was done over these
colic for 16 years and was addicted to morphine. He was afe- tubes. Kehr became justly famous for his introduction of biliary
brile the day after the operation, had little pain, and was intubation and was probably the most outstanding biliary
smoking a cigar; the patient was walking at 12 days and left the surgeon of his day. In 1913 he published a treatise entitled
hospital 6 weeks later, pain free and gaining weight (Traverso, Surgery of the Biliary Tract, which for more than 40 years was
1976). Report of this case (Langenbuch, 1882) led to a con- the most respected text on the subject. Kehr (1908a, 1908b)
troversy over cholecystostomy as championed by Tait. Langen- described the resection of cancerous gallbladders, including
buch’s operation was the new cholecystectomy. hepatic resection, and he resected hepatic tumors and aneu-
In 1886 Gaston had tabulated 33 cholecystostomy opera- rysms of the hepatic artery. He also performed the first hepati-
tions with a mortality rate of 27%, compared with 8 cholecys- coenteric anastomosis.
tectomies (5 by Langenbuch) with one death recorded, a Before the development of Kehr’s tubes, choledochotomies
mortality rate of 12%. By 1890, 47 cholecystectomies had been often became biliary fistulae, either because of residual supra-
done by 27 surgeons, and in 1897, the number had increased ampullary stones, or because the surgeon had inadvertently
to nearly 100 operations with a mortality rate of less than 20% opened the bile duct proximal to a cancer. German and Aus-
(Gaston, 1897). trian surgeons were the first to perform supraduodenal choledo-
The most important recent advance in surgery of the gall- choduodenostomy (Riedel, 1892; Sprengel, 1891). Sprengel’s
bladder came 100 years after Langenbuch’s first cholecystec- operation described a side-to-side choledochoduodenostomy,
tomy. Laparoscopic cholecystectomy was conceived and first and it subsequently became popular in Europe and the United
performed in Germany by Muhe in 1985. Between 1985 and States (Madden et al, 1965). Cholecystenterostomy also was
1987, he performed 94 laparoscopic cholecystectomies (Muhe, developed initially by von Winiwarter (1882) and was used later
1986, 1991). Mouret, in Lyon, France, performed the first by many surgeons, including Oddi (1888).
hepaticojejunostomy at the French Congress of Surgery in

1908. Dahl (1909), in what he called “a new operation on the
bile ducts,” advocated the Roux anastomosis in biliary surgery.
With the advent of cholecystectomy and choledochostomy
came the inevitable sequelae of residual bile duct stones and
iatrogenic lesions. Initially, these were treated by various forms
of intubation, similar to the techniques advocated by Kehr, but
many leading surgeons, such as Moynihan and Mayo, used
hepaticoduodenal anastomoses (Estefan et al, 1977). Dahl
(1909) used the Roux-en-Y loop, but more recently, other
authors (Madden et al, 1965; Schein & Gliedman, 1981) have
preferred choledochoduodenostomy as described previously.
Operative biliary drainage in upper biliary tract cancer and
in patients with severe scarring in the porta hepatis is always
difficult. Kehr (1913) successfully performed three operations
in which he fixed a jejunal loop to a cut section of a hepatos-
tomy wound; similar procedures were done by others, as cited
by Praderi (1982). Longmire and Sanford (1948) performed a
similar operation, resecting the end of the left lobe of the liver
and anastomosing it to a jejunal loop. This technique, although
popular for a short time, was complicated by stenosis and has
fallen out of favor.
Goetze (1951) developed a procedure of transanastomotic
drainage with a catheter that went through the hepatic paren-
chyma, the anastomosis, and the jejunal loop. Goetze’s contri-
bution was apparently forgotten but was reinvented and
popularized by others, including Praderi in 1962 (Praderi,
1982). Praderi argued strongly that Goetze’s procedure was the
answer to difficult operations in the high biliary tract, but he
conceded that the emergence of transhepatic percutaneous
cholangiography (Okuda et al, 1974) allowed the development
of transhepatic percutaneous methods for intubation and dila-
tion of the biliary tract, which are now common procedures
FIGURE 0.7 Hans Kehr (1862–1916), surgeon. throughout the world.
The greatest advances in techniques for repair of biliary
injuries came from two sources. At the Lahey clinic in the
Transduodenal surgery was not long in developing. In 1895 United States, three generations of famous surgeons, Lahey,
Kocher wrote an article on internal choledochoduodenostomy Cattell, and Warren, perfected the reconstruction of the
to remove supraampullary choledochal calculi; by 1899 he had common bile duct as an immediate or delayed procedure and
performed the operation 20 times. MacBurney (1898) pub- splinted their anastomoses with a variety of tubes (Estefan et al,
lished his experience with duodenostomy and papillotomy in 1977). The results of these techniques were unsatisfactory, with
patients with impacted periampullary calculi. These early pro- numerous recurrent strictures. However, after the detailed
cedures of a choledochotomy and choledochoenteric anasto- study of the anatomy by Couinaud in France (1954), Hepp and
mosis for the treatment of jaundice and of stones in the biliary Couinaud (1956) and Soupault and Couinaud (1957) devel-
tract are still used today, with frequent application of the same oped techniques for direct biliary-enteric anastomosis, either to
principles but with aid of the endoscope. the left hepatic duct or to the segment III duct of the left liver.
At about the same time, surgeons began to operate on These surgical techniques have become widely accepted and
cancer of the papilla. In Baltimore in 1900, Halsted resected a are used throughout the world (Bismuth & Corlette, 1956;
portion of the duodenum that included a tumor and reim- Bismuth et al, 1978; Voyles & Blumgart, 1982; Warren &
planted the common duct, at the same time performing a Jefferson, 1973).
cholecystostomy. He then reoperated to remove the gallblad-
der. Mayo (1901) reported an operation on a 49-year-old man Pancreas
with papillary cancer. Mayo opened the duodenum, removed Like liver surgery, pancreatic surgery developed largely as a
the tumor, and carried out a choledochoduodenostomy. This result of responding to wounds inflicted in wars. Claessen
was the first transduodenal ampullectomy. Kausch (1912) in (1842), Ancelet (1866), Da Costa (1858), and Nimier (1893)
Germany and Hartmann (1923) in France gave accounts of documented the early development of pancreatic surgery. In
resections of ampullary tumors. In Lausanne, Switzerland, 1923 Hartmann wrote an extensive review of the French,
Roux, a disciple of Kocher, had described preparation of a German, and English literature. Early efforts at elective pancre-
“jejunal loop” for use in gastric surgery, and it soon became atic surgery largely revolved around drainage of cysts (Thiersch,
an adjunct in biliary surgery (Roux, 1897). This Roux-en-Y 1881). In 1883 Gussenbauer marsupialized a pancreatic pseu-
procedure was soon used by Monprofit (1904) in the per docyst, and the patient survived. Other surgeons, including
formance of cholecystojejunostomy, and he proposed it for Senn (1886), soon performed similar procedures.
Direct anastomosis of the pancreatic duct to the gastroin- organ, without touching the duodenum. Codivilla of Italy is
testinal tract followed in the early part of the 20th century, reported to have resected the duodenum and head of the pan-
when Coffey (1909) performed an anastomosis of the tail of the creas and performed a cholecystojejunostomy in 1898, but the
pancreas to the small bowel. Ombredanne (1911) anastomosed patient died in the postoperative period (Codivilla, 1898). The
a pancreatic cyst to the duodenum, and Jedlicka (1923) anas- first surgeon to perform cephalic pancreaticoduodenectomy
tomosed a cyst to the posterior wall of the stomach. Chester- successfully was Kausch, a professor of surgery in Breslau and
man (1943) performed a cystojejunostomy, and König (1946) later in Berlin. On June 15, 1909, he performed a cholecysto-
carried out the same operation to a Roux-en-Y loop of jejunum. jejunostomy on a jaundiced 49-year-old man; on August 21,
Operations for pancreatic tumors were being done at about 1909, Kausch reoperated, performing a posterolateral gastro-
the same time. Ruggi (1890) reported a resection of a large enterostomy and resection of the head of the pancreas with the
lesion of the tail of the pancreas, and Briggs (1890) performed tumor, the pylorus, and the first and second part of the duode-
a similar operation. Biondi (1897) reported resection of a num. He anastomosed the third portion of the duodenum to
tumor arising from the inferior part of the head of the pancreas, the pancreatic stump. Kausch (1912) published a report enti-
and the patient was still alive 18 months later. tled “Cancer of the Duodenal Papilla and Its Radical Treat-
Whipple (Fig. 0.8) and associates (1935) published a tech- ment.” He gathered in that publication all reports that included
nique for cephalic duodenopancreatectomy, done in two stages, excisions of the papilla. Tenani (1922) reproduced this opera-
for cancer of the ampulla of Vater. In the first operation, they tion with success, also in two stages but with one difference:
performed a cholecystogastrostomy and gastroenterostomy. In Tenani performed a choledochojejunostomy, and the patient
the second and subsequent procedure, they resected the head survived 3 years. Kausch and Tenani, although cited by Whipple
of the pancreas with a portion of the duodenum without anas- (1941), are generally ignored in the English and American
tomosis to the stump of the pancreas, which they sutured. literature.
Whipple (1941) subsequently performed this operation in one Pancreatic resection was extended to manage cancers of the
stage and reported 41 cases, with a mortality rate of 27%. common bile duct and of the duodenum. Recognition of endo-
Eventually the technique of this operation was perfected. crine tumors of the pancreas later led to operations for these
Although English-speaking surgeons throughout the world con- conditions. Wilder and colleagues (1927) reported the first case
tinue to call this operation “Whipple’s operation,” the proce- of resection of an insulinoma arising from the islet cells of the
dure had been done many years earlier. pancreas. Mayo operated on a patient in whom he found a
Sauve (1908) reviewed the literature on pancreatectomy and tumor of the pancreas with metastases to the liver, and an
reported that several surgeons had resected small tumors from extract from one of the metastases produced an insulin-type
the head of the pancreas, and larger ones from the body of the reaction when injected into a rabbit. Graham and Hartmann
(1934) reported subtotal pancreatectomy for hypoglycemia,
and total pancreatectomy was performed by Priestley and col-
leagues (1944) for a patient in whom there was proven hyper-
insulinism, but in whom no tumor of the pancreas was evident.
The patient was cured by the operation, and the small tumor
causing the syndrome was discovered during the pathology
examination.
Later, Fallis and Szilagyi (1948) performed total pancreatec-
tomy for cancer in the hope that removal of the whole gland
would reduce rates of morbidity and mortality and perhaps lead
to better results. ReMine and colleagues (1970) and Brooks
and Culebras (1976) adopted this approach, but early sugges-
tions regarding efficacy were not fulfilled, and total pancreatec-
tomy for pancreatic cancer has largely been abandoned. Fortner
(1973) described the even more extensive procedure of regional
pancreatectomy. The operation included an extensive total
pancreatectomy and resection of the pancreatic segment of the
portal vein—and sometimes the hepatic and superior mesen-
teric arteries, if these vessels were invaded by tumor—together
with subtotal gastrectomy and regional lymph node dissection.
Fortner’s results have not been reproduced by others.
Surgery for pancreatitis also developed over the same period,
and the severity of acute pancreatitis was recognized. Ockinczyc
(1933) revealed the frustration of the times in the surgical
management of pancreatitis. He advocated “the use of drainage
of the pancreas and hope.” Because the mortality rate of
emergency surgery reached 78%, the conservative approach to
acute pancreatitis was advocated. Nevertheless, some patients
with acute pancreatitis still came to operation because of the
necessity to treat gallstone-related pancreatitis or to manage
complications, such as abscess and pseudocyst (Cattell &
FIGURE 0.8 Allen O. Whipple (1881–1963), surgeon. Warren, 1953).
The development of intensive care, antibiotics, and better which Prometheus’s liver grew back nightly after the eagle’s
metabolic management of patients with acute pancreatitis led daily and apparently bloodless “resections.” This oft-cited
to improvements in outcome, although the death rates remained myth does not necessarily mean that the ancient Greeks knew
unacceptably high. Ranson and colleagues (1976) and Imrie about liver regeneration; however, more recent evidence pro-
(1978) made important contributions in that they defined the vides dramatic proof of the liver’s power to grow back rapidly
evaluation of the severity of an attack. The surgical treatments (Blumgart et al, 1971).
of abscess and peritonitis were made systematic. The role of Knowledge of the liver’s lobar and segmental structures,
gallstones in acute pancreatitis was defined, particularly by functional reserves, and capacity to regenerate, as well as tech-
Acosta and Ledesma (1974), and endoscopic papillotomy to niques to prevent intraoperative hemorrhage are essential for
extract calculi impacted at the papilla of Vater was also described. major surgical resection. This knowledge also is relevant to
Surgical procedures aimed at the treatment of chronic pan- transplantation, particularly to split-liver homograft transplan-
creatitis had been developing since the beginning of the 20th tation and in the use of living, related donors.
century. As mentioned previously, Coffey (1909) had anasto- Early descriptions of operations on the liver usually con-
mosed the tail of the pancreas to a loop of jejunum, and in cerned complete or partial avulsion of some portion that was
1953, Link reported external drainage of a pancreatic duct in protruding externally as a result of abdominal trauma. Berta in
a patient who lived for a prolonged period but expelled calculi 1716 amputated a portion of liver protruding from the abdomen
periodically from the fistula tract. Roget (1958) described 50 after a wound, and Von Bruns in 1870 (Beck, 1902) removed
cases treated this way, but only a few patients were cured by a lacerated part of the liver of a soldier wounded in battle.
drainage. Doubilet and Mulholland (1948) attempted sphinc- Despite these anecdotal descriptions, true hepatobiliary opera-
terotomy in the treatment of pancreatitis. As a result of the tions were not possible until the advent of anesthesia (Fig. 0.9)
success of cystojejunostomy in the treatment of pancreatic and antisepsis (Lister, 1867a, 1867b; Fig. 0.10). Although
pseudocyst mentioned earlier, Cattell in 1947 used a side-to- Warvi (1945) recounted that Couzins performed a liver resec-
side anastomosis between a loop of jejunum and the pancreatic tion in 1874, Langenbuch (1888) recorded the first planned
duct in an attempt to relieve pain in a patient with cancer of hepatic resection in Germany. In the United States, Tiffany
the head of the pancreas. Subsequently, Longmire and associ- (1890) resected a liver tumor, and Lucke (1891) reported a
ates (1956) carried out end-to-end pancreaticojejunostomy to liver resection for cancer. Keen (1899), in the United States,
a Roux-en-Y loop of jejunum; however, this procedure proved reported 76 liver resections, of which 37 were for benign or
unsuccessful. Du Val (1957) performed a similar operation in malignant tumors.
an attempt to drain the pancreatic duct, but this too was About the time of these surgical developments, Rex (1888)
unsuccessful. and later Cantlie (1897) did detailed anatomic studies of the
Leger and Brehand (1956) and von Puestow and Gillespy liver and its intrahepatic architecture. These studies established
(1958) performed pancreatic ductal drainage in chronic pan- the lobar and segmental structure of the liver and of the Glis-
creatitis patients for the relief of pain. Longitudinal opening of sonian sheath enclosing structures that enter or leave the organ
the pancreactic duct allowed better decompression after anas- at the porta hepatis. They delineated the planes within the
tomosis to a jejunal loop. Mercadier (1964) performed a similar parenchyma of the organ that were relatively devoid of major
operation without the need for splenectomy, and further techni- blood vessels and bile ducts. These descriptions would make
cal variations were introduced more recently by Prinz and possible controlled hepatic resection.
Greenlee in 1981 and by Frey and Smith in 1987. Mallet-Guy
(1952) and Mercadier (1964) performed resection of the left
pancreas for chronic pancreatitis. This approach subsequently
was extended, so that only a small portion of the head of the
pancreas remained. A more modern approach aimed at con-
serving pancreatic parenchyma has since been developed (Beger
et al, 1985) and was further promoted by Buchler and col-
leagues (Di Sebastiano et al, 2007).
Liver Surgery
Liver surgery has evolved from being almost nonexistent to
comprising a repertoire of operations that can safely remove
nearly any amount of liver tissue (Chakravorty & Wanebo,
1987; Fortner & Blumgart, 2001). These operations are now
performed at numerous hospitals and medical centers through-
out the world. Courageous surgeons were enabled by extraor-
dinary developments in anesthesiology, blood transfusion,
infectious disease, and radiologic imaging.
Features of the anatomy and physiology of the liver that
allow major resection were known in ancient times. The struc-
tural arrangement of the liver into lobes was apparent to indi-
viduals preparing animals for food or for ceremony or to those
preparing humans for mummification. Two other properties FIGURE 0.9 The First Operation with Ether, by Robert Cutler Hinckley,
of the liver, functional reserve and rapid regeneration, were 1893. (Courtesy of Boston Medical Library in the Francis A. Countway
suggested by the early Greek mythology of Prometheus, in Library of Medicine.)
(1972a, 1972b). Although almost totally abandoned in recent

times, such clamps still are occasionally employed. Garre
(1907) mentioned attempts to provide perforated plates to be
used on either side of a wound of the liver, through which
sutures could be passed to allow compression of liver tissue and
bleeding vessels. This technique also was conceived more
recently for the management of bleeding after elective liver
resection (Wood et al, 1976) and for the control of bleeding
from liver injury (Berne & Donovan, 2000).
Garre (1907) also emphasized the use of packing bleeding
hepatic wounds, and he advocated that packing ought to be
considered more frequently in cases of injury; this, too, was
prophetic. Having passed through the phases of vascular liga-
tion and hepatic resection for wounds of the liver, modern
surgeons now rely much more frequently on packing to control
oozing and hemorrhage for ragged and exposed liver lacerations
(Berne & Donovan, 2000). The use of such packing recognizes,
but does not specifically state, that major hemorrhage in the
liver does not usually occur from the cut surface of the paren-
chyma but rather occurs from the main hepatic veins or lacera-
tions to the vena cava.
In a seminal contribution in 1908, Pringle described a
method of temporarily compressing the portal inflow vessels so
as to reduce liver bleeding, but all eight patients he described
died during surgery or shortly thereafter. Later, Pringle
described the method as being uniformly successful in animals
FIGURE 0.10 Joseph Lister (1827–1927), surgeon. and used the technique successfully on a patient. This tech-
nique has been used ever since to control hepatic inflow, and
although initially used to occlude the inflow for periods of 1
Surgeons quickly learned to fear the liver’s friability and hour or more, it is now employed in an intermittent fashion.
capacity for bleeding and its propensity for biliary leakage after The early understanding of liver structure and function and
operation. Elliot (1897) wrote that the liver is “so friable, so of the first operations on the liver in the latter half of the 19th
full of gaping vessels and so evidently incapable of being sutured century set the scene for the rapid advances of the 20th century.
that it seemed impossible to successfully manage large wounds Not only did anatomic description reach a high level of accu-
of its substance.” Kousnetzoff and Pensky (1896) reported, racy relevant to the operating surgeon, advances in operative
however, passing ligatures in the liver substance at a sufficient technique and extraordinary developments in diagnostic
distance from the margins of the wound to ensure that they imaging as outlined previously allowed the flowering of hepa-
would not slip, and reported that, by pulling these up tightly, tobiliary surgery as a specialty.
it was possible to allow them to cut into the liver parenchyma
and compress the blood vessels. In writing about surgical Major Hepatic Resection
approaches for parenchymal transection of the liver and the Wendell (1911) was the first to perform a successful right
arrest of bleeding during operations on the liver, Garre (1907) hepatic lobar resection using hilar ligation. Nevertheless, his
paid respect to Kousnetzoff and Pensky’s work, in which they achievement was lost to his colleagues for many years, and it
had essentially shown that the vessels in human liver are no less was not until the detailed anatomic studies of Hjortsjo in 1951,
resistant than arteries and veins of similar caliber in other parts Healey and Schroy in 1953, and Couinaud in 1954 that the
of the body, and that they were suitable for ligature. These basic concept of segmental anatomy as originally described by Rex
techniques for suture of the liver substance and ligation of and Cantlie was finally accepted. Couinaud’s (Fig. 0.11)
vessels as a means of controlling hemorrhage have persisted to seminal work in 1954 simplified the understanding of segmen-
modern times and find recent application in the control of the tal liver anatomy and made it easily applicable by numbering
pedicles of the liver within its parenchyma, as described by Ton the liver segments I through VIII. Goldsmith and Woodburne
(1979), and more recently in approaches to the control of the (1957) used a different nomenclature but subdivided the liver
pedicles of the liver described by Couinaud (1954), Patel and similarly. Subsequent refinements in the description of the liver
Couinaud (1952a, 1952b), Takasaki and others (1986), and as were offered by Couinaud (1981) in a monograph, which con-
developed and practiced by Launois and Jamieson (1992). tributed greatly to the understanding of practical surgical
Although ligatures and suture ligation remain the basic main- anatomy.
stays for the control of intrahepatic vessels, stapling techniques In 1950 Honjo published a case of anatomic right hepatec-
are now used for the same purpose (Fong & Blumgart, 1997). tomy; the operation was performed in Kyoto, Japan, on March
Garre (1907) described the use of Doyen’s elastic stomach 7, 1949, and was later reported by Honjo and Araki (1955) in
clamp to provide compression of the liver before it is transected, English. In 1952 Lortat-Jacob (Fig. 0.12) and Robert per-
and he referred to the work of Masnata and Lollini in Bologna, formed a true anatomic liver resection with preliminary vascular
who used a similar clamp. This method also has been used control. These reports opened the way for the development of
more recently by Storm and Longmire (1971) and Balasegaram liver surgery. The Lortat-Jacob operation captured much
FIGURE 0.11 Claude Couinaud (1922–2008), surgical anatomist. FIGURE 0.12 Jean Louis Lortat-Jacob (1908–1992), surgeon.
attention, and many further case reports followed. As recounted mL of blood replacement. The surgical bluntness of this
by Fortner and Blumgart (2001), at the Southern Surgical approach and the absence of the customary surgical principles
Association meeting in the United States on December 10, and techniques precluded widespread acceptance of the tech-
1952, Quattlebaum (1953) described a right hepatic lobectomy nique. Similarly, a prominent figure in the early development of
for hepatoma. The operation had been done about 4 months liver surgery, Brunschwig (1953; Serrea & Brunschwig, 1955)
after that of Lortat-Jacob and Robert. Apparently unaware of advocated nonanatomic resection, but his results—with a very
Quattlebaum’s achievement, and thinking he would be the first high operative mortality rate for so-called simple right lobectomy,
to do a so-called right hepatic lobectomy in the United States, mostly from uncontrolled hemorrhage—led him to accept the
Pack at Memorial Sloan Kettering Cancer Center in New York approach of Lortat-Jacob. Brunschwig’s report of long-term
performed the operation on a 40-year-old woman on December survivors after hepatic resection for advanced cancer challenged
14, 1952 (Pack et al, 1955). the prevailing skepticism about resecting liver tumors.
Pack and colleagues (1962) recognized quickly the oppor- The laparothoracotomy approach of Lortat-Jacob and
tunity to study liver regeneration. Their studies were the first Robert in 1952 quickly became the standard for major hepatic
to describe and document regeneration of the human liver after resection. Opening the chest added operating time and increased
major resection and suggested that complete regeneration morbidity but gave needed exposure to the liver. Later, the
occurs over 3 to 6 months. Later, Lin and Chen (1965) studied introduction of costal arch retractors made thoracotomy exten-
metabolic function and regeneration of the cirrhotic liver and sion obsolete for most hepatic resections. Steps generally con-
found no discernible regeneration after resection. Blumgart and sidered unnecessary, such as T-tube drainage of the common
colleagues (1971) described the speed of liver regeneration as bile duct, were eliminated in the 1970s. Opinions were divided
recovery of liver size within 10 to 11 days after major hepatic in the 1970s and 1980s about the superiority of either intrahe-
resection for trauma. patic or extrahepatic management of the major hepatic veins.
Contrary to the trends toward anatomic precision, Lin Closed drainage was substituted for multiple Penrose drains
(1958) advocated a finger-fracture technique for liver resection, and packing, although these were still advocated by some in
removing a hepatic lobe in 10 minutes with an average of 2000 the 1980s. Many hepatic resections have been done without
drainage, as described by Franco and associates (1989) and laparoscopic liver resection remain under active investigation.
subsequently at Memorial Sloan Kettering Cancer Center by Robotic operative approaches to liver resection have been
Blumgart and colleagues (Fong et al, 1996). devised (Choi et al, 2012).
Methods for anatomically based segmental liver resection
that conserves parenchyma were developed and have proved Liver Tumors
valuable in hepatic resection for tumor. Numerous methods to The management of liver tumors has a relatively short history.
transect the liver parenchyma had been advocated, including After the initial reports of hepatic resection for tumors per-
the knife handle (Quattlebaum, 1953), nitrogen knife (Serrea formed by Langenbuch (1888), Tiffany (1890), Lucke (1891),
& Brunschwig, 1955), electrocautery (Fortner et al, 1978), and Keen (1899), no significant reports were published until
microwave tissue coagulator (Tabuse, 1979), waterjet (Papa- Foster’s review in 1970. He reported a multi-institutional
christou & Barters, 1982), ultrasonic aspiration dissector survey conducted on 296 adults who had liver resection for
(Hodgson & DelGuercio, 1984), and harmonic scalpel; other primary cancer and were followed for at least 5 years or until
techniques are described on an almost daily basis, and simple death. The operative mortality rate was 24%. Foster also
crushing of the liver tissue remains in use by many. studied the records of 115 patients who had undergone hepatic
Bleeding has remained a problem, and preliminary vascular resection for metastatic colorectal cancer. Operative mortality
control is difficult or impossible in some clinical settings. Inflow was 17.3%, and 21% survived for at least 5 years. Hemorrhage
occlusion using the Pringle maneuver for longer than 15 to 20 during operation was still a problem, and the operative mortal-
minutes was considered hazardous. Local and general hypo- ity rate was essentially the same or higher than the cure rate so
thermia was used in an effort to increase the liver’s tolerance of that skeptics remained unconvinced.
the presumed ischemic anoxia associated with the procedure In the 1970s, major hepatic resection for tumors became
(Longmire & Marable, 1961). More complete vascular control more frequent, and this coincided with improvements in anes-
was obtained by Heaney and colleagues (1966), without hypo- thesia, surgical technique, and postoperative support, and sig-
thermia, by employing aortic cross clamping and temporary nificantly, with the development of better imaging modalities.
occlusion of the porta hepatis and vena cava for 24 minutes In the United States, Wilson and Adson (1976), Fortner and
without ill effect. Fortner and colleagues (1971) described colleagues (1974), and Thompson and colleagues (1983) were
isolation-hypothermic perfusion of the liver, making prolonged major figures in this area of endeavor. In Asia, Ong (1977) and
vascular isolation for complicated resections possible. Huguet others were early explorers in the field, and other major centers
and associates (1978) challenged the long-held belief that the made significant technical improvements with encouraging
liver could tolerate only 15 to 20 minutes of normothermic results (Attiyeh et al, 1978; Cady et al, 1979; Starzl et al,
perfusion by extending the period to 65 minutes. This observa- 1975).
tion was subsequently confirmed (Bismuth et al, 1989; Huguet Morbidity and mortality rates decreased dramatically, and
et al, 1992, 1994). As described previously, however, major major centers began to provide encouraging preliminary long-
bleeding during liver resection usually arises from the major term survival figures. As described by Fortner and Blumgart
hepatic veins or vena cava, and development of techniques of (2001), by the 1980s the early reluctance by many surgeons to
low central venous pressure anesthesia during liver resection, accept the therapeutic benefit of hepatic resection for tumors
as first described by Blumgart and colleagues (Cunningham faded before the onslaught of encouraging reports from a
et al, 1994, Melendez et al, 1998), have proved simple and variety of institutions in the United States, Europe, and South-
efficacious and have rendered vascular isolation techniques east Asia.
rarely necessary (Jarnagin et al, 2002). Methods for limiting the extent of the parenchymal transec-
Apart from hepatic resection, many other techniques for the tion developed along with the emergence of surgical resection
enucleation of tumors and of hydatid cysts have been devel- for tumors in cirrhotic and noncirrhotic livers. Although sug-
oped. Garre (1907) had first noted that if he kept close to the gested originally by Pack and Miller (1961) and by McBride
line of the dense membrane around hydatid cysts—so as not to and Wallace (1972), Asian and European surgeons soon
enter the tissues of the liver too deeply, where large vessels may assumed a leadership role in developing isolated segmental
be encountered—it was possible to resect such lesions with resection and described various operations (Bismuth et al,
minimal blood loss. Enucleative techniques are now used by 1982; Hasegawa et al, 1989; Scheele, 1989; Ton, 1979). The
contemporary surgeons, not only for the removal of hydatid development of reliable intraoperative ultrasound allowed
cysts but also for the enucleation of giant hemangiomata (Baer further developments in liver resection (Bismuth et al, 1982;
et al, 1992; Blumgart et al, 2000; Hochwald & Blumgart, Makuuchi et al, 1985), and subsegmental resection was devel-
2000). As described in this book, enucleation also may be used oped. More recently, the segmental approach was reported and
for resection of certain neoplasms of the liver, such as fibronod- popularized in the United States, particularly by Blumgart and
ular hyperplasia. others (Billingsley et al, 1998; DeMatteo et al, 2000). The dif-
The widespread use of laparoscopic cholecystectomy led to ficulty of the approach to the caudate lobe was addressed, and
the development of techniques for laparoscopic resection of the the surgical anatomy and technique for caudate resection,
liver. First used by Gagner and colleagues (1992), laparoscopic either as an isolated segment or in combination with major
partial hepatectomy has been used by many surgeons. Ferzli resections, was described by Blumgart and colleagues (Lerut
and colleagues (1995) and Azagra and colleagues (1996) devel- et al, 1990; Bartlett et al, 1996).
oped the operation for anatomic resection of the left lateral Probes to perform cryosurgical ablation were developed
segment, and most such resections have involved this proce- (Ravikumar & Kaleya, 2000), and cryosurgery-assisted seg-
dure. More recently, major hepatectomy has been performed mental resection has been described (Polk et al, 1995). Radio-
using laparoscopic techniques (Gigot et al, 2002). Experience frequency ablation for liver tumors also has been evaluated
remains limited, however, and the indications and place of (Curley et al, 1999; Wood et al, 2000). Surgeons now perform
controlled tumor ablation or hepatic resection for tumors using The treatment of metastatic liver tumors by hepatic resec-
robotic techniques and refinements in the development of tion was viewed initially as highly dubious. Many regarded the
virtual reality techniques as originally suggested by Marescaux removal of tumor from the liver by hepatic resection as an
and colleagues in 1998 and by Satava and colleagues in 1999 irrational approach. The natural history of colorectal cancer
(Marescaux et al, 1998; Satava, 1999). Image-guided surgery offered the opportunity for resection of liver metastases
opens great possibilities for developments in hepatic surgery in more often than any other cancer in Western countries. The
the future (Lang et al, 2004). multiinstitutional report from Foster (1970) in the United
The evolution of liver surgery for cancer was influenced States suggested a 5 year survival rate of 25%, and multiple
profoundly by the geographic distribution of the disease. Sur- other reports supported that view. However, the question
geons in the United States and Europe were confronted much remained: How long could patients have lived without surgical
more often with metastatic cancer or hepatocellular carcinoma operations?
(HCC) in otherwise normal livers. In the 1980s and early A landmark article from Glasgow (Wood et al, 1976)
1990s, surgeons in the United States had no great drive to described the natural history of colorectal hepatic metastases.
conserve normal liver tissue. By contrast, surgeons in Asia were It was shown that survival depended on the extent of the
confronted frequently with HCC in patients with hepatitis B disease, and that multiple bilobar metastases indicated a much
and cirrhosis of the liver. Estimating the functional reserve of more serious prognosis than single or multiple nodules on one
the cirrhotic liver became a major concern, and methods for side of the liver. In Wood’s study, only one patient survived for
limiting the resection of nontumorous liver were developed. 5 years (1%). It was clear that if the operative mortality rate
Screening programs for the detection of HCC at an early stage declined, as it had in many reported series, the issue was closed.
were a natural corollary. A major study from Blumgart’s group at Memorial Sloan Ket-
Primary liver cancer (HCC) occurs mainly in areas where tering Cancer Center (Fong et al, 1999) described the results
viral hepatitis is endemic. Because of the prevalence of hepatitis in 1001 consecutive patients with a 5 year overall survival rate
B infection in Southeast Asia, most of the initial, significant, of 37% and an operative mortality rate of 3.4%. Five year
published experiences examining the treatment of primary survival reached more than 50% in patients with favorable
HCC are from Southeast Asia. Improved diagnostic and surgi- prognostic factors. Similar results are reported from many insti-
cal approaches in Japan permitted the Liver Cancer Study tutions worldwide. Tumor recurrence with the liver the only
Group of Japan (1990) to report a 5 year survival rate of 57.5% site of recurrence is common, and such hepatic recurrences are
for HCC; a high proportion of the cases were small, encapsu- often amenable to repeated resections, with risks and results
lated tumors. From China, Tang (1993) reported a 5 year similar to those of the initial operation (Elias et al, 1992; Fong
survival rate of 40.2%. et al, 1994; Scheele & Stangl, 2000).
In the West, experience with HCC differed from that More recently, the Memorial Sloan Kettering group reported
reported in the East in material ways: Tumors were generally a series of 1600 patients with metastatic colorectal cancer
large at presentation, the incidence of viral hepatitis was much treated since 1985 (House et al, 2010). They demonstrated an
lower, and many patients did not have underlying cirrhosis. improvement in results over time between 1999 and 2004, with
There were several reports of surgical mortality rates for hepatic a 5 year survival rate of 51% in 563 patients and an operative
resection in cirrhotic patients, ranging from 15% to 100%. In mortality rate of only 1%.
recent years, spread of hepatitis B and hepatitis C has changed Pichlmayr and colleagues (1990) did the first ex situ tumor
the demographics of the disease in the United States and in resection on the liver, removing the liver for bench surgery, then
Western Europe. This change has allowed a much greater surgi- autotransplanting it back into the patient. The liver was pre-
cal experience, and better techniques and improved patient served by hypothermic perfusion, as had been described earlier
selection have led to vastly improved results. Blumgart’s group for the in situ procedure (Fortner et al, 1974). This approach
(Fong et al, 1999) reported a mortality rate of 5% in 100 resec- may prove to have merit in highly selected cases, but it has not
tions, 57 being major resections, in patients with HCC and been accepted generally.
cirrhosis, and a 5-year survival rate of 37% was reported among Although initially completely ineffective, systemic chemo-
patients operated on since 1990. therapy improved dramatically with the development of new
Today the surgical mortality rate in noncirrhotic patients, agents, and its use as a neoadjuvant agent and as adjuvant
even for the most extensive resections, is uniformly less than chemotherapy after resection has been extensively explored.
5% in major centers. The overall 5 year survival rate is approxi- The use of neoadjuvant chemotherapy to convert unresectable
mately 35% in patients with HCC reported from Western tumors to resectable ones was reported by Bismuth and col-
series, in which small HCC is relatively rare (Bismuth et al, leagues (1996), and extensive studies were reported regarding
1986; Ringe et al, 1991). In a study reported by Blumgart and the possible benefits of chemotherapy after hepatic resection
others (Fong et al, 1999), patients who had HCC with a using hepatic arterial infusion of chemotherapy (Kemeny et al,
median tumor size of 10 cm but no cirrhosis had hepatic resec- 1999). Kemeny’s study followed the demonstration in 1982 by
tion with a surgical mortality rate of 3.7% and a 5 year survival Ensminger and colleagues of an implantable pump for hepatic
rate of 42%. Although survival rates have not changed substan- arterial infusion therapy. Others have studied regression of
tially, a great improvement has been seen in rates of morbidity advanced refractory cancer in the liver using isolated hepatic
and mortality. perfusion chemotherapy (Alexander et al, 2000).
In recent years a move toward hepatic transplantation has As described in this book, hepatic resection also has been
been seen in selected patients with small HCCs within a cir- shown to yield satisfactory results in terms of relief of symptoms
rhotic liver. Patients with small HCCs and compromised liver and prolongation of life in the management of hepatic metas-
function seem to be ideal candidates for total hepatectomy and tases from neuroendocrine cancer and sarcoma. Hepatic resec-
allotransplantation. tion for biliary and gallbladder cancer also has developed in a
remarkable fashion. Launois and associates in France (1979)

reported the first major series of patients in whom hepatic
resection was performed to ensure clearance of hilar cholangio-
carcinoma. Working contemporaneously, Fortner and col-
leagues (1976) and Blumgart and Beazley and associates
(Beazley et al, 1984; Blumgart et al, 1984) reported similar
results, and in some patients, hepatic resection was accompa-
nied by resection of the portal vein and portal venous recon-
struction (Blumgart et al, 1984). Multiple other reports have
confirmed that hepatic resection offers advantages in the man-
agement of hilar cholangiocarcinoma and helps achieve nega-
tive tumor margins that are associated with long-term survival
(Hadjis et al, 1990; Jarnagin et al, 2001; Klempnauer et al,
1997; Nimura et al, 1990). Surgeons in Southeast Asia and
Japan have contributed considerably in this area and in particu-
lar have contributed to the development of caudate lobe resec-
tion and vascular reconstruction (Mizumoto & Suzuki, 1988;
Nimura et al, 1990).
Pack and colleagues (1955) advocated a similar radical
approach for gallbladder cancer. They believed that an extended
right hepatic lobectomy had its greatest applicability for gall-
bladder cancer, and they recommended an en bloc lymph node
resection of the porta hepatis to be included with the operation.
Brasfield (1956) reported a microscopic focus of metastatic
cancer in a patient without gross liver involvement and helped
establish the need for hepatic resection for gallbladder cancer.
More recent reports indicate that for more invasive lesions and
for patients seen after initial cholecystectomy, at which cancer
was inadvertently found, an extensive resection seems to be
effective (Fong et al, 2000). Fortner and colleagues (1970)
performed total hepatectomy and orthotopic liver transplanta-
tion in two such patients with short-term survival.
Tumor ablation has been a major feature of the recent
history of liver surgery for tumors. Cryosurgery (Adam et al,
1997; Crews et al, 1997; Zhou et al, 1988) has been used for FIGURE 0.13 Thomas Starzl (1926–), surgeon.
treatment of small tumors and for the management of recurrent
disease after resection or for the management of unresectable host-versus-graft immune response (Gibson & Medawar, 1943;
tumors in combination with chemotherapy. More recently, Medawar, 1944). For further details, the reader is referred to
radiofrequency ablation of tumors has been developed, and its the publications of Starzl (2001) and of Murray and Hills
use in primary and metastatic cancer was reported (Curley (2005). This portion of the review of the history of liver surgery
et al, 1999; Wood et al, 2000) and is now established. The is concerned with the history of liver transplantation.
technique can be used at open operation or percutaneously. In a magnificent leap forward, Starzl (Fig. 0.13) and associ-
Ethanol injection has been used in the management of small ates (1968), in the United States, did the first successful total
HCCs (Ebara et al, 1990; Tanikawa, 1991). It is of particular hepatectomy with orthotopic liver transplantation. Calne and
interest because it is inexpensive, widely available, and easily Williams (1968), in the United Kingdom, reported similar
used with a low complication rate even in poor-risk patients, studies. The contributions of both groups to the development
and it can be repeated frequently. Hepatic arterial embolization of liver transplantation and immunosuppressive therapy have
can also be used to ablate tumors and may be particularly effec- been monumental. Fortner and colleagues (1970) reported the
tive in primary HCCs and in metastatic tumors arising from first successful heterotopic (auxiliary) liver homograft. This
primary neuroendocrine sources (Allison et al, 1977; Allison, contribution was the forerunner of the further development of
1978). heterotopic liver transplantation and of split-liver and living
related-donor transplants. Liver transplantation underwent
Liver Transplantation continued evolution despite the modestly effective immunosup-
Starzl (2001) has provided a fascinating overview of the origins pressive therapy initially available. One-year survival rates
of clinical organ transplantation. A historical consensus devel- slowly reached 50% by 1979 (Calne & Williams, 1979; Starzl
opment conference was convened in March 1999 at the Uni- et al, 1979). Chronic infection, rejection, and surgical infec-
versity of California, Los Angeles, to identify the principal tions diminished the small group to only a few long-term sur-
milestones leading to the clinical use of various transplantation vivors. The report in 1979 of immunosuppression with
procedures; the conclusions were published in the July 2000 cyclosporine by Calne and associates (1979) transformed
issue of the World Journal of Surgery (Groth & Longmire, 2000). the field rapidly. A consensus conference held in 1983 declared
The early history of the evolution of organ transplantation fol- that liver transplantation was now an acceptable therapy and
lowed the first convincing evidence that organ rejection is a no longer an experimental procedure. Zeevi and colleagues
(1987) reported on another powerful immunosuppressive (1990) were the first to report living-donor liver transplantation
agent, tacrolimus. using segments II and III of the liver; Yamaoka and colleagues
Serious complications of immunosuppressive therapy for (1994) reported using the right lobe of the liver. Use of split-
liver and other organ transplants have been described, includ- liver grafts and living, related donors for grafts of both the right
ing de novo malignancy (Penn, 1988; Penn & Brunson, 1988; and left lobes has been driven by cultural restrictions and by a
Penn & Starzl, 1972). In 1988 Iwatsuki and associates reported shortage of donor organs. This has been a major development,
an overall 54% survival rate of patients at 5 years. Development and the potential for mortality and morbidity not only of recipi-
was rapid, and by 1992, more than 3000 orthotopic liver trans- ents but also of the live donors has remained a major ethical
plantations were done annually in the United States (United concern.
Network for Organ Sharing). Liver transplantation for tumors has evolved considerably in
Total hepatectomy and liver transplantation were initially the field of primary HCC. Although liver resection remains the
disappointing for liver cancer. Iwatsuki and colleagues (1988) treatment of choice for HCC in patients with good liver func-
reported that three of every four patients who lived at least 2 tion, similar patients with compromised liver function and
months after transplantation for cancer had recurrence, and patients with hepatitis C and a small tumor in a portion of the
adjuvant chemotherapy had no demonstrable benefit. Ringe liver geographically unfavorable for resection are now consid-
and associates (1989) reported a 5 year survival rate of 15.2% ered best treated by liver transplantation. Liver transplantation
for such patients, and Calne and colleagues (1986) had similar has come to be a recognized therapy for many patients with
results. The best results and apparent cure were obtained when small HCCs (Bismuth et al, 1999), a wide range of benign
a cancer was an incidental finding in a liver removed for non- disease, and in particular for patients with compromised liver
cancerous disease (e.g., alcoholic cirrhosis). Geevarghese and function as a result of cirrhosis of the liver, Budd-Chiari syn-
associates (1998) reported a 1 year survival rate of 85% and a drome, polycystic liver and kidney disease, sclerosing cholan-
5 year survival rate of 78% for such cases. Liver transplantation gitis, and for patients with a wide variety of other parenchymal
for HCC was transformed with a publication by Mazzaferro and metabolic liver diseases. Liver transplantation for patients
and colleagues (Milan criteria, Mazzaferro et al, 1996), which with hilar cholangiocarcinoma is occasionally indicated, and
showed excellent long-term survival in patients with limited good results have been achieved in selected patients treated
disease and generated the patient selection criteria that are still with aggressive neoadjuvant chemotherapy and radiotherapy
widely used today. (Rea et al, 2005). Some patients with widespread neuroendo-
Current adjuvant chemotherapy may improve the results. crine metastatic disease of the liver may benefit from liver
Olthoff and colleagues (1995) reported a 3 year survival rate of transplantation. Patients with metastatic disease from adeno-
45% using fluorouracil, doxorubicin, and cisplatin, but only carcinoma have had poor results, and transplantation is no
three survivors had cancers larger than 5 cm. Laine and associ- longer used for this indication.
ates (1999) reported five children with advanced hepatoma who In writing this historical account, I have drawn freely on the
were alive at a mean of 4.6 years after transplantation that was work of the following excellent publications: Chakravorty and
preceded by induction chemotherapy. Wanebo (1987), Glenn (1971), Praderi (1982), and Starzl
Improved immunosuppression has allowed surgical tech- (2001). I have no doubt that there will be some inaccuracies
niques to blossom. Bismuth and Houssin (1984) reported and disputed claims as to “firsts,” but I have attempted to relate
reduced-size orthotopic liver grafts in children, and Pichlmayr a fascinating surgical story. I apologize for any disagreement I
and colleagues (1988) reported the use of one donor liver for may ignite.
two recipients (split-liver transplantation). Raia and associates
(1989), in a letter to the Lancet, and Strong and colleagues References are available at expertconsult.com.
Introduction Hepatobiliary and pancreatic surgery: historical perspective 15.e1
REFERENCES and biliary tract, ed 3, Philadelphia, 2000, Saunders, pp 1639–

1715.
Acosta JM, Ledesma CL: Gallstone migration as a cause of acute Bobbs J: Case of lithotomy of the gallbladder, Trans Med Soc Indiana
pancreatitis, N Engl J Med 290:484–487, 1974. 68, 1868.
Adam R, et al: Place of cryosurgery in the treatment of malignant liver Boyden E: The pars intestinalis of the common bile duct, as viewed
tumors, Ann Surg 225:39–50, 1997. by the older anatomists (Vesalius, Glisson, Bianchi, Vater, Haller,
Adams F: The genuine works of hippocrates, Baltimore, 1939, Williams Santorini et al), Anat Rec 66:217, 1936.
& Wilkins. Brasfield RD: Prophylactic right hepatic lobectomy for carcinoma of
Alexander B: Translation of seats and causes of disease, facsimile of London, the gallbladder, Am J Surg 91:829–832, 1956.
ed 1769, New York, 1960, Hafner. Briggs E: Tumor of the pancreas, laparotomy, recovery, St Louis Med
Alexander HR Jr, et al: Current status of isolated hepatic perfusion with Surg J 58:154, 1890.
or without tumor necrosis factor for the treatment of unresectable Brooks JR, Culebras JM: Cancer of the pancreas: palliative operation,
cancers confined to liver, Oncologist 5:416–424, 2000. Whipple procedure or total pancreatectomy, Am J Surg 131:516–
Allison DJ, et al: Treatment of carcinoid liver metastases by hepatic- 520, 1976.
artery embolisation, Lancet 2:1323–1325, 1977. Brunschwig A: Surgery of hepatic neoplasms, with special reference to
Allison DJ: Therapeutic embolization, Br J Hosp Med 20:707–715, secondary malignant neoplasms, Cancer 6:725–742, 1953.
1978. Bryan PJ, et al: Correlation of CT, gray scale ultrasonography, and
Ancelet E: Étude sur les maladies du pancréas, Paris, 1866, Savy. radionuclide imaging of the liver in detecting space-occupying pro-
Attiyeh FF, et al: Hepatic resection for metastasis from colorectal cesses, Radiology 124:387–393, 1977.
cancer, Dis Colon Rectum 21:160–162, 1978. Buxbaum A: Über die Photographie von Gallensteinen in vivo, Wein
Azagra JS, et al: Laparoscopic anatomical left lateral segmentectomy: Med Presse 39:534, 1898.
technical aspects, Surg Endosc 10:758–761, 1996. Cady B, et al: Elective hepatic resection, Am J Surg 137:514–521,
Baer HU, et al: Enucleation of giant hemangiomas of the liver: techni- 1979.
cal and pathologic aspects of a neglected procedure, Ann Surg Calne RY, Williams R: Liver transplantation in man: I. observations
216:673–676, 1992. on technique and organization in five cases, BMJ 4:535–540,
Bakes J: Die Choledochopapilloskopie, Arch Klin Chir 126:473–483, 1968.
1923. Calne RY, Williams R: Liver transplantation, Curr Probl Surg 16:1–44,
Balasegaram M: Hepatic surgery, J R Coll Surg Edinb 17:85–89, 1972a. 1979.
Balasegaram M: Hepatic surgery: a review of a personal series of 95 Calne RY, et al: Cyclosporin A initially as the only immunosuppressant
major resections, Aust N Z J Surg 42:1–10, 1972b. in 34 recipients of cadaveric organs: 32 kidneys, 2 pancreases, and
Bartlett D, et al: Complete resection of the caudate lobe of the liver: 2 livers, Lancet 2:1033–1036, 1979.
technique and results, Br J Surg 83:1076–1081, 1996. Calne RY, et al: Liver transplantation in the adult, World J Surg
Beazley RM, et al: Clinicopathological aspects of high bile duct cancer: 10:422–431, 1986.
experience with resection and bypass surgical treatments, Ann Surg Cantlie J: On a new arrangement of the right and left lobes of the liver,
199:623–636, 1984. J Anat Physiol Proc Anat Soc Great Brit Ire 32:4–9, 1897.
Beck C: Surgery of the liver, J Am Surg Assoc 78:1063, 1902. Cattel RD: Anastomosis of the duct of Wirsung: its use in palliative
Beger HG, et al: Duodenum-preserving resection of the head of the operations for cancer of the head of the pancreas, Surg Clin North
pancreas in patients with severe chronic pancreatitis, Surgery 97:467– Am 27:636–643, 1947.
473, 1985. Cattell RB, Warren W: Surgery of the pancreas, Philadelphia, 1953, WB
Berne TV, Donovan AJ: Liver and bile duct injury. In Blumgart LH, Saunders.
Fong Y, editors: Surgery of the liver and biliary tract, ed 3, Philadel- Celsus AC, editor: De medicina, loeb classical library ed, Cambridge, MA,
phia, 2000, Saunders, pp 1277–1301. 1935, Harvard University Press.
Berta G: cited by Chen TS, Chen PS, 1984: Understanding the liver: a Chakravorty RC, Wanebo HJ: Historic preamble: liver and biliary
history, Westport, CT, 1716, Greenwood Press, p 293. cancer. In Wanebo HJ, editor: Science and practice of surgery, vol 8,
Billingsley KG, et al: Segment-oriented hepatic resection in the man- Hepatic and biliary cancer, New York, 1987, Marcel Dekker, pp
agement of malignant neoplasms of the liver, J Am Coll Surg xiii–xxvii.
187:471–481, 1998. Chesterman JT: Treatment of pancreatic cysts, Br J Surg 30:234–235,
Biondi G: Contributo clinico e sperimentale al achirurgia del pancreas, 1943.
Clin Chir 5:132, 1897. Choi GH, et al: Robotic liver resection: technique and results of 30
Bismuth H, Corlette MB: Intrahepatic cholangiojejunostomy: an consecutive procedures, Surg Endosc 26:2247–2258, 2012.
operation for biliary obstruction, Surg Clin North Am 36:849–863, Claessen F, 1842: Krankheiten der Bauchspeicheldruse. Cologne,
1956. Germany, Schauberg.
Bismuth H, Houssin D: Reduced-sized orthotopic liver graft in hepatic Classen M, Demling L: Endoscopic sphincterotomy of the papilla of
transplantation in children, Surgery 95:367–370, 1984. Vater and extraction of stones from the choledochal duct (author’s
Bismuth H, et al: Long-term results of hepaticojejunostomy Roux-en- transl), Dtsch Med Wochenschr 99:496–497, 1974.
Y, Surg Gynecol Obstet 146:161–167, 1978. Codivilla A: Rendiconto statistico della sezione chirurgica dell’ospedale di
Bismuth H, et al: Major and minor segmentectomies “réglées” in liver imola, 1898.
surgery, World J Surg 6:10–24, 1982. Coffey RC: Pancreato-enterostomy and pancreatectomy: a preliminary
Bismuth H, et al: Liver resections in cirrhotic patients: a Western report, Ann Surg 50:1238–1264, 1909.
experience, World J Surg 10:311–317, 1986. Cotton PB, et al: Endoscopic transpapillary radiographs of pancreatic
Bismuth H, et al: Major hepatic resection under total vascular exclu- and bile ducts, Gastrointest Endosc 19:60–62, 1972a.
sion, Ann Surg 210:13–19, 1989. Cotton PB, et al: Cannulation of papilla of Vater via fiber-duodenoscope:
Bismuth H, et al: Resection of nonresectable liver metastases from assessment of retrograde cholangiopancreatography in 60 patients,
colorectal cancer after neoadjuvant chemotherapy, Ann Surg Lancet 1:53–58, 1972b.
224:509–522, 1996. Couinaud C: Bases anatomiques des hepatectomies gauche et droite
Bismuth H, et al: Liver transplantation for hepatocellular carcinoma, réglées, J Chir 70:933–966, 1954.
Semin Liver Dis 19:311–322, 1999. Couinaud C: Le foie: étude anatomique et chirurgicale, Paris, 1957,
Blumgart LH, et al: Observations on liver regeneration after right Masson.
hepatic lobectomy, Gut 12:922–928, 1971. Couinaud C: Controlled hepatectomies and controlled exposure of the
Blumgart LH, et al: Endoscopy and retrograde choledochopancreatog- intrahepatic bile ducts: anatomic and technical study, Paris, 1981,
raphy in the diagnosis of the jaundiced patient, Lancet 2:1269–1273, Author.
1972. Crews KA, et al: Cryosurgical ablation of hepatic tumors, Am J Surg
Blumgart LH, et al: Surgical approaches to cholangiocarcinoma at 174:614–617, 1997.
confluence of hepatic ducts, Lancet 1:66–70, 1984. Cunningham JD, et al: One hundred consecutive hepatic resections:
Blumgart LH, et al: Liver resection for benign disease and for liver and blood loss, transfusion, and operative technique, Arch Surg
biliary tumors. In Blumgart LH, Fong Y, editors: Surgery of the liver 129:1050–1056, 1994.
15.e2 Introduction Hepatobiliary and pancreatic surgery: historical perspective
Curley SA, et al: Radiofrequency ablation of unresectable primary and Gagner M, et al: Laparoscopic partial hepatectomy for liver tumor,
metastatic hepatic malignancies: results in 123 patients, Ann Surg Surg Endosc 6:99, 1992.
230:1–8, 1999. Garre C: On resection of the liver, Surg Gynecol Obstet 5:331, 1907.
Da Costa JM: Cancer of the pancreas, North Am Med-Chir Rev 2:883, Gaston J: Surgery of the gall bladder and ducts, vol IX, New York, 1897,
1858. William Wood.
Dahl R: Eine Neue operation des gallenwege, Zb F Chir 36:266, 1909. Geevarghese SK, et al: Outcomes analysis in 100 liver transplantation
Damadian R: Tumor detection by nuclear magnetic resonance, Science patients, Am J Surg 175:348–353, 1998.
171:1151–1153, 1971. Gibson T, Medawar PB: The fate of skin homografts in man, J Anat
DeMatteo RP, et al: Anatomic segmental hepatic resection is superior 77:299–310, 1943.
to wedge resection as an oncologic operation for colorectal liver Gigot JF, et al: Laparoscopic liver resection for malignant liver tumors:
metastases, J Gastrointest Surg 4:178–184, 2000. preliminary results of a multicenter European study, Ann Surg
Demling L, Classen M: Duodenojejunoscopy, Dtsch Med Wochenschr 236:90–97, 2002.
95:1427–1428, 1970. Glenn F: Biliary tract disease since antiquity, Bull N Y Acad Med
Di Sebastiano P, et al: Management of chronic pancreatitis: conserva- 47:329–350, 1971.
tive, endoscopic, and surgical. In Blumgart LH, et al, editors: Surgery Goetze O: [Transhepatic permanent drainage in high-seated stenosis
of the liver, biliary tract, and pancreas, vol II, ed 4, Philadelphia, 2007, of the common bile duct]. Eine transhepatische Dauer Drainage bei
Saunders, pp 1341–1386. der hohen Gallengang stenose, Langenbecks Arch Klin Chir Ver Dtsch
Doubilet H, Mulholland JH: The surgical treatment of recurrent acute Z Chir 270:97–101, 1951.
pancreatitis by endocholedochal sphincterotomy, Surg Gynecol Obstet Goldsmith NA, Woodburne RT: The surgical anatomy pertaining to
86:295–306, 1948. liver resection, Surg Gynecol Obstet 105:310–318, 1957.
Dubois E, et al: Cholecystectomy by coelioscopy, Presse Med 18:980– Graham EA, Cole WH: Roentgenologic examination of the gallbladder,
982, 1989. JAMA 82:613, 1924.
Dubois Ε, et al: Coelioscopic cholecystectomy: preliminary report of Graham EA, Hartmann AF: Subtotal resection of the pancreas for
36 cases, Ann Surg 211:60–62, 1990. hypoglycemia, Surg Gynecol Obstet 59:474–479, 1934.
Du Val MK: Pancreaticojejunostomy for chronic pancreatitis, Surgery Green R: A translation of Galen’s hygiene, Springfield, IL, 1951,
41:1019–1028, 1957. Thomas.
Ebara M, et al: Percutaneous ethanol injection for the treatment of Grossman ZD, et al: RN Imaging, CT and gray-scale ultrasonography
small hepatocellular carcinoma: study of 95 patients, J Gastroenterol of the liver: a comparative study, J Nucl Med 18:327–332, 1977.
Hepatol 5:616–626, 1990. Groth CG, Longmire WP: Historical landmarks in clinical transplanta-
Elias D, et al: Another failure in the attempt of definition of the indication, World J Surg 24:755–843, 2000.
tions to the resection of liver metastases of colorectal origin, J Chir Gussenbauer G: Zur operativen Behandlung der Pancreascysten, Arch
Surg (Paris) 129:59–65, 1992. Klin Chir 29:355, 1883.
Elliot J: Surgical treatment of tumor of the liver with report of a case, Hadjis NS, et al: Outcome of radical surgery in hilar cholangiocarci-
Ann Surg 26:83, 1897. noma, Surgery 107:597–604, 1990.
Ensminger W, et al: Effective control of liver metastases from colon Halsted W: Contributions to the surgery of the bile passages, especially
cancer with an implanted system for hepatic arterial chemotherapy, of the common bile duct, Johns Hopkins Hosp Bull 11:1, 1900.
Proc Am Clin Oncol 1:94, 1982. Hartmann H: Chirurgie des voies biliares, Paris, 1923, Masson.
Estefan A, et al: Anastomosis colangiodigestivas en cancer bilar, Cir Hasegawa H, et al: Central bisegmentectomy of the liver: experience
Uruguay 47:51, 1977. in 16 patients, World J Surg 13:786–790, 1989.
Fallis LS, Szilagyi DE: Observations on some metabolic changes after Healey J, Schroy P: The anatomy of the bile ducts within the human
total pancreatoduodenectomy, Ann Surg 128:639–667, 1948. liver: an analysis of the prevailing patterns of branching and their
Ferzli G, et al: Laparoscopic resection of a large hepatic tumor, Surg major variants, Arch Surg 66:599–616, 1953.
Endosc 9:733–735, 1995. Heaney JP, et al: An improved technique for vascular isolation of the
Fong Y, Blumgart LH: Useful stapling techniques in liver surgery, liver: experimental study and case reports, Ann Surg 163:237–241,
J Am Coll Surg 185:93–100, 1997. 1966.
Fong Y, et al: Repeat hepatic resections for metastatic colorectal Hemingway AP, Allison DJ: Complications of embolization: analysis of
cancer, Ann Surg 220:657–662, 1994. 410 procedures, Radiology 166:669–672, 1988.
Fong Y, et al: Drainage is unnecessary after elective liver resection, Am Hepp J, Couinaud C: [Approach to and use of the left hepatic duct in
J Surg 171:158–162, 1996. reparation of the common bile duct]. L’abord et l’utilisation du canal
Fong Y, et al: Clinical score for predicting recurrence after hepatic hépatique gauche dans les réparations de la voie biliare principale,
resection for metastatic colorectal cancer: analysis of 1001 consecu- Presse Med 64:947–948, 1956.
tive cases, Ann Surg 230:309–318, 1999. Hjortsjo C: The topography of the intrahepatic duct systems, Acta Anat
Fong Y, et al: Gallbladder cancer: comparison of patients presenting 11:599–615, 1951.
initially for definitive operation with those presenting after prior Hochwald SN, Blumgart LH: Giant hepatic hemangioma with
noncurative intervention, Ann Surg 232:557–569, 2000. Kasabach-Merritt syndrome: is the appropriate treatment enucle-
Fortner J: Regional resection of cancer of the pancreas: a new surgical ation or liver transplantation?, HPB Surg 11:413–419, 2000.
approach, Surgery 73:307–320, 1973. Hodgson WJ, DelGuercio LR: Preliminary experience in liver surgery
Fortner JG, Blumgart LH: A historic perspective of liver surgery for using the ultrasonic scalpel, Surgery 95:230–234, 1984.
tumors at the end of the millennium, J Am Coll Surg 193:210–222, Honjo I: Resection of the right lobe of the liver, Shujutsu 4:345–349,
2001. 1950.
Fortner JG, et al: Orthotopic and heterotopic liver homografts in man, Honjo I, Araki C: Total resection of the right lobe of the liver, J Int
Ann Surg 172:23–32, 1970. Coll Surg 1:23–28, 1955.
Fortner JG, et al: A new concept for hepatic lobectomy: experimental House M, et al: Survival after hepatic resection for metastatic colorectal
studies and clinical application, Arch Surg 102:312–315, 1971. cancer: trends in outcomes for 1,600 patients during two decades at
Fortner JG, et al: Major hepatic resection using vascular isolation and a single institution, J Am Coll Surg 210:744–752, 752–755, 2010.
hypothermic perfusion, Ann Surg 180:644–652, 1974. Huguet C, et al: Normothermic hepatic vascular exclusion for exten-
Fortner JG, et al: Surgical management of carcinoma of the junction sive hepatectomy, Surg Gynecol Obstet 147:689–696, 1978.
of the main hepatic ducts, Ann Surg 184:68–73, 1976. Huguet C, et al: Technique of hepatic vascular exclusion for extensive
Fortner JG, et al: Major hepatic resection for neoplasia: personal expe- liver resection, Am J Surg 163:602–605, 1992.
rience in 108 patients, Ann Surg 188:363–371, 1978. Huguet C, et al: Hepatic resection with ischemia of the liver exceeding
Foster JH: Survival after liver resection for cancer, Cancer 26:493–502, one hour, J Am Coll Surg 178:454–458, 1994.
1970. Imrie CW: Medical treatment of acute pancreatitis, Gastroenterology
Franco D, et al: Hepatectomy without abdominal drainage: results of 75:766–767, 1978.
a prospective study in 61 patients, Ann Surg 210:748–750, 1989. Iwatsuki S, et al: Experience in 1,000 liver transplants under
Frey C, Smith G: Description and rationale of a new operation for cyclosporine-steroid therapy: a survival report, Transplant Proc
chronic pancreatitis, Pancreas 2:701–707, 1987. 20(Suppl 1):498–504, 1988.
Jarnagin WR, et al: Staging, resectability, and outcome in 225 patients Longmire WP, Marable SA: Clinical experiences with major hepatic
with hilar cholangiocarcinoma, Ann Surg 234:507–519, 2001. resections, Ann Surg 154:460–474, 1961.
Jarnagin WR, et al: Improvement in perioperative outcome after Longmire WP, Sanford M: Intrahepatic cholangiojejunostomy with
hepatic resection: analysis of 1,803 consecutive cases over the past partial hepatectomy for obstruction, Surgery 24:264, 1948.
decade, Ann Surg 236:397–406, 2002. Longmire WP, et al: Experience with resection of the pancreas in
Jastrow M: The liver in antiquity and the beginnings of anatomy, Trans treatment of chronic relapsing pancreatitis, Ann Surg 14:681,
Col Phila 39:117–138, 1908. 1956.
Jedlicka R: Eine neue Operations methode der Pankreascysten, Zentrabl Lortat-Jacob J, Robert H: Hépatectomie droite réglée, Presse Med
Chir 50:132, 1923. 60:549–551, 1952.
Jourdan M: De la choledochotomie stenheil, Paris, 1895. Lucke T: Entfernung des Linken krebs Leber Lappans, Zentralbl Chir
Kausch W: Das Carcinoma der Papilla duodeni und seine radikale 18:115, 1891.
Entfernung, Beitr Klin Chir 78:29, 1912. MacBurney C: Removal of biliary calculi from the common duct by
Keen W: Report of a case of resection of the liver for the removal of a the duodenal route, Ann Surg 28:481, 1898.
neoplasm, with a table of seventy-six cases of resection of the liver Madden JL, et al: Obstructive (surgical) jaundice: an analysis of 140
for hepatic tumors, Ann Surg 30:267, 1899. consecutive cases and a consideration of choledochoduodenostomy
Kehr H: Die Behandlung der kalkulosen Cholangitis durch die direkte in its treatment, Am J Surg 109:89–99, 1965.
Drainage des Ductus Hepaticus, Munch Med Wochenschr 44:1128, Major R: History of medicine, Springfield, IL, 1954, Thomas.
1897. Makuuchi M, et al: Ultrasonically guided subsegmentectomy, Surg
Kehr H: Drei jahre gallensteinechirurgie, München, 1908a, Lehmann. Gynecol Obstet 161:346–350, 1985.
Kehr H: Gallensteine, Congr Soc Int Chir 20:341, 1908b. Mallet-Guy P: Pancréatectomie gauche pour pancréatite chronique
Kehr H: Zur Verbesserung der Hepaticus Drainage, Zbl B Chir 40:1017, récidivante, Lyon Chir 47:385–399, 1952.
1912. Marescaux J, et al: Virtual reality applied to hepatic surgery simulation:
Kehr H: Chirurgie der gallenwege, Stuttgart, 1913, Ferdinand Enke. the next revolution, Ann Surg 228:627–634, 1998.
Kemeny N, et al: Hepatic arterial infusion of chemotherapy after resec- Mayo W: Cancer of the common bile duct: report of a case of carci-
tion of hepatic metastases from colorectal cancer, N Engl J Med noma of the duodenal end of the common duct with successful
341:2039–2048, 1999. excision, St Paul Med J 3:374, 1901.
Klempnauer J, et al: What constitutes long-term survival after surgery Mazzaferro V, Regalia E, et al: Liver transplantation for the treatment
for hilar cholangiocarcinoma?, Cancer 79:26–34, 1997. of small hepatocellular carcinomas in patients with cirrhosis, N Engl
Kocher T: Ein Fall von Choledocho-Duodenostomia interna wegen J Med 334:693–699, 1996.
Gallenstein, Korrespondenbl Schweizer Arzte 1:192, 1895. Mazzariello RM: La extraccíon instrumental de los calculos biliares
Kocher T: Mobilsierung des Duodenum und Gastroduodenostomie, residuals, Bol Soc Arg Cir 27:640–660, 1966.
Zentrabl Chir 30:33, 1903. Mazzariello RM: Transcholecystic extraction of residual calculi in
König E: Die innere Anastomose in de Behandlung der Pankreascyten, common bile duct, Surgery 75:338–347, 1974.
Der Chirurg 17:24, 1946. McBride CM, Wallace S: Cancer of the right lobe of the liver: a variety
Kousnetzoff M, Pensky J: Sur la resection partielle du foie, Rev Chir of operative procedures, Arch Surg 105:289–296, 1972.
16:954, 1896. McCune WS, et al: Endoscopic cannulation of the ampulla of Vater: a
Laine J, et al: Successful liver transplantation after induction chemo- preliminary report, Ann Surg 167:752–756, 1968.
therapy in children with inoperable, multifocal primary hepatic McIver MA: An instrument for visualization of the interior of the
malignancy, Transplantation 67:1369–1372, 1999. common bile duct at operation, Surgery 9:112–114, 1941.
Lang H, et al: Extended left hepatectomy–modified operation planning Medawar PB: The behavior and fate of skin autografts and skin homo-
based on three-dimensional visualization of liver anatomy, Langen- grafts in rabbits, J Anat 78:176–199, 1944.
becks Arch Surg 389:306–310, 2004. Melendez JA, et al: Perioperative outcomes of major hepatic resections
Langenbuch C: Ein Fall von extirpation der Gallenblase wegen chro- under low central venous pressure anesthesia: blood loss, blood
nischer Cholelithiasis, Heilung Berl Klin Wocheschr 19:725, 1882. transfusion, and the risk of postoperative renal dysfunction, J Am
Langenbuch C: Ein Fall von Resektion eines linksseitig Schnur Lappens Coll Surg 187:620–625, 1998.
der Leber, Heilung Klin Wochenschr 25:37, 1888. Mercadier M: Les pancréatectomies presque totales de gauche à droite:
Langenbuch C: Ein Ruckblick auf die Entwicklung der Chirurgie des nouvelle tentative de traitement chirurgical de la pancréatite, Mem
Gallen Systems, Verh Dtsch Ges Chirurgie 68, 1896. Acad Chir (Paris) 90:84–88, 1964.
Launois B, Jamieson GG: The posterior intrahepatic approach for Mirizzi P: La colangiografia durante las operaciones de las vias biliares,
hepatectomy or removal of segments of the liver, Surg Gynecol Obstet Bol Soc Cir Cuenos Bires 16:1133, 1932.
174:155–158, 1992. Mirizzi P: Operative cholangiography, Surg Gynecol Obstet 65:702–710,
Launois B, et al: Carcinoma of the hepatic hilus: surgical management 1937.
and the case for resection, Ann Surg 190:151–157, 1979. Mizumoto R, Suzuki H: Surgical anatomy of the hepatic hilum
Lauterbur PC: Image formation by induced local interactions: exam- with special reference to the caudate lobe, World J Surg 12:2–10,
ples employing nuclear magnetic resonance, Nature 242:190–191, 1988.
1973. Mondet AF: Technic of blood extraction of calculi in residual lithasis
Leger L, Brehand J: Chirurgie de pancreas, Paris, 1956, Masson. of the choledochus, Bol Trab Soc Cir B Aires 46:278–290, 1962.
Lerut J, et al: Resection of the caudate lobe of the liver, Surg Gynecol Monprofit A: Une nouvelle methode de cholecystenterostomose: la
Obstet 171:160–162, 1990. cholecystentero-stomie en Y, Arch Provinc Chir 13:380, 1904.
Lin T: Study of a lobectomy of the liver, J Formosa Med Assoc 57:742– Morgagni G: Seats and causes of disease, facsimile of London 1769 ed, New
759, 1958. York, 1960, Hafner.
Lin TY, Chen CC: Metabolic function and regeneration of cirrhotic Morgenstern L: The murder of Johann Georg Wirsung, Surgery 57:906,
and non-cirrhotic livers after hepatic lobectomy in man, Ann Surg 1965.
162:959–972, 1965. Muhe E: Die erste Cholecystectomie durch das Laparoskop, Langen-
Link G: Pancreatostomy for chronic pancreatitis with calculi in the duct becks Arch Chir 369:804, 1986.
of Wirsung and diffuse calcinosis of the pancreatic parenchyma, Ann Muhe E: Laparoscopic cholecystectomy: late results, Langenbecks Arch
Surg 138:287, 1953. Chir (Suppl):416–423, 1991.
Lister J: On a new method of treating compound fracture, abscess, and Murray JE, Hills W: The first successful organ transplants in man,
observations on the condition of suppuration, Lancet i:326, 357, 387, J Am Coll Surg 200:5–9, 2005.
507, 1867a. Nimier H: Notes sur la chirurgie du pancréas, Rev Chir 13:41, 1893.
Lister J: On the antiseptic principle in the practice of surgery, Br J Surg Nimura Y, et al: Hepatic segmentectomy with caudate lobe resection
ii:246, 1867b. for bile duct carcinoma of the hepatic hilus, World J Surg 14:535–
Liver Cancer Study Group of Japan: Primary liver cancer in Japan: 543, 1990.
clinicopathologic features and results of surgical treatment, Ann Surg Ockinczyc P: Technique opératoire du pancréas et de la rate, Paris, 1933,
211:277–287, 1990. Doin.
15.e4 Introduction Hepatobiliary and pancreatic surgery: historical perspective
Oddi R: Effeti dell’estirpazione della cistifellca, Arq Ital Biolog 10:425, Riedel B: Erfahrungen über die gallensteinkrankheit mit und ohne ikterus,
1888. Berlin, 1892, Hirschwald.
Ohto M, Tsuchiya Y: Non-surgically available percutaneous transhe- Ringe B, et al: The role of liver transplantation in hepatobiliary malig-
patic cholangiography: technique and cases, Medicina (Tokyo) nancy: a retrospective analysis of 95 patients with particular regard
6:735–739, 1969. to tumor stage and recurrence, Ann Surg 209:88–98, 1989.
Oi I, et al: Endoscopic pancreatocholangiography, Endoscopy 2:103– Ringe B, et al: Surgical treatment of hepatocellular carcinoma: experi-
106, 1970. ence with liver resection and transplantation in 198 patients, World
Okuda K, et al: Nonsurgical, percutaneous transhepatic cholangiogra- J Surg 15:270–285, 1991.
phy: diagnostic significance in medical problems of the liver, Am J Roget C: La pancréatectomie gauche d’amont dans le traitement de la
Dig Dis 19:21–36, 1974. lithiase du canal de wirsung. thesis, Lyon, 1958.
Olthoff KM, et al: Adjuvant chemotherapy improves survival after liver Rosner F: Pigeons as a remedy (segulah) for jaundice, N Y State J Med
transplantation for hepatocellular carcinoma, Ann Surg 221:734– 92:189–192, 1992.
741, 1995. Roux C: De la gastro-entérostomie, Rev Gynécol Chir Abdom 1:67–122,
Ombredanne L: Kysto duodenostomie pancréatique, Bulletin de la 1897.
Societe Nationale de la Chirurgie 37:977, 1911. Ruggi G: Intorno ad un caso di carcinoma primitivo del páncreas,
Ong GB: Techniques and therapies for primary and metastatic liver curato e guarito con l’asportazione del tumore, Giornale Internazio-
cancer, Curr Probl Cancer 2:1–48, 1977. nale della Scienza Medica 12:81, 1890.
Pack GT, Miller TR: Middle hepatic lobectomy for cancer, Cancer Satava RM: Emerging technologies for surgery in the 21st century, Arch
14:1295–1300, 1961. Surg 134:1197–1202, 1999.
Pack GT, et al: Total right hepatic lobectomy for cancer of the gallblad- Sauve L: Des pancréatectomies et specialement de la pancréatectomie
der: report of three cases, Ann Surg 91:829–832, 1955. céphalique, Rev Chir (Paris) 37:335, 1908.
Pack GT, et al: Regeneration of human liver after major hepatectomy, Scheele J: Segment oriented resection of the liver: rationale and tech-
Surgery 52:617–623, 1962. nique. In Lygidakis NJ, Tygat G, editors: Hepatobiliary and pancreatic
Papachristou DN, Barters R: Resection of the liver with a water jet, malignancies, Stuttgart, 1989, Thieme, pp 219–246.
Br J Surg 69:93–94, 1982. Scheele J, Stangl R: Segment-orientated anatomical liver resections. In
Patel J, Couinaud C: Au sujet du “premier cas d’hépaticomie droite Blumgart LH, Fong Y, editors: Surgery of the liver and biliary tract,
réglée”, Mem Acad Chir 78:253–254, 1952a. ed 3, Edinburgh, 2000, Churchill Livingstone, pp 1743–1764.
Patel J, Couinaud C: Remarques techniques sur l’hépaticomie droite Schein CJ, Gliedman ML: Choledochoduodenostomy as an adjunct to
par ligature premiere des pedicle vasculaires, Presse Med 60:1819– choledocholithotomy, Surg Gynecol Obstet 152:797–804, 1981.
1820, 1952b. Senn N: Surgery of the pancreas as based upon experiments and clini-
Penn I: Secondary neoplasms as a consequence of transplantation and cal researches, Trans Am Surg Assoc 4:99, 1886.
cancer therapy, Cancer Detect Prev 12:39–57, 1988. Serrea P, Brunschwig A: Freezing of liver parenchyma with liquid
Penn I, Brunson ME: Cancers after cyclosporine therapy, Transplant nitrogen for hemostasis in excisional liver surgery: an experimental
Proc 20(Suppl 3):885–892, 1988. study, Cancer 8:1234–1238, 1955.
Penn I, Starzl TE: Malignant tumors arising de novo in immunosup- Sims J: Remarks on cholecystotomy in dropsy of the gallbladder, BMJ
pressed organ transplant recipients, Transplantation 14:407–417, 1:811, 1878.
1972. Singer CJ, Long ER: Translation of benivieni’s “the hidden causes of
Perissat J, et al: Gallstones: laparoscopic treatment: cholecystectomy, disease.”, Springfield, IL, 1954, Charles C. Thomas.
cholecystostomy, and lithotropsy, Surg Endosc 4:1–5, 1990. Soupault R, Couinaud C: Sur un procédé nouveau de dérivation biliare
Pichlmayr R, et al: Transplantation of a donor liver to 2 recipients intrahépatique: la cholangio-jéjunostomie gauche sans sacrifice
(splitting transplantation): a new method in the further development hépatique, Presse Med 65:1157, 1957.
of segmental liver transplantation, Langenbecks Arch Chir 373:127– Spencer WG: Translation of celsus’s “de medicina,” loeb classical library
130, 1988. ed, Cambridge, MA, 1935, Harvard University Press.
Pichlmayr R, et al: Technique and preliminary results of extracorporeal Sprengel O: Über einen Fall von Exstirpation der Gallenblase mit
liver surgery (bench procedure) and of surgery on the in situ perfused anlegung einer Communication zwischen Ductus choledochus und
liver, Br J Surg 77:21–26, 1990. Duodenum, Arch Klin Chir 42:550, 1891.
Polk W, et al: A technique for the use of cryosurgery to assist hepatic Starzl TE: The birth of clinical organ transplantation, J Am Coll Surg
resection, J Am Coll Surg 180:171–176, 1995. 192:431–446, 2001.
Praderi R: One hundred years of biliary surgery, Surg Gastroenterol Starzl TE, et al: Orthotopic homotransplantation of the human liver,
1:269, 1982. Ann Surg 168:392–415, 1968.
Priestley JT, et al: Total pancreatectomy for hyperinsulinism due Starzl TE, et al: Hepatic trisegmentectomy and other liver resections,
to islet cell adenoma: survival and cure at 16 months after Surg Gynecol Obstet 141:429–437, 1975.
operational presentation of metabolic studies, Ann Surg 119:221, Starzl TE, et al: Fifteen years of clinical liver transplantation, Gastro-
1944. enterology 77:375–388, 1979.
Pringle JH: Notes on the arrest of hepatic hemorrhage due to trauma, Storm FK, Longmire WP Jr: A simplified clamp for hepatic resection,
Ann Surg 48:514–519, 1908. Surg Gynecol Obstet 133:103–104, 1971.
Prinz R, Greenlee H: Pancreatic duct drainage in 100 patients with Strong RW, et al: Successful liver transplantation from a living donor
chronic pancreatitis, Ann Surg 194:313–320, 1981. to her son, N Engl J Med 322:1505–1507, 1990.
Quattlebaum J: Massive resection of the liver, Ann Surg 137:787–796, Tabuse K: A new operative procedure of hepatic surgery using a micro-
1953. wave tissue coagulator, Nippon Geka Hokan 48:160–172, 1979.
Quenu A, Duval P: Les angiocholites aiguës, Aiguës Congres Soc Int Chir Tait L: Cholecystotomy versus cholecystectomy, BMJ 1:1224, 1885.
20:453, 1908. Takasaki K, et al: New developed systematized hepatectomy by Glis-
Raia S, et al: Liver transplantation from live donors, Lancet 2:497, sonian pedicle transaction method, Syujutu (Operation) 40:7–14,
1989. 1986.
Rains H: Gallstones: causes and treatment, London, 1964, Heinemann. Tang ZY: Liver cancer surgery, Semin Surg Oncol 9:291–292, 1993.
Ranson JHC, et al: Prognostic signs and non-operative peritoneal Tanikawa K: Advancement of ethanol injection therapy for hepatocel-
lavage in acute pancreatitis, Surg Gynecol Obstet 143:595, 1976. lular carcinoma, Shokoki Geka 14:1389–1395, 1991.
Ravikumar TS, Kaleya RN: Cryotherapy for liver tumors. In Blumgart Tenani O: Contributo alla Chirurgia della Papilla di Vater, Policlinico
LH, Fong Y, editors: Surgery of the liver and biliary tract, ed 3, Phila- 29:291, 1922.
delphia, 2000, Saunders, pp 1577–1589. Thiersch A: Operative Drainage einer Cyste des Pankreas, Berl Klin
Rea DJ, et al: Liver transplantation with neoadjuvant chemoradiation Wochenschr 18:591, 1881.
is more effective than resection for hilar cholangiocarcinoma, Ann Thompson HH, et al: Major hepatic resection: a 25-year experience,
Surg 242:451–461, 2005. Ann Surg 197:375–388, 1983.
ReMine W, et al: Total pancreatectomy, Ann Surg 172:595–604, 1970. Thornton H: Removal of a calculus from the common bile duct—2
Rex H: Beitrage zur Morphologie der Saugerleber, Morphol Jahrb inches long and 31/2 inches in circumference—without suturing the
14:517–617, 1888. duct, Lancet 2:22, 1898.
Tiffany L: The removal of a solid tumor from the liver by laparotomy, Whipple AO: Rationale of radical surgery of cancer of the pancreas and
Md Med J 23:531, 1890. ampullary region, Ann Surg 114:612–615, 1941.
Ton TT: Les resections majeurs et mineurs du foie, Paris, 1979, Masson. Whipple AO, et al: Treatment of carcinoma of the ampulla of Vater,
Traverso LW: Carl Langenbuch and the first cholecystectomy, Am J Ann Surg 102:765–769, 1935.
Surg 132:81–82, 1976. Wilder RM, et al: Carcinoma of the islands of the pancreas, hyperin-
Vautrin A: De l’obstruction calculeuse du choledoque, Rev Chir 16:446, sulinism and hypoglycemia, JAMA 89:348–355, 1927.
1896. Wilson SM, Adson MA: Surgical treatment of hepatic metastases from
von Puestow CB, Gillespy WJ: Management of pancreatic cysts and colorectal cancers, Arch Surg 111:330–334, 1976.
pancreatic lithiasis, Am Surg 20:355–362, 1958. Wood CB, et al: A retrospective study of the natural history of patients
von Winiwarter A: Ein Fall von Gallenretention bedingt durch imper- with liver metastases from colorectal cancer, Clin Oncol 2:285–288,
meabilität des Ductus choledochus anlegung einer Gallenblasen 1976.
Darmfistel, heilung, Prager Mediz Wonchensf 7:201, 1882. Wood M: Presidential address: eponyms in biliary tract surgery, Am J
Voyles CR, Blumgart LH: A technique for the construction of high Surg 138:746–754, 1979.
biliary-enteric anastomoses, Surg Gynecol Obstet 154:885–887, 1982. Wood TF, et al: Radiofrequency ablation of 231 unresectable hepatic
Walker R: Francis Glisson and his capsule, Ann R Coll Surg Engl tumors: indications, limitations, and complications, Ann Surg Oncol
38:71–91, 1966. 7:593–600, 2000.
Warren KW, Jefferson MP: Prevention and repair of strictures of the Yamaoka Y, et al: Liver transplantation using a right lobe graft from a
extra-hepatic bile ducts, Surg Clin North Am 152:797–804, 1973. living related donor, Transplantation 57:1127–1130, 1994.
Warvi W: Primary tumors of the liver, Surg Gynecol Obstet 80:643–650, Zeevi A, et al: Immunosuppressive effect of FK-506 on in vitro lym-
1945. phocyte alloactivation: synergism with cyclosporine A, Transplant
Weigall A: Alexander the great, London, 1933, Putnam’s Sons. Proc 19(Suppl 6):40–44, 1987.
Wendell W: Beiträge zur Chirurgie der Leber, Arch Klin Chir 95:887– Zhou XD, et al: Clinical evaluation of cryosurgery in the treatment of
892, 1911. primary liver cancer: report of 60 cases, Cancer 61:1889–1892, 1988.
CHAPTER 1
Embryologic development of the liver, biliary tract,
and pancreas
Yousef El-Gohary, Kai Zhao, and George K. Gittes
OVERVIEW OF LIVER AND BILIARY derm and the surrounding mesenchyme. The primitive gut
TRACT DEVELOPMENT tube, which is derived from the endodermal germ layer during
gastrulation, is divided into the foregut, midgut, and hindgut
Developmental biologists have often marveled at the enormous domains, each of which gives rise to specialized regions (Grapin-
regenerative capacity of the liver after injury, and such growth Botton, 2005). This specialization is initiated by transcription
represents one of the fastest by mammalian tissues. Presum- factors expressed in different regions (Table 1.1). For example,
ably, this process is based on the recapitulation of embryonic the coexpression of pancreatic and duodenal homeobox gene 1
signals in the liver, but our understanding of the mechanisms (PDX1) and pancreas-specific transcription factor (PTF1A) in
is relatively poor. The liver is considered the largest internal the endoderm gives rise to the pancreas (see Table 1.1;
organ and consists of diverse cell types that arise from various Fig. 1.2) (Grapin-Botton, 2005; Moore-Scott et al, 2007) (see
embryologic origins. It is a vital organ that has an array of Table 1.1). The definitive endoderm is formed at the ventral
diverse functions, including endocrine, exocrine, and essential side of the vertebrate embryo during gastrulation. Evagination
metabolic functions. The two principal cell types of the liver of the endoderm at the anterior end of the embryo generates
are the hepatocytes, which comprise nearly 70% of the mass of the ventral foregut, which will eventually give rise to the liver,
the adult organ and are responsible for the majority of the lung, thyroid, and the ventral pancreas. The dorsal region of
metabolic liver functions, and the cholangiocytes. Both cells are the definitive endoderm develops into the intestines and the
derived from the embryonic endoderm. Other cell types of the dorsal pancreas (see Fig. 1.2).
liver include hematopoietic, Kuppfer, stromal, and stellate This complex dialogue between the endoderm and meso-
cells, which are of mesodermal origin (Fig. 1.1). Despite their derm appears to be critical for the final patterning of the gut
homogenous appearance, hepatocytes do not all function iden- tube, where several signaling pathways have been implicated in
tically; they perform various tasks depending on their physical the regulation of foregut endoderm development, promoting
location within a hepatic lobule, the primary functional unit of organ specification along its anterior-posterior axis for organs
the liver. For instance, periportal hepatocytes are responsible such as the thyroid, lung, liver, and pancreas. The plasticity of
for the urea cycle enzymes, whereas pericentral hepatocytes the endoderm was demonstrated by experimentally recombin-
express glutamine synthase and utilize ammonia to generate ing the posterior mesoderm with early foregut endoderm,
glutamine. Thus, liver development entails not only hepatocyte leading to repression of liver and pancreatic development in
and cholangiocyte differentiation but also cellular differentia- favor of intestinal development (Kumar et al, 2003; Wells &
tion within the hepatocyte population (see Fig. 1.1). Melton, 2000).
During the third week of gestation, liver primordium first Specific signals from the surrounding mesoderm in the
appears as an outgrowth of the ventral foregut endoderm at the foregut region lead to hepatic specification and subsequent
caudal end of the foregut. The proliferation of the epithelial morphogenesis (Gualdi et al, 1996). The molecular pathway
cells in this liver bud leads to its outgrowth and branching into linking endodermal patterning to the initiation of liver and
the surrounding mesenchyme, giving rise to the liver and intrapancreatic development has been partially elucidated by recent
hepatic biliary tree. As it grows caudally, traversing the septum studies in the Xenopus model, supporting a role for FGF4 and
transversum, the persistent connection between the branching WNT signaling from the posterior mesoderm in inhibiting
epithelium and the foregut develops into the extrahepatic bile foregut fate and promoting hindgut formation; these signals are
ducts and gallbladder (Sadler & Langman, 2006). The bipo- inhibited in the anterior endoderm to allow foregut develop-
tential hepatoblasts eventually differentiate into hepatocytes ment (McLin et al, 2007).
and cholangiocytes. The final liver structure continues to The β-catenin signaling pathway is essential in liver and
develop through the postnatal period. In addition to giving rise pancreas development. Specifically, repression of β-catenin (a
to the liver and biliary tract, the proximal endoderm (foregut) downstream mediator of canonical WNT signaling) in endo-
also gives rise to respiratory epithelium and the glandular cells derm is necessary to initiate reciprocal signaling from the meso-
of the thyroid, thymus, and pancreas due to its pluripotent derm, leading to hepatic induction. This role for β-catenin is
nature (Fig. 1.2). illustrated by the fact that forced β-catenin expression in the
anterior endoderm leads to downregulation of the hematopoi-
ENDODERMAL PATTERNING etically expressed homeobox gene (HHEX) and inhibition of
liver formation, whereas forced repression of β-catenin in the
Embryologically, the liver, biliary tract and pancreas develop posterior endoderm (future hindgut that normally expresses
through a series of reciprocal interactions between the endo- β-catenin) induced ectopic HHEX expression and ectopic liver
17
18 PART 1 LIVER, BILIARY, AND PANCREATIC ANATOMY AND PHYSIOLOGY
bud initiation (McLin et al, 2007). HHEX is a target of et al, 2009); and intestine (ENDOCUT) (Costa et al, 2003).
β-catenin and is one of the earliest foregut markers (Thomas Conversely, inhibiting β-catenin in the posterior mesoderm led
et al, 1998), which is essential for normal liver and ventral to ectopic expression of other liver and pancreas markers
pancreas development in mice (Bort et al, 2004; Keng et al, (FOR1, PDX1, elastase, and amylase) along with the inhibition
2000). In addition to repressing HHEX, β-catenin also down- of the intestinal marker ENDOCUT (McLin et al, 2007).
regulates other foregut markers for liver (FOR1) (Seo Conversely, inhibiting the WNT receptor FZD7 resulted in
et al, 2002); pancreas (PDX1); lung/thyroid (NKX2.1) (Goss loss of the liver primordium in Xenopus embryo. This result
indicates that low levels of WNT signaling are necessary to
maintain foregut fate, with hepatic specification a dynamic
process.
Stem cell
Studies in mouse and chick models demonstrate the need
for FGF2 signals from the cardiogenic mesoderm and bone
morphogenic proteins (BMPs) from the septum transversum
mesenchyme (Rossi et al, 2001; Zhang et al, 2004) for liver
Ectoderm Mesoderm
development. In addition, WNT2b has also been shown to be
Endoderm
Midgut Foregut Hindgut TABLE 1.1 Specialization of the Different Domains of the
Primitive Gut Tube is Initiated by Organ-Specific Transcription
Factors Expressed in Different Regions
Pancreas Lung Gallbladder
Liver Transcription Factor Organ
SOX 2 Esophagus and stomach

NKX2.1 Lungs
Islets
PDX1 and PTF1A Pancreas
Acinar and ducts Hepatocyte Biliary epithelium
HHEX Liver
FIGURE 1.1. Cell lineage schematic for hepatic and pancreatic devel- CDXC Small bowel
opment from a multipotent progenitor stem cell. (Courtesy Mariam CDXA Large bowel and rectum
Hobeldin.)
Notochord
Spinal cord
Forebrain
Primitive foregut tube
Lungs and thyroid: NKX2.1 Pericardial sac
Esophagus and stomach: SOX2
Pancreas: PDX1
Liver: HHEX
Large intestine: CDXA

mariam hobeldin
Small intestine: CDXC

Signaling gradient: WNT + FGF4
FIGURE 1.2. Schematic representation of the regional specification of the primitive foregut tube (E8.5) with the various transcription factors. WNT
and FGF4 (a fibroblast growth factor) are secreted from the posterior mesoderm in a gradient and repress foregut fate and promote hindgut develop-
ment. WNT and GFG4 inhibition in the anterior endoderm promote foregut fate. (Courtesy Mariam Hobeldin.)
Chapter 1 Embryologic development of the liver, biliary tract, and pancreas 19
an important signal for liver specification in zebrafish (Ober expression was seen (Lee et al, 2005a). This result indicates the
et al, 2006). necessity for FOXA1 and FOXA2 for hepatogenesis. In vivo
Other examples of transcription factors implicated in foregut DNA-binding studies revealed that the liver-specific ALB has
patterning include the HOX genes. HOXD13 is expressed in an important upstream regulatory binding site for FOXA factors
the hindgut of mouse and chick embryos (Dolle et al, 1991; (Bossard & Zaret, 1998; Gualdi et al, 1996). Before FOXA or
Roberts et al, 1995; Yokouchi et al, 1995). Knock out mice GATA4 binding, the ALB gene is transcriptionally silent, with
homozygous and null mutated for HOXD13 develop with a closed chromatin. After binding with FOXA and GATA4, the
hindgut defects (Kondo et al, 1996). Similarly, when HOXD13 chromatin domain is thought to become exposed (Cirillo et al,
is misexpressed in chick embryo midgut mesoderm, hindgut 2002), thus increasing the ability of the gene to be activated. By
features are seen in the midgut epithelium (Roberts et al, 1998). the E9.5 gestation age, other transcription factors CCAAT/
The exact mechanism of HOX genes in inducing endodermal enhancer binding protein-β and nuclear factor 1 bind to sites
patterning remains unknown, as the exact downstream targets adjacent to the FOXA site, and as a result, the albumin gene
are not clear. In addition to HOX genes, retinoic acid signaling becomes active (Zaret, 2002). Therefore, FOXA binding to
has been implicated in this foregut patterning, probably through chromatin is the critical step in hepatic competence by increas-
regulating HOX gene expression. Specifically, retinoic acid sig- ing gene expression and allowing binding of other transcription
naling appears to specify the position of the pancreatic domain factors.
in the foregut of Xenopus, zebra fish, and mice (Chen et al,
2004; Huang et al, 1998; Martin et al, 2005; Stafford et al, Hepatic Induction
2004). Goss and coworkers (2009) demonstrated that WNT2 FGF from the cardiogenic mesoderm and BMPs from the
and WNT2b play an essential role in specifying lung endoderm septum transversum mesenchyme (STM) have been implicated
without affecting the specification of other foregut-derived in the induction of liver fate in mouse and chick embryonic
tissues. WNT2/2b double knock out mutants display complete endoderm. In vitro studies demonstrated that when ventral
lung agenesis. NKX2.1, the earliest known transcription factor foregut and cardiogenic mesoderm were cocultured in the pres-
marker for the developing lung, was also lost from the anterior ence of fibroblast growth factor inhibitors, liver induction was
foregut, confirming the loss of trachea and lung development. inhibited (Jung et al, 1999). Culturing foregut endoderm
Despite our growing knowledge of molecular signals that without the cardiogenic mesoderm in the presence of low con-
govern endodermal patterning of the primitive foregut tube, centrations of exogenous FGF2 (2 to 5 ng/mL) rescued hepatic
most of the pathways remain poorly understood; exactly how gene expression (Jung et al, 1999) and simultaneously sup-
these pathways confer regional identity remains unresolved. pressed the expression of the pancreas program (PDX1)
(Deutsch et al, 2001), whereas a high concentration of FGF2
Hepatic Competence (10 to 500 ng/mL) induced NKX2-1, an early marker for respi-
Hepatic competence, or the ability to form liver from the foregut ratory epithelium, but not albumin gene expression (Serls et al,
endoderm, is considered the first of a two-step process for the 2005). When foregut endoderm was cocultured with noggin
specification of liver in vertebrates. Competence is a prerequi- (NOG), a BMP inhibitor, inhibited hepatic gene induction was
site for the endoderm to respond to specific signals, such as observed, an effect that was reversed when BMP2 or BMP4
FGFs from the surrounding mesoderm, which then leads to the were added. Despite these results, however, embryos that were
second step, the induction of liver-specific genes such as albumin homozygous mutants for BMP4 still exhibited normal hepatic
(ALB), α-fetoprotein (AFP), and hepatocyte nuclear factor 4α gene induction (Rossi et al, 2001; Smith & Harland, 1992).
(HNF4α). This “competence” is facilitated through the FOXA These data indicate that the cardiac mesoderm, a source for
gene transcription factor family that includes FOXA1 and FGF signaling, is crucial for the hepatic induction from the
FOXA2 (forkhead box proteins A1 and A2) and the GATA- ventral foregut endoderm and subsequent morphogenesis,
binding proteins 4 and 6. FOXA gene expression precedes the whereas BMP aides in the process (Fig. 1.3).
induction of the hepatic program by FGF signals in the endo-
derm, and FOXA2 binding reverses chromatin-mediated Morphogenesis of the Hepatic Bud
repression of AFP gene transcription in vitro (Crowe et al, Following hepatic specification (E8.5 to E9.0), the “liver” starts
1999). Deletion of FOXA1 and FOXA2 led to the absence of to express liver-specific genes (ALB, AFP, HNF4α), and even-
liver bud formation with the loss of AFP expression in the tually form the liver bud. Hepatic bud morphogenesis is facili-
ventral foregut, which indicates that hepatic specification had tated by two transcription factors, HHEX (hematopoietically
failed to initiate (Lee et al, 2005a). Similarly, deletion of either expressed homeobox, discussed earlier) (Crompton et al, 1992)
Gata4 or Gata6 in mouse embryos resulted in failed liver expan- and PROX1 (prospero-related homeobox 1) (Oliver et al,
sion, with the hepatic endoderm still expressing hepatic genes 1993). HHEX is expressed in the anterior endoderm at E7.0,
(Watt et al, 2007; Zhao et al, 2005). The forkhead box (FOX) which eventually gives rise to the liver as well as the ventral
proteins are an extensive family of transcription factors that pancreas. HHEX-null embryos grow without a liver or thyroid
share homology with the winged helix/fork head DNA-binding and develop forebrain defects at E11.5; however, evidence of
domains (Kaestner et al, 2000) that play important roles in endodermal epithelial thickening suggests a possible defect in
regulating expression of genes involved in cellular differentia- differentiation as evidence of possible hepatic induction (Mar-
tion, proliferation, transformation, and metabolic homeostasis tinez Barbera et al, 2000).
(Duncan, 2000; Wang et al, 2001; Zaret, 1999). In vitro explant In a separate study on HHEX-null embryos, liver genes ALB
cultures demonstrated that FGF was sufficient to induce ventral and PROX1 were expressed in the ventral foregut endoderm at
foregut endodermal cells to differentiate into hepatoblasts around E8.5, with the thickening of the hepatic endoderm
(Gualdi et al, 1996). However, when FOXA1/FOXA2-deficient region at E9.0 being smaller compared with heterozygote
endoderm was cultured with exogenous FGF2, no liver embryos. In addition, a significantly lower proliferation rate was
STM The role of endothelial cells in liver organogenesis has also

BMP been studied, with the liver vasculature providing a vital source
VEGFR2 for hematopoiesis in early life. Endothelial cells are positioned
HGF
c-MET as a loose necklace of cells interceding between the thickening
2 hepatic epithelium and the STM (Matsumoto et al, 2001).
Embryos that were null mutants for VEGFR2 (also known as
FOREGUT KDR) had failure of delamination and subsequent migration by
ENDODERM the hepatoblasts (Matsumoto et al, 2001). These results imply
HEX that the endothelial cells interact with nascent hepatic cells and
PROX1 aid in liver bud outgrowth.
β1-Integrin
HLX
SMAD 2/3
Liver Bud Growth
The STM cells are closely related to the ventral endoderm and
mariam hobeldin
FGF
contribute to hepatic induction and growth. This epithelial-
mesenchymal interaction is essential for liver bud formation,
1 expansion and differentiation. During this phase (E9.5 to E15),
FIGURE 1.3. Schematic for liver ontogeny controlled by factors the liver bud undergoes a tremendous amount of growth and
released from the cardiogenic mesenchyme 1, and septum transversum becomes an important site for hematopoiesis. Signals regulating
mesenchyme (STM, 2, which induce hepatic specification. Signaling this stage arise from both the hepatic mesenchyme and the STM.
molecules (eg, VEGFR2, HGF) and transcription factors (eg, PROX, HLX) These signals include FGF and BMP, which promote liver
from STM and endothelial cells leads to delamination and migration of growth in addition to aiding in hepatic specification. BMP4 is
endodermal cells into the STM to subsequently form the liver bud. BMP,
strongly expressed in the STM and continues to be expressed at
Bone morphogenetic protein; HGF, hepatocyte growth factor; c-MET,
E9.0, during which the liver bud migrates into the STM (Rossi
HGF receptor; VEGFR2, vascular endothelial growth factor receptor 2.
(Courtesy Mariam Hobeldin.) et al, 2001). BMP4-null mutant mice embryos had a delay in the
growth of the liver bud, which indicates that BMP constitutes
an important growth signal for the liver (Rossi et al, 2001).
Another factor that has been implicated in liver growth is
seen in the prospective hepatic domain when compared with HLX, which is expressed prominently in the visceral mesen-
the control group, demonstrated by bromodeoxyuridine (BrdU) chyme (STM), into which the liver will expand (Hentsch et al,
staining, with no evidence apoptosis (Bort et al, 2004). The 1996). HLX-null embryos exhibited severe liver hypoplasia
basal membrane layer, which is rich in laminin, surrounds the without affecting liver specification (Hentsch et al, 1996). A
hepatic endoderm and degrades around E9.0 to E9.5, so that similar finding was observed in embryos that were null mutants
the hepatocytes start migrating into the STM to form the liver for hepatocyte growth factor (HGF) (Schmidt et al, 1995). In
bud. This degradation is facilitated by the hepatoblasts, which contrast to HLX−/− embryos, apoptosis was the underlying
under normal conditions downregulate E-cadherin. However, cause for the severe liver hypoplasia in HGF-null embryos. HGF
in the PROX1-null mutant embryos, the progenitor liver cells is produced by the cells lining the sinusoids, which are of mes-
fail to migrate into the STM as a result of excess E-cadherin enchymal origin, mediating its effects through the c-MET tyro-
and basement membrane proteins laminin and collagen 4. The sine kinase receptor produced by hepatocytes (Schmidt
basal lamina fails to degrade, and the cells remain trapped in et al, 1995). Transforming growth factor-β (TGF-β) signaling
the hepatic diverticulum, with an overall reduction in liver size is also involved in mediating their signals through SMAD2 and
(Sosa-Pineda et al, 2000). The bulk of the liver lobe lacks SMAD3, which translocate to the nucleus to either upregulate
hepatocytes, suggesting that the mesenchymal component con- or downregulate gene expression. Embryos that were heterozy-
tributes most of the liver mass in PROX1-null mutant embryos. gous mutants for SMAD2 and SMAD3 exhibited severe liver
A similar phenotype is seen with ONECUT1 (also called HNF6) hypoplasia as a result of a decrease in β1-integrin expression
and ONECUT2 double mutants, which are required for basal (Weinstein et al, 2001). Intriguingly, this phenotype was rescued
lamina degradation (Margagliotti et al, 2007). Furthermore, a when HGF, a potent hepatotrophic growth factor, was added
pharmacologic inhibition of matrix metalloproteinases (MMPs), to the culture medium. Presumably, this was a result of β1-
extracellular matrix remodeling enzymes usually expressed by integrin expression (Weinstein et al, 2001), which plays an
the hepatoblasts and STM cells, inhibit hepatoblast migration important role in hepatocytic adhesion to the extracellular
in culture (Margagliotti et al, 2008). To further illustrate the matrix; TGF-β and HGF are known to induce β1-integrin
importance of the extracellular matrix in hepatic bud morpho- expression (Kagami et al, 1996; Kawakami-Kimura et al, 1997).
genesis, hepatoblasts deficient in laminin receptor β1-integrin
fail to colonize the liver bud (Fassler & Meyer, 1995). These OVERVIEW OF HEPATOBLAST
β1-integrins are among those that act as receptors for extracel- DIFFERENTIATION
lular matrix proteins, such as laminins and collagens (Hynes,
1992). Hepatoblasts begin to differentiate into mature hepatocytes and
In summary, PROX1 and ONECUT factors are important biliary epithelial cells at about E13.5. Before differentiation, the
in regulating delamination and in controlling hepatoblast hepatoblasts express genes for adult hepatocytes (albumin,
migration through regulating MMPs and hepatoblast interac- HNF4α), biliary epithelial cells (KRT19), and fetal liver (AFP)
tions with the extracellular matrix. Without an appropriate (Lemaigre, 2003; Shiojiri et al, 1991).
extracellular matrix, cell migration into the STM will be Hepatoblasts in proximity to the portal vein form a bilayer
disrupted. architecture and eventually differentiate into biliary epithelial
E12 HEPATOBLASTS
AFP/ALB
HNF4
CK19
PORTAL
MESENCHYME IMMATURE IMMATURE
BILE DUCT CELLS HEPATOCYTE
Jagged-1/Notch CK19 ALB
EGF HNF6 HNF4 HNF1α
HGF HNF1β C/EBP
FOXF1
E18
OSM
DEX
HGF
WNT
BILE DUCT MATURE

CELLS HEPATOCYTE
CK19 ALB
HNF4
C/EBP
PERINATAL
mariam hobeldin
FIGURE 1.4. Overview of differentiation of hepatoblasts into mature hepatocytes and bile duct cells. Bipotential hepatoblasts initially express markers
for both hepatocytes (ALB, AFP) and biliary epithelial cells (KRT19). Hepatoblasts in the periportal mesenchymal area upregulate KRT19 and down-
regulate hepatogenic transcription factors (HNF4, CEBP) to form mature biliary epithelial cells, further signaling from the portal mesenchyme (NOTCH,
HGF) and duct remodeling. The rest of the hepatoblasts upregulate hepatogenic factors. Other factors (OSM, dexamethasone) help in maturation of
mature hepatocytes. HGF, Hepatocyte growth factor; EGF, epidermal growth factor; OSM, oncostatin M; DEX, dexamethasone; ALB, albumin;
AFP, α-fetoprotein; HNF, hepatocyte nuclear factor. (Courtesy Mariam Hobeldin.)
cells by upregulating biliary specific cytokeratin-19 (KRT19) are derived from bipotential hepatoblasts that can differentiate
and downregulating the other hepatic genes. This bilayer into either hepatocytes or biliary epithelial cells. This theory is
around the portal vein begins to form focal dilatations incorpo- based on the observation that immature hepatoblasts coexpress
rated into the portal mesenchyme to form intrahepatic biliary markers of both hepatocytes (ALB) and biliary epithelial cells
duct at E17.0 until birth. Areas of the ductal bilayer plate not (KRT19). The biliary specific marker KRT19 becomes strongly
involved in the formation of the ducts progressively regress. expressed at a later gestational age, as the cells become ductal
The remaining hepatoblasts differentiate into mature hepato- cells, whereas other cells transiently express the hepatocyte
cytes, arranged in hepatic chords with the bile canaliculi on the markers ALB and AFP as they develop into mature hepatocytes
apical surfaces (Lemaigre, 2003). In the mature hepatobiliary (Shiojiri et al, 1991). This theory was further supported when
system, the bile is produced by the hepatocytes and is secreted embryonic liver, before the intrahepatic biliary ducts form, was
into the canaliculi, which are connected to the network of transplanted into the testis of syngeneic animals, giving rise to
intrahepatic biliary ducts. The bile then flows to the hepatic both hepatocytes and typical bile ducts (Shiojiri et al, 1991).
ducts, transits through the cystic duct, and is stored in the Suzuki and colleagues (Suzuki et al, 2002) used in vitro studies
gallbladder; eventually, the bile is excreted into the bowel via on hepatic “stem” cells from E13 embryonic livers, identified
the common bile duct. The biliary epithelial cells delineate the with self-renewing capability and multilineage differentiation
lumen of the intrahepatic, extrahepatic biliary tree (hepatic, potential, to demonstrate that these cells could form differenti-
cystic and common bile duct), and the gallbladder (Fig. 1.4). ated hepatocytes, biliary epithelial cells, pancreatic, and intes-
tinal cells.
Biliary Epithelial Cell Differentiation and Formation The first step in triggering the initiation of the transition
of the Ductal Plate from hepatoblast to a biliary epithelial cell is thought to be
The exact origin of the biliary epithelial cells has been greatly facilitated through the ONECUT transcription factor hepato-
debated; however, the popular school of thought is that they cyte nuclear factor 6 (HNF6), which is expressed in the biliary
epithelial cells of the developing intrahepatic biliary ducts and for the development of organs derived from foregut endoderm
in hepatoblasts, gallbladder primordium, and the extrahepatic such as the pancreas, liver, gallbladder and lung (Kalinichenko
bile ducts (Landry et al, 1997; Rausa et al, 1997). HNF6−/− et al, 2002). FOXF1 expression is restricted to the gallbladder
embryos displayed severe biliary anomalies. Extrahepatically, mesenchyme and STM. In FOXF1+/−embryos, the gallblad-
this mutation resulted in the absence of gallbladder and the der develops severe structural abnormalities with significant
normal bile ducts; instead, there was an enlarged structure reduction in size with reduced mesenchymal cell numbers, an
connecting the liver to the duodenum. Intrahepatically, absent biliary epithelial cell layer, and a deficient external
however, it caused abnormal differentiation of biliary epithelial smooth muscle layer. The reduction in mesenchymal cell
cells, resulting in cholestasis. Closer histologic examination numbers was attributed to the reduction in vascular cell adhe-
revealed an increased number of KRT19-positive cells com- sion molecule and cell adhesion α5-integrin, both of which
pared with control cells at E13.5, with the development of are essential for mesodermal formation (Mahlapuu et al,
abnormal large cysts at E15.5 to E16.5 that contained an epi- 2001; Yang et al, 1993). Defective smooth muscle layer for-
thelium of KRT19-expressing cells. These abnormal cysts are mation was attributed to diminished levels of platelet-derived
similar to those seen in Caroli disease, an autosomal recessive growth factor receptor α, which is essential for smooth muscle
disorder with ductal plate malformation and ectasias. The cell differentiation (Jacob et al, 1994). FOXF1 is not expressed
abnormal increase in KRT-positive cells at E13.5 lacked any in intrahepatic biliary duct mesenchyme in wild-type mice,
proliferative marker, suggesting that they are postmitotic, and no defects were seen in the intrahepatic biliary ducts of
resulting from hepatoblasts that have differentiated toward a FOXF1+/− mice; however, FOXF1 mRNA levels in the liver
biliary lineage prematurely (Clotman et al, 2002). In addition, were increased, suggesting that it may be a compensatory
the excess KRT-positive biliary epithelial cells formed cordlike mechanism to prevent defects in the liver (Kalinichenko et al,
extensions within the liver parenchyma, compared with the 2002). All these suggest that FOXF1 is crucial for the devel-
control group, which is restricted to the vicinity of the portal opment of the extrahepatic biliary duct and gallbladder with
vein. This observation supports the role of HNF6 in control- the mesenchyme playing an essential role in biliary epithelial
ling the differentiation of hepatoblasts into biliary epithelial cell differentiation. The interaction between the biliary epithe-
cells and the morphogenesis of the intrahepatic biliary ducts, lial cell and the extracellular matrix is also thought to contrib-
confining biliary epithelial cells to the periportal area. A ute to biliary epithelial cell differentiation.
similar morphologic defect of intrahepatic biliary ducts was Integrins are membrane receptors for extracellular matrix
observed in HNF1β−/− embryos, which suggests that HNF6 proteins where they play an important role in mediating the
controls intrahepatic biliary duct development via HNF1β interaction between differentiating cells and the extracellular
(Clotman et al, 2002). In contrast to the intrahepatic ducts matrix (Couvelard et al, 1998; Hynes, 1992). Hepatoblasts
that are derived from bipotential hepatoblasts, the cholangio- express integrin heterodimers containing the β1 subunit (α1β1,
cytes that line the extrahepatic bile ducts are derived from the α5β1, α6β1, and α9β1), and when hepatoblasts differentiate
common ventral pancreatobiliary bud. The bile duct fate of into immature intrahepatic biliary epithelial cells while in
these primitive embryonic bud–derived cells is determined by contact with the mesenchyme, the morphology of the integrins
the transcription factor SOX17, which is coexpressed with changes. The primitive intrahepatic biliary epithelial cells
PDX1 in these pancreaticobiliary progenitor cells (Spence et upregulate α6β1 expression and loses α1β1 while acquiring
al, 2009). The cell fate decision between pancreas-lineage several other integrin dimmers not previously expressed on
PDX1-positive cells versus biliary primordium SOX17- hepatoblasts such as α2β1, α3β1, αVβ1, and α6β4 (Couvelard
positive cells is determined by hairy and enhancer of split 1 et al, 1998). Intrahepatic biliary epithelial cells are contacted
(HES1) (see later in the PANCREAS section). Deleting by a basement membrane containing collagen, enactin, and
SOX17 at E8.5 resulted in ectopic expression of pancreatic laminin (Desmet, 1985); however, hepatocytes are surrounded
tissue in the common bile duct with PDX1-positive cells in by the perisinusoidal matrix, which is devoid of laminin or
the liver bud along with the loss of biliary structures. Con- enactin (Schuppan, 1990). The increase in α6β1 expression
versely, overexpression of SOX17 suppressed pancreas devel- along with the acquisition of biliary-specific expression of α2β1,
opment and promoted ectopic biliary-like tissue in the α3β1,and α6β4, which are integrin receptors for laminin, cor-
PDX1-positive domain tissue (Spence et al, 2009). Further- related with the deposition of laminin at the contact points of
more, it has been demonstrated that SOX17 regulates insulin the portal mesenchyme with the ductal plate (Couvelard et al,
secretion postnatally, with the mice becoming prone to devel- 1998).
oping diabetes with the deletion of SOX17 gene in the pan-
creas (Jonatan et al, 2014). Remodeling of the Ductal Plate
With regard to mesenchymal-epithelial induction of liver As mentioned earlier, the double-layered ductal plate around
primordium and gallbladder, studies have suggested that the the portal vein begins to form focal dilatations incorporated
mesenchyme contributes to biliary epithelial cell differentia- into the portal mesenchyme to form intrahepatic biliary ducts,
tion, where differentiation of hepatoblasts into biliary epithe- while the parts of the ductal bilayer plate not involved in the
lial cells was stimulated when cocultured with hepatic or lung formation of the ducts progressively regress. This mechanism
mesenchyme (Shiojiri & Koike, 1997). It was also noted in the of regression is thought to be carried out by apoptosis (Sergi
study of HNF6−/− mice that biliary epithelial cell differentia- et al, 2000; Terada & Nakanuma, 1995). Cell-matrix interac-
tion occurred at the interface between the portal mesenchyme tions are also thought to contribute to the remodeling process.
and the liver parenchyma (Clotman et al, 2002). Tenascin was found to be expressed in the mesenchyme around
A recent study revealed that the forkhead box f1 (FOXF1) the biliary epithelial cells of primitive ducts migrating into the
transcription factor may play an important role in the mesenchyme. In contrast, tenascin was absent in the mesen-
mesenchymal-epithelial signaling, an interface that is required chyme of peripheral ducts (Terada & Nakanuma, 1994).
based on observations of a number of human diseases termed

ENDOTHELIAL “ductal plate malformations.” These diseases include biliary
CELLS atresia, Caroli disease, and Meckel and Alagille syndromes, in
which abnormal biliary ducts are associated with anomalies of
E12 the portal mesenchyme and of the portal blood vessels (Lemai-
gre, 2003). This association was also demonstrated in studies
HEPATOBLAST
on NOTCH pathway defects. In Alagille syndrome, an auto-
somal dominant disease, bile ducts are absent in the portal
tract, associated with an increased number of arteries and fibro-
sis. Haploinsufficiency of Jagged-1 (JAG1), a NOTCH recep-
PORTAL
MESENCHYME tor ligand, is associated with Alagille syndrome, where JAG1 is
persistently expressed in the ductal epithelium in humans (Li
et al, 1997; Louis et al, 1999). It is also expressed in the endo-
thelial cells of the developing portal vasculature (Crosnier et al,
2000). The animal model for Alagille syndrome was replicated
E14 in double-heterozygous mice for mutations in the JAG1 and
BILIARY NOTCH2 genes (J1N2+/−). JAG1 protein was expressed in the
EPITHELIUM hepatic vasculature, and NOTCH2 was expressed in a subset
of hepatoblasts surrounding the portal vein, hepatic artery, and
bile ducts. Interestingly, although neither JAG1 nor NOTCH2
protein was expressed in the bile duct epithelium of these mice,
JAG1 is expressed in ductal epithelium in humans (Louis et al,
1999; McCright et al, 2002). The differences in JAG1 expres-
IMMATURE
HEPATOCYTE
sion between human and mouse most likely reflect species
specificity rather than technical artifacts, as mice that are het-
E18 erozygous or homozygous for the JAG1 gene did not show the
liver symptoms of Alagille syndrome (Xue et al, 1999). Similar
FOCAL
biliary abnormalities were observed in NOTCH2-null mutant
DILATATIONS mice (McCright et al, 2002).
Hepatocyte Differentiation
During later stages of development, hepatocytes undergo a
transition period from a hematopoietic support role to a mature
MATURE adult hepatocyte. This change occurs under the control of the
PERINATAL HEPATOCYTE transcription factor CEBP with HNF4, the latter being a crucial
mariam hobeldin
BILE
factor in hepatocyte differentiation. Loss of HNF4 function led
DUCT to the disruption of the expression of several genes associated
with a mature hepatocyte phenotype. In HNF4−/− mice, hepa-
tocytes failed to express many mature hepatic enzymes and
FIGURE 1.5. Overview of intrahepatic bile duct formation around the lacked normal morphology, leading to low glycogen storage,
portal vein. Hepatoblasts in the portal mesenchyme vicinity begin to disrupted sinusoids, and gap junction disruption. These results
acquire the biliary epithelial cell marker (KRT19) and downregulate demonstrate that HNF4 is essential for liver differentiation.
hepatic genes, first forming a single layer at E14, then a bilayer at E18 Other factors that have been shown to promote hepatocyte
with focal dilatations that eventually form the intrahepatic bile ducts. The differentiation include oncostatin M (OSM), an interleukin-6
rest of the bilayer regresses. (Courtesy Mariam Hobeldin.) (IL-6) family cytokine, HGF, and WNT (Michalopoulos et al,
2003; Tan et al, 2008). Although as discussed earlier in the
chapter, the repression of WNT signaling in the foregut endo-
Tubulogenesis of ductal cells is thought to be contributed derm is necessary for hepatic specification, its role is reversed
by soluble factors secreted from hepatocytes or biliary epithelial at later gestational stages to promoting hepatocyte differentia-
cells. When coculturing human biliary epithelial cells with tion. β-Catenin, a central component of the canonical WNT
hepatocytes, a marked ductular morphogenic response was pathway, is essential for normal development; its aberrant acti-
induced, and the biliary epithelial cells formed well-organized vation in liver was associated with tumors, including hepatic
luminal ducts. This result was reproduced when biliary epithe- adenomas and hepatocellular cancers (de La Coste et al, 1998;
lial cells were grown in a conditioned medium from previous Peifer & Polakis, 2000; Zucman-Rossi et al, 2006). Livers that
hepatocyte and biliary epithelial coculturing (Auth et al, 2001); were deficient in β-catenin displayed decreased numbers of
however, it remains unclear whether soluble factors contribute hepatocytes, with hepatoblasts that lacked maturation, prolif-
to tubulogenesis in in vivo studies (Fig. 1.5). eration, and function (Tan et al, 2008). In addition, CEBPα,
a fundamental regulator of hepatocyte differentiation and mat-
Developmental Relationship Between the Ducts, uration (Tan et al, 2008), was decreased. Also noted in these
Vessels and Mesenchyme of the Portal Tract livers was a complete absence of CK-19–positive intrahepatic
A functional relationship appears to exist between the contents biliary ducts, suggesting that β-catenin may play a role in biliary
to the portal tract (bile duct, hepatic artery, and portal vein) differentiation (Tan et al, 2008). In vitro studies showed that
regenerative process, suggesting that the biliary and pancreatic

ductal tree (SOX9-expressing domains) contain a previously
unappreciated pool of progenitors. It has been suggested that
these progenitor-like SOX-positive cells reside in the glands
(also known as peribiliary glands, or PBGs) of the extrahepatic
and large intrahepatic bile ducts. These progenitor-like cells are
responsible for the renewal of surface epithelium, generating
mature cells such as cholangiocytes, goblet cells, and hepato-
cytes. Although it is well demonstrated that the canal of Hering
is a likely stem cell niche in the adult liver, harboring human
hepatic stem cells, PBGs may be a newly identified reservoir.
Clinically, human hepatic stem cells are considered to be the
origin of some hepatocarcinomas and intrahepatic cholangio-
carcinomas (Cardinale et al, 2012; Carpino et al, 2012).
However, the PBG may be a site for oncogenic initiation.
Specifically, endodermal-like stem cells within PBGs may be
the cells of origin for mucin-producing cholangiocarcinoma
cells (Carpino et al, 2012). Furthermore, cholangiocarcinomas
express several markers in common with PBG cells, such as
EpCAM, OCT4, and CD133 (Komuta et al, 2008). There
Schematic Illustration have been several studies postulating the exact source of liver
by Mariam Hobeldin regeneration after injury, including hepatocytes (Schaub et al,
2014), ductal cells (Furuyama et al, 2011), PBGs (Carpino et
al, 2012), or an identified progenitor cell source (Huch, 2015;
FIGURE 1.6. SOX9-positive progenitor cells residing in the intestinal Huch et al, 2013). In addition, it was also demonstrated
crypts that are interconnected with the biliary and pancreatic ductal tree. through lineage tracing that hepatocytes dedifferentiate into a
(Courtesy Mariam Hobeldin.) bile duct cell progenitor after injury, before differentiating back
into functionally mature hepatocytes (Tarlow et al, 2014). A
mechanism that is similar to pancreatic acinar cells dedifferen-
OSM, produced by hematopoietic cells in fetal liver, as well as tiating into ductlike progenitor cells in response to stress (Shi
HGF, induce hepatic differentiation in the presence of dexa- et al, 2013), as well as the dedifferentiation-redifferentiation
methasone, both working through different pathways (Kamiya pathway seen in pancreatic cells (El-Gohary et al, 2014; Puri
et al, 2001). It was demonstrated that embryonic livers at E14.5 et al, 2015; Talchai et al, 2012). However, further studies are
expressed glucose-6-phosphatase (G6Pase), tyrosine amino needed to reconcile these differences into a unifying theory for
transferase (TAT), and accumulated glycogen, all signs of a the source of liver regeneration.
mature and differentiated liver when cultured with HGF or
OSM. This was further supported when TAT levels, as well as PANCREAS
glycogen storage, were significantly reduced in livers derived
from mice null mutant for gp130, the common receptor subunit Overview of Pancreatic Development
of IL-6 family cytokines (Kamiya et al, 2001). CEBPα is also The term pancreas is derived from the Greek, meaning all flesh.
critical for the acquisition and maintenance of hepatocyte dif- It consists of two morphologically distinct tissues, exocrine and
ferentiation, with knockout mice exhibiting defects in liver endocrine, which must derive from one simple epithelium. This
growth and architecture as well as increased cell proliferation feature has intrigued developmental biologists for years. The
(Flodby et al, 1996). CEBPα is also thought to be a key factor morphologic development of the pancreas is dictated by its two
in controlling the switch in the differentiation of bipotential major functions: production of digestive enzymes by the exo-
hepatoblasts to become either biliary epithelial cells or hepato- crine acinar tissue and regulation of blood chemistry by the
cytes. CEBPα starts to be expressed in the endodermal liver endocrine tissue. Some have even described it as two organs in
primordium at E9.5, and its expression in the nuclei of hepa- one due to the function and organization of these two distinct
toblasts and hepatocytes becomes stronger with development. tissues in the pancreas.
During biliary cell differentiation, CEBPα expression was sup- The endocrine pancreas is organized in islets of Langerhans,
pressed in periportal biliary cell progenitors, suggesting that its which consist of five cell subtypes, α, β, δ, ε, and PP cells that
suppression may be a prerequisite to biliary cell differentiation secrete glucagon, insulin, somatostatin, ghrelin, and pancreatic
from hepatoblasts (Shiojiri et al, 2004). polypeptide, respectively. All totaled, these cells occupy only
Lineage-tracing experiments have recently shown that the 2% of the adult pancreatic mass (Gittes, 2009). The exocrine
hepatic biliary tree, as well as the pancreatic ductal tree, and pancreas, on the other hand, produces digestive enzymes, is
intestinal crypts, which technically are all a continuous epithe- composed of acinar and ductal epithelial cells, and accounts for
lial lining (Fig. 1.6), harbor a common pool of progenitor cells, nearly 98% of the adult pancreatic mass.
SOX9 positive, that can all generate a continuous supply of
hepatocytes, acini, and all of the mature intestinal cell types, Basic Pancreatic Embryology
respectively, under physiologic conditions (Furuyama et al, During early development, the pancreas is formed by the dorsal
2011). In hepatocyte differentiation from the lineage tagged and ventral bud, each originating from the endodermal lining
SOX positive, bile duct cells increased during the hepatic of the caudal part of the primitive foregut tube after regional
Branching morphogenesis
mariam hobeldin
Pancreatic bud Salivary gland bud Ureteric bud
FIGURE 1.7. Pancreatic branching morphogenesis is different than in other organ systems, such as lungs, kidneys, and salivary glands, which
branch in a 90-degree pattern to be fully exposed to the mesenchyme. However, the pancreas has an acute branching pattern; thus some areas
are excluded from mesenchymal contact. Mesenchyme contains factors that regulate pancreatic growth and differentiation. (Courtesy Mariam
Hobeldin.)
specification with the transcription factor PDX1, followed by rise to mature islets (Jensen, 2004; Pictet et al, 1972). In a
PTF1A (Grapin-Botton, 2005; Moore-Scott et al, 2007; Sadler similar gestational window, the exocrine pancreatic precursors
& Langman, 2006). The first morphologic evidence of pancreas undergo rapid branching morphogenesis and acinar cell dif-
development is in the form of mesenchymal condensation over- ferentiation (see Fig. 1.7). During this secondary wave of dif-
lying the dorsal aspect of the primitive foregut tube at the level ferentiation, the simple stratified epithelium of the early
of the duodenal anlagen, distal to the stomach. Shortly after, pancreatic bud restructures into a branched network, with pro-
at about E9.5 of gestation in mice and day 26 of gestation in genitor cells becoming segregated to the distal “tips” expressing
humans, the dorsal bud begins to evaginate into the overlying acinar-specific enzymes, with transcription factors including
mesenchyme (Kallman & Grobstein, 1964; Pictet et al, 1972). PTF1A, and proximal “trunks” expressing duct markers such
Approximately 12 hours later in mice, and 6 days after dorsal as SOX9 and endocrine precursor markers such as NEUROG3
bud evagination in humans, the ventral bud begins to arise. (Kesavan et al, 2009; Villasenor et al, 2010). All the various
The morphologic development of the ventral pancreatic bud pancreatic cell types are eventually derived from PDX1+ cells,
is similar to the dorsal pancreas; however, it is markedly differ- whereas cell progenies from NEUROG3-positive cells will
ent at the molecular level. After this initial bud evagination, the delaminate from the main “trunk” and specifically form clusters
pancreatic bud undergoes elongation of the stalk region, a of islet cells. It was originally thought that toward the end of
precursor to the main pancreatic duct and branching morpho- gestation, different transcription factors, such as in PDX1-
genesis of the apical region of the bud. Unlike growth patterns positive, PTF1A-positive, and SOX9-positive cells, become
seen in the developing lung, kidney, and salivary glands, in gradually restricted to β cells, acinar cells, and duct cells,
which the branching morphogenesis undergoes a typical respectively. However, with the advent of lineage tracing experi-
90-degree pattern, the pancreas grows in an acute branching ments using the Cre-LoxP and the new Dre-Rox system, this
pattern that can lead to the exclusion, or “squeezing out,” of has proven not to be the case. For instance, PDX1-positve cells
mesenchyme from between the closely apposed branches of have now been shown to be expressed in pancreatic duct glands
epithelium. This exclusion of mesenchyme may influence (Strobel et al, 2010) (see later) and also give rise to the extra-
epithelial-mesenchymal interactions and lineage selection (Fig. hepatic biliary system (Spence et al, 2009).
1.7). The ventral pancreatic bud, which arises from the base of
the hepatic diverticulum, rotates dorsally, eventually fusing Endodermal Patterning of the Pancreas
with the dorsal pancreatic bud (sixth to seventh week of gesta- As mentioned earlier in the chapter, the pancreas develops
tion in humans, E12 to E13 in mice) contributing to the forma- through an epithelial-mesenchymal interaction. Before this, the
tion of the uncinate process and inferior part of the head of the dorsal foregut becomes competent to receive propancreatic
pancreas, with the rest of the pancreas arising from the dorsal signals from the notochord (FGFs) (Wells & Melton, 2000) and
pancreatic bud. The entire ventral pancreatic bud and the distal from the anterior foregut (retinoic acids and BMPs) (Stafford
part of the dorsal pancreatic duct fuse together to form the main and Prince, 2002; Tiso et al, 2002). The prepancreatic region
pancreatic duct of Wirsung. The remaining proximal part of of the definitive foregut comes into contact with the notochord,
the dorsal pancreatic duct is either obliterated or persists as a whereas the ventral pancreas contacts the lateral plate meso-
small accessory pancreatic duct of Santorini (Sadler & Langman, derm (Kumar et al, 2003).
2006).
During development, the pancreas undergoes a major Dorsal and Ventral Pancreatic Bud Development
amplification of the endocrine cell population through two The notochord is crucial to the development of the dorsal
waves of differentiation: an early primary transition (pre-E13.5 pancreas, where it remains in contact with the dorsal prepan-
in mice), marked by the expression of the homeobox transcrip- creatic endoderm until the paired dorsal aortae fuse in the
tion factor PDX1 at E8.5, followed by the secondary transition midline (≈E8.0). The dorsal pancreas failed to form when
(E13.5 to E16.5 in mice), which is marked by a dramatic the notochord was removed from early chick embryos in vitro
increase in endocrine and exocrine cell differentiation, giving (Kim et al, 1997), but no effect was seen on ventral pancreas
development, further emphasizing the disparate molecular con- TGF-β Signaling

trols for dorsal versus ventral pancreas development. The The TGF-β superfamily has also been implicated in many
ventral pancreas develops under the control of signals from the developmental processes, including early pancreatic develop-
overlying cardiogenic mesenchyme, which produces prohepatic ment. TGF-β isoforms (TGF-β1, -β2, and -β3) are localized
signals (FGFs) to induce liver formation. Lack of prohepatic to the pancreatic embryonic epithelium as early as E12.5 and
FGF signaling in regions of the cardiogenic mesenchyme will progressively become focused in acinar cells (Crisera et al,
lead to the endoderm by “default” differentiating into ventral 1999). The ligands mediate their actions through TGF-β
pancreas (Deutsch et al, 2001). Sonic Hedgehog (SHH) in the receptor type II (TBR-II), a heterodimer receptor; however, in
prospective pancreatic endoderm is suppressed by the noto- late gestation it localizes to the pancreatic ducts (Tulachan et
chord, which is necessary for dorsal pancreatic development, al, 2007). TBR-II inhibition has been shown to recruit an inap-
and SHH is otherwise expressed along the entire primitive gut propriately high number of ductal cells to become endocrine
tube epithelium. When the notochord was deleted in chick progenitor cells, suggesting that TBR-II signaling is important
embryo cultures, ectopic SHH was seen in the pancreatic for regulating the flow of endocrine progenitors out of ducts
region of the foregut endoderm, and the pancreas failed to (Lee et al, 1995; Sanvito et al, 1995).
develop (Hebrok et al, 1998). This difference between the Activins, part of the TGF-β superfamily, also play a role in
dorsal and ventral pancreatic bud development was under- pancreatic development. When exogenous activin was added to
scored in a human with aortic coarctation. The patient lacked embryonic pancreas explant cultures, it inhibited branching
the pancreatic body and tail, but not the head of the pancreas, morphogenesis, a process associated with ductal and acinar
with the latter arising from the ventral bud, which develops differentiation (Ritvos et al, 1995). This observation was further
independently of the aortae (Kapa et al, 2007). Here, the pro- supported when follistatin, an activin inhibitor that is expressed
spective pancreatic endoderm presumably lost the inducing in the pancreatic mesenchyme by E12.5, was added to the
signal from the “narrowed” aorta, leading to dorsal pancreatic pancreatic epithelium; it mirrored the mesenchymal effect, pro-
agenesis. This also further underscores the important role that moting exocrine and repressing endocrine development
endothelial cells play in pancreatic development and their deli- (Miralles et al, 1998). These observations would support a role
cate and complex interaction with the prepancreatic endoderm of activin in promoting and regulating endocrine differentiation
(see “Endothelial Cells” section). and suppressing exocrine differentiation in the early pancreas.
Along similar lines, transgenic mice that express a dominant-
Pancreatic Mesenchyme negative activin receptor display reduced levels of differentiated
The pancreatic epithelium becomes enveloped by mesen- islet cells (Kim et al, 2000; Shiozaki et al, 1999; Yamaoka et
chyme, which contains important factors that are pivotal for al, 1998). Hence investigators were interested to study the
pancreatic morphogenesis. The mesenchyme has a general underlying mechanism of activin in inducing insulin-positive
proexocrine effect on the pancreas through cell contact as differentiation. It was noted when AR42J cells, which are cells
well as proendocrine effect mediated by a diffusible factor derived from pancreatic acinar cells expressing both exocrine
secreted from the mesenchyme (Li et al, 2004). Mesenchymal and neuroendocrine properties, were cultured with activin A,
contact with the epithelium both enhances NOTCH signaling– it converted the cells into insulin-producing cells after the
induced HES1, which favors acinar lineage, and also inhibits induction of NGN3, an key determinant for endocrine differ-
neurogenin 3 (NEUROG3) expression, leading to the sup- entiation and lineage selection, and PAX4, a paired homeobox
pression of endocrine differentiation (Duvillie et al, 2006). transcription factor critical for β-cell differentiation (Mashima
The “default” differentiation of the pancreatic epithelium in et al, 1996; Zhang et al, 2001). It was also shown that activin
the absence of mesenchyme is endocrine (Gittes et al, 1996). A regulates glucagon gene expression by preventing α-cell lines
Thus the mesenchyme has both an inductive effect on the proliferating and differentiating through the inhibition of Arx,
exocrine tissue development and a repressive effect on the a homeobox-containing gene that is critical in promoting the
development of endocrine cells, most likely being mediated by acquisition of α-cell fate from endocrine progenitors (Mamin
soluble factors. & Philippe, 2007). Thus activin A favors β-cell differentiation
The interaction between the mesenchyme and the pancre- through stimulation of PAX4 and NGN3 and inhibition of
atic epithelium is regulated by several factors, the most impor- ARX expression.
tant of which are the FGFs, which are known to be expressed As was discussed earlier with the dominant-negative mutant
in several epithelial-mesenchymal interface regions of the pan- of the activin type II receptors displaying islet hypoplasia, the
creas and are particularly important in regulating branching growth differentiation factor 11 (GDF11) was demonstrated to
morphogenesis (Hogan, 1999). FGFs 1, 7, and 10 mediate be the key ligand to induce β-cell differentiation. GDF11-null
their effect through FGF receptor 2B (FGFR2B) to induce mutant mice pancreas developed with more Ngn3-positive
epithelial proliferation, favoring exocrine differentiation (Celli endocrine progenitor cells and glucagon positive α cells with
et al, 1998; Dichmann et al, 2003; Elghazi et al, 2002). Null less mature insulin positive β cells (Harmon et al, 2004). A
mutation for FGF10 or FGFR2B leads to blunting of branching similar phenotype was also seen in activin receptor type IIB
pancreatic morphogenesis and lack of proliferation in endocrine mutants. Mice heterozygous for Smad2 also showed increase in
cells, indicating that FGF10 induces proliferation of epithelial NGN3-positive cells with a subsequent decrease in β-cell mass,
cells and prevents endocrine differentiation (Bhushan et al, indicating that GDF11 acts through SMAD2, an intracellular
2001; Pulkkinen et al, 2003). These roles of FGFs are primarily mediator of TGF signaling, to induce its effects in recruiting β
for ongoing development of the pancreatic buds, in contrast to cells (Goto et al, 2007).
the role of FGF from the cardiogenic mesenchyme, which BMP signaling has a role in pancreatic development;
inhibits ventral pancreatic bud formation in favor of liver devel- however, studies have given different results. Transgenic mice
opment (Deutsch et al, 2001). with PDX1-BMP6 developed complete pancreatic agenesis
(Dichmann et al, 2003), while overexpressing BMP4-enhanced pancreatic development. It appears that the downstream factors
ductal cell proliferation (Hua et al, 2006). of NOTCH targets are variably expressed at a certain time in
The intracellular mediators of TGF-β signaling, SMADs, development of the pancreas. The early pattern of downstream
have gained interest among developmental biologists. Specifi- molecules is a key mediator of NOTCH signaling in determin-
cally, the intracellular mediators of TGF-β signaling, SMAD2 ing endocrine versus exocrine differentiation of progenitor cells
and SMAD3, along with their inhibitor SMAD7, have been and in maintaining a stem cell state in some progenitor cells.
found to play an intricate role in regulating pancreatic endo- Later in gestation, the downstream molecules for NOTCH
crine maturation and development. Genetic inactivation of signaling change to foster cell maturation and differentiation
SMAD2 and SMAD3 led to both a significant expansion of the beyond this initial lineage decision between endocrine and exo-
embryonic endocrine compartment and a more robust islet crine differentiation.
proliferation in adult mouse pancreas after partial pancreatec-
tomy. Whereas genetic inactivation of SMAD7 led to a signifi- HEDGEHOG Signaling
cant decrease in the endocrine compartment with little β-cell As mentioned earlier, SHH suppression in the prospective pan-
proliferation after pancreatectomy in the adult mouse pancreas creatic endoderm is a prerequisite need for pancreas formation;
(El-Gohary et al, 2013; El-Gohary et al, 2014). Furthermore, however, HEDGEHOG signaling in pancreatic development
SMAD4 was specifically identified to be mutated in 50% of appears to be complex. SHH-null mutant mice do not exhibit
pancreatic cancers (Hahn et al, 1996). However, a key role for an expanded pancreatic field; in contrast, Indian Hedgehog
SMAD4 has not been identified yet in pancreatic development (IHH)-null mutant mice develop with a small pancreas, indicat-
(Bardeesy et al, 2006; Simeone et al, 2006). These results ing a propancreatic role for IHH (Hebrok et al, 2000). However,
identify the SMAD2/3/7 network as a legitimate target for regu- when combining a SHH-null mutation with a heterozygous
lating β-cell proliferation, and for generating more β-cells for mutation for IHH, an annular pancreas develops. Mice that are
future treatment of diabetes. null mutants for Hedgehog Patched receptor Ptc lack PDX1
and glucagon expression at E9.5, supporting the need to sup-
NOTCH Signaling press SHH for early pancreatic development (Hebrok et al,
A key step in pancreatic development is the lineage decision by 2000). Similarly, transgenic overexpression of SHH or IHH
pancreatic progenitor cells between the endocrine and exocrine under PAX4 promoter led to significant decrease in endocrine
lineage. NOTCH signaling has been identified as a master and exocrine tissue mass (Smith et al, 2000). Furthermore,
regulator of this fate decision switch (Apelqvist et al, 1999). a novel and distinct anatomic compartment has been described
However, the exact mechanism for NOTCH signaling in pan- within the pancreatic ducts, known as the pancreatic
creatic lineage selection remains elusive, and ambiguity still duct glands (McMinn and Kugler, 1961; Strobel et al, 2010;
surrounds the exact role of NOTCH signaling through Wilcox et al, 2013) (Fig. 1.8). They are present throughout the
A B
C D
FIGURE 1.8. A, Whole-mount image of the main common pancreatic duct containing the pancreatic duct glands (PDGs). Stained with Dolichos
Biflorus Agglutin (DBA), with a higher magnification seen in (B). C, Cross-section of the main common pancreatic duct, with duct glands seen in the
periphery. D, These blind-ending pouches communicate with the lumen. (Courtesy Mariam Hobeldin.)
pancreatic ductal network in both humans and rodents, with Zhou et al, 1999) and ARIP cells (Hui et al, 2001) into insulin-
little known about the function of these duct glands other than expressing cells. Together, these studies strongly support the
that they are blind-ending pouches that unexpectedly express role of glucagon signaling, through its receptor, in initiating
Sonic Hedgehog, HES1, and PDX1 (Strobel et al, 2010). Both early insulin differentiation.
HES1 and PDX1 are markers of pancreatic progenitors (Gittes, Current treatments for diabetes, such as insulin replacement
2009), and their specific expression in these blind-ending therapy, do not address the actual physical loss of β cells, a
pouches, reminiscent of intestinal crypts, which are known to hallmark of the disease. Hence extensive efforts are focused on
harbor intestinal stem cells, possibly indicate that they may trying to establish alternate sources for the replenishment of the
have a role in pancreas regeneration. depleted β-cell population. For many years there has been a
dogma that once a cell reaches its terminal differentiation, it
WNT can no longer switch cell fate. Therefore the issue of whether
As mentioned earlier in the chapter, β-catenin repression in the β-cell neogenesis does occur within the mature adult pancreas
anterior endoderm is necessary to initiate liver and pancreas has been vigorously disputed in the literature. However, evi-
development. Consistent with this, Pdx1-Wnt1 and Pdx1- dence from recent studies has demonstrated the plasticity that
Wnt5a transgenic mice developed pancreatic agenesis and exists between α and βcells. Forced PDX1 expression in NGN3
severe hypoplasia, respectively, confirming that early WNT sig- progenitor cells resulted in nearly all of the α cells converting
naling is detrimental for pancreatic development (Heller et al, to β cells by postnatal day 12, without any obvious evidence of
2002). Similar to NOTCH signaling, the role of WNT signal- the α cells regenerating (Yang et al, 2011). Furthermore, near
ing beyond endodermal patterning is complex and depends on total ablation of β cells (about 99%) using a diphtheria toxin
time and place of WNT signaling. β-Catenin, which is mostly transgenic model led to α cell conversion to β cells via a bihor-
expressed in the mesenchyme (Heller et al, 2002), when ablated monal cell stage (glucagon positive/insulin positive) after 10
from PDX1-positive cells resulted in the loss of acinar tissue at months (Thorel et al, 2010). Also, ectopic expression of PDX1
birth, confirming earlier studies that exocrine tissue develop- or PAX4 has been shown to induce α cells or α-cell progenitors
ment is mesenchyme dependent (Murtaugh et al, 2005; Wells to convert to β cells (Collombat et al, 2009; Piccand et al,
et al, 2007). Others found a role for Wnt signaling in promoting 2014). To illustrate that α cells can genuinely act as a source
postnatal pancreatic growth (Heiser et al, 2006), illustrating for new β cells in the future, Thorel and coworkers (2011) used
the complex and multiple roles that it plays in pancreatic the diphtheria toxin model to ablate 98% of α cells, with the
development. transgenic mice able to maintain a euglycemic state. The previ-
ous notion that terminally differentiated cells in the pancreas
Endothelial Cells lack plasticity is no longer accepted, with several studies dem-
Recently, an important role for endothelial cells in pancreatic onstrating that their fate is flexible (Chera et al, 2014; Puri
endocrine development has been demonstrated. As mentioned et al, 2015).
earlier, the notochord separates from the gut endoderm by the
fusion of the dorsal aortas at about E8.0. Lammert and cowork- Extracellular Matrix
ers demonstrated that the removal of the dorsal aorta from The pancreatic epithelium is contained within a continuous
Xenopus embryos led to the absence of pancreatic endocrine sheath of basement membrane, which constitutes the epithelial-
development (Lammert et al, 2001). In addition, aortic endo- mesenchymal interface and plays an important role in guiding
thelial cells were able to induce pancreatic bud formation and pancreatic development (Hisaoka et al, 1993). Basement mem-
subsequent insulin cell differentiation in the adjacent endo- brane has also been shown to play an important role in regulat-
derm. Of importance, ectopic VEGFA in the intestine and ing branching morphogenesis in many other organs. Laminin-1
stomach led to ectopic formation of islets in the stomach was found to induce duct formation in isolated E11 mouse
(Lammert et al, 2001). Of interest, it appears that the ventral pancreatic epithelium (Gittes et al, 1996) and to mediate the
pancreas development seems to be independent of the endo- proexocrine effect of the mesenchyme and the pro–β-cell role
thelium, despite its close proximity to the vitelline veins (Yoshi- in later gestation (Jiang et al, 2001; Li et al, 2004). Cadherins
tomi & Zaret, 2004). (calcium-dependent adhesion molecules), critical for cell-cell
adhesions, play an essential role in migration and differentiation
Glucagon of pancreatic endocrine progenitor cells. E-cadherin and
The first endocrine cells to appear during pancreatic develop- R-cadherin are initially localized to the ducts, but they become
ment are the glucagon-containing α-granules at about E9.0 in downregulated as cells move out of the ducts to form islets
the mouse, and recent studies have shown that glucagon signal- (Dahl et al, 1996; Sjodin et al, 1995). N-CAM, another cell-
ing is necessary for early differentiation of insulin expressing adhesion molecule expressed in mature α and PP cells (Cirulli
cells, which appear at about E13.0 (Pictet et al, 1972; Prasadan et al, 1994), when knocked out led to improper aggregation of
et al, 2002). Glucagon hormone is generated from proglucagon endocrine cells within the islet (Esni et al, 1999). Thus cell-
by the action of prohormone convertase 2 (PC2). When PC2 adhesion molecules regulate the adhesive properties of endo-
is knocked out, mutant mice are generated being without glu- crine cells, and these adhesive properties are essential for
cagon and displaying a delay in islet cell differentiation and endocrine cell aggregation into islets.
maturation but still retaining the large amplification of insulin-
positive cells (“secondary-wave”) later in gestation (Vincent et Transcription Factors
al, 2003). Exogenous addition of exendin-4, a glucagon-like The role of transcription factors has been extensively studied
peptide-1 (GLP-1) analogue, was able to rescue the delay in in pancreatic development, and much of our understanding of
early insulin differentiation (Prasadan et al, 2002). Exendin-4 pancreatic development has been facilitated by the develop-
was also shown also to convert AR42J cells (Yew et al, 2004; ment of the various transcription factor–specific knock-out
mice. Especially with the advent of the Cre-LoxP system, with no pancreatic endocrine cells form (Gradwohl et al, 2000); thus
or without tamoxifen-inducible (Cre-ERT2), allowing spatio- it appears to be a critical and essential factor for endocrine
temporal control of lineage labeling (Guo et al, 2013). The differentiation.
most heavily studied of the transcription factors is PDX1, an
early marker for pancreatic progenitors that later “restricts” to RFX6. RFX6 is a transcription factor downstream of NGN3 that
the β-cells and is clearly required for maintenance of β-cell fate has recently been identified as a key proendocrine regulator that
and function (Gao et al, 2014). It is expressed in the prepan- directs islet cell differentiation. It is initially expressed broadly
creatic region of the primitive foregut tube at E8.5, then in gut endoderm, particularly in PDX1-positive cells in the
expands to be expressed in distal stomach, common bile duct, prospective pancreatic region, and then becomes restricted to
and duodenum by E10.5 to 11.5 (Guz et al, 1995; Offield et the endocrine lineage, in postmitotic islet progenitor cells. Null-
al, 1996). Pdx1 is initially expressed throughout the epithelium; mutant mice for RFX6 fail to generate all islet cell types except
however, its expression becomes suppressed in cells as they pancreatic polypeptide cells (insulin, glucagon, somatostatin,
commit to the endocrine (Jensen et al, 2000) lineage or ducts and ghrelin). A human syndrome of neonatal diabetes (patients
(Gu et al, 2002). Its expression reappears as cells differentiate who lack pancreatic endocrine cells) with bowel atresia were
to insulin-positive β-cell lineage. More recently, Pdx1 has been shown to have mutations in the RFX6 gene (Smith et al, 2010).
shown to be expressed in pancreatic duct glands in addition to Thus RFX6 is dependent on NGN3 and is a unique regulator
β cells (Strobel et al, 2010). PDX1-null mutant mice and of islet cell development. It is also essential in maintaining func-
humans developed pancreatic agenesis (Jonsson et al, 1994; tional identity in adult pancreatic β cells (Piccand et al, 2014).
Stoffers et al, 1997a; Stoffers et al, 1997b; Yee et al, 2001).
Delayed PDX1 inhibition using tetracycline regulatable trans- PAX6. PAX6 is also another marker of endocrine lineage, but
genic knock-in system demonstrated severe blunting of pancre- unlike NEUROG3, it is not absolutely necessary for endocrine
atic development with complete absence of acini and β cells formation because null-mutant mice for PAX6 still formed
(Holland et al, 2002), indicating that PDX1 continues to play endocrine cells, albeit at a reduced rate (St-Onge et al, 1997).
a role in pancreas development beyond regional specification PAX6 expression is retained in cells that are committed to the
during endodermal patterning. endocrine lineage as early as E9.5 in glucagon-positive cells or
in insulin-positive cells at E12.5, thereafter beginning hormone
PTF1A. Pancreas-specific transcription factor 1A (PTF1A) is expression (St-Onge et al, 1997). Transgenic expression of
an early marker of pancreatic progenitor cells, expressed PAX6 under the PDX1 promoter led to islet and ductal hyper-
slightly later than PDX1 at around E9.5; by E18.5 it becomes plasia, indicating that it may have roles in the formation of
progressively restricted to acinar cells (Cras-Meneur et al, ducts and in islet growth (Yamaoka et al, 2000). However,
2009). PTF1A-null mutant mice develop a phenotype similar when PAX6 was ectopically expressed in β cells under the
to that seen with PDX1-null mutants, with severe pancreatic insulin promoter, it led to β-cell apoptosis, indicating that
hypoplasia and absent acini and ducts; however, endocrine PAX6 promotes proliferation only in pancreatic progenitors,
cells still develop. Of interest, the endocrine cells migrate out not in mature cells.
through the pancreatic mesenchyme to populate the spleen
(Kawaguchi et al, 2002; Krapp et al, 1998). A PTF1A non- PAX4/ARX
functioning mutation has been seen in humans born without PAX4 is another endocrine cell marker and is expressed as early
a pancreas (Sellick et al, 2004). Furthermore, PTF1A has as E9.5; it then peaks at about E13 to E15, coinciding with the
complex interactions with downstream targets of NOTCH to secondary transition of insulin-positive cells (Wang et al, 2004).
control exocrine lineage selection and differentiation. Overex- Its importance was further delineated when β cells and δ cells
pressing PTF1A in pancreatic progenitors represses the pro- failed to develop in PAX4-null mutant mice. Early embryonic
endocrine marker NKX6.1, stopping endocrine differentiation insulin-positive cells were present in these mice, indicating that
(Puri et al, 2015). This further emphasizes the role of PTF1A PAX4 is needed for the formation of mature β cells (Sosa-
as a promoter of acinar fate while suppressing the endocrine Pineda et al, 1997; Wang et al, 2004). In these mice, the β and
program fate. δ cells seems to have been replaced by cells coexpressing glu-
cagon and ghrelin, indicating that PAX4 may act as transcrip-
NEUROG3. NEUROG3 is one of the earliest markers specific to tional inhibitor for this inhibitor (Wang et al., 2008). PAX4
the pancreatic endocrine lineage, and it is a key driver of endo- seems also to inhibit the expression of ARX, a homeobox-
crine differentiation (Li et al, 2014). It is first expressed at E9.0 containing gene that enhances glucagon-positive cell differen-
and peaks at around E15.5; it is thought to be antagonized by tiation (Collombat et al, 2005). ARX, which is downstream of
NOTCH signaling in cells with an acinar fate (Gradwohl et al, NEUROG3 and is absent in NEUROG3-null mutant mice,
2000; Jensen et al, 2000). NEUROG3 cells then proliferate, when overexpressed in PDX1-positive progenitor cells, diverted
giving rise to proendocrine postmitotic cells expressing tran- most β- and δ-cell precursors toward α- and PP-cell lineages,
scription factors RFX6, NEUROD, NKX6.1, and PAX6. without a change in the overall number of endocrine cells.
NEUROG3 shuts off at around E17.5 (Gu et al, 2002; Jensen When ARX was knocked out, no α cells were seen with the
et al, 2000), and forced overexpression of NEUROG3 leads to appearance of the β-cell marker (Collombat et al, 2003; Court-
cells prematurely committing to an endocrine lineage, with the ney et al, 2013; Wilcox et al, 2013 ). These findings further
development of only glucagon-positive clusters (Johansson et emphasize the opposing roles for PAX4 and ARX in α and PP
al, 2007). Overexpressing NEUROG3 at different time points cells versus β and δ cells (Collombat et al, 2007). In summary,
later in gestation (E11 to E12) resulted in the formation of PAX4 and ARX mutually repress each other, with PAX4 pro-
insulin- and pancreatic polypeptide–positive cells (Johansson moting β-cell fate at the expense of α cells, whereas ARX
et al, 2007). However, when NEUROG3 is deleted from cells, promotes α cells at the expense of β cells.
FIGURE 1.9. Overview of the pancreatic endocrine and exocrine cell lineage. (Courtesy Mariam Hobeldin.)
NKX2.2. NKX2.2 is expressed as early as E9.5, is coexpressed NKX6.1 AND NKX6.2. Single- and double-null mutant studies for
with PDX1, acting as a marker of multipotent pancreatic pro- NKX6.1 and NKX6.2 have been performed and suggest that
genitor cells, and eventually becomes restricted to NEUROG3- both molecules, which are expressed early in pancreatic endo-
positive cells, persisting in all endocrine lineages except for δ dermal cells, play central roles in pancreatic development, espe-
cells (Chiang & Melton, 2003; Sussel et al, 1998). NKX2.2- cially in β-cell development (Henseleit et al, 2005). NKX6.1,
null mutant mice develop with no β cells, reduced PP cells, an which is downstream of PDX1 and is an endocrine progenitor
80% reduction in α cells, and no effect on δ cells. However, a marker, becomes restricted to insulin-positive cells; whereas
large number of ghrelin-positive ε cells were found with no NKX6.2 is restricted to glucagon-positive and amylase-positive
glucagon coexpression, indicating that NKX2.2 normally cells; after E15.5, it disappears (Sander et al, 2000). In NKX6.1-
induces insulin-positive differentiation and represses ε-cell for- null mutant mice, an 85% reduction in β cells is seen, whereas
mation (Sussel et al, 1998). NKX6.1 and NKX6.2 double-null mutants developed a 92%
and 65% reduction in β cells and α cells, respectively, suggest- as described earlier (Haumaitre et al, 2005; Jacquemin et al,
ing that NKX6.1 has an important role in generating β cells. 2003). HNF6 is expressed in the foregut-midgut endoderm
Deletion of NKX 6.1 in the adult pancreas leads to β cells junction at E8.0 under the control of HNF1β (Poll et al, 2006).
converting into somatostatin-expressing δ cells (Schaffer et al, HNF6 in turn regulates the expression of FOXA2, which is
2013). NKX6.2 plays an important role in α-cell formation and essential for gut formation (Weinstein et al, 1994). FOXA2,
can compensate for NKX6.1 absence in β-cell formation however, is not necessary for pancreas formation because its
(Henseleit et al, 2005; Sander et al, 2000). β-Cell loss was deletion did not affect pancreatic development (Lee et al,
rescued with transgenic expression of NKX6.1 or NKX6.2 2005b). As discussed earlier, in the bile duct formation, HNF6
under the PDX1 promoter (Nelson et al, 2007). also acts upstream from HNF1β, where HNF6 activates HNF1β
to initiate the formation of endocrine progenitors (Clotman
MAFA AND MAFB. The transcription factor MAFA is a critical et al, 2002).
regulator of the insulin gene, and it also regulates glucose-
responsive expression of insulin (Matsuoka et al, 2003; SRY (SEX-DETERMINING REGION Y)-BOX: SOX9
Nishimura et al, 2015). MAFA is first expressed in insulin- SOX9 acts as a marker and maintenance factor for pancreatic
positive cells during the “secondary transition” wave of insulin- progenitor cells. It is initially expressed throughout the early
cell differentiation; however, it is not absolutely necessary for pancreatic epithelium, then later becomes restricted to trunk
β-cell formation, because MAFA-null mutant mice have a progenitors and mature ducts, eventually giving rise to all pan-
normal proportion of insulin-positive cells at birth (Nishimura creatic cell types. It is seen as necessary for maintaining cells in
et al, 2009). MAFB is expressed early in pancreatic endocrine a progenitor state (Akiyama et al, 2005). SOX9 mutants display
development and, as the pancreas develops, it switches off in pancreatic hypoplasia as a result of depletion of the pancreatic
insulin-positive–forming cells as they form mature β cells, to progenitor pool (Seymour, 2014). It has also been suggested
then express MAFA (Nishimura et al, 2006). It was demon- that SOX9 mediates HNF6 and the control of NEUROG3-cell
strated that early expression of MAFA under the PDX1 pro- formation, because conditional deletion of SOX9 led to the
moter was detrimental for pancreatic development, with a formation of cystic structures similar to that seen in HNF6−/−
severe reduction in pancreatic mass and proliferation of pro- embryos, associated with absent pancreatic endocrine cells
genitors, suggesting that MAFA is a key regulator for the matu- (Seymour et al, 2007). Furthermore, it was found that SOX9 is
ration of pancreatic β cells rather than specification (Nishimura transiently coexpressed with NEUROG3 and frequently
et al, 2009). The same conclusion was drawn in MAFB-null colocalizes with HES1, which acts as a precursor to NEUROG3-
mutant mice, in which there was a delay in the development of positive cells. It was suggested that SOX9 helps mediate
early insulin- and glucagon-positive cells, with a reduction by the transition of endocrine progenitors from nonendocrine
almost half of their total mass (Artner et al, 2007). HES1-positive, NEUROG3-negative state to a HES1-negative,
NEUROG3-positive endocrine-committed state (Seymour
HNF CASCADE. Three HNF factors: HNF6/ONECUT1, et al, 2007). Thus SOX9 acts a key mediator for the commit-
HNF1β, and HNF3β (renamed FOXA2), appear to have an ment of NEUROG3-positive endocrine progenitors (Seymour,
interwoven, complex set of interactions that all have an impact 2014). More recently, it has been shown that SOX9 is a marker
on pancreatic development, starting from the earliest pancreatic of a common pool of progenitors cells in the intestinal crypts
specification all the way through to mature pancreatic cell dif- and the biliary and pancreatic ductal trees that are intercon-
ferentiation and function. They are all expressed in the pancre- nected in a contiguous epithelial lining (Furuyama et al, 2011)
atic epithelium at about E9.0 to E10.0, with HNF6 appearing (Fig. 1.9). SOX9 is also seen as a useful marker in clinical
to be the key determinant of pancreatic specification (Maestro pathology (Shroff et al, 2014) and key player in the initiation of
et al, 2003). This conclusion comes from observations that pancreatic cancer (Grimont et al, 2015; Seymour, 2014).
HNF6−/− embryos developed ventral pancreatic agenesis, with
an atrophic dorsal pancreas along with severe biliary anomalies References are available at ExpertConsult.com.
Chapter 1 Embryologic development of the liver, biliary tract, and pancreas 31.e1
REFERENCES Crompton MR, et al: Identification of a novel vertebrate homeobox

gene expressed in haematopoietic cells, Nucleic Acids Res 20:5661–
Akiyama H, et al: Osteo-chondroprogenitor cells are derived from Sox9 5667, 1992.
expressing precursors, Proc Natl Acad Sci U S A 102:14665–14670, Crosnier C, et al: JAGGED1 gene expression during human embryo-
2005. genesis elucidates the wide phenotypic spectrum of Alagille syn-
Apelqvist A, et al: Notch signalling controls pancreatic cell differentia- drome, Hepatology 32:574–581, 2000.
tion, Nature 400:877–881, 1999. Crowe AJ, et al: Hepatocyte nuclear factor 3 relieves chromatin-
Artner I, et al: MafB is required for islet beta cell maturation, Proc Natl mediated repression of the alpha-fetoprotein gene, J Biol Chem
Acad Sci U S A 104:3853–3858, 2007. 274:25113–25120, 1999.
Auth MK, et al: Morphogenesis of primary human biliary epithelial Dahl U, et al: Cadherins regulate aggregation of pancreatic beta-cells
cells: induction in high-density culture or by coculture with autolo- in vivo, Development 122:2895–2902, 1996.
gous human hepatocytes, Hepatology 33:519–529, 2001. de La Coste A, et al: Somatic mutations of the beta-catenin gene are
Bardeesy N, et al: Smad4 is dispensable for normal pancreas develop- frequent in mouse and human hepatocellular carcinomas, Proc Natl
ment yet critical in progression and tumor biology of pancreas Acad Sci U S A 95:8847–8851, 1998.
cancer, Genes Dev 20:3130–3146, 2006. Desmet VJ: Intrahepatic bile ducts under the lens, J Hepatol 1:545–559,
Bhushan A, et al: Fgf10 is essential for maintaining the proliferative 1985.
capacity of epithelial progenitor cells during early pancreatic organo- Deutsch G, et al: A bipotential precursor population for pancreas and
genesis, Development 128:5109–5117, 2001. liver within the embryonic endoderm, Development 128:871–881,
Bort R, et al: Hex homeobox gene-dependent tissue positioning is 2001.
required for organogenesis of the ventral pancreas, Development Dichmann DS, et al: Expression and misexpression of members of the
131:797–806, 2004. FGF and TGFbeta families of growth factors in the developing
Bossard P, Zaret KS: GATA transcription factors as potentiators of gut mouse pancreas, Dev Dyn 226:663–674, 2003.
endoderm differentiation, Development 125:4909–4917, 1998. Dolle P, et al: The Hox-4.8 gene is localized at the 5′ extremity of the
Cardinale V, et al: The biliary tree—a reservoir of multipotent stem Hox-4 complex and is expressed in the most posterior parts of the
cells, Nat Rev Gastroenterol Hepatol 9:231–240, 2012. body during development, Mech Dev 36:3–13, 1991.
Carpino G, et al: Biliary tree stem/progenitor cells in glands of Duncan SA: Transcriptional regulation of liver development, Dev Dyn
extrahepatic and intraheptic bile ducts: an anatomical in situ study 219:131–142, 2000.
yielding evidence of maturational lineages, J Anat 220:186–199, Duvillie B, et al: The mesenchyme controls the timing of pancreatic
2012. beta-cell differentiation, Diabetes 55:582–589, 2006.
Celli G, et al: Soluble dominant-negative receptor uncovers essential Elghazi L, et al: Role for FGFR2IIIb-mediated signals in controlling
roles for fibroblast growth factors in multi-organ induction and pat- pancreatic endocrine progenitor cell proliferation, Proc Natl Acad Sci
terning, EMBO J 17:1642–1655, 1998. U S A 99:3884–3889, 2002.
Chen Y, et al: Retinoic acid signaling is essential for pancreas develop- El-Gohary Y, et al: Smad signaling pathways regulate pancreatic endo-
ment and promotes endocrine at the expense of exocrine cell dif- crine development, Dev Biol 378:83–93, 2013.
ferentiation in Xenopus, Dev Biol 271:144–160, 2004. El-Gohary Y, et al: A smad signaling network regulates islet cell pro-
Chera S, et al: Diabetes recovery by age-dependent conversion of liferation, Diabetes 63:224–236, 2014.
pancreatic delta-cells into insulin producers, Nature 514:503–507, Esni F, et al: Neural cell adhesion molecule (N-CAM) is required for
2014. cell type segregation and normal ultrastructure in pancreatic islets,
Chiang MK, Melton DA: Single-cell transcript analysis of pancreas J Cell Biol 144:325–337, 1999.
development, Dev Cell 4:383–393, 2003. Fassler R, Meyer M: Consequences of lack of beta 1 integrin gene
Cirillo LA, et al: Opening of compacted chromatin by early develop- expression in mice, Genes Dev 9:1896–1908, 1995.
mental transcription factors HNF3 (FoxA) and GATA-4, Mol Cell Flodby P, et al: Increased hepatic cell proliferation and lung abnormali-
9:279–289, 2002. ties in mice deficient in CCAAT/enhancer binding protein alpha,
Cirulli V, et al: Expression of neural cell adhesion molecule (N-CAM) J Biol Chem 271:24753–24760, 1996.
in rat islets and its role in islet cell type segregation, J Cell Sci 107(Pt Furuyama K, et al: Continuous cell supply from a Sox9-expressing
6):1429–1436, 1994. progenitor zone in adult liver, exocrine pancreas and intestine, Nat
Clotman F, et al: The onecut transcription factor HNF6 is required Genet 43:34–41, 2011.
for normal development of the biliary tract, Development 129:1819– Gao T, et al: Pdx1 maintains beta cell identity and function by repress-
1828, 2002. ing an alpha cell program, Cell Metab 19:259–271, 2014.
Collombat P, et al: The simultaneous loss of Arx and Pax4 genes pro- Gittes GK: Developmental biology of the pancreas: a comprehensive
motes a somatostatin-producing cell fate specification at the expense review, Dev Biol 326:4–35, 2009.
of the alpha- and beta-cell lineages in the mouse endocrine pancreas, Gittes GK, et al: Lineage-specific morphogenesis in the developing
Development 132:2969–2980, 2005. pancreas: role of mesenchymal factors, Development 122:439–447,
Collombat P, et al: Embryonic endocrine pancreas and mature beta 1996.
cells acquire alpha and PP cell phenotypes upon Arx misexpression, Goss AM, et al: Wnt2/2b and beta-catenin signaling are necessary and
J Clin Invest 117:961–970, 2007. sufficient to specify lung progenitors in the foregut, Dev Cell 17:290–
Collombat P, et al: Opposing actions of Arx and Pax4 in endocrine 298, 2009.
pancreas development, Genes Dev 17:2591–2603, 2003. Goto Y, et al: Genetic interactions between activin type IIB receptor
Collombat P, et al: The ectopic expression of Pax4 in the mouse pan- and Smad2 genes in asymmetrical patterning of the thoracic organs
creas converts progenitor cells into alpha and subsequently beta cells, and the development of pancreas islets, Dev Dyn 236:2865–2874,
Cell 138:449–462, 2009. 2007.
Costa RM, et al: Novel gene expression domains reveal early patterning Gradwohl G, et al: Neurogenin3 is required for the development of the
of the Xenopus endoderm, Gene Expr Patterns 3:509–519, 2003. four endocrine cell lineages of the pancreas, Proc Natl Acad Sci U S
Courtney M, et al: The inactivation of Arx in pancreatic alpha-cells A 97:1607–1611, 2000.
triggers their neogenesis and conversion into functional beta-like Grapin-Botton A: Antero-posterior patterning of the vertebrate diges-
cells, PLoS Genet 9:e1003934, 2013. tive tract: 40 years after Nicole Le Douarin’s PhD thesis, Int J Dev
Couvelard A, et al: Expression of integrins during liver organogenesis Biol 49:335–347, 2005.
in humans, Hepatology 27:839–847, 1998. Grimont A, et al: SOX9 regulates ERBB signalling in pancreatic cancer
Cras-Meneur C, et al: Presenilins, Notch dose control the fate of pan- development, Gut 64:1790–1799, 2015.
creatic endocrine progenitors during a narrow developmental Gu G, et al: Direct evidence for the pancreatic lineage: NGN3+ cells
window, Genes Dev 23:2088–2101, 2009. are islet progenitors and are distinct from duct progenitors, Develop-
Crisera CA, et al: The ontogeny of TGF-beta1, -beta2, -beta3, and ment 129:2447–2457, 2002.
TGF-beta receptor-II expression in the pancreas: implications for Gualdi R, et al: Hepatic specification of the gut endoderm in vitro: cell
regulation of growth and differentiation, J Pediatric Surg 34:689–693, signaling and transcriptional control, Genes Dev 10:1670–1682,
discussion 693-684, 1999. 1996.
31.e2 PART 1 LIVER, BILIARY, AND PANCREATIC ANATOMY AND PHYSIOLOGY
Guo P, et al: Specific transduction and labeling of pancreatic ducts by Kagami S, et al: Transforming growth factor-beta (TGF-beta) stimu-
targeted recombinant viral infusion into mouse pancreatic ducts, Lab lates the expression of beta1 integrins and adhesion by rat mesangial
Invest 93:1241–1253, 2013. cells, Exp Cell Res 229:1–6, 1996.
Guz Y, et al: Expression of murine STF-1, a putative insulin gene Kalinichenko VV, et al: Haploinsufficiency of the mouse Forkhead Box
transcription factor, in beta cells of pancreas, duodenal epithelium f1 gene causes defects in gall bladder development, J Biol Chem
and pancreatic exocrine and endocrine progenitors during ontogeny, 277:12369–12374, 2002.
Development 121:11–18, 1995. Kallman F, Grobstein C: Fine structure of differentiating mouse pan-
Hahn SA, et al: DPC4, a candidate tumor suppressor gene at human creatic exocrine cells in transfilter culture, J Cell Biol 20:399–413,
chromosome 18q21.1, Science 271:350–353, 1996. 1964.
Harmon EB, et al: GDF11 modulates NGN3+ islet progenitor cell Kamiya A, et al: Oncostatin M and hepatocyte growth factor induce
number and promotes beta-cell differentiation in pancreas develop- hepatic maturation via distinct signaling pathways, FEBS Lett
ment, Development 131:6163–6174, 2004. 492:90–94, 2001.
Haumaitre C, et al: Lack of TCF2/vHNF1 in mice leads to pancreas Kapa S, et al: Dorsal pancreas agenesis and polysplenia/heterotaxy
agenesis, Proc Natl Acad Sci U S A 102:1490–1495, 2005. syndrome: a novel association with aortic coarctation and a review
Hebrok M, et al: Notochord repression of endodermal Sonic hedgehog of the literature, JOP 8:433–437, 2007.
permits pancreas development, Genes Dev 12:1705–1713, 1998. Kawaguchi Y, et al: The role of the transcriptional regulator Ptf1a in
Hebrok M, et al: Regulation of pancreas development by hedgehog converting intestinal to pancreatic progenitors, Nat Genet 32:128–
signaling, Development 127:4905–4913, 2000. 134, 2002.
Heiser PW, et al: Stabilization of beta-catenin impacts pancreas growth, Kawakami-Kimura N, et al: Involvement of hepatocyte growth factor
Development 133:2023–2032, 2006. in increased integrin expression on HepG2 cells triggered by adhe-
Heller RS, et al: Expression patterns of Wnts, Frizzleds, sFRPs, and sion to endothelial cells, Br J Cancer 75:47–53, 1997.
misexpression in transgenic mice suggesting a role for Wnts in pan- Keng VW, et al: Homeobox gene Hex is essential for onset of mouse
creas and foregut pattern formation, Dev Dyn 225:260–270, 2002. embryonic liver development and differentiation of the monocyte
Henseleit KD, et al: NKX6 transcription factor activity is required for lineage, Biochem Biophys Res Commun 276:1155–1161, 2000.
alpha- and beta-cell development in the pancreas, Development Kesavan G, et al: Cdc42-mediated tubulogenesis controls cell specifi-
132:3139–3149, 2005. cation, Cell 139:791–801, 2009.
Hentsch B, et al: Hlx homeo box gene is essential for an inductive Kim SK, et al: Activin receptor patterning of foregut organogenesis,
tissue interaction that drives expansion of embryonic liver and gut, Genes Dev 14:1866–1871, 2000.
Genes Dev 10:70–79, 1996. Kim SK, et al: Notochord to endoderm signaling is required for pan-
Hisaoka M, et al: Pancreatic morphogenesis and extracellular matrix creas development, Development 124:4243–4252, 1997.
organization during rat development, Differentiation 53:163–172, Komuta M, et al: Clinicopathological study on cholangiolocellular
1993. carcinoma suggesting hepatic progenitor cell origin, Hepatology
Hogan BL: Morphogenesis, Cell 96:225–233, 1999. 47:1544–1556, 2008.
Holland AM, et al: Experimental control of pancreatic development Kondo T, et al: Function of posterior HoxD genes in the morphogen-
and maintenance, Proc Natl Acad Sci U S A 99:12236–12241, esis of the anal sphincter, Development 122:2651–2659, 1996.
2002. Krapp A, et al: The bHLH protein PTF1-p48 is essential for the for-
Hua H, et al: BMP4 regulates pancreatic progenitor cell expansion mation of the exocrine and the correct spatial organization of the
through Id2, J Biol Chem 281:13574–13580, 2006. endocrine pancreas, Genes Dev 12:3752–3763, 1998.
Huang D, et al: A conserved retinoic acid responsive element in the Kumar M, et al: Signals from lateral plate mesoderm instruct endo-
murine Hoxb-1 gene is required for expression in the developing gut, derm toward a pancreatic fate, Dev Biol 259:109–122, 2003.
Development 125:3235–3246, 1998. Lammert E, et al: Induction of pancreatic differentiation by signals
Huch M: Regenerative biology: the versatile and plastic liver, Nature from blood vessels, Science 294:564–567, 2001.
517:155–156, 2015. Landry C, et al: HNF-6 is expressed in endoderm derivatives and
Huch M, et al: In vitro expansion of single Lgr5+ liver stem cells nervous system of the mouse embryo and participates to the cross-
induced by Wnt-driven regeneration, Nature 494:247–250, 2013. regulatory network of liver-enriched transcription factors, Dev Biol
Hui H, et al: Glucagon-like peptide 1 induces differentiation of islet 192:247–257, 1997.
duodenal homeobox-1-positive pancreatic ductal cells into insulin- Lee CS, et al: The initiation of liver development is dependent on Foxa
secreting cells, Diabetes 50:785–796, 2001. transcription factors, Nature 435:944–947, 2005a.
Hynes RO: Integrins: versatility, modulation, and signaling in cell adhe- Lee MS, et al: Accumulation of extracellular matrix and developmental
sion, Cell 69:11–25, 1992. dysregulation in the pancreas by transgenic production of transform-
Jacob A, et al: Retinoic acid-mediated activation of HNF-3 alpha ing growth factor-beta 1, Am J Pathol 147:42–52, 1995.
during EC stem cell differentiation, Nucleic Acids Res 22:2126–2133, Lee CS, et al: Foxa2 is required for the differentiation of pancreatic
1994. alpha-cells, Dev Biol 278:484–495, 2005b.
Jacquemin P, et al: The Onecut transcription factor HNF-6 (OC-1) is Lemaigre FP: Development of the biliary tract, Mech Dev 120:81–87,
required for timely specification of the pancreas and acts upstream 2003.
of Pdx-1 in the specification cascade, Dev Biol 258:105–116, 2003. Li L, et al: Alagille syndrome is caused by mutations in human Jagged1,
Jensen J: Gene regulatory factors in pancreatic development, Dev Dyn which encodes a ligand for Notch1, Nat Genet 16:243–251, 1997.
229:176–200, 2004. Li Z, et al: Multifaceted pancreatic mesenchymal control of epithelial
Jensen J, et al: Independent development of pancreatic alpha- and beta- lineage selection, Dev Biol 269:252–263, 2004.
cells from neurogenin3-expressing precursors: a role for the notch Li W, et al: In vivo reprogramming of pancreatic acinar cells to three
pathway in repression of premature differentiation, Diabetes 49:163– islet endocrine subtypes, eLife 3:e01846, 2014.
176, 2000. Louis AA, et al: Hepatic jagged1 expression studies, Hepatology
Jiang FX, et al: Regulation of laminin 1-induced pancreatic beta-cell 30:1269–1275, 1999.
differentiation by alpha6 integrin and alpha-dystroglycan, Mol Med Maestro MA, et al: Hnf6 and Tcf2 (MODY5) are linked in a gene
7:107–114, 2001. network operating in a precursor cell domain of the embryonic pan-
Johansson KA, et al: Temporal control of neurogenin3 activity in pancreas, Hum Mol Genet 12:3307–3314, 2003.
creas progenitors reveals competence windows for the generation of Mahlapuu M, et al: The forkhead transcription factor Foxf1 is required
different endocrine cell types, Dev Cell 12:457–465, 2007. for differentiation of extra-embryonic and lateral plate mesoderm,
Jonatan D, et al: Sox17 regulates insulin secretion in the normal and Development 128:155–166, 2001.
pathologic mouse beta cell, PLoS ONE 9:e104675, 2014. Mamin A, Philippe J: Activin A decreases glucagon and arx gene
Jonsson J, et al: Insulin-promoter-factor 1 is required for pancreas expression in alpha-cell lines, Mol Endocrinol 21:259–273, 2007.
development in mice, Nature 371:606–609, 1994. Margagliotti S, et al: The Onecut transcription factors HNF-6/OC-1
Jung J, et al: Initiation of mammalian liver development from endo- and OC-2 regulate early liver expansion by controlling hepatoblast
derm by fibroblast growth factors, Science 284:1998–2003, 1999. migration, Dev Biol 311:579–589, 2007.
Kaestner KH, et al: Unified nomenclature for the winged helix/ Margagliotti S, et al: Role of metalloproteinases at the onset of liver
forkhead transcription factors, Genes Dev 14:142–146, 2000. development, Dev Growth Differ 50:331–338, 2008.
Chapter 1 Embryologic development of the liver, biliary tract, and pancreas 31.e3
Martin M, et al: Dorsal pancreas agenesis in retinoic acid-deficient Roberts DJ, et al: Sonic hedgehog is an endodermal signal inducing
Raldh2 mutant mice, Dev Biol 284:399–411, 2005. Bmp-4 and Hox genes during induction and regionalization of the
Martinez Barbera JP, et al: The homeobox gene Hex is required in chick hindgut, Development 121:3163–3174, 1995.
definitive endodermal tissues for normal forebrain, liver and thyroid Rossi JM, et al: Distinct mesodermal signals, including BMPs from the
formation, Development 127:2433–2445, 2000. septum transversum mesenchyme, are required in combination for
Mashima H, et al: Betacellulin and activin A coordinately convert hepatogenesis from the endoderm, Genes Dev 15:1998–2009, 2001.
amylase-secreting pancreatic AR42J cells into insulin-secreting cells, Sadler TW, Langman J: Langman’s medical embryology, ed 10, Philadel-
J Clin Invest 97:1647–1654, 1996. phia, 2006, Lippincott Williams & Wilkins.
Matsumoto K, et al: Liver organogenesis promoted by endothelial cells Sander M, et al: Homeobox gene Nkx6.1 lies downstream of Nkx2.2
prior to vascular function, Science 294:559–563, 2001. in the major pathway of beta-cell formation in the pancreas, Develop-
Matsuoka TA, et al: Members of the large Maf transcription family ment 127:5533–5540, 2000.
regulate insulin gene transcription in islet beta cells, Mol Cell Biol Sanvito F, et al: TGF-beta 1 overexpression in murine pancreas
23:6049–6062, 2003. induces chronic pancreatitis and, together with TNF-alpha, triggers
McCright B, et al: A mouse model of Alagille syndrome: Notch2 as a insulin-dependent diabetes, Biochem Biophys Res Commun 217:1279–
genetic modifier of Jag1 haploinsufficiency, Development 129:1075– 1286, 1995.
1082, 2002. Schaffer AE, et al: Nkx6.1 controls a gene regulatory network required
McLin VA, et al: Repression of Wnt/beta-catenin signaling in the ante- for establishing and maintaining pancreatic Beta cell identity, PLoS
rior endoderm is essential for liver and pancreas development, Devel- Genet 9:e1003274, 2013.
opment 134:2207–2217, 2007. Schaub JR, et al: Evidence against a stem cell origin of new hepatocytes
McMinn RM, Kugler JH: The glands of the bile and pancreatic ducts: in a common mouse model of chronic liver injury, Cell Rep 8:933–
autoradiographic and histochemical studies, J Anat 95:1–11, 1961. 939, 2014.
Michalopoulos GK, et al: HGF-, EGF-, and dexamethasone-induced Schmidt C, et al: Scatter factor/hepatocyte growth factor is essential
gene expression patterns during formation of tissue in hepatic organ- for liver development, Nature 373:699–702, 1995.
oid cultures, Gene Expr 11:55–75, 2003. Schuppan D: Structure of the extracellular matrix in normal and
Miralles F, et al: Follistatin regulates the relative proportions of endo- fibrotic liver: collagens and glycoproteins, Semin Liver Dis 10:1–10,
crine versus exocrine tissue during pancreatic development, Develop- 1990.
ment 125:1017–1024, 1998. Sellick GS, et al: Mutations in PTF1A cause pancreatic and cerebellar
Moore-Scott BA, et al: Identification of molecular markers that are agenesis, Nat Genet 36:1301–1305, 2004.
expressed in discrete anterior-posterior domains of the endoderm Seo YW, et al: FOR, a novel orphan nuclear receptor related to farne-
from the gastrula stage to mid-gestation, Dev Dyn 236:1997–2003, soid X receptor, J Biol Chem 277:17836–17844, 2002.
2007. Sergi C, et al: Contribution of apoptosis and apoptosis-related proteins
Murtaugh LC, et al: Beta-catenin is essential for pancreatic acinar but to the malformation of the primitive intrahepatic biliary system in
not islet development, Development 132:4663–4674, 2005. Meckel syndrome, Am J Pathol 156:1589–1598, 2000.
Nelson SB, et al: The transcription factors Nkx6.1 and Nkx6.2 possess Serls AE, et al: Different thresholds of fibroblast growth factors pattern
equivalent activities in promoting beta-cell fate specification in the ventral foregut into liver and lung, Development 132:35–47,
Pdx1+ pancreatic progenitor cells, Development 134:2491–2500, 2005.
2007. Seymour PA: Sox9: a master regulator of the pancreatic program, Rev
Nishimura W, et al: Expression of MafA in pancreatic progenitors is Diabet Stud 11:51–83, 2014.
detrimental for pancreatic development, Dev Biol 333:108–120, 2009. Seymour PA, et al: SOX9 is required for maintenance of the pancreatic
Nishimura W, et al: A switch from MafB to MafA expression accom- progenitor cell pool, Proc Natl Acad Sci U S A 104:1865–1870, 2007.
panies differentiation to pancreatic beta-cells, Dev Biol 293:526–539, Shi G, et al: Maintenance of acinar cell organization is critical to pre-
2006. venting Kras-induced acinar-ductal metaplasia, Oncogene 32:1950–
Nishimura W, et al: MafA is critical for maintenance of the mature beta 1958, 2013.
cell phenotype in mice, Diabetologia 58:566–574, 2015. Shiojiri N, Koike T: Differentiation of biliary epithelial cells from the
Ober EA, et al: Mesodermal Wnt2b signalling positively regulates liver mouse hepatic endodermal cells cultured in vitro, Tohoku J Exp Med
specification, Nature 442:688–691, 2006. 181:1–8, 1997.
Offield MF, et al: PDX-1 is required for pancreatic outgrowth and Shiojiri N, et al: Cell lineages and oval cell progenitors in rat liver
differentiation of the rostral duodenum, Development 122:983–995, development, Cancer Res 51:2611–2620, 1991.
1996. Shiojiri N, et al: Suppression of C/EBP alpha expression in biliary cell
Oliver G, et al: Prox 1, a prospero-related homeobox gene expressed differentiation from hepatoblasts during mouse liver development,
during mouse development, Mech Dev 44:3–16, 1993. J Hepatol 41:790–798, 2004.
Peifer M, Polakis P: Wnt signaling in oncogenesis and embryogenesis— Shiozaki S, et al: Impaired differentiation of endocrine and exocrine
a look outside the nucleus, Science 287:1606–1609, 2000. cells of the pancreas in transgenic mouse expressing the truncated
Piccand J, et al: Rfx6 Maintains the functional identity of adult pan- type II activin receptor, Biochim Biophys Acta 1450:1–11, 1999.
creatic beta cells, Cell Rep 9:2219–2232, 2014. Shroff S, et al: SOX9: a useful marker for pancreatic ductal lineage of
Pictet RL, et al: An ultrastructural analysis of the developing embryonic pancreatic neoplasms, Hum Pathol 45:456–463, 2014.
pancreas, Dev Biol 29:436–467, 1972. Simeone DM, et al: Islet hypertrophy following pancreatic disruption
Poll AV, et al: A vHNF1/TCF2-HNF6 cascade regulates the transcrip- of Smad4 signaling, Am J Physiol 291:E1305–E1316, 2006.
tion factor network that controls generation of pancreatic precursor Sjodin A, et al: Mouse R-cadherin: expression during the organogen-
cells, Diabetes 55:61–69, 2006. esis of pancreas and gastrointestinal tract, Exp Cell Res 221:413–425,
Prasadan K, et al: Glucagon is required for early insulin-positive dif- 1995.
ferentiation in the developing mouse pancreas, Diabetes 51:3229– Smith WC, Harland RM: Expression cloning of noggin, a new dorsal-
3236, 2002. izing factor localized to the Spemann organizer in Xenopus embryos,
Pulkkinen MA, et al: The IIIb isoform of fibroblast growth factor Cell 70:829–840, 1992.
receptor 2 is required for proper growth and branching of pancreatic Smith SB, et al: Rfx6 directs islet formation and insulin production in
ductal epithelium but not for differentiation of exocrine or endocrine mice and humans, Nature 463:775–780, 2010.
cells, Mech Dev 120:167–175, 2003. Smith SB, et al: Autoregulation and maturity onset diabetes of the
Puri S, et al: Plasticity and dedifferentiation within the pancreas: young transcription factors control the human PAX4 promoter,
development, homeostasis, and disease, Cell Stem Cell 16:18–31, J Biol Chem 275:36910–36919, 2000.
2015. Sosa-Pineda B, et al: The Pax4 gene is essential for differentiation of
Rausa F, et al: The cut-homeodomain transcriptional activator HNF-6 insulin-producing beta cells in the mammalian pancreas, Nature
is coexpressed with its target gene HNF-3 beta in the developing 386:399–402, 1997.
murine liver and pancreas, Dev Biol 192:228–246, 1997. Sosa-Pineda B, et al: Hepatocyte migration during liver development
Ritvos O, et al: Activin disrupts epithelial branching morphogenesis in requires Prox1, Nat Genet 25:254–255, 2000.
developing glandular organs of the mouse, Mech Dev 50:229–245, Spence JR, et al: Sox17 regulates organ lineage segregation of ventral
1995. foregut progenitor cells, Dev Cell 17:62–74, 2009.
Stafford D, et al: A conserved role for retinoid signaling in vertebrate Watt AJ, et al: Development of the mammalian liver and ventral pan-
pancreas development, Devt Genes Evol 214:432–441, 2004. creas is dependent on GATA4, BMC Dev Biol 7:37, 2007.
Stafford D, Prince VE: Retinoic acid signaling is required for a critical Weinstein M, et al: Smad proteins and hepatocyte growth factor control
early step in zebrafish pancreatic development, Curr Biol 12:1215– parallel regulatory pathways that converge on beta1-integrin to
1220, 2002. promote normal liver development, Mol Cell Biol 21:5122–5131,
Stoffers DA, et al: Early-onset type-II diabetes mellitus (MODY4) 2001.
linked to IPF1, Nat Genet 17:138–139, 1997a. Weinstein DC, et al: The winged-helix transcription factor HNF-3 beta
Stoffers DA, et al: Pancreatic agenesis attributable to a single nucleo- is required for notochord development in the mouse embryo, Cell
tide deletion in the human IPF1 gene coding sequence, Nat Genet 78:575–588, 1994.
15:106–110, 1997b. Wells JM, et al: Wnt/beta-catenin signaling is required for development
St-Onge L, et al: Pax6 is required for differentiation of glucagon- of the exocrine pancreas, BMC Dev Biol 7:4, 2007.
producing alpha-cells in mouse pancreas, Nature 387:406–409, Wells JM, Melton DA: Early mouse endoderm is patterned by soluble
1997. factors from adjacent germ layers, Development 127:1563–1572,
Strobel O, et al: Pancreatic duct glands are distinct ductal compart- 2000.
ments that react to chronic injury and mediate Shh-induced meta- Wilcox CL, et al: Pancreatic alpha-cell specific deletion of mouse Arx
plasia, Gastroenterology 138:1166–1177, 2010. leads to alpha-cell identity loss, PLoS ONE 8:e66214, 2013.
Sussel L, et al: Mice lacking the homeodomain transcription factor Xue Y, et al: Embryonic lethality and vascular defects in mice lacking
Nkx2.2 have diabetes due to arrested differentiation of pancreatic the Notch ligand Jagged1, Hum Mol Genet 8:723–730, 1999.
beta cells, Development 125:2213–2221, 1998. Yamaoka T, et al: Hypoplasia of pancreatic islets in transgenic mice
Suzuki A, et al: Clonal identification and characterization of self- expressing activin receptor mutants, J Clin Invest 102:294–301,
renewing pluripotent stem cells in the developing liver, J Cell Biol 1998.
156:173–184, 2002. Yamaoka T, et al: Diabetes and pancreatic tumours in transgenic mice
Talchai C, et al: Pancreatic beta cell dedifferentiation as a mechanism expressing Pax 6, Diabetologia 43:332–339, 2000.
of diabetic beta cell failure, Cell 150:1223–1234, 2012. Yang JT, et al: Embryonic mesodermal defects in alpha 5 integrin-
Tan X, et al: Beta-catenin deletion in hepatoblasts disrupts hepatic deficient mice, Development 119:1093–1105, 1993.
morphogenesis and survival during mouse development, Hepatology Yang YP, et al: Context-specific alpha- to-beta-cell reprogramming by
47:1667–1679, 2008. forced Pdx1 expression, Genes Dev 25:1680–1685, 2011.
Tarlow BD, et al: Bipotential adult liver progenitors are derived from Yee NS, et al: Zebrafish pdx1 morphant displays defects in pancreas
chronically injured mature hepatocytes, Cell Stem Cell 15:605–618, development and digestive organ chirality, and potentially identifies
2014. a multipotent pancreas progenitor cell, Genesis 30:137–140, 2001.
Terada T, Nakanuma Y: Expression of tenascin, type IV collagen and Yew KH, et al: Interplay of glucagon-like peptide-1 and transforming
laminin during human intrahepatic bile duct development and in growth factor-beta signaling in insulin-positive differentiation of
intrahepatic cholangiocarcinoma, Histopathology 25:143–150, 1994. AR42J cells, Diabetes 53:2824–2835, 2004.
Terada T, Nakanuma Y: Detection of apoptosis and expression of Yokouchi Y, et al: Coordinated expression of Abd-B subfamily genes
apoptosis-related proteins during human intrahepatic bile duct of the HoxA cluster in the developing digestive tract of chick embryo,
development, Am J Pathol 146:67–74, 1995. Dev Biol 169:76–89, 1995.
Thomas PQ, et al: Hex: a homeobox gene revealing peri-implantation Yoshitomi H, Zaret KS: Endothelial cell interactions initiate dorsal
asymmetry in the mouse embryo and an early transient marker of pancreas development by selectively inducing the transcription factor
endothelial cell precursors, Development 125:85–94, 1998. Ptf1a, Development 131:807–817, 2004.
Thorel F, et al: Normal glucagon signaling and beta-cell function after Zaret K: Developmental competence of the gut endoderm: genetic
near-total alpha-cell ablation in adult mice, Diabetes 60:2872–2882, potentiation by GATA and HNF3/fork head proteins, Dev Biol
2011. 209:1–10, 1999.
Thorel F, et al: Conversion of adult pancreatic alpha-cells to beta-cells Zaret KS: Regulatory phases of early liver development: paradigms of
after extreme beta-cell loss, Nature 464:1149–1154, 2010. organogenesis, Nat Rev Genet 3:499–512, 2002.
Tiso N, et al: BMP signalling regulates anteroposterior endoderm pat- Zhang YQ, et al: Changes in the expression of transcription factors in
terning in zebrafish, Mech Dev 118:29–37, 2002. pancreatic AR42J cells during differentiation into insulin-producing
Tulachan SS, et al: TGF-beta isoform signaling regulates secondary cells, Diabetes 50(Suppl 1):S10–S14, 2001.
transition and mesenchymal-induced endocrine development in the Zhang W, et al: Regulation of Hex gene expression and initial stages
embryonic mouse pancreas, Dev Biol 305:508–521, 2007. of avian hepatogenesis by Bmp and Fgf signaling, Dev Biol 268:312–
Villasenor A, et al: Epithelial dynamics of pancreatic branching mor- 326, 2004.
phogenesis, Development 137:4295–4305, 2010. Zhao R, et al: GATA6 is essential for embryonic development of the
Vincent M, et al: Abrogation of protein convertase 2 activity results in liver but dispensable for early heart formation, Mol Cell Biol 25:2622–
delayed islet cell differentiation and maturation, increased alpha-cell 2631, 2005.
proliferation, and islet neogenesis, Endocrinology 144:4061–4069, Zhou J, et al: Glucagon-like peptide 1 and exendin-4 convert pancre-
2003. atic AR42J cells into glucagon- and insulin-producing cells, Diabetes
Wang Q, et al: Ghrelin is a novel target of Pax4 in endocrine progeni- 48:2358–2366, 1999.
tors of the pancreas and duodenum, Dev Dyn 237:51–61, 2008. Zucman-Rossi J, et al: Genotype-phenotype correlation in hepatocel-
Wang J, et al: The concerted activities of Pax4 and Nkx2.2 are essential lular adenoma: new classification and relationship with HCC, Hepa-
to initiate pancreatic beta-cell differentiation, Dev Biol 266:178–189, tology 43:515–524, 2006.
2004.
Wang X, et al: Increased levels of forkhead box M1B transcription
factor in transgenic mouse hepatocytes prevent age-related prolifera-
tion defects in regenerating liver, Proc Natl Acad Sci U S A 98:11468–
11473, 2001.
CHAPTER 2
Surgical and radiologic anatomy of the liver, biliary
tract, and pancreas
Leslie H. Blumgart, Lawrence H. Schwartz, and Ronald P. DeMatteo*
ANATOMY OVERVIEW the pancreas as a retroperitoneal structure passing upward

behind the foramen of Winslow posterior to the right hilar struc-
Precise knowledge of the architecture of the liver, biliary tract, tures of the liver. The renal veins lie in front of the arteries and
and pancreas and the related blood vessels and lymphatic drain- join the IVC at almost a right angle on the left and obliquely on
age is essential for the successful performance of hepatopancre- the right. The IVC is embraced in a groove on the posterior
aticobiliary surgical operations. surface of the liver. The IVC comes to lie on the right crus of
the diaphragm, behind the bare area of the liver; it extends to
LIVER the central tendon of the diaphragm, which it pierces on a level
with the body of T8, behind and higher than the beginning of
The liver lies protected under the lower ribs, closely applied to the abdominal aorta. While the IVC courses upward, it is sepa-
the undersurface of the diaphragm and on top of the inferior rated from the right crus of the diaphragm by the right celiac
vena cava (IVC) posteriorly (Fig. 2.1). Most of the liver bulk ganglion and, higher up, by the right phrenic artery. The right
lies to the right of the midline, where the lower border lies near adrenal vein is a short vessel that enters the IVC behind the bare
the right costal margin. The liver extends as a wedge to the left area. There may be a small accessory right adrenal vein on the
of the midline, between the anterior surface of the stomach and right that enters into the confluence of the right renal vein and
the left dome of the diaphragm. The upper surface is boldly the IVC. Also, occasionally, a right adrenal vein drains directly
convex and molded to the diaphragm, and the surface projec- into the posterior liver. The lumbar veins drain posterolaterally
tion on the anterior body wall extends up to the fourth inter- into the IVC below the level of the renal veins, but above this
costal space on the right and to the fifth intercostal space on level, there are usually no vena caval tributaries posteriorly.
the left. The convexity of the upper surface slopes down to a
posterior surface that is triangular in outline. The liver is Hepatic Veins
invested with peritoneum except on the posterior surface, The hepatic veins (Figs. 2.2 to 2.4) drain directly from the
where the peritoneum reflects onto the diaphragm, forming the upper part of the posterior surface of the liver at an oblique
right and left triangular ligaments. The undersurface of the liver angle directly into the vena cava. The right hepatic vein, which
is concave and extends down to a sharp anterior border. The is larger than the left and middle hepatic veins, has a short
posterior surface of the liver is triangular in outline with its base extrahepatic course of approximately 1 to 2 cm. The left and
to the right, and here the liver lying between the upper and middle hepatic veins may drain separately into the IVC but are
lower “leaves” of the triangular ligaments is bare and devoid of usually joined, after a short extrahepatic course, to form a
peritoneum. The peritoneum reflects onto the right posterior common venous channel approximately 2 cm in length that
liver from the medial aspect of Gerota’s fascia, which is associ- traverses to the left part of the anterior surface of the IVC below
ated with the right kidney. The right adrenal gland lies beneath the diaphragm. In addition to the three major hepatic veins,
this reflection. The anterior border lies under cover of the right there is the umbilical vein, which is single in most cases and
costal margin, lateral to the right rectus abdominis muscle, but runs beneath the falciform ligament between the middle and
it slopes upward to the left across the epigastrium. Anteriorly, left hepatic veins; it empties into the terminal portion of the left
the convex surface of the liver lies against the concavity of the hepatic vein, although, rarely, it drains into the middle hepatic
diaphragm and is attached to it by the falciform ligament, left vein or directly into the confluence of the middle and left
triangular ligament, and upper layer of the right triangular hepatic veins. In approximately 15% of patients, an accessory
ligament. right hepatic vein is present inferiorly (see Fig. 2.3). Hepatic
venous drainage of the caudate lobe is directly into the IVC, as
Retrohepatic Inferior Vena Cava described later.
The IVC runs to the right of the aorta on the bodies of the This classic description of the anatomy of the liver is suffi-
lumbar vertebrae, diverging from the aorta as it passes upward. cient for gross appreciation and for mobilization of the liver to
Below the liver, the IVC lies behind the duodenum and head of allow access for repair of injuries, liver transplantation, or the
placement of probes onto or into the liver substance. Hidden
beneath this external gross appearance is a detailed internal
anatomy, an understanding of which is essential to the perfor-
*The authors acknowledge Dr. Lucy E. Hann who coauthored this chapter in mance of precise hepatectomy. This internal anatomy has been
the fifth edition of this book. Much of her initial contribution is included here. called the functional anatomy of the liver.
32
Chapter 2 Surgical and radiologic anatomy of the liver, biliary tract, and pancreas 33
Gallbladder
Quadrate lobe
Umbilical
fissure
Left lobe
Right lobe
Gallbladder Caudate lobe

A B
IVC
Lumen of MHV Lumen of RHV
and LHV
Left triangular LHV
ligament Right lobe
Left lobe
Right adrenal
vein
Ligamentum Right triangular
venosum ligament
Ligamentum
teres
LEFT LOBE RIGHT LOBE

C
FIGURE 2.1. A, The liver as seen in situ has two main lobes, a large right and a smaller left, and conventional description places their line of fusion
on the upper surface of the liver along the attachment of the falciform ligament at the inferior extent of which the ligamentum teres enters the umbili-
cal fissure. B, With the liver flipped upward, the inferior surface of the right lobe is seen as the transverse hilar fissure, which constitutes the posterior
limit of this lobe. The portion of the right lobe located anterior to the fissure is called the quadrate lobe, which is limited on the left by the umbilical
fissure and on the right by the gallbladder fossa. Posterior to the hilar transverse fissure is a fourth lobe, the caudate lobe, which hugs the inferior
vena cava (IVC) and extends upward on its left side. Thus the liver comprises two main lobes and two smaller lobes, separated by visible, well-defined
fissures on the liver surface. C, The posterior aspect of the liver is shown. The IVC lies snugly in a deep groove within the bare area; the hepatic
veins open directly into it. Within this bare area, the right suprarenal gland lies adjacent to the IVC, and the adrenal vein drains into the right of the
IVC. The remainder of the bare area of liver is directly in contact with the diaphragm. To the left of the IVC, the caudate lobe slopes upward from
the inferior to the posterior surface of the liver and is demarcated on the left by a fissure, within which lies the ligamentum venosum. The gastrohe-
patic omentum is attached to the ligamentum venosum, placing the caudate lobe within the lesser sac of the peritoneum. The left lobe of the liver
is situated anteriorly in the supracolic compartment of the peritoneal cavity. The posterior surface of the left lobe is narrow; there is a very fine bare
area on this side. While the vena cava traverses upward in the groove on the posterior surface of the liver, it is shielded on the right side by a layer
of fibrous tissue that passes from the posterior edge of the liver backward toward the lumbar vertebrae and fans out posteriorly, especially in the
upper part. Behind the IVC, a prolongation of this fibrous layer joins a less marked fibrous extension from the lateral edge of the caudate lobe. This
layer of fibrous tissue, sometimes called the ligament of the vena cava, must be divided on the right, to allow surgical exposure of the IVC and the
right hepatic vein, and on the left, to allow mobilization of the caudate lobe. Occasionally, the liver tissue embraces the vena cava completely, so
that it runs within a tunnel of parenchyma. LHV, Left hepatic vein; MHV, middle hepatic vein; RHV, right hepatic vein.
M
L
R
IVC
FIGURE 2.4. The anterior surfaces of the major extrahepatic veins and
the inferior vena cava are retroperitoneal and masked behind the layers
of the falciform ligament, while it splits and passes to the right and left
triangular ligaments. The left and middle hepatic veins usually join within
the liver, and not outside the liver as depicted here for visual simplicity.
FIGURE 2.2. Transverse ultrasound image of the hepatic vein conflu-
ence shows the left (L), middle (M), and right (R) hepatic veins as they
join the inferior vena cava (IVC).
M
M
IVC
IVC
R
R
A B
R
PV IVC
IVC A
C D
FIGURE 2.3. Two inferior accessory right hepatic veins. A, Contrast-enhanced computed tomographic (CT) image of the hepatic vein confluence.
IVC, inferior vena cava; M, middle hepatic vein;R, Right hepatic vein. B, A small right inferior accessory vein (arrow) enters the IVC below the hepatic
venous confluence. C, The second, larger right inferior accessory right hepatic vein (arrow) is seen more inferiorly. PV, Portal vein. D, CT coronal
reconstruction image shows the right hepatic vein (R) and one right inferior accessory vein (arrow). A, Aorta.
subdivisions are referred to as segments in the description of

Functional Surgical Anatomy Goldsmith and Woodburne (1957), but in Couinaud’s nomen-
The internal architecture of the liver is composed of a series of clature (1957), they are termed sectors.
segments that combine to form sectors separated by scissurae The right portal scissura separates the right liver into two
that contain the hepatic veins (Fig. 2.5), as described by Couin- sectors: anteromedial (anterior) and posterolateral (posterior).
aud (1957). Together or separately, these constitute the visible With the body supine, this scissura is almost in the frontal
lobes described previously. The internal structure has been plane. The right hepatic vein runs within the right scissura. The
clarified by the publications of McIndoe and Counseller (1927), left portal scissura divides the left liver into two sectors, but the
Ton That Tung (1939, 1979), Hjörtsjö (1931), Healey and left portal scissura is not within the umbilical fissure because
Schroy (1953), Goldsmith and Woodburne (1957), Couinaud this fissure is not a portal scissura, and instead it contains a
(1957), and Bismuth and colleagues (1982). Essentially, the portal pedicle. The left portal scissura is located posterior to
three main hepatic veins within the scissurae divide the liver the ligamentum teres and within the left liver, along the course
into four sectors, each of which receives a portal pedicle. The of the left hepatic vein.
main portal scissura contains the middle hepatic vein and pro- Although the description by Couinaud has been used widely,
gresses from the middle of the gallbladder bed anteriorly to the it is being replaced by an alternative terminology suggested by a
left of the vena cava posteriorly. The right and left parts of the committee of the International Hepato-Pancreatico-Biliary
liver, demarcated by the main portal scissura, are independent Association in 2000 (Strasberg, 2005). The main difference is
in terms of portal and arterial vascularization and biliary drain- that in the alternative terminology, Couinaud’s sectors are
age (Fig. 2.6). These right and left livers are themselves divided
into two by the remaining portal scissurae. These four
VII
RIGHT MAIN LEFT
VIII
SCISSURA SCISSURA SCISSURA
II
V I
III
IV
VI
Umbilical
fissure
Right portal pedicle Left portal pedicle II
VIII
RIGHT LIVER LEFT VII
FIGURE 2.5. The portal vein, hepatic artery, and draining bile ducts
are distributed within the liver in a beautifully symmetric pedicular pattern, III
which belies the asymmetric external appearance. Each segment (I to
VIII) is supplied by a portal triad composed of a branch of the portal vein
and hepatic artery and drained by a tributary of the right or left main
hepatic ducts. The four sectors demarcated by the three main hepatic IV
veins are called the portal sectors (now referred to as sections in the V
Brisbane terminology); these portions of parenchyma are supplied by
independent portal pedicles. The hepatic veins run between the sectors
in the portal scissurae; the scissurae containing portal pedicles are called VI
the hepatic scissurae. The umbilical fissure corresponds to a hepatic
scissura. The internal architecture of the liver consists of two hemilivers,
the right and the left liver separated by the main portal scissura, also
known as Cantlie’s line. It is preferable to call them the right and left liver, B
rather than the right and left lobes because the latter nomenclature is FIGURE 2.6. The functional division of the liver and its segments
erroneous; there is no visible mark that permits identification of a true according to Couinaud’s nomenclature. A, As seen in the patient. B, In
hemiliver. the ex vivo position.
referred to as sections (Table 2.1) (see Chapter 103B for differ- This prolongation of the left portal pedicle turns anteriorly and
ences in the terminology of the various hepatic resections). Also, caudally within the umbilical fissure, giving branches of supply
note that the left medial section, in the terminology of Strasberg to segment II first and then segment III and recurrent branches
(2005), is composed of one segment (i.e., segment IV). (“feedback vessels”) to segment IV (Fig. 2.8; see Fig. 2.6).
At the hilus of the liver, the right portal triad pursues a short Beneath segment IV, the pedicle is composed of the left branch
course of approximately 1 to 1.5 cm before entering the sub- of the portal vein and the left hepatic duct, but it is joined at
stance of the right liver (Fig. 2.7). In some cases, the right the base of the umbilical fissure by the left branch of the hepatic
anterior and posterior pedicles arise independently, and their artery.
origins may be separated by 2 cm. In some cases, it appears as The branching of the portal pedicle at the hilus (Fig. 2.9),
if the left portal vein arises from the right anterior branch (see the distribution of the branches to the caudate lobe (segment
Fig. 2.40). On the left side, however, the portal triad crosses I) on the right and left sides, and the distribution to the seg-
over approximately 3 to 4 cm beneath segment IV (formerly ments of the right (segments V through VIII) and left (segments
called the quadrate lobe), embraced in a peritoneal sheath at the II through IV) hemiliver follow a remarkably symmetric pattern
upper end of the gastrohepatic ligament and separated from the and, as described by Scheele (1994), allow separation of
undersurface of segment IV by connective tissue (hilar plate). segment IV into segment IVa superiorly and segment IVb
TABLE 2.1 Brisbane Terminology of Liver Anatomy

and Resections
Couinaud
Anatomic Term Segments Surgical Resection
Right hemiliver/ 5-8 Right hepatectomy

right liver
RAPV Left hemiliver/left 2-4 Left hepatectomy
liver
LPV
Right anterior 5, 8 Right anterior sectionectomy
section
RPPV Right posterior 6, 7 Right posterior sectionectomy
MPV section
Left medial section 4 Left medial sectionectomy or
Resection of segment 4
IVC
Left lateral section 2, 3 Left lateral sectionectomy or
Bisectionectomy 2, 3
A 4, 5, 6, 7, 8 Right trisectionectomy or
Extended right hepatectomy
2, 3, 4, 5, 8 Left trisectionectomy or
Extended left hepatectomy
4 3
U
2
p
B
FIGURE 2.7. A, Transverse sonogram at the level of the portal vein
bifurcation. The main portal vein (MPV) bifurcates into the left and right
portal veins (LPV and RPV). The RPV bifurcates shortly into the right IVC
anterior (RAPV) and right posterior (RPPV) branches, but the LPV has a
longer horizontal course within the hilar plate. The inferior vena cava
(IVC) is seen posteriorly. B, Coronal view of computed tomographic FIGURE 2.8. Transverse sonogram shows the branching pattern of
angioportography. Reconstruction shows the right hepatic vein (open the left portal vein (P), which courses horizontally and into the umbilical
arrow) and the portal vein (large arrow); anterior and posterior sectional fissure. The umbilical portion of the left portal vein (U) gives branches to
branches of the RPV (small arrows) are seen to arise directly and sepa- the left hepatic segments (II to IV). The left hepatic vein (arrow) and
rately from the main portal trunk. inferior vena cava (IVC) also are shown.
MHV Ligamentum venosum
II/III
IV
LPV
Lesser
omentum
VII
L
I
IVC
RA R
FIGURE 2.11. The main bulk of the caudate lobe (segment I; dark
area) lies to the left of the inferior vena cava (IVC); the left and inferior
margins are free in the lesser omental bursa. The gastrohepatic (lesser)
omentum separates the left portion of the caudate from segments II and
RP
III of the liver, while it passes between them to be attached to the liga-
mentum venosum. The left portion of the caudate lobe inferiorly tra-
verses to the right between the left portal vein (LPV) and IVC as the
caudate process, where it fuses with the right lobe of the liver. Note the
position of the middle hepatic vein (MHV).
2.10 to 2.11). The caudate is intimately related to major vas-

FIGURE 2.9. Contrast-enhanced computed tomographic image of the
cular structures. On the left, the caudate lies between the IVC
portal vein bifurcation. L, Left portal vein; R, right portal vein; RA, right
anterior portal vein; RP, right posterior portal vein.
posteriorly and the left portal triad inferiorly and the IVC and
the middle and left hepatic veins superiorly (Fig. 2.12). The
portion of the caudate on the right varies but is usually quite
small. The anterior surface within the parenchyma is covered
by the posterior surface of segment IV, the limit an oblique
plane slanting from the left portal vein to the left hepatic vein.
Thus there is a caudate lobe with a constantly present left
portion and a right portion of variable size. This portion of
the caudate on the right is adjacent to the recently described
segment IX, which lies between it and segment XIII. The
authors find segment IX of little practical clinical
p
significance.
The caudate is supplied by blood vessels and drained by
IVC
a biliary tributaries from the right and left portal triad. Small
vessels from the portal vein and tributaries joining the biliary
ducts also are found. The right portion of the caudate, includ-
ing the caudate process, predominantly receives portal venous
blood from the right portal vein or from the bifurcation of the
main portal vein, whereas on the left side, the portal supply
arises from the left branch of the portal vein almost exclusively.
FIGURE 2.10. Contrast-enhanced computed tomographic scan of the
Similarly, the arterial supply and biliary drainage of the right
liver shows the intimate relationship of the caudate lobe (arrow), inferior
vena cava (IVC), portal vein (p), and aorta (a). portion is most commonly associated with the right posterior
sectional vessels and the left portion with the left main vessels.
The hepatic venous drainage of the caudate is unique in that it
is the only hepatic segment that drains directly into the IVC.
inferiorly (see Fig. 2.6). This arrangement of subsegments These veins can sometimes drain into the posterior aspect of
mimics the distribution to segments V and VIII on the right the vena cava, if a significant retrocaval caudate component is
side. The umbilical vein provides drainage of at least parts of present.
segment IVb after ligation of the middle hepatic vein, and it is In the usual and common circumstance, the posterior edge
important in the performance of segmental resection. of the caudate lobe on the left has a fibrous component, which
The caudate or segment I is the dorsal portion of the liver fans out and attaches lightly to the crural area of the dia-
lying posteriorly and embraces the retrohepatic IVC (Figs. phragm; but it extends posteriorly, behind the vena cava, to
IVC
MHV
LHV
II/III
Ligamentum
venosum
*
RPV IVC
v
a
PV
LPV
FIGURE 2.12. The caudate lobe (shaded) and segments II and III,
rotated to the patient’s right. Superiorly, the left portion of the caudate
lobe is linked by a deep anterior portion, embedded in the parenchyma
immediately under the middle hepatic vein (MHV), reaching inferiorly to
the posterior margin of the hilus of the liver and fusing anterolaterally to
the inferior vena cava (IVC) on the right side to segments VI and VII of
the right liver. The major blood supply arises from the left branch of the A
left portal vein (LPV) and the left hepatic artery close to the base of the
umbilical fissure of the liver. The hepatic veins (MHV, LHV) are short in
course and drain from the caudate directly into the anterior and left
aspect of the vena cava. LHV, Left hepatic vein; RPV, right portal vein;
PV, main trunk of portal vein.
link with a similar component of fibrous tissue (called the

venal caval ligament) that protrudes from the posterior surface
of segment VII and embraces the vena cava (see Figs. 2.1C
and 2.11). In 50% of patients, this ligament is replaced by
hepatic tissue, in whole or in part, and the caudate may com- p
pletely encircle the IVC and may contact segment VII on the
right side; a significant retrocaval component may prevent a *
left-sided approach to the caudate veins. The caudal margin
v
of the caudate lobe can have a papillary projection that occa-
sionally may attach to the rest of the lobe via a narrow con- a
nection. It is bulky in 27% of cases and can be mistaken for
an enlarged lymph node on computed tomography (CT) scan
(Fig. 2.13).
To summarize:
1. The liver is divided into two hemilivers by the main hepatic
scissura, where the middle hepatic vein runs.
2. The left liver is divided into two sections. The
B
Brisbane 2000 nomenclature describes the left lateral
section (segments 2 and 3) and the left medial section FIGURE 2.13. Computed tomographic image of the caudate lobe with
(segment 4). papillary process. A, Caudate lobe (asterisk) positioned between the left
3. The right liver is divided into an anterior section (segments portal vein (arrow) and inferior vena cava (v). a, Aorta. B, Papillary
5 and 8) and posterior section (segments 6 and 7). process of the caudate (p) represents the lower medial extension of the
4. Segment 1, the caudate lobe, lies posteriorly and embraces caudate (asterisk) and may mimic a periportal lymph node (Arrow) indi-
the IVC, its intraparenchymal anterior surface abutting the cates left portal vein.
posterior surface of segment 4 and merging with segments
6 and 7 on the right (Fig. 2.14; see Fig. 2.11).
Further details of segmental anatomy important in sectional
or segmental resection are described in Chapters 103B and
108B.
Surgical Implications and Exposure

All methods for precise partial hepatectomy depend on control
of the inflow vasculature and draining bile ducts and the outflow
hepatic veins of the portion of liver to be excised, which may
IVa be a segment, a subsegment, or an entire lobe. The remnant
II remaining after partial hepatectomy must be provided with an
excellent portal venous inflow, hepatic arterial supply, and
I
VIII biliary drainage and unimpeded hepatic venous outflow. The
classification of the various partial hepatic resection procedures,
incisions and exposure, necessary mobilization of the liver, and
VII the methods of control of the structures within the portal triads
and of the hepatic veins are described in detail in Chapters 103
and 108B.
BILIARY TRACT
A Biliary exposure and precise dissection are the most important
steps in any biliary operative procedure. A thorough under-
standing of biliary anatomy is necessary.
Intrahepatic Bile Duct Anatomy

The right and left livers are drained by the right and the left
hepatic ducts, whereas the caudate lobe is drained by several
IVb III ducts that join both the right and left hepatic ducts. The intra-
hepatic ducts are tributaries of the corresponding hepatic ducts,
which form part of the major portal triads that penetrate the
V I liver, invaginating Glisson’s capsule at the hilus. Bile ducts
usually are located above the corresponding portal branches,
whereas hepatic arterial branches are situated inferiorly to the
VI veins. Each branch of the intrahepatic portal veins corresponds
to bile duct tributaries that join to form the right and left
hepatic ductal systems, converging at the liver hilus to consti-
tute the common hepatic duct. The umbilical fissure divides
the left liver, passing between segments III and IV, which may
B be bridged by a tongue of liver tissue. The ligamentum teres
passes through the umbilical fissure to join the left branch of
the portal vein.
The left hepatic duct drains the three segments—II, III,
and IV—that constitute the left liver (Fig. 2.15). The duct
that drains segment III is located slightly behind the left horn
of the umbilical recess. It is joined by the tributary from
IVb III segment IVb to form the left duct, which is similarly joined
by the duct of segment II and the duct of segment IVa, where
the left branch of the portal vein turns forward and caudally.
V
I The left hepatic duct traverses beneath the left liver at the
base of segment IV, just above and behind the left branch of
the portal vein; it crosses the anterior edge of that vein and
VI joins the right hepatic duct to constitute the hepatic ductal
confluence. In its transverse portion, it receives a few small
branches from segment IV.
The right hepatic duct drains segments V, VI, VII, and VIII
and arises from the junction of two main sectional duct tribu-
taries. The posterior or lateral duct and the anterior or medial
duct are each accompanied by a corresponding vein and artery.
C The right posterior sectional duct has an almost horizontal
FIGURE 2.14. Hepatic segmental anatomy as shown by computed course and constitutes the confluence of the ducts of segments
tomography at A, the level of the hepatic veins, B, at the portal vein VI and VII (Fig. 2.16). The duct then runs to join the right
bifurcation, and C, below the hepatic hilus. anterior sectional duct, as it descends in a vertical manner. The
right anterior sectional duct is formed by the confluence of the
ducts draining segments V and VIII. Its main trunk is located
to the left of the right anterior sectional branch of the portal
II
VIII V IV
VI
III
III
IV
V
V
VII II
Right Left
A B
Left hepatic duct
II
VIII
VII
IV
III
V
Common hepatic
duct
VI
C Right hepatic duct Common bile

duct
FIGURE 2.15. A, Biliary drainage of the two functional hemilivers. Note the position of the right anterior and right posterior sections. The caudate
lobe drains into the right and left ductal system. B, Inferior aspect of the liver. The biliary tract is represented in black, and the portal branches are
represented in white. Note the biliary drainage of segment IV (segment VIII is not represented because of its cephalad location). C, T-tube cholan-
giogram shows the most common arrangement of hepatic ducts.
vein, which pursues an ascending course. The junction of these system only. In about 7%, the drainage is into the right hepatic
two main right biliary channels usually occurs above the right system.
branch of the portal vein. The right hepatic duct is short and
joins the left hepatic duct to constitute the confluence lying in Extrahepatic Biliary Anatomy and Vascular Anatomy of
front of the right portal vein and forming the common hepatic the Liver and Pancreas
duct.
The caudate lobe (segment I) has its own biliary drainage The extrahepatic bile ducts are represented by the extrahepatic
(Healey & Schroy, 1953). The caudate lobe is divided into right segments of the right and left hepatic ducts, joining to form the
and left portions and a caudate process. In 44% of individuals, biliary confluence and the main biliary channel draining to the
three separate ducts drain these three parts of the lobe, whereas duodenum. (Figs. 2.17 and 2.18). The confluence of the right
in another 26%, a common duct lies between the right portion and left hepatic ducts occurs at the right of the hilar fissure of
of the caudate lobe proper and the caudate process and an the liver, anterior to the portal venous bifurcation and overlying
independent duct that drains the left part of the caudate lobe. the origin of the right branch of the portal vein. The extrahe-
The site of drainage of these ducts varies. In 78% of patic segment of the right duct is short, but the left duct has a
cases, drainage of the caudate lobe is into the right and left much longer extrahepatic course. The biliary confluence is
hepatic ducts, but in 15%, drainage is by the left hepatic ductal separated from the posterior aspect of segment IVB of the liver
Right posterior
VIII sectoral duct
A VII
Left hepatic duct
VI
Right anterior
sectoral duct
B
FIGURE 2.16. A, Biliary and vascular anatomy of the right liver. Note the horizontal course of the posterior sectional duct and the vertical course
of the anterior sectional duct. B, Transtubal cholangiogram shows a common normal variant: the right posterior sectional duct drains into the left
hepatic duct. In this case, the posterior duct is anterior to the posterior sectional duct. Frequently in this variant, the posterior duct passes posteriorly
to the anterior sectional pedicle.
by the hilar plate, which is the fusion of connective tissue et  al, 1981). The commonly accepted working definition of
enclosing the biliary and vascular elements with the Glisson the triangle of Calot recognizes, however, the inferior surface
capsule (Fig. 2.19). Because of the absence of any major vas- of the right lobe of the liver as the upper border and the cystic
cular interposition, it is possible to open the connective tissue duct as the lower border (Wood, 1979). Dissection of the
constituting the hilar plate at the inferior border of segment IV triangle of Calot is of key significance during cholecystectomy,
and, by elevating it, to display the biliary confluence and left because in this triangle runs the cystic artery, often the right
hepatic duct (Fig. 2.20). branch of the hepatic artery, and occasionally a bile duct,
which should be displayed before cholecystectomy (see
Main Bile Duct and Sphincter of Oddi Chapter 35). If there is a replaced or accessory common or
The main bile duct, the mean diameter of which is approxi- right hepatic artery, it usually runs behind the cystic duct to
mately 6  mm, is divided into two portions: the upper is called enter the triangle of Calot (Fig. 2.21).
the common hepatic duct and is situated above the cystic duct, The common variations in the relationship of the hepatic
which joins it to form the lower portion, the common bile artery and origin and course of the cystic artery to the biliary
duct (CBD). The common duct courses downward anterior apparatus are shown in Figure 2.22. Ignorance of these varia-
to the portal vein, in the free edge of the lesser omentum; it tions may provoke unexpected hemorrhage or biliary injury
is closely applied to the hepatic artery, which runs upward on (Champetier et al, 1982) during cholecystectomy and may
its left, giving rise to the right branch of the hepatic artery, result in bile duct injury during efforts to secure hemostasis
which crosses the main bile duct usually posteriorly, although (see Chapter 42). The union between the cystic duct and the
in approximately 20% of cases, it crosses anteriorly. The cystic common hepatic duct may be located at various levels. At its
artery, arising from the right branch of the hepatic artery, may lower extrahepatic portion, the CBD traverses the posterior
cross the common hepatic duct posteriorly or anteriorly. The aspect of the pancreas, running in a groove or tunnel. The
common hepatic duct constitutes the left border of the triangle retropancreatic portion of the CBD approaches the second
of Calot, the other corners of which were originally described portion of the duodenum obliquely, accompanied by the ter-
as the cystic duct below and the cystic artery above (Rocko minal part of the pancreatic duct of Wirsung.
a b
c
i
d
j e
f
g
h
k
FIGURE 2.18. Endoscopic retrograde choledochopancreatogram

showing the pancreatic duct (arrow), gallbladder, and biliary tree.
V VIII
VII
C
FIGURE 2.17. Anterior aspect of the biliary anatomy and of the head
of the pancreas: right hepatic duct (a), left hepatic duct (b), common II
hepatic duct (c), hepatic artery (d), gastroduodenal artery (e, cystic duct VI
(f), retroduodenal artery (g), common bile duct (h), neck of the gallbladder B
(i), body of the gallbladder (j), fundus of the gallbladder (k). Note particu-
larly the position of the hepatic bile duct confluence anterior to the right III
branch of the portal vein, the posterior course of the cystic artery behind IV
the common hepatic duct, and the relationship of the neck of the gall-
bladder to the right branch of the hepatic artery. Note also the relation- A
ship of the major vessels (portal vein, superior mesenteric vein, and
superior mesenteric artery) to the head of the pancreas.
FIGURE 2.19. Anatomy of the plate system. A, Cystic plate, above the
Gallbladder and Cystic Duct gallbladder. B, Hilar plate, above the biliary confluence and at the base
The gallbladder is a reservoir located on the undersurface of of segment IV. C, Umbilical plate, above the umbilical portion of the portal
vein. Large, curving arrows indicate the plane of dissection of the cystic
the right lobe of the liver, within the cystic fossa; it is separated
plate during cholecystectomy and of the hilar plate during approaches
from the hepatic parenchyma by the cystic plate, which is com-
to the left hepatic duct.
posed of connective tissue that extends to the left as the hilar
plate (see Fig. 2.19). Sometimes the gallbladder is deeply
embedded in the liver, but occasionally it occurs on a mesen-
teric attachment and may be susceptible to volvulus. The gall- The cystic duct arises from the neck or infundibulum of the
bladder varies in size and consists of a fundus, a body, and a gallbladder and extends to join the common hepatic duct. Its
neck (Fig. 2.23). The fundus usually, but not always, reaches lumen usually measures approximately 1 to 3 mm, and its
the free edge of the liver and is closely applied to the cystic length varies, depending on the type of union with the common
plate. The cystic fossa is a precise anterior landmark to the main hepatic duct. The mucosa of the cystic duct is arranged in spiral
liver incisura. The neck of the gallbladder makes an angle with folds known as the valves of Heister (Wood, 1979). Although the
the fundus and creates Hartmann’s pouch, which may obscure cystic duct joins the common hepatic duct in its supraduodenal
the common hepatic duct and constitute a real danger point segment in 80% of cases, it may extend downward to the ret-
during cholecystectomy. roduodenal or retropancreatic area. Occasionally, the cystic
A B
Umbilical
fissure
Ligamentum
teres
Gallbladder
fossa
C
FIGURE 2.20. A, Relationship between the posterior aspect of segment IV and the biliary confluence. The hilar plate (arrow) is formed by the fusion
of the connective tissue enclosing the biliary and vascular elements with the Glisson capsule. B, Biliary confluence and left hepatic duct exposed by
lifting segment IV upward after incision of the Glisson capsule at its base. This technique, lowering of the hilar plate, generally is used to display a
dilated bile duct above an iatrogenic stricture or hilar cholangiocarcinoma. C, Line of incision (left) to allow extensive mobilization of segment IV. This
maneuver is of particular value for high bile duct strictures and in the presence of liver atrophy or hypertrophy. The procedure consists of lifting
segment IV upward (A and B), then not only opening the umbilical fissure but also incising the deepest portion of the gallbladder fossa. Right, Inci-
sion of the Glisson capsule to gain access to the biliary system (arrow). (B, From Hepp J, Couinaud C: L’abord et l’utilisation du canal hépatique
gauche dans les reparations de la voie biliare principale. Presse Med 64:947-948, 1956.)
duct may join the right hepatic duct or a right hepatic sectional 16% the right anterior sectional duct, and in 4% the right
duct (Fig. 2.24). posterior sectional duct, may approach the main bile duct in
this fashion. In 6%, a right sectional duct may join the left
BILIARY DUCTAL ANOMALIES hepatic duct (the posterior duct in 5% and the anterior duct in
1%. In 3%, there is an absence of the hepatic duct confluence,
Full knowledge of the frequent variations from the described and the right posterior sectional duct may join the neck of the
normal biliary anatomy is required when any hepatobiliary gallbladder, or it may be entered by the cystic duct in 2%
procedure is performed (Fig. 2.25). The constitution of a (Couinaud, 1957). In any event, these multiple biliary ductal
normal biliary confluence by union of the right and left hepatic variations at the hilus are important to recognize in resection
ducts, as described previously, is reported in only 72% of and reconstructive surgery of the biliary tree at the hilus and
patients (Healey & Schroy, 1953). There is a triple confluence during partial hepatectomy and cholecystectomy.
of the right anterior and posterior sectional ducts and the left Intrahepatic bile duct variations also are common (Fig.
hepatic duct in 12% of individuals (Couinaud, 1957), and a 2.26) (Healey & Schroy, 1953). The main right intrahepatic
right sectional duct joins the main bile duct directly in 20%. In duct variations are represented by an ectopic drainage of
A B
CBD HA
CD
Replaced or
accessory
RHA
C
FIGURE 2.21. Hepatic artery variations shown by angiography. A, Replaced common hepatic artery arises from the superior mesenteric trunk. B,
Left, The hepatic artery (large arrowhead) arises from the celiac axis. The small arrowheads indicate a drainage catheter in the bile duct. Right, An
accessory right hepatic artery (large arrowhead) is arising from the superior mesenteric artery and lies lateral to the catheter (small arrowheads) in
the common bile duct (CBD). C, The accessory right hepatic artery usually courses upward in the groove posterolateral to the CBD, appearing on
the medial side of the triangle of Calot, usually running just behind the cystic duct (CD). This common variation occurs in about 25% of individuals.
HA, Hepatic artery; RHA, right hepatic artery.
segment V in 9%, of segment VI in 14%, and of segment VIII In 67% of patients (Healey & Schroy, 1953) a classic distri-
in 20%. In addition, a subvesical duct has been described in bution of the main left intrahepatic biliary ductal system exists.
20% to 50% of cases. This duct, sometimes deeply embedded The main variation in this region is represented by a common
in the cystic plate, joins either the common hepatic duct or the union between the ducts of segments III and IV in 25%, and
right hepatic duct. It does not drain any specific liver territory, in only 2% does the duct of segment IV join the common
never communicates with the gallbladder, and is not a satellite hepatic duct independently. Several anomalies of drainage of
of an intrahepatic branch of the portal vein or hepatic artery. the intrahepatic ducts into the neck of the gallbladder or cystic
Although not of major anatomic significance, injury may occur duct have been reported (Fig. 2.27) (Albaret et al, 1981;
during cholecystectomy if the cystic plate is not preserved. This Couinaud, 1957), and these must be kept in mind during cho-
may lead to a postoperative biliary leak. lecystectomy (see Chapter 33).
a b c
d e f
g h
A
FIGURE 2.22. The main variations of the cystic artery: typical course
(a), double cystic artery (b), cystic artery crossing anterior to main bile
duct (c), cystic artery originating from the right branch of the hepatic
artery and crossing the common hepatic duct anteriorly (d), cystic artery
originating from the left branch of the hepatic artery (e), cystic artery
originating from the gastroduodenal artery (f), cystic artery arising from
the celiac axis (g), cystic artery originating from a replaced right hepatic
artery (h).
GB
PV
IVC
B
FIGURE 2.23. Longitudinal sonogram shows the relationship of the
FIGURE 2.24. A, T-tube cholangiogram shows a very low insertion of
liver, gallbladder (GB), portal vein (PV), inferior vena cava (IVC), hepatic
a right sectional duct into the common hepatic duct (arrow). B, Endo-
artery (curved arrow) and common bile duct (straight arrow).
scopic retrograde choledochopancreatogram shows a low right sec-
tional duct (large arrow), into which is draining the cystic duct (small
arrow), an uncommon but important normal variant.
ra ra duct but two fundi (Hobby, 1970), and duplication of the

rp lh rp lh gallbladder with two cystic ducts all have been described. A
double cystic duct may drain a unilocular gallbladder (Perel-
man, 1961), and congenital diverticulum of the gallbladder
A 57% B 12% with a muscular wall may also be found (Eelkema et al, 1958).
More frequently reported are anomalies of position of the gall-
bladder, which may be in an intrahepatic position, completely
ra ra
lh surrounded by normal liver tissue, or it may be found on the
rp lh left of the liver (Newcombe & Henley, 1964).
The mode of union of the cystic duct with the common
rp hepatic duct may be angular, parallel, or spiral. An angular
union is the most frequent and is found in 75% of patients
16% 4%
C 20% (Kune, 1970). The cystic duct may run a parallel course to the
C1 16% C2 common hepatic duct in 20%, with connective tissue ensheath-
ra
ing both ducts. Finally, the cystic duct may approach the CBD
ra in a spiral fashion. The absence of a cystic duct is probably an
rp acquired anomaly, representing a cholecystocholedochal fistula.
rp lh lh
D 6% BILE DUCT BLOOD SUPPLY

5% 1%
The bile duct may be divided into three segments: hilar, supra-
D1 D2 duodenal, and retropancreatic. The blood supply of the supra-
IV III duodenal duct is essentially axial (Fig. 2.29) (Northover &
ra IV III
Terblanche, 1979). Most vessels to the supraduodenal duct
rp ra
arise from the superior pancreaticoduodenal artery, right branch
II II of the hepatic artery, cystic artery, gastroduodenal artery, and
rp
E 3% I I retroduodenal artery. On average, eight small arteries, each
measuring approximately 0.3 mm in diameter, supply the
2% 1%
supraduodenal duct. The most important of these vessels run
E1 E2 along the lateral borders of the duct and have been called the
3 o’clock and 9 o’clock arteries. Of the blood vessels vascularizing
ra the supraduodenal duct, 60% run upward from the major infe-
rp rior vessels, and only 38% of arteries run downward, originating
lh from the right branch of the hepatic artery and other vessels.
Only 2% of the arterial supply is nonaxial, arising directly from
the main trunk of the hepatic artery, as it courses up parallel
F 2% to the main biliary channel. The hilar ducts receive a copious
supply of arterial blood from surrounding vessels, forming a
rich network on the surface of the ducts in continuity with the
FIGURE 2.25. Main variations of the hepatic duct confluence. plexus around the supraduodenal duct. The source of blood
A, Typical anatomy of the confluence. B, Triple confluence. C, Ectopic supply to the retropancreatic CBD is from the retroduodenal
drainage of a right sectional duct into the common hepatic duct. C1, artery, which provides multiple small vessels running around
Right anterior (ra) duct draining into the common hepatic duct; C2, right the duct to form a mural plexus.
posterior (rp) duct draining into the common hepatic duct. D, Ectopic The veins draining the bile ducts are satellites to the cor-
drainage of a right sectional duct into the left hepatic ductal system. D1, responding described arteries, draining into 3 o’clock and 9
Right posterior sectional duct draining into the left hepatic (lh) ductal
o’clock veins along the borders of the common biliary channel.
system; D2, right anterior sectional duct draining into the left hepatic
Veins draining the gallbladder empty into this venous system,
ductal system. E, Absence of the hepatic duct confluence. F, Absence
of right hepatic duct and ectopic drainage of the right posterior duct into not directly into the portal vein, and the biliary tree seems to
the cystic duct. (From Couinaud C: Le Foi: Études Anatomogiques et have its own portal venous pathway to the liver.
Chirurgicales. Paris, 1957, Masson.)
ANATOMY OF BILIARY EXPOSURE

Biliary-Vascular Sheaths and Exposure of the Hepatic
Bile Duct Confluence
ANOMALIES OF THE GALLBLADDER AND Fusion of the Glisson capsule with the connective tissue sheaths
CYSTIC DUCT surrounding the biliary and vascular elements at the inferior
aspect of the liver constitute the plate system (see Figs. 2.19
Many anomalies of the accessory biliary apparatus have been and 2.20), which includes the hilar plate above the biliary con-
described (Fig. 2.28) (Gross, 1936). Although rare, agenesis of fluence, the cystic plate related to the gallbladder, and the
the gallbladder (Boyden, 1926; Rachad-Mohassel et al, 1973; umbilical plate situated above the umbilical portion of the left
Rogers et al, 1975), bilobar gallbladders with a single cystic portal vein (Couinaud, 1957). Hepp and Couinaud (1956)
VIII D seg IV
VII
A seg V
VI a 67% b 1%
II II
91%
VIII VII VIII III

VII III
VI VI c 1%
II
5% 4%
III
VII
d 25% e 1% II
B seg VI II
VIII 86%
V
VII VII III III
VII
VIII VIII
VIII f 1%
II
V 10% V 2% V 2%
III
VII VII
VI g 4%
VI II
V
80% V 20%
C seg VIII
III
FIGURE 2.26. The main variations of the intrahepatic ductal system. A, Variations of segment V. B, Variations of segment VI. C, Variations of
segment VIII. D, Variations of segment IV. There is no variation of drainage of segments II, III, and VII. seg, Segment.
described a technique whereby lifting the segment IV upward Frequently, by a simultaneous opening of the deepest portion
and incising the Glisson capsule at its base offers a good expo- of the gallbladder fossa and the umbilical fissure (see Fig.
sure of the hepatic hilar structures (see Fig. 2.20). This tech- 2.20C), good exposure of the biliary duct confluence, and
nique is referred to as lowering of the hilar plate. It can be carried especially the right hepatic duct, can be obtained without the
out safely because only exceptionally (in 1% of cases) is there necessity for full hepatotomy.
any major vascular interposition between the hilar plate and the
inferior aspect of the liver, although tiny venules are common. Umbilical Fissure and Segment III (Ligamentum
This maneuver is of particular value in exposing the extrahe- Teres) Approach
patic segment of the left hepatic duct because it has a long
course beneath the segment IV. It is not as effective in exposing The round ligament, which is the remnant of the obliterated
the extrahepatic right duct or its secondary branches, which are umbilical vein, runs through the umbilical fissure to connect
short. The technique is of major importance for the identifica- with the left branch of the portal vein. The round ligament is
tion of proximal biliary mucosa during bile duct repair after sometimes deeply embedded in the umbilical fissure. At the
injury. Basically, an incision is required at the posterior edge of junction of the round ligament and the termination of the left
segment IV, where the Glisson capsule is attached to the hilar portal vein, elongations containing channels that are elements
plate. The upper surface of the hilar plate can be separated from of the left portal system course into the liver. The bile ducts of
the hepatic parenchyma and, by lifting segment IV upward, the left lobe of the liver (Figs. 2.30 and 2.31A) are located above
display of the hepatic duct convergence, which is always extra- the left branch of the portal vein and lie behind these elonga-
hepatic, is effected. Bile duct incision allows performance of a tions, whereas the corresponding artery is situated below the
mucosa-to-mucosa anastomosis. Rarely, it may be hazardous to vein. Dissection of the round ligament on its left side and divi-
approach the biliary confluence in this manner, especially when sion of one or two vascular elongations of segment III allow
anatomic deformity has been created by atrophy or hypertrophy display of the pedicle or anterior branch of the duct of segment
of liver lobes and in patients in whom there appears to be a III (Fig. 2.32). In the event of biliary obstruction with intrahe-
very deep hilus that is displaced upward and rotated laterally. patic biliary ductal dilation, a dilated segment III duct is
VI
A B C d
RP a
D
b f
RP
g
c
h
E F
FIGURE 2.27. The main variations of ectopic drainage of the intrahe-
patic ducts into the gallbladder and cystic duct. A, Drainage of the cystic
duct into the biliary confluence. B, Drainage of cystic duct into the left
hepatic duct, associated with no biliary confluence. C, Drainage of
segment VI duct into the cystic duct. D, Drainage of the right posterior
(RP) sectional duct into the cystic duct. E, Drainage of the distal part of
the right posterior sectional duct into the neck of the gallbladder. F,
Drainage of the proximal part of the right posterior sectional duct into FIGURE 2.29. The bile duct blood supply. Note the axial arrangement
the body of the gallbladder. of the vasculature of the supraduodenal portion of the main bile duct
and the rich network enclosing the right and left hepatic ducts: right
branch of the hepatic artery (a), 9 o’clock artery (b), retroduodenal artery
(c), left branch of the hepatic artery (d), hepatic artery (e), 3 o’clock artery
A (f), common hepatic artery (g), gastroduodenal artery (h).
1 2 3
C
IV
B
II
1
1 2
1 2
III
E
FIGURE 2.30. Biliary and vascular anatomy of the left liver. Note the
location of the segment III duct above the corresponding vein. The
75% 20% 5% anterior branch of the segment IV duct is not represented.
a b c
FIGURE 2.28. Main variations in gallbladder and cystic duct anatomy.
A, Duplicated gallbladder. B, Septum of the gallbladder. C, Diverticulum
of the gallbladder. D, Variations in cystic ductal anatomy. E, Different
types of union of the cystic duct and common hepatic duct: angular
union (a), parallel union (b), spiral union (c).
II
b
a
III
b
IV
c
c
d
A B
FIGURE 2.31. A, The biliary and vascular anatomy of the left liver. Note the relationship of the left horn of the umbilical recess with the segment
III ductal system: left portal vein (a), left hepatic duct (b), segment III system—note that the duct (black) lies adjacent to the portal venous branch
indicated (c), ligamentum teres (d). B, Segment III ductal approach: exposure of the left horn of the umbilical recess (a), division of the left horn of
the umbilical recess including segment III portal vein branches (b), exposure and opening of segment III duct: hepaticojejunostomy to the segment
III ductal system (c; see also Chapters 31 and 42).
A B
FIGURE 2.32. A, The liver is split to the left of the ligamentum teres in the umbilical fissure. It may be necessary to remove a small wedge of liver
tissue (c). B, Segment III duct is exposed at the base of the liver split, above and behind its accompanying vein, and is ready for anastomosis (see
also Chapters 31 and 42).
A B
FIGURE 2.33. A, Anterior sectional approach. If necessary, the liver substance is opened through a short distance in the line of the right anterior
sectional pedicle. B, The duct is displayed anterior and to the left side of the corresponding vein. This can be facilitated by using a posterior pedicular
approach as described by Launois (see Chapter 103B).
generally easily located above the left branch of the portal vein. of the mobilization of segment IV after opening of the principal
It is often preferable to split the normal liver tissue just to the scissura and the umbilical fissure as described previously.
left of the umbilical fissure to widen the fissure further, which
allows access to the ductal system with no need to divide any EXTRAHEPATIC VASCULATURE
elements of the portal blood supply to segment III (see Fig.
2.32). Celiac Axis and Blood Supply of Liver, Biliary Tract,
and Pancreas
Surgical Approaches to the Right Hepatic Biliary
Ductal System The usual classic description of the arterial blood supply of the
Because of the lack of precise anatomic landmarks, exposure liver, biliary system, and pancreas is found in only approxi-
of the right intrahepatic ductal system is much more hazardous mately 60% of patients (Figs. 2.34 to 2.36). The right and the
and imprecise than that of the left. In some cases of hilar left hepatic arteries, the former in the right of the hilus of the
cholangiocarcinoma, the planned surgical procedure—partial liver and the latter in the left at the base of the umbilical fissure,
hepatectomy (see Chapters 51B and 103B) or segment III become enclosed in the sheath of peritoneum, forming the right
duct bypass (see Chapters 31 and 42)—seems impossible at and left portal triads. In this sheath, further branching to the
operation. In such a critical operative situation, intrahepatic right anterior and posterior sections of the liver and on the left
right ductal system drainage is an option. Anatomically, the to segments II, III, and IV occurs within the respective pedicles,
anterior sectional duct and its branches run on the left side which also come to enclose the portal vein branches and the
of the corresponding portal vein. In essence, the end of the tributary bile ducts from these sections and segments. The arte-
liver scissura, within which lies the right branch of the portal rial supply of the CBD was described earlier; it arises from
vein, is opened through a short distance. The anterior sectional branches of the hepatic artery, the gastroduodenal artery, and
duct is displayed on the left aspect of the vein, and the dilated the pancreaticoduodenal arcades.
duct is opened longitudinally and sewn to a Roux-en-Y loop For practical surgical issues, the most important relation-
of jejunum (Fig. 2.33). Although this technique is rarely used, ships in the anatomy of the pancreas concern the arterial blood
it may be valuable in selected cases. Preferably, the right-sided supply and the venous drainage. The dorsal pancreatic artery
pedicles can be encircled and exposed by the technique used is a major branch, usually arising from the splenic artery, but
for pedicle exposure and control described for right-sided liver it can arise directly from the hepatic artery. When splenectomy
resection. is performed, it is important to establish the site of origin of the
dorsal pancreatic artery to avoid distal pancreatic ischemia. The
Exposure of the Bile Ducts by Liver Resection superior mesenteric artery (SMA) arises from the aorta poste-
This chapter does not detail exposure of the bile ducts by resec- riorly behind the pancreas and runs forward and upward to run
tion of liver substance. In essence, a segment of the left lobe first behind and then to the left of the superior mesenteric vein
may be amputated to expose the segment II or III ducts, or a (SMV) (see Fig. 2.35).
similar procedure may be carried out after removal of the infe-
rior tip of the right lobe. Finally, in some instances, removal of Variations in the Hepatic Artery
segment IV may be carried out to effect exposure of the biliary As a result of the complex embryologic development of the
confluence. This procedure really represents a simple extension celiac axis and SMA, wide variations in the arterial supply of
MH LH Left gastric
Aorta LGA SA
PV
RH HA
Celiac trunk RGA DP
GDA
Cystic Splenic SPDA
Proper hepatic
Pancreas
Supraduodenal Post PDA
MCA
Gastroduodenal MCV
Right Common hepatic GEA
gastric
FIGURE 2.34. The celiac trunk is a short, thick artery originating from Ant PDA
the aorta just below the aortic hiatus of the diaphragm and extending
horizontally and forward above the pancreas, where it divides into the IPDA
left gastric, common hepatic, and splenic arteries. An inferior phrenic SMA
artery, usually arising from the aorta or the splenic artery, occasionally SMV
arises from the celiac trunk. The left gastric artery curves toward the FIGURE 2.35. The primary arteries that supply the pancreas are the
stomach and extends along its lesser curve, forming anastomoses with gastroduodenal artery (GDA), which arises usually from the common
the right gastric artery. The splenic artery, the largest of the three celiac hepatic artery (HA) while it crosses the portal vein (PV) above the pan-
branches, takes a tortuous course to the left, behind and along the creas proper, and the dorsal pancreatic artery (DP), arising from the
upper border of the pancreas and at the hilus of the spleen, where it splenic artery (SA). The superior pancreaticoduodenal arteries (SPDAs)
splits into numerous terminal branches. The splenic artery usually arise from the GDA and join the inferior pancreaticoduodenal arteries
approaches and runs superiorly to the splenic vein. An uncommon but (IPDAs) from the superior mesenteric artery (SMA), forming two arcades
dangerous abnormality can occur when the splenic artery runs inferiorly along the anterior and posterior aspects of the head of the pancreas.
and behind the splenic vein, close to the splenic vein–mesenteric vein The GDA, after giving rise to the pancreaticoduodenal artery (PDA),
confluence. The left gastroepiploic artery and the short gastric arteries passes forward and to the left as the right gastroepiploic artery (GEA).
originate from one of these terminal branches. The common hepatic The GDA is a good landmark for the identification of the portal vein above
artery passes forward into the retroperitoneum then curves to the right the pancreas, and surgical division of the GDA just at its origin from the
to enter the right margin of the lesser omentum, just above the pancreas, common HA gives much greater access to the anterior surface of
and ascends; it approaches the common bile duct (CBD) on its left side the portal vein at this site. The right gastric artery (RGA) also usually
and runs usually anterior to the portal vein. While it turns upward just arises from the common HA just distal to the GDA, but it can arise from
above the pancreas, it gives rise to the gastroduodenal artery, which various sites. The GDA commonly divides into a larger right GEA and
also may originate from the right hepatic artery. This descends to supply smaller SPDA. The right GEA runs forward between the first part of the
the anterior, superior, and posterior surfaces of the first inch of the duodenum and pancreas; the SPDA divides into anterior and posterior
duodenum. The gastroduodenal artery can be duplicated and often has branches. The anterior superior PDA continues downward on the ante-
a small branch running with it toward the pylorus. The right gastric artery rior surface of the head of the pancreas to anastomose with the IPDA,
passes to the left along the lesser curve of the stomach, and anasto- which arises from the SMA. The posterior superior PDA behaves simi-
mosis is to the left gastric artery. The continuation of the common larly. Ant, Anterior; LGA, left gastric artery; MCA, middle colic artery;
hepatic artery, beyond the origin of the gastroduodenal artery and right MCV, middle colic vein; Post, posterior; SMV, superior mesenteric vein.
gastric artery, is known as the proper hepatic artery and usually soon
divides into a right and a left branch. The left branch extends vertically,
directly toward the base of the umbilical fissure, and usually gives off a
branch known as the middle hepatic artery (MH), which is directed
toward the right of the umbilical fissure and is destined to supply
segment IV of the liver. A further branch of the left hepatic artery (LH)
courses to the left to supply the caudate lobe, and further smaller
caudate branches arise from the left and right hepatic artery. The right
hepatic artery (RH) usually passes behind the common hepatic duct and
enters the cystic triangle of Calot; in some cases, it passes in front of
the bile duct, which is important in surgical exposure of the CBD. The p
cystic artery usually arises from the right hepatic artery but has many c
variations. IVC a s
FIGURE 2.36. Computed tomographic image of the main portal vein

shows the hepatic artery (solid arrows) coursing anterior to the portal
vein (p). The interlobar fissure (open arrow), splenic vein (s), celiac axis
(c), aorta (a), and inferior vena cava (IVC) are also shown.
the liver are found (Fig. 2.37). These variations are important posteriorly to the right renal vein and the anterior surface of
to recognize. Failure to show all arteries feeding the liver at the IVC. The neck and body of the pancreas lie atop the SMA
angiography may not only result in errors of diagnosis but may and the splenic vessels and their branches, the left renal vein,
also seriously mislead the surgeon or the interventional radiolo- and, more laterally, the left kidney. The right gastroepiploic vein
gist. In most cases, the hepatic artery arises from the celiac axis commonly drains into the anterior surface of the SMV just at
as described earlier, but it may be entirely replaced by a common the inferior border of the pancreas; this can often be involved
hepatic artery that originates from the SMA. In this instance, by tumor, as can the anterior branch of the inferior pancreati-
the hepatic artery passes posterior and then lateral to the portal coduodenal vein; the middle colic vein may also join at this
vein while it ascends and lies posterolateral to the CBD in the point. In mobilizing the SMV, these vessels are ligated so as to
hepatoduodenal ligament, where it is susceptible to operative avoid bothersome hemorrhage. Abnormalities of the IVC are
injury if not recognized. This applies to a right replaced or an uncommon, with duplication of the vena cava and a left-sided
accessory hepatic artery. Other variations in the origin of the vena cava seen rarely.
common hepatic artery include its origin directly from the aorta Several variations in anatomy and rare congenital anomalies
and the persistence of a primitive embryologic link between the of the portal vein are of surgical significance (Figs. 2.40 to
celiac and superior mesenteric systems. These variations are of 2.43). For example, performance of right hepatic resection,
considerable importance in controlling the arterial blood supply with division of what appears to be the right portal vein in a
to the liver during hepatic resection, liver transplantation, devas- patient with absence of the left portal vein (see Figs. 2.42 and
cularization of the liver, placement of intraarterial hepatic infu- 2.43) can be fatal. Agenesis of the right branch of the portal
sion devices, and in the resection of the head of the pancreas. vein is associated with agenesis of the right hemiliver and left
liver hypertrophy. This may be associated with biliary and
Portal Vein hepatic venous anatomic anomalies, which can compromise
The portal vein (Fig. 2.38) is formed behind the neck of the surgical approaches to the liver and to biliary repair (Fields
pancreas by confluence of the superior mesenteric and splenic et al, 2008).
veins (Fig. 2.39). The venous drainage of the pancreas usually Portal venous blood is derived from the venous drainage of
runs parallel to the arterial supply. There are anterior and the stomach, small bowel, spleen, and pancreas, and this drain-
posterior and superior and inferior pancreaticoduodenal veins age is important when considering surgery of the pancreas and
that drain to the portal vein and the SMV. The left gastric vein in patients with portal hypertension; it is described in detail
and the inferior mesenteric vein (IMV) usually drain into the along with the description of the anatomy of the pancreas.
splenic vein, but they can drain directly into the portal vein,
whereas the various small splenic tributaries drain directly to PANCREAS
the splenic vein.
The anatomic relationship of the pancreas to the SMV, the The pancreas is a posteriorly situated retroperitoneal organ that
splenic vein, and the portal vein (see Fig. 2.38) is important in lies transversely (Fig. 2.44). The organ is composed of a head,
pancreatic resection (see Chapter 66). The uncinate process can neck, body, and tail. The head is encompassed by the duode-
extend behind the SMV to well behind the SMA (see Fig. 2.35). num, whereas the tail rests in the splenic hilum (Figs. 2.45 and
Access to the portal vein behind the pancreas usually is obtained 2.46). A portion of the head inferiorly is termed the uncinate
from below by elevating the pancreas from the surface of the process and is intimately related to the SMV and SMA. Poste-
SMV just before it joins the splenic vein. With the exception of riorly, the pancreas is related to the IVC, aorta, left renal vein
the inferior pancreaticoduodenal veins, which enter the SMV and kidney, and spleen. The portion lateral to the portal vein
at the inferior border of the pancreas, it is uncommon to see averages 56.4% of the total weight. The pancreatic capsule is
branches from the pancreas run directly posteriorly into the loosely attached to the surface of the pancreas and is contiguous
SMV. Fixation here is usually by some inflammatory or neo- with the anterior layer of the mesocolon such that it can be
plastic process. Superiorly, the portal vein runs behind the dissected in continuity if necessary. The mesenteric attach-
pancreas and is identified first in the gap between the curvature ments to the pancreas tend to be contiguous (see Fig. 2.46).
of the splenic vein, splenic artery, common hepatic artery, and The arterial blood supply and venous drainage and the relation-
gastroduodenal artery. Division of the gastroduodenal artery ships to the CBD are described and illustrated earlier (see Figs.
provides much greater access to the superior surface of the 2.17, 2.29, 2.34, 2.35, and 2.38).
portal vein. If difficulty is encountered in this area, division of
the CBD, usually above the cystic duct, can provide excellent Pancreatic Duct
access to the right lateral aspect and anterior surface of the The duct of Wirsung, beginning in the distal tail as a confluence
portal vein. The SMA can be approached behind the pancreas of small ductules, runs through the body to the head, where it
above the point at which it is embraced by the uncinate process usually passes downward and backward in close juxtaposition
at the origin from the aorta. This allows dissection of the most to the CBD (see Fig. 2.45). The sphincter of Oddi (Fig. 2.47)
proximal part of the SMA. has been thoroughly studied (Boyden, 1957; Delmont, 1979)
Occasionally, the middle colic artery and other vessels of and consists of a unique cluster of smooth muscle fibers distin-
supply to the colon can arise from the more proximal SMA, guishable from the adjacent smooth muscle of the duodenal
such that they pass through the pancreas; this abnormality wall. The papilla of Vater at the termination of the CBD is a
should be searched for carefully. Division of the middle colic small, nipple-like structure that protrudes into the duodenal
artery is usually not a problem, however, because the colon is lumen and is marked by a longitudinal fold of duodenal mucosa.
well supplied with blood, and ischemia typically does not occur. The duct of Wirsung runs downward and parallel to the CBD
Of special importance to the surgeon is the direct relation- for approximately 2 cm and joins it within the sphincter
ship of the head of the pancreas to the duodenum and Text continued on p. 58
FIGURE 2.37. In approximately 25% of individuals, the right hepatic artery arises
partially or completely from the superior mesenteric artery (A, C, E); in a similar propor-
tion of patients, the left hepatic artery may be partially or completely replaced by a
branch arising from the left gastric artery, coursing through the gastrohepatic omentum
to enter the liver at the base of the umbilical fissure (D, F). Rarely, the right or left
hepatic arteries originate independently from the celiac trunk or branch after a very
short common hepatic artery origin from the celiac, and the gastroduodenal artery
may originate from the right hepatic artery (B, C). Multidetector computed tomography
(CT) angiogram demonstrating an accessory right hepatic artery (arrow) arising from
the superior mesenteric artery (G). Multidetector CT angiogram demonstrating a A B
replaced left hepatic artery arising from the left gastric artery (H). Another common
arterial variant is the hepatic trifurcation (I).
C D
E F
G I
H
RAS LPV
Ligamentum teres
IVb
RAS III
RPS IVa
II
LPV
RPS
PV
RPV
CV
B
RAS LPV
SV
SMV
IMV PV
RPS
A C
FIGURE 2.38. A, The superior mesenteric vein (SMV) at the root of the lesser omentum is usually a single trunk; two, or sometimes even three,
branches may unite as the vessel enters the tunnel beneath the neck of the pancreas (shaded) to form a superior mesenteric trunk. This trunk
ascends behind the neck of the pancreas and is joined by the splenic vein (SV), which enters it from the left to form the portal vein (PV), which
emerges from the retroperitoneal upper border of the neck of the pancreas and ascends toward the liver within the free edge of the lesser omentum,
lying behind the bile duct and the hepatic artery and surrounded by the lymphatics and nodes of the lesser omentum. During this course, it receives
blood through the coronary vein (CV), which communicates with esophageal venous collaterals, which connect with the gastric vein and the esopha-
geal plexus. Sometimes a separate right gastric vein enters the PV in this area. A superior pancreaticoduodenal vein often enters the PV just above
the level of the pancreas, and several smaller veins enter the SMV and PV from the right side beneath the neck of the pancreas. As the PV ascends
behind the common bile duct and common hepatic duct, it approaches the hilus of the liver and bifurcates into two branches, a larger right (RPV)
and a smaller left portal vein (LPV). The branch on the left courses below the left hepatic duct to enter the umbilical fissure, in company with the left
hepatic artery, and subsequently branches to supply the left liver segments (II-IV). Just before its entry into the umbilical fissure, it gives off a major
caudate vein, segment I, which runs posteriorly and laterally to the left. Sometimes this vein consists of two or more branches; the right portal branch,
which is much shorter in length before its entry into the liver, divides at the extremity of the hilus into the right anterior (RAS) and posterior (RPS)
sectional branches and is accompanied by the respective arterial branches and biliary tributaries. B, The division of the portal vein may arise more
proximally, however, and C, the right anterior and posterior sectional portal veins may arise independently from the portal venous trunk. IMV, Inferior
mesenteric vein.
L
RA
M
p s
RP
sm
FIGURE 2.39. Magnetic resonance imaging of the splenoportal con-

fluence: postcontrast, T1-weighted, three-dimensional gradient-echo
coronal maximum intensity projection. Shown are the splenic vein (s),
portal vein (p), and superior mesenteric vein (sm).
FIGURE 2.40. Contrast-enhanced computed tomographic scan of

variant portal vein branching with trifurcation pattern. L, Left portal vein;
M, main portal vein; RA, right anterior portal vein; RP, right posterior portal
vein. (From Covey AM, et al: Incidence, patterns, and clinical relevance of
variant portal vein anatomy. Am J Roentgenol 183:1055-1064, 2004.)
A C
FIGURE 2.41. A, The portal vein anterior to the head of the pancreas and the duodenum may be in an abnormal position. B, Another rare but
interesting anomaly is the entrance of the portal vein into the inferior vena cava. C, Very rare, the entrance of a pulmonary vein into the portal vein.
FIGURE 2.42. In a congenital absence of the left branch of the portal

vein as described by Couinaud, the right branch courses through the right
lobe of the liver supplying it and curves within the liver substance to supply FIGURE 2.43. Computed tomographic scan in a patient with Caroli’s
the left lobe, which in such instances is usually smaller than normal. disease shows a large right portal trunk. The left branch of the portal
vein is absent, with findings confirmed at operation for left hepatic
lobectomy.
D
S
v
v
a
i a
r IVC
A
A B
FIGURE 2.44. A, Magnetic resonance imaging of the pancreas, oblique axial reconstruction, T1-weighted three-dimensional gradient-echo tech-
nique. Aorta (a), inferior vena cava (i), common bile duct (CBD) (arrow), inferior mesenteric vein at the splenoportal confluence (v), superior mesenteric
artery (arrowhead), and left renal vein (r) are shown. B, Normal pancreatic anatomy. Postcontrast computed tomographic scan at the level of the
pancreas. A, Aorta; a, superior mesenteric artery; D, duodenum; IVC, inferior vena cava; S, stomach; v, superior mesenteric vein;. Long arrow, CBD;
short arrow, inferior pancreaticoduodenal artery; open arrow, gastroduodenal artery.
Aorta
NECK BODY
HEAD POSITION OF THE COMMON BILE DUCT
Superior mesenteric
vein and artery
Uncinate
process
B
UNCINATE PROCESS
a b c
(i) (ii)
A C
FIGURE 2.45. A, The head of the pancreas is globular with an extension, the uncinate process, which curves behind the superior mesenteric
vessels and may end even before it embraces the superior mesenteric vein (a), or it may pass completely behind between the aorta and the left of
the patient’s superior mesenteric artery (b, c). All variations are commonly seen. Posteriorly, the head of the pancreas lies in juxtaposition to the
inferior vena cava at the level of the entry of the left and right renal veins. The head of the pancreas forms a narrow neck in front of the superior
mesenteric and splenic vein confluence. The neck joins to the body of the gland, which forms a narrow tail. B, The common bile duct (CBD) passes
through the pancreas, either directly in the substance of the gland or initially with a posterior groove. C, The duct of Wirsung courses from left to
right within the pancreas, curves downward approaching the CBD, and runs parallel with but separated from it by the transampullary septum to
enter the duodenum, 7 to 10 cm distal to the pylorus, at the papilla of Vater after traversing the sphincter of Oddi. An accessory duct, the duct of
Santorini, runs more proximally in the head of the pancreas and usually terminates in the duodenum at an accessory papilla. Multiple variations of
the ductal system occur, depending on the extent of development of the duct of Santorini, such that rarely the accessory duct can enter the duo-
denum inferior to the main duct. It can be in communication with the main duct directly (i), or it can occur in duplicate version known as pancreas
divisum (ii). The duct of Santorini drains the body and tail of the organ, and the duct of Wirsung drains the head and the uncinate process.
Aorta
Portal vein
Vena cava Aorta
Peritoneum
Neck of Superior mesenteric

Middle
pancreas artery
colic
vessels Left renal vein
Uncinate process
of pancreas
Middle
Superior mesenteric artery colic
artery Third part of
Superior mesenteric vein duodenum
Superior
mesenteric
artery
A B
Pancreas posterior view
Portal Common
vein bile duct
Splenic Splenic Hepatic Pancreatic

artery vein artery duct
Superior
mesenteric
artery
Uncinate
FIGURE 2.46. A, The anterior surface of the pancreas, covered by the posterior process Superior
layer of the omental bursa or lesser peritoneal sac, can often be obliterated by adhe- mesenteric vein
sions. The transverse mesocolon arises from the lower border of the pancreas and Left Right
envelops the middle colic vessels as they arise from the superior mesenteric vessels Related Related to left kidney, Related to
just beneath the pancreatic neck. B, The relationship of the pancreatic neck and unci- to spleen left renal vein and inferior
nate process to the aorta and superior mesenteric artery. Note the position of the left adrenal gland vena cava
and aorta
renal vein and duodenum. C, The posterior relationships of the duodenal loop and
pancreas. Note the relationship to the inferior vena cava, aorta, and hilum of the spleen. C
b
c
d
e
f
FIGURE 2.47. Schematic representation of the sphincter of Oddi: notch (a), biliary sphincter (b), transampullary septum (c), pancreatic sphincter
(d), membranous septum of Boyden (e), common sphincter (f), smooth muscle of duodenal wall (g).
MRCP: PANCREAS DIVISUM
B
Posterior Anterior
FIGURE 2.48. A, Magnetic resonance imaging cholangiography (MRCP), T2-weighted coronal image at the level of the ampulla, shows the duo-
denum (D) and the pancreatic head with common bile duct (curved arrow) and pancreatic duct (straight arrow). B, MRCP of pancreas divisum.
Anterior projection shows variant anatomy with the duct of Santorini (vertical arrow) between the duodenum above the duct of Wirsung (angled
arrow). The two ducts are separate; the duct of Santorini drains mainly the neck and body of the pancreas, and the duct of Wirsung drains mainly
the uncinate process portion of the head of the pancreas.
segment in 70% to 85% of patients; it enters the duodenum Annular Pancreas

independently in 10% to 13% of patients and is replaced by Annular pancreas is the development of a ring of pancreatic
the duct of Santorini in 2% of patients (Fig. 2.48; see Fig. tissue that surrounds and often embraces the duodenum
2.45). Rarely, the duct of Santorini and the duct of Wirsung (see Chapter 1). This ring may contain a large duct and
are separate, which is known as pancreas divisum (see Figs. can be firmly affixed to the duodenal musculature. The
2.45Cii and 2.48B). The islets of Langerhans, which provide duodenum beneath this annulus is often stenosed such that
the endocrine component of the gland, are scattered through- dividing this ring does not always relieve chronic duodenal
out the pancreas. obstruction. This accounts for the common process of
Splanchnic nerves
Vagal nerves
Celiac ganglion
FIGURE 2.50. Note the distribution of sympathetic and parasympa-

thetic nerves to the liver and pancreas from the celiac ganglion, mainly
A in association with major arteries.
applying duodenojejunostomy to relieve strictures caused by

such an annulus.
LYMPHATIC DRAINAGE
Liver and Pancreas
The lymphatic drainage of the liver and gallbladder is mainly
to nodes in the hepatoduodenal ligament and along the hepatic
artery; this is shown in Figure 2.49. The lymphatic drainage of
i PV the pancreas is predominantly to the nodes that lie in juxtaposi-
tion to the arteries and veins (Fig. 2.50B).
HA BD
SA
NERVE SUPPLY TO THE LIVER AND PANCREAS
The nerve supply to the liver and pancreas (see Fig. 2.50) is
from branches of the celiac ganglion. It is composed of sympa-
thetic and parasympathetic elements.
SV
IMV
References are available at expertconsult.com.
SMA SMV
B ii
FIGURE 2.49. A, The liver drains principally to hepatoduodenal nodes
at the hilus and along the hepatic artery and portal vein. The gallbladder
drains partly to the liver, but it also drains via the cystic node to nodes
of the hepatoduodenal ligament and to suprapancreatic nodes.
B, Numerous nodes (i) lie along the superior mesenteric vein along the
borders of the pancreas, draining back into the splenic hilar nodes; along
the superior border of the pancreas to the superior pancreatic nodes;
and to the celiac trunk and nodes at the base of the common hepatic
artery. A large node commonly lodges in intimate association with the
surface of the superior border of the pancreas and the right side of the
common hepatic artery. This node often needs to be dissected and
elevated to gain access to the anterior surface of the portal vein. Removal
of this node often improves access, as does division of the gastroduo-
denal artery. Posterior pancreaticoduodenal nodes (ii) lie along the pos-
terior pancreatic duodenal arterial arcade. BD, Bile duct; HA, hepatic
artery; IMV, interior mesenteric vein; PV, portal vein; SA, splenic artery;
SMA, superior mesenteric artery; SMV, superior mesenteric vein; SV,
splenic vein.
Chapter 2 Surgical and radiologic anatomy of the liver, biliary tract, and pancreas 59.e1
REFERENCES Hjörtsjö CH: The topography of the intrahepatic duct systems, Acta
Anat (Basel) 11:599–615, 1931.
Albaret P, et al: À propos des caneaux hépatiques directement abou- Hobby JAE: Bilobed gallbladder, Br J Surg 57:870–872, 1970.
chés dans la voie biliaire accessorie, Ann Chir 35:88–92, 1981. Kune GA: The influence of structure and function in the surgery of
Bismuth H, et al: Major and minor segmentectomies—“réglées”—in the biliary tract, Ann R Coll Surg Engl 47:78–91, 1970.
liver surgery, World J Surg 6:10–24, 1982. McIndoe AH, Counseller VX: A report on the bilaterality of the liver,
Boyden EA: The accessory gallbladder: an embryological and compara- Arch Surg 15:589, 1927.
tive study of aberrant biliary vesicles occurring in man and the Newcombe JF, Henley FA: Left-sided gallbladder: a review of the lit-
domestic mammals, Am J Anat 38:177–231, 1926. erature and a report of a case associated with hepatic duct carci-
Boyden EA: The anatomy of the choledochoduodenal junction in man, noma, Arch Surg 88:494–497, 1964.
Surg Gynecol Obstet 104:641–652, 1957. Northover JMA, Terblanche J: A new look at the arterial blood supply
Champetier J, et al: Aberrant biliary ducts (vasa aberrantia): surgical of the bile duct in man and its surgical implications, Br J Surg
implications, Anat Clin 4:137–145, 1982. 66:379–384, 1979.
Couinaud C: Le foi: études anatomogiques et chirurgicales, Paris, 1957, Perelman H: Cystic duct reduplication, JAMA 175:710–711, 1961.
Masson. Rachad-Mohassel MA, et al: Duplication de la vésicule biliaire, Arch
Delmont J: Le sphincter d’Oddi: anatomie traditionelle et fonction- Fr Mal App Dig 62:679–683, 1973.
nelle, Gastroentrol Clin Biol 3:157–165, 1979. Rocko JM, et al: Calot’s triangle revisited, Surg Gynecol Obstet 153:410–
Eelkema HH, et al: Partial duplication of the gallbladder, diverticulum 414, 1981.
type: report of a case, Radiology 70:410–412, 1958. Rogers HI, et al: Congenital absence of the gallbladder with choledo-
Fields RC, et al: Biliary injury after laparoscopic cholecystectomy in a cholithiasis: literature review and discussion of mechanisms, Gastro-
patient with right liver agenesis: case report and review of the litera- enterology 48:524–529, 1975.
ture, J Gastrointest Surg 12:1577–1581, 2008. Scheele J, Stangl R: Segment-orientated anatomical liver resections. In
Goldsmith NA, Woodburne RT: Surgical anatomy pertaining to liver Blumgart LH, editor: Surgery of the liver and biliary tract, ed 2, New
resection, Surg Gynecol Obstet 195:310–318, 1957. York, 1994, Churchill Livingstone, pp 1557–1578.
Gross RE: Congenital abnormalities of the gallbladder: a review of 148 Strasberg SM, et al: Nomenclature of hepatic anatomy and resections:
cases with report of a double gallbladder, Arch Surg 32:131–162, a review of the Brisbane 2000 system, J Hepatobiliary Pancreat Surg
1936. 12:351–355, 2005.
Healey JE, Schroy PC: Anatomy of the biliary ducts within the human Ton That Tung: La vascularisation veineuse du foie et ses applications aux
liver: analysis of the prevailing pattern of branchings and the major resections hépatiques, Hanoi, 1939, Thèse.
variations of the biliary ducts, AMA Arch Surg 66:599–616, 1953. Ton That Tung: Les resections majeures et mineures du foie, Paris, 1979,
Hepp J, Couinaud C: L’abord et l’utilisation du canal hépatique gauche Masson.
dans les reparations de la voie biliare principale, Presse Med 64:947– Wood D: Eponyms in biliary tract surgery, Am J Surg 138:746–754,
948, 1956. 1979.
CHAPTER 3
Assessment of hepatic function: implications for the
surgical patient
Paul J. Karanicolas
PREOPERATIVE CONSIDERATIONS VOLUMETRIC ASSESSMENT

OF LIVER REMNANT
The limits of hepatic resectability are constantly expanding
with increased understanding of hepatic anatomy and refine- Extent of liver resection (i.e., the number of segments resected)
ments in surgical technique. In past years, partial hepatectomy is strongly correlated with risk of postoperative liver insuffi-
was limited to anatomic resection and small-wedge resections, ciency. Although this is intuitive and easily assessed, it is actu-
with a general consensus that two contiguous segments of ally the volume of liver remaining (i.e., the FLR) that is more
hepatic parenchyma having adequate vascular inflow/outflow predictive of outcome and thus critical to accurately measure.
and biliary drainage was the minimum threshold for safe resec- Furthermore, assessment of number of segments remaining is
tion (Adams et al, 2013; Charnsangavej et al, 2006). This not sufficient due to substantial variability among patients in
conventional definition served the surgical community well but segmental anatomy and liver volume. In most patients, the right
has required refinement for two reasons. First, a variety of side of the liver represents more than half of the total liver
techniques have been developed that allow more extensive volume (TLV); however, there is a broad range, from 49% to
resection than this definition suggests, including induced 82%, with the left side of the liver conversely ranging from 17%
hypertrophy of the future liver remnant (FLR) (e.g., two-stage to 49% (Abdalla et al, 2004). Thus formal radiologic assess-
hepatectomy, portal vein embolization [PVE], associating liver ment of volumetrics is required to accurately assess the FLR
partition and portal vein ligation for staged hepatectomy), and for anticipated major (i.e., >4 segments) liver resection.
nonanatomic parenchymal-sparing resections (see Chapter
108). Indeed, through these and other techniques, it may be Techniques of Volumetric Assessment
possible to safely resect tumors from all segments of the liver Formal measurement of liver volumes is most commonly
while maintaining adequate postoperative liver function. accomplished by using computed tomography (CT) or mag-
Second, patients selected for partial hepatectomy are increas- netic resonance imaging (MRI) (Heymsfield et al, 1979; Huynh
ingly treated with preoperative chemotherapy or have other et al, 2014; Karlo et al, 2010; Suzuki et al, 2013). Other
risk factors for background liver injury; in these patients, the imaging modalities may also be used, but CT and MRI are
minimal requirement of two contiguous segments of liver is commonly obtained in patient care for characterization of
likely too liberal and puts patients at unacceptable risk of lesions and operative planning, and therefore additional tests
posthepatectomy liver failure (see Chapters 71, 92, 100, and are typically not needed. Cross-sectional images obtained from
103). either of these modalities are sequentially marked with the
Given the trend toward more extended, nonanatomic liver planned resection line, following which the surface area is
resections in patients at risk for background liver disease, thor- derived and multiplied by the slice thickness (Fig. 3.1).
ough assessment of hepatic function is crucial. Liver function Due to the variability in total liver size based on patient body
following hepatic resection is dependent on the quantity and habitus, the FLR volume is typically expressed as a ratio of FLR
quality of the FLR. Thus the optimal assessment of fitness for to TLV. Although the measurement of the FLR is fairly stan-
liver resection would ideally incorporate some measure of FLR dard, there are several variations to calculate the TLV. The
volume and function. This is particularly important in patients simplest and most intuitive technique involves manually tracing
at risk for, or with documented evidence of, background the borders of the liver in a variety of planes and using software
liver disease, including heavy alcohol consumption, hepatitis, to calculate the total volume in the same manner as the FLR
cirrhosis, nonalcoholic steatohepatitis, and chemotherapy- calculation. There are several limitations to this technique.
associated liver injury such as sinusoidal obstruction syndrome, Most notably, because resection is usually considered on the
steatosis, and chemotherapy-associated steatohepatitis. Sur- basis of hepatic tumors, the volume of the tumors is implicitly
geons contemplating major liver resection in patients with any included in the measurement of the TLV. This is problematic,
of these risk factors should ensure that some measure of liver because the tumor volume does not contribute to hepatic func-
function, in addition to FLR volume, is considered. This tion and so provides a falsely elevated value of the TLV and
chapter reviews these two critical components of FLR assess- hence a falsely diminished anticipated FLR ratio. Manually
ment in detail. measuring the volume of each tumor and subtracting it from
60
Chapter 3 Assessment of hepatic function: implications for the surgical patient 61
techniques are accurate and can be performed by surgeons and

trainees on personal computers with results comparable to
those from experienced radiologists (Dello et al, 2011; van der
Vorst et al, 2010). A recent study from Simpson and coworkers
(2014) using a semiautomated computer software system
(Scout, Pathfinder Technologies; Nashville, Tennessee) high-
lighted the value of this approach.
Volumetric Thresholds
Despite the refinement in methods to measure the FLR, the
clinical application of the information gathered remains contro-
versial. It has long been clear that patients with lower FLR are
at increased risk of hepatic dysfunction, but the exact threshold
below which resection should not be performed is debated.
587.805 cm3 Several studies have attempted to address this fundamental
question, yielding different conclusions (Ferrero et al, 2007;
Kishi et al, 2009; Lin et al, 2014; Pulitano et al, 2014; Schindl
FIGURE 3.1. Volumetric assessment based on magnetic resonance et al, 2005; Shoup et al, 2003). The variable results may be
imaging. attributable to the heterogeneity of included patients (some
having background liver disease and others healthy livers),
methods used to calculate the FLR (TLV vs. TELV), indica-
the TLV to yield the total functioning liver volume can correct tions for PVE, and definitions of hepatic dysfunction. Further-
this but is very labor intensive and prone to measurement error more, only two studies analyzed their results using a formal
(Kubota et al, 1997). The direct measurement technique of receiving operator characteristic (ROC) curve to determine the
TLV is further limited by the fact that the parenchyma beyond optimal FLR threshold, and both studies were limited by small
tumors may be abnormal due to biliary or vascular obstruction. sample sizes (Ferrero et al, 2007; Schindl et al, 2005). Allowing
These limitations typically do not apply to the assessment of for these admittedly crucial differences, the optimal cutoff for
the FLR, which usually does not contain tumors. patients with a normal background liver appears to be between
An alternative method referred to as the total estimated liver 20% and 30%.
volume (TELV) was first proposed by Urata and colleagues Patients who have received preoperative chemotherapy are
(1995) in Japan for use in liver transplantation. Rather than at risk of background liver injury that impairs regeneration fol-
measuring the TLV directly and subtracting the volume of liver lowing partial hepatectomy (Dello et al, 2014; Narita et al,
tumors, this technique estimates the TLV based on body surface 2011; Kele et al, 2013) (see Chapters 71 and 100). There is
area (BSA). The formula was subsequently modified to apply to general consensus that patients treated with extensive preopera-
Western patients, based on the observation that Urata’s formula tive chemotherapy or who have evidence of background liver
underestimated TELV by an average of 323 cm3 (Heinemann injury require a larger FLR to allow safe hepatectomy, although
et al, 1999; Vauthey et al, 2000). The resulting equation, the exact threshold is again controversial. Two studies exam-
TELV = −794 + 1267 × BSA, has been extensively studied and ined this question and performed formal ROC curve analyses,
found to yield a precise estimate of TLV across institutions with reporting optimal thresholds of 31% and 48.5%, respectively
different CT scanners and three-dimensional reconstruction (Ferrero et al, 2007; Narita et al, 2012). The largest study
techniques (Vauthey et al, 2002). When the TELV is used as the includes 194 patients undergoing extended hepatectomy on the
denominator to calculate the FLR ratio (i.e., FLR/TELV), the right side, stratified by extent of preoperative chemotherapy,
resultant ratio is referred to as the standardized FLR (sFLR). with long-duration chemotherapy defined as greater than 12
The measured TLV was compared with the TELV in a weeks (86 patients) (Shindoh et al, 2013b). Using a minimum
recent study of 243 patients who underwent major liver resec- P-value approach, the authors concluded that the optimal
tion (three or more segments) (Ribero et al, 2013). There was cutoff value of FLR for preventing postoperative liver insuffi-
a strong correlation between the two measures across the popu- ciency in these patients was 30%.
lation; however, in overweight patients (body mass index [BMI] The optimal FLR threshold in patients with documented
> 25), TELV was significantly higher, yielding a lower sFLR in underlying liver disease is even less certain, given the additional
these patients. Based on the surgeons’ thresholds, 47 patients variability of defining the extent of background liver injury (see
were deemed to have insufficient liver volume for resection Chapters 71 and 103D). Some authors advocate for PVE in all
using TLV compared with 73 patients using TELV. According patients with chronic liver disease prior to right side hepatec-
to institutional practices at the time, patients who had sufficient tomy, and others apply a conservative threshold as high as 40%
liver volume based on TLV underwent resection, and the subset (Farges et al, 2003; Suda et al, 2009). Given the importance
of patients who had insufficient volume based on TELV had of background liver function, additional functional tests to
significantly worse outcomes than the patients who had suffi- assess the liver remnant should be considered prior to embark-
cient volume based on both calculations. The authors conclude ing on major hepatectomy in the setting of significant back-
that TELV (i.e., sFLR) is a better measure of postoperative ground liver disease.
hepatic insufficiency risk.
Over the past several years, a number of more sophisticated Response to Portal Vein Embolization
software packages have been developed to simplify the process In patients at increased risk of posthepatectomy liver failure,
of volumetric assessment. Several studies suggest that these hypertrophy of the FLR may be induced by preoperative
ipsilateral PVE (Hemming et al, 2003) (see Chapter 108C). (alanine aminotransferase, aspartate aminotransferase, and
Cross-sectional imaging is typically repeated 2 to 6 weeks fol- alkaline phosphatase), and markers of hepatic metabolism [bili-
lowing PVE, and the FLR (or sFLR) may be recalculated. The rubin] and synthetic function (albumin and international nor-
post-PVE FLR can be interpreted with the same thresholds malized ratio [INR]). Aberrations in any of these laboratory
previously discussed, although the degree of hypertrophy (DH), measures should prompt further investigation of background
defined as absolute difference between FLR before and after liver dysfunction, although none of them are sensitive or spe-
PVE, may be more informative (Ribero et al, 2007). In addition cific enough for surgeons to rely on exclusively.
to its therapeutic intent, PVE functions as a diagnostic test The Child-Turcotte and subsequent Child-Pugh scoring
analogous to a cardiac stress test; patients who do not experi- systems were developed to predict the risk of death in patients
ence substantial growth in the FLR following PVE should be undergoing surgical management of portal hypertension. Child-
suspected of harboring background liver disease and approached Pugh has since been applied to predict risk in patients with
with caution. hepatic cirrhosis undergoing a variety of other procedures. The
Recognizing that the DH is contingent upon the duration Child-Pugh score is easily calculated from three readily avail-
from PVE to reimaging, surgeons have proposed incorporating able laboratory tests (bilirubin, albumin, and INR) and two
some measure of growth rate into consideration. The kinetic clinical findings (ascites and encephalopathy). The Child-Pugh
growth rate (KGR) may be calculated by dividing the DH by score is a good marker of global liver function in a patient with
the number of weeks elapsed since PVE (Shindoh et al, 2013a). cirrhosis and may help in selection of patients appropriate for
In one study of 107 patients who underwent liver resection for resection, particularly in the setting of hepatocellular carci-
colorectal liver metastases with an sFLR volume of greater than noma. In general, surgery is reasonable to consider in patients
20%, KGR was a more accurate predictor of postoperative with class A cirrhosis, should be approached cautiously in
hepatic insufficiency than absolute sFLR or DH (area under patients with class B cirrhosis, and should be avoided in patients
the curve [AUC] 0.830) (Shindoh et al, 2013a). In this study, with class C cirrhosis. In patients without cirrhosis, the Child-
patients with a KGR less than 2% per week suffered a 21.6% Pugh score will almost always be normal even when there is
hepatic insufficiency rate and a 8.1% 90-day mortality rate, substantial background liver dysfunction; in this setting, it does
compared with no hepatic insufficiency or 90-day mortality in not predict postoperative liver dysfunction, and other tests are
patients with a KGR greater than 2% per week. In a similar needed.
study of 153 patients who underwent major hepatectomy after The Model for End-Stage Liver Disease (MELD) score is
PVE, post-PVE absolute FLR correlated poorly with liver a mathematical equation frequently used in liver transplanta-
failure (Leung et al, 2014). Both DH and KGR were good tion to allocate organs. The MELD score is similar to the
predictors of liver failure (AUC 0.80 and 0.79, respectively). Child-Pugh score in that it incorporates simple laboratory
Notably, posthepatectomy liver failure did not develop in any investigations, including serum bilirubin, creatinine, and INR,
patients with a KGR greater than 2.66% per week. although it is more cumbersome to calculate. It was initially
In summary, for patients with insufficient FLR (or sFLR) validated for the prediction of short-term survival in patients
to safely undergo hepatectomy, response to PVE provides a with cirrhosis and has subsequently been validated for long-
good measure of the remnant liver’s ability to hypertrophy. term survival as well. In patients with cirrhosis undergoing
Post-PVE FLR should be interpreted in combination with partial hepatectomy, a MELD score greater than 8 is a strong
some measure of extent of hypertrophy (either DH or KGR), predictor of perioperative mortality and decreased long-term
to optimally predict a patient’s risk of posthepatectomy liver survival (Delis et al, 2009; Hsu et al, 2009; Teh et al, 2005).
insufficiency. In contrast, in patients without documented background liver
injury, a MELD score is not strongly associated with inferior
FUNCTIONAL ASSESSMENT outcomes (Rahbari et al, 2011; Schroeder et al, 2006; Teh
OF LIVER REMNANT et al, 2008).
Thus, in patients with cirrhosis being considered for partial
Although a thorough assessment of the anticipated FLR volume hepatectomy, Child-Pugh and MELD scores provide good
is required prior to embarking on major hepatectomy, a com- measures of global liver function. Surgeons should approach
plete assessment should ideally also account for the quality of patients with Child-Pugh class B/C or MELD score greater
the background liver that will be preserved. The optimal than 8 with caution, and consider alternative treatment
method to assess hepatic function would be accurate, noninva- approaches. Clinical scoring systems are not sensitive enough
sive, inexpensive, specific to the remnant portion of the liver, to detect background liver injury and subsequent risk of post-
and widely reproducible. Unfortunately, none of the techniques operative liver dysfunction in patients without cirrhosis; other
currently available fulfill all of these criteria, and therefore methods of functional liver assessment are needed in these
none are frequently utilized in routine assessment. However, patients.
several newer techniques show promise, and with further inves-
tigation may find a role in routine assessment of liver function
Measurement of Hepatic Uptake, Metabolism,
(Table 3.1).
and Elimination
Clinical Scoring Systems Indocyanine Green Clearance
The simplest, most widely available method to assess liver func- Indocyanine green (ICG) clearance is the quantitative measure
tion relies on laboratory investigations either in isolation or of hepatic function most used worldwide. ICG is a water-
combined into clinical scoring systems. Clinicians are familiar soluble tricarbocyanine dye that binds to albumin and distrib-
with conventional liver laboratory tests routinely used in clinical utes rapidly and uniformly in the blood after intravenous
practice, including enzymatic measures of hepatocyte injury injection. ICG is exclusively cleared from the bloodstream by
TABLE 3.1 Comparison of Tests Available for Assessment of Liver Function

Modality Rationale Advantages Limitations
Volume of liver Lower liver remnant volume is Can be easily calculated with Does not incorporate measure of
remnant (MRI, associated with worse conventional imaging underlying liver function
CT) outcomes Can be performed by surgeons Threshold for safe resection in setting of
Incorporates planned resection background liver disease unclear
Response of liver Failure of liver to hypertrophy Can be easily calculated with Requires invasive procedure that may not
remnant to in response to portal vein conventional imaging be necessary in some patients
portal vein embolization indicates Can be performed by surgeons
embolization underlying liver injury Incorporates planned resection
Clinical scoring Scoring systems are Easy to calculate Not sensitive enough for background liver
systems associated with poor Noninvasive dysfunction
(Child-Pugh, outcomes following other
MELD, etc.) procedures
ICG clearance ICG is metabolized by the Good measure of underlying liver Time consuming
liver, poor ICG clearance is function Measures total liver function, not specific
indicative of underlying liver to remnant
dysfunction Altered based on environmental conditions
99m
Hepatobiliary Tc- GSA binds to Provides anatomic and functional High interrater and interinstitution variability
scintigraphy hepatocyte receptors; information Limited availability
99m
Tc- IDA derivatives are May be specific to remnant liver
metabolized by the liver; May be combined with SPECT
poor uptake of either are
indicative of liver
dysfunction
Other measures of Metabolized almost Correlated with other measures Not widely available
metabolic exclusively by the liver of total liver function Limited data related to clinical outcomes
function (P450), poor clearance High interrater variability
(lidocaine, indicates underlying liver Time consuming
galactose, 13C dysfunction Measures total liver function, not specific
breath tests) to remnant
Altered based on environmental conditions
MRI with Taken up and cleared by Routinely available Not directly correlated with postresection
Gd-EOB-DTPA hepatocytes, poor uptake Frequently used in preoperative clinical outcomes
contrast indicates liver dysfunction assessment
May be specific to remnant liver
Provides other information
Transient Provides assessment of liver Noninvasive User dependent
elastography fibrosis Fast Not correlated with clinical outcomes
CT, Computed tomography; Gd-EOB-DTPA, gadolinium ethoxybenzyl dimeglumine; ICG, indocyanine green; MRI, magnetic resonance imaging; SPECT, ; 99mTc- GSA, technetium-99m–labeled galactosyl
serum albumin; 99mTc- IDA, technetium-99m–labeled iminodiacetic acid.
the liver in a similar manner to bilirubin and toxins, and then capillarization. Further, ICG clearance testing is a measure of
excreted unchanged into bile. Thus ICG clearance tests reflect global liver function, so if there is heterogeneous uptake in the
blood flow-dependent clearance, hepatocyte uptake and biliary liver (e.g., the portion being resected does not function as well
excretion. due to tumor, biliary obstruction, etc.), the results may be mis-
The conventional measurement of ICG clearance involves leading. Finally, ICG testing does not incorporate the extent of
intravenous injection of ICG, followed by serial collection resection, or conversely, the volume of the remnant that will
of venous blood at 5 minute intervals for 15 minutes. ICG remain. Researchers have attempted to mitigate some of these
clearance can also be measured noninvasively by pulse- limitations by creating scoring systems and decision trees that
spectrophotometry, which allows real-time monitoring of liver incorporate ICG (Du et al, 2011; Kim et al, 2015).
function (Okochi et al, 2002; Sakka et al, 2000;). The results
of ICG tests may be expressed as the percentage of ICG Nuclear Imaging Techniques
retained in the circulation 15 minutes after injection (ICG- Theoretically, nuclear imaging represents an attractive preop-
R15), the plasma disappearance rate (ICG-PDR), and the erative hepatic assessment, combining anatomic considerations
elimination rate constant (ICG-k). Several studies have identi- (FLR volume) with both total and regional liver functional
fied an association between elevated ICG-R15 and posthepa- assessment. Several scintigraphic tests have been developed
tectomy complications, with proposed threshold values of over the past few decades, but the most widely used radiophar-
ICG-R15 ranging from 14% to 20% (Das et al, 2001; Fan maceutical imaging methods for liver functional assessment are
et al, 1995; Lau et al, 1997). technetium-99m–labeled galactosyl serum albumin (GSA)
Despite the theoretical attractiveness of ICG clearance as scintigraphy and hepatobiliary scintigraphy (HBS) with 99mTc-
a simple measure of hepatic function, several limitations labeled iminodiacetic acid (IDA) derivatives. Both of these
have hampered enthusiasm for its widespread use. The results of methods provide quantitative data on the total and regional
ICG clearance tests are not reliable in patients with hyperbiliru- hepatic function, although they are based on different principles
binemia, or in patients with intrahepatic shunting or sinusoidal and therefore interpretation varies.
99m
Tc-GSA is an analogue of a glycoprotein (ascites sialo- (Oellerich et al, 2001; Reichen, 1993). Unfortunately, the test
glycoprotein) that binds to receptors on the hepatocyte cell is limited by poor reliability and the need for frequent monitor-
membrane and is taken up by the hepatocytes. Chronic liver ing; therefore its present application in preoperative assessment
disease results in diminished hepatocyte glycoprotein receptors of liver function is investigational only. Galactose elimination
and subsequent accumulation of plasma glycoproteins. To capacity also accurately reflects metabolic function of the liver
perform dynamic scintigraphy, an intravenous bolus of 99mTc- but is similarly limited by practical constraints and alterations
GSA is administered, and images are obtained by using a due to environmental conditions (Ranek et al, 1976). Finally,
gamma camera positioned over the heart and liver. Several there are a variety of 13C breath tests available that follow the
parameters may be calculated to document the extent of hepatic same principles, including the 13C-methacetin breath test
99m
Tc-GSA uptake, including the hepatic uptake ratio (LHL15 (LiMAx; Humedics, Berlin) (Cieslak et al, 2014; Stockmann
[receptor index: uptake ratio of the liver to the liver plus heart et al, 2009). Unfortunately, these tests also suffer from poor
at 15 min]) and the blood clearance ratio (HH15 [blood clear- interobserver reliability and alterations based on physiologic
ance index: uptake ratio of the heart at 15 min to that at conditions so should be viewed as experimental at present
3 min]). In patients with cirrhosis, 99mTc-GSA uptake corre- (Afolabi et al, 2013).
sponds well with other conventional liver function tests, includ-
ing ICG clearance, and predicts histologic severity of disease Magnetic Resonance Imaging Hepatic Agents
better than ICG clearance in a substantial proportion of patients MRI with contrast enhancement offers high-resolution cross-
(Kwon et al, 1995; Nanashima et al, 2004). Several small sectional assessment of background liver anatomy and accurate
studies have demonstrated an association between poor 99mTc- characterization of hepatic tumors. MRI is more sensitive and
GSA uptake and postoperative complications after liver resec- specific than CT for the detection of primary and metastatic liver
tion (Kim et al, 1997; Nanashima et al, 2004; Takeuchi et al, neoplasms and is used routinely at most centers before embark-
1999). However, 99mTc-GSA uptake is limited by interoperator ing on liver resection (Zech et al, 2014). Gadolinium ethoxyben-
and interinstitutional differences and does not provide a zyl dimeglumine (Gd-EOB-DTPA) is a liver-specific contrast
measure of regional liver function (Koizumi et al, 1997). To agent that has as much as 50% hepatobiliary excretion in a
address this limitation, 99mTc-GSA scintigraphy may be com- normal liver (Van Beers et al, 2012). Gd-EOB-DTPA improves
bined with static single-photon emission computed tomography– the detection and characterization of focal liver lesions and
CT (SPECT-CT) to allow a three-dimensional measurement diffuse liver disease. Given the hepatic uptake and elimination of
of 99mTc-GSA uptake. Results of dynamic SPECT-CT may Gd-EOB-DTPA, contrast-enhanced MRI may also provide
help to predict postoperative liver failure; however, this method functional assessment of the background liver (Fig. 3.2).
suffers from the same interobserver variability and environmen- Several small studies have demonstrated correlation
tal factors as dynamic 99mTc-GSA scintigraphy (Beppu et al, between Gd-EOB-DTPA uptake on MRI and conventional
2011; Iimuro et al, 2010; Satoh et al, 2003). Thus, although measures of liver function (Nilsson et al, 2013; Nishie et al,
this is a promising tool with some utility in the preoperative 2012; Saito et al, 2014). There are several theoretical and
assessment of patients, further work is needed to standardize practical advantages to using Gd-EOB-DTPA–enhanced MRI
measurement before wide implementation. to assess liver function. First, MRI is routinely available and
99m
Tc-mebrofenin is an organic IDA derivative with similar frequently used in the preoperative assessment of these
properties to ICG: It has high hepatic uptake, low displacement patients, so no additional testing is required. Second, func-
by bilirubin, and low urinary excretion. The test is administered tional assessment may be focused on the planned FLR in
in an identical manner to 99mTc-GSA scintigraphy, using a cases of heterogeneous uptake, rather than calculating uptake
gamma camera and calculating similar parameters and ratios. for the whole liver as is the case in most other functional
However, the uptake ratio is divided by the patient’s BSA to quantitative tests. Finally, MRI provides visual assessment of
compensate for differences in metabolic requirements. 99mTc- background liver injury, including steatosis and fibrosis, which
mebrofenin HBS correlates well with ICG clearance, and may further assist in preoperative decision making. However,
appears to be a good marker of post resection liver function enthusiasm for the use of MRI for functional liver assessment
(Bennink et al, 2004; de Graaf et al, 2010a; Dinant et al, 2007; must be tempered by the lack of prospective data demonstrat-
Erdogan et al, 2004). HBS may also be combined with ing a clear relationship between Gd-EOB-DTPA uptake and
SPECT-CT to allow three-dimensional assessment of the FLR, clinical outcomes postresection. Further research is needed to
particularly in patients undergoing PVE (de Graaf et al, 2010b). determine whether MRI with Gd-EOB-DTPA contrast will
The main limitation of HBS is, again, interobserver and inter- indeed be the panacea of functional liver assessment that it
institution variability. Although these techniques offer great appears to be.
advantages compared with more conventional methods, further
research is needed to ensure that results are reproducible across Transient Elastography
different settings before wider application. Ultrasound transient elastography (TE) has been reported as a
test to estimate the extent of liver fibrosis. Ultrasound TE has
Other Measures of Metabolic Function the clear advantages of being noninvasive and fast but is limited
In addition to ICG, several other compounds are metabolized by significant interobserver variability and anatomic variations
almost exclusively by the liver cytochrome P450 system and (Kawamoto et al, 2006; Sandrin et al, 2003). Furthermore,
have been investigated as potential markers of hepatic function. although results of ultrasound TE are correlated with extent of
For example, lidocaine is metabolized to monoethylglycinexyli- fibrosis, the impact of abnormal TE on clinical outcomes fol-
dide (MEGX) primarily in the liver. The MEGX test has lowing resection is unknown. Thus further research is needed
been studied in transplantation and critical care medicine and before incorporating this imaging modality into routine clinical
appears to correlate with other measures of hepatic metabolism practice.
Texture Analysis
Texture analysis is an established technique that characterizes
regions of interest in an image based on spatial variations in pixel
intensity. On CT imaging, texture analysis can potentially quan-
tify regional variations in enhancement that cannot be assessed
by inspection. Several studies have shown potential utility of this
new technique for tumor diagnosis and characterization and
prognostication. In a recent study by Simpson and coworkers
(2015) analyzing patients submitted to major hepatic resection,
texture variables of preoperative CT scans showed promise for
predicting postoperative hepatic failure, and may represent a
new means of preoperative risk stratification.
A
CONCLUSION
Hepatobiliary surgeons now have a variety of tools at their
disposal to assist with preoperative assessment of hepatic func-
tion. The gold standard remains volumetric-based assessment
of the FLR with cross-sectional imaging (CT or MRI). In
patients in whom there is concern about insufficient liver
volume or background liver injury, response to PVE provides
a functional assessment of the FLR in addition to its therapeu-
tic role. Quantitative measures of hepatic uptake, metabolism,
and elimination, including ICG clearance, nuclear scintigraphy,
and MRI hepatic-specific contrast agents, may have a role in
assessment of patients with borderline FLR volume or back-
ground liver disease but are currently limited by physiologic
alterations and poor reproducibility. Further refinement of
these techniques may allow the development of algorithms or
B decision aids that incorporate both volumetric and functional
assessment of the FLR.
FIGURE 3.2. Magnetic resonance imaging with gadolinium ethoxy-
benzyl dimeglumine contrast on two patients demonstrating normal References are available at expertconsult.com.
uptake (A) and diffusely decreased uptake (B).
Chapter 3 Assessment of hepatic function: implications for the surgical patient 65.e1
REFERENCES Karlo C, et al: CT- and MRI-based volumetry of resected liver speci-
men: comparison to intraoperative volume and weight measure-
Abdalla EK, et al: Total and segmental liver volume variations: implica- ments and calculation of conversion factors, Eur J Radiol 75(1):
tions for liver surgery, Surgery 135(4):404–410, 2004. e107–e111, 2010.
Adams RB, et al: Selection for hepatic resection of colorectal liver Kawamoto M, et al: Assessment of liver fibrosis by a noninvasive
metastases: expert consensus statement, HPB (Oxford) 15(2):91– method of transient elastography and biochemical markers, World J
103, 2013. Gastroenterol 12(27):4325–4330, 2006.
Afolabi P, et al: Clinical utility of 13C-liver-function breath tests for Kele PG, et al: The impact of hepatic steatosis on liver regeneration
assessment of hepatic function, Dig Dis Sci 58(1):33–41, 2013. after partial hepatectomy, Liver Int 33(3):469–475, 2013.
Bennink RJ, et al: Preoperative assessment of postoperative remnant Kim HJ, et al: Volumetric analysis and indocyanine green retention rate
liver function using hepatobiliary scintigraphy, J Nucl Med 45(6):965– at 15 min as predictors of post-hepatectomy liver failure, HPB
971, 2004. (Oxford) 17(2):159–167, 2015.
Beppu T, et al: Liver functional volumetry for portal vein embolization Kim YK, et al: Prediction of postoperative decompensated liver func-
using a newly developed 99mTc-galactosyl human serum albumin tion by technetium-99m galactosyl-human serum albumin liver scin-
scintigraphy SPECT-computed tomography fusion system, J Gastro- tigraphy in patients with hepatocellular carcinoma complicating
enterol 46(7):938–943, 2011. chronic liver disease, Br J Surg 84(6):793–796, 1997.
Charnsangavej C, et al: Selection of patients for resection of hepatic Kishi Y, et al: Three hundred and one consecutive extended right
colorectal metastases: expert consensus statement, Ann Surg Oncol hepatectomies: evaluation of outcome based on systematic liver volu-
13(10):1261–1268, 2006. metry, Ann Surg 250(4):540–548, 2009.
Cieslak KP, et al: New perspectives in the assessment of future remnant Koizumi M, et al: An easy and reproducible semi-automatic method
liver, Dig Surg 31(4–5):255–268, 2014. for the evaluation of 99mTc-galactosyl human serum albumin, Ann
Das BC, et al: Analysis of 100 consecutive hepatectomies: risk factors Nucl Med 11(4):345–348, 1997.
in patients with liver cirrhosis or obstructive jaundice, World J Surg Kubota K, et al: Measurement of liver volume and hepatic functional
25(3):266–272, discussion 272-263, 2001. reserve as a guide to decision-making in resectional surgery for
de Graaf W, et al: Assessment of future remnant liver function using hepatic tumors, Hepatology 26(5):1176–1181, 1997.
hepatobiliary scintigraphy in patients undergoing major liver resec- Kwon AH, et al: Use of technetium 99m diethylenetriamine-pentaacetic
tion, J Gastrointest Surg 14(2):369–378, 2010a. acid-galactosyl-human serum albumin liver scintigraphy in the evalu-
de Graaf W, et al: (99m)Tc-mebrofenin hepatobiliary scintigraphy with ation of preoperative and postoperative hepatic functional reserve for
SPECT for the assessment of hepatic function and liver functional hepatectomy, Surgery 117(4):429–434, 1995.
volume before partial hepatectomy, J Nucl Med 51(2):229–236, Lau H, et al: Evaluation of preoperative hepatic function in patients
2010b. with hepatocellular carcinoma undergoing hepatectomy, Br J Surg
Delis SG, et al: Model for end-stage liver disease (MELD) score, as a 84(9):1255–1259, 1997.
prognostic factor for post-operative morbidity and mortality in cir- Leung U, et al: Remnant growth rate after portal vein embolization is
rhotic patients, undergoing hepatectomy for hepatocellular carci- a good early predictor of post-hepatectomy liver failure, J Am Coll
noma, HPB (Oxford) 11(4):351–357, 2009. Surg 219(4):620–630, 2014.
Dello SA, et al: Influence of preoperative chemotherapy on CT volu- Lin XJ, et al: The critical value of remnant liver volume-to-body weight
metric liver regeneration following right hemihepatectomy, World J ratio to estimate posthepatectomy liver failure in cirrhotic patients,
Surg 38(2):497–504, 2014. J Surg Res 188(2):489–495, 2014.
Dello SA, et al: Prospective volumetric assessment of the liver on a Nanashima A, et al: Relationship between indocyanine green test and
personal computer by nonradiologists prior to partial hepatectomy, technetium-99m galactosyl serum albumin scintigraphy in patients
World J Surg 35(2):386–392, 2011. scheduled for hepatectomy: Clinical evaluation and patient outcome,
Dinant S, et al: Risk assessment of posthepatectomy liver failure using Hepatol Res 28(4):184–190, 2004.
hepatobiliary scintigraphy and CT volumetry, J Nucl Med 48(5):685– Narita M, et al: Sinusoidal obstruction syndrome compromises liver
692, 2007. regeneration in patients undergoing two-stage hepatectomy with
Du ZG, et al: A new scoring system for assessment of liver function portal vein embolization, Surg Today 41(1):7–17, 2011.
after successful hepatectomy in patients with hepatocellular carci- Narita M, et al: What is a safe future liver remnant size in patients
noma, Hepatobiliary Pancreat Dis Int 10(3):265–269, 2011. undergoing major hepatectomy for colorectal liver metastases and
Erdogan D, et al: Preoperative assessment of liver function: a compari- treated by intensive preoperative chemotherapy? Ann Surg Oncol
son of 99mTc-Mebrofenin scintigraphy with indocyanine green 19(8):2526–2538, 2012.
clearance test, Liver Int 24(2):117–123, 2004. Nilsson H, et al: Gd-EOB-DTPA-enhanced MRI for the assessment of
Fan ST, et al: Hospital mortality of major hepatectomy for hepatocel- liver function and volume in liver cirrhosis, Br J Radiol 86(1026):
lular carcinoma associated with cirrhosis, Arch Surg 130(2):198–203, 20120653, 2013.
1995. Nishie A, et al: Quantitative analysis of liver function using superpara-
Farges O, et al: Portal vein embolization before right hepatectomy: magnetic iron oxide- and Gd-EOB-DTPA-enhanced MRI: compari-
prospective clinical trial, Ann Surg 237(2):208–217, 2003. son with Technetium-99m galactosyl serum albumin scintigraphy,
Ferrero A, et al: Postoperative liver dysfunction and future remnant Eur J Radiol 81(6):1100–1104, 2012.
liver: where is the limit? Results of a prospective study, World J Surg Oellerich M, Armstrong VW: The MEGX test: a tool for the real-
31(8):1643–1651, 2007. time assessment of hepatic function, Ther Drug Monit 23(2):81–92,
Heinemann A, et al: Standard liver volume in the Caucasian popula- 2001.
tion, Liver Transpl Surg 5(5):366–368, 1999. Okochi O, et al: ICG pulse spectrophotometry for perioperative liver
Hemming AW, et al: Preoperative portal vein embolization for extended function in hepatectomy, J Surg Res 103(1):109–113, 2002.
hepatectomy, Ann Surg 237(5):686–691, discussion 691–683, 2003. Pulitano C, et al: Preoperative assessment of postoperative liver
Heymsfield SB, et al: Accurate measurement of liver, kidney, and function: the importance of residual liver volume, J Surg Oncol
spleen volume and mass by computerized axial tomography, Ann 110(4):445–450, 2014.
Intern Med 90(2):185–187, 1979. Rahbari NN, et al: The predictive value of postoperative clinical risk
Hsu KY, et al: Predicting morbidity and mortality after hepatic resec- scores for outcome after hepatic resection: a validation analysis in
tion in patients with hepatocellular carcinoma: the role of Model for 807 patients, Ann Surg Oncol 18(13):3640–3649, 2011.
End-Stage Liver Disease score, World J Surg 33(11):2412–2419, Ranek L, et al: Galactose elimination capacity as a prognostic index in
2009. patients with fulminant liver failure, Gut 17(12):959–964, 1976.
Huynh HT, et al: Computerized liver volumetry on MRI by using 3D Reichen J: MEGX test in hepatology: the long-sought ultimate quan-
geodesic active contour segmentation, AJR Am J Roentgenol 202(1): titative liver function test? J Hepatol 19(1):4–7, 1993.
152–159, 2014. Ribero D, et al: Portal vein embolization before major hepatectomy and
Iimuro Y, et al: Preoperative estimation of asialoglycoprotein receptor its effects on regeneration, resectability and outcome, Br J Surg
expression in the remnant liver from CT/99mTc-GSA SPECT fusion 94(11):1386–1394, 2007.
images correlates well with postoperative liver function parameters, Ribero D, et al: Measured versus estimated total liver volume to pre-
J Hepatobiliary Pancreat Sci 17(5):673–681, 2010. operatively assess the adequacy of the future liver remnant: which
method should we use? Ann Surg 258(5):801–806, discussion 806– Stockmann M, et al: Prediction of postoperative outcome after hepa-
807, 2013. tectomy with a new bedside test for maximal liver function capacity,
Saito K, et al: Measuring hepatic functional reserve using low temporal Ann Surg 250(1):119–125, 2009.
resolution Gd-EOB-DTPA dynamic contrast-enhanced MRI: a pre- Suda K, et al: Risk factors of liver dysfunction after extended hepatic
liminary study comparing galactosyl human serum albumin scintig- resection in biliary tract malignancies, Am J Surg 197(6):752–758,
raphy with indocyanine green retention, Eur Radiol 24(1):112–119, 2009.
2014. Suzuki K, et al: Computerized segmentation of liver in hepatic CT and
Sakka SG, et al: Comparison of invasive and noninvasive measure- MRI by means of level-set geodesic active contouring, Conf Proc
ments of indocyanine green plasma disappearance rate in critically IEEE Eng Med Biol Soc 2013:2984–2987, 2013.
ill patients with mechanical ventilation and stable hemodynamics, Takeuchi S, et al: Predicting survival and post-operative complications
Intensive Care Med 26(10):1553–1556, 2000. with Tc-GSA liver scintigraphy in hepatocellular carcinoma, Hepa-
Sandrin L, et al: Transient elastography: a new noninvasive method for togastroenterology 46(27):1855–1861, 1999.
assessment of hepatic fibrosis, Ultrasound Med Biol 29(12):1705– Teh SH, et al: Hepatic resection of hepatocellular carcinoma in patients
1713, 2003. with cirrhosis: Model of End-Stage Liver Disease (MELD) score
Satoh K, et al: 99mTc-GSA liver dynamic SPECT for the preoperative predicts perioperative mortality, J Gastrointest Surg 9(9):1207–1215,
assessment of hepatectomy, Ann Nucl Med 17(1):61–67, 2003. discussion 1215, 2005.
Schindl MJ, et al: The value of residual liver volume as a predictor of Teh SH, et al: Model for End-Stage Liver Disease score fails to predict
hepatic dysfunction and infection after major liver resection, Gut perioperative outcome after hepatic resection for hepatocellular car-
54(2):289–296, 2005. cinoma in patients without cirrhosis, Am J Surg 195(5):697–701,
Schroeder RA, et al: Predictive indices of morbidity and mortality after 2008.
liver resection, Ann Surg 243(3):373–379, 2006. Urata K, et al: Calculation of child and adult standard liver volume for
Shindoh J, et al: Kinetic growth rate after portal vein embolization liver transplantation, Hepatology 21(5):1317–1321, 1995.
predicts posthepatectomy outcomes: toward zero liver-related mor- Van Beers BE, et al: Primovist, Eovist: what to expect? J Hepatol
tality in patients with colorectal liver metastases and small future 57(2):421–429, 2012.
liver remnant, J Am Coll Surg 216(2):201–209, 2013a. van der Vorst JR, et al: Virtual liver resection and volumetric analysis
Shindoh J, et al: Optimal future liver remnant in patients treated with of the future liver remnant using open source image processing
extensive preoperative chemotherapy for colorectal liver metastases, software, World J Surg 34(10):2426–2433, 2010.
Ann Surg Oncol 20(8):2493–2500, 2013b. Vauthey JN, et al: Body surface area and body weight predict total liver
Shoup M, et al: Volumetric analysis predicts hepatic dysfunction in volume in Western adults, Liver Transpl 8(3):233–240, 2002.
patients undergoing major liver resection, J Gastrointest Surg Vauthey JN, et al: Standardized measurement of the future liver
7(3):325–330, 2003. remnant prior to extended liver resection: methodology and clinical
Simpson AL, et al: Liver planning software accurately predicts postop- associations, Surgery 127(5):512–519, 2000.
erative liver volume and measures early regeneration, J Am Coll Surg Zech CJ, et al: Randomized multicentre trial of gadoxetic acid-
219:199–207, 2014. enhanced MRI versus conventional MRI or CT in the staging of
Simpson AL, et al: Texture analysis of preoperative CT images for colorectal cancer liver metastases, Br J Surg 101(6):613–621, 2014.
prediction of postoperative hepatic insufficiency: a preliminary study,
J Am Coll Surg 220:339–346, 2015.
CHAPTER 4
Pancreatic Physiology and Functional Assessment
Alessandro Paniccia and Richard D. Schulick
The pancreas is a complex retroperitoneal gland with both STRUCTURE

endocrine (e.g., glucose homeostasis) and exocrine (e.g., nutri-
ent digestion) functions. An adult human pancreas measures There are at least five major cell types in each islet of Langer-
approximately 15 cm in length and weighs between 60 to hans: α, β, δ, F, and ε cells. In humans, pancreatic α cells,
100 g; however, its size can vary due to aging or pathologic which principally secrete glucagon, represent approximately
conditions (e.g., pancreatitis, neoplasia) (Gray, 2000; Syed 35% of all islet cells. Pancreatic β cells, which are responsible
et al, 2012). It is of endodermal origin and arises from two for the production and secretion of insulin and amylin, repre-
independent primordia: a ventral bud (derived from the hepatic sent approximately 55% of islet cells. Pancreatic δ cells, which
diverticulum) and a dorsal bud (derived from the developing principally secrete somatostatin, represent less than 10% of the
duodenum). Around the fifth week of gestation, the ventral bud islet cells, and pancreatic F cells, which secrete pancreatic poly-
rotates clockwise with the developing duodenum to fuse with peptide (PP), account for less than 5%. Finally, the ε cells,
the dorsal pancreatic bud (Cano et al, 2014; Pan & Wright, which secrete human islet cells, account for less than 1% of
2011). Ultimately, the ventral bud will form the inferior pan- human islet cells (Jain & Lammert, 2009).
creatic head and the uncinate process. The dorsal bud will The distribution and cellular composition of the different
constitute the majority of the pancreatic gland, representing the cells types within the islet varies among species. Previous animal
superior pancreatic head, the body, and the tail of the adult models with rabbits, rats, and mice demonstrated that β cells
pancreas. During this process, the main ducts of the ventral occupy the core of the islet of Langerhans and that non–β cells
and dorsal pancreatic buds fuse to form the main pancreatic are distributed toward the outside of the islet (Cabrera et al,
duct (duct of Wirsung). The major pancreatic duct drains most 2006). Recent studies in humans, however, have demonstrated
of the organ’s secretions through the major duodenal papilla a different cytoarchitecture, where the majority of α, β, and δ
(ampulla of Vater). A separate draining duct, arising from the cells are distributed along the islet blood vessels without a
dorsal pancreatic bud, usually persists and forms the minor specific order (Cabrera et al, 2006). Furthermore, approxi-
pancreatic duct (duct of Santorini). The minor duct drains a mately 70% of human β cells appear to be in contact with
portion of the pancreatic head secretions into the duodenum non–β-islet cells, suggesting a predisposition for paracrine
through the minor papilla, located 2 cm anterosuperior of the interaction (Barker et al, 2013). Regional location of the islet
major papilla (Boron & Boulpaep, 2012). within the human pancreas is also important to islet cytoarchi-
The pancreas receives a rich arterial vascular supply from tecture (Barker et al, 2013). Islets located in the body and tail
branches of the celiac and superior mesenteric artery. The of the pancreas have a higher proportion of α cells and a lower
venous drainage follows the arterial supply, with venous efflu- proportion of F cells, whereas islets located in the uncinate
ents ultimately draining into the portal vein (Moore et al, process have a higher proportion of F cells and a lower propor-
2013). Furthermore, the pancreas is supplied by several neural tion of α cells. Notably, β cells and δ cells are present in nearly
sources, including sympathetic fibers from the splanchnic equal proportions throughout the pancreas (Boron & Boulpaep,
nerves, parasympathetic fibers from the vagus nerve, and pep- 2012).
tidergic neurons (releasing amines and peptides) (Mussa & The islets are rich in axonal terminals and blood capillaries
Verberne, 2013; Rodriguez-Diaz, et al, 2011a) (see Chapter 2). that participate in extensive neurohumoral and nonneuronal
paracrine regulation. Recent studies, utilizing three-dimensional
ENDOCRINE PANCREAS reconstruction of the axonal terminal field, revealed that the
autonomic innervation to the human islet of Langerhans is dif-
The islets of Langerhans are the functional units of the endo- ferent from that previously identified in rodents (Rodriguez-
crine pancreas and have a paramount role in maintaining Diaz, et al, 2011a). Contrary to what was previously understood,
glucose homeostasis. In light of their complex cytoarchitecture human β cells are poorly innervated by the parasympathetic
structure and regulatory system, they are currently defined as cholinergic system (Rodriguez-Diaz, et al, 2011a). Instead,
a microorgan(s) within the pancreas (Barker et al, 2013). The sympathetic neural terminals penetrate into the human islet of
pancreas of a healthy adult has approximately one million islets Langerhans to innervate the smooth muscle cells of the blood
that are evenly distributed throughout the pancreatic gland and vessels, allowing fine regulation of islet blood flow. Conse-
account for 1% to 2% of the organ’s mass. Each islet ranges in quently, sympathetic nerves indirectly influence downstream
size from 50 to 300 µm in diameter and contains a few hundred endocrine cells by regulating the local blood flow containing
to a few thousand endocrine cells (Boron & Boulpaep, 2012). secreted endocrine hormone (Rodriguez-Diaz, et al, 2011a).
66
Chapter 4 Pancreatic Physiology and Functional Assessment 67
The islets of Langerhans receive approximately 20% of the The main glucose transporters on pancreatic β cells are
pancreatic arterial flow, with distribution significantly influ- GLUT-2, which are highly expressed in rodents (van de Bunt
enced by the different phases of digestion (Bonner-Weir, 1993). & Gloyn, 2012), and GLUT-1 and GLUT-3, both recently
Furthermore, an insuloacinar portal system responsible for recognized to be expressed at high levels on human β cells (Fig.
draining blood and secreted hormones from the islets of Lang- 4.1) (McCulloch et al, 2011). The GLUT transporters allow
erhans to the acinar element of the pancreas has been described equalization of the intracellular and extracellular glucose con-
in several species, including humans (Merkwitz et al, 2013). centrations. If the glucose concentration exceeds 5 mmol/L,
Therefore hormones secreted by the islet of Langerhans are intracellular β-cell glucokinase enzymes activate to allow fine
directly transported to the acinar cells, where they can exert a regulation of insulin secretion (Tan, 2014). Acting as a “glucose
local regulatory function. In addition, a local regulatory effect sensor,” these enzymes are responsible for the phosphorylation
on endocrine and exocrine pancreatic function is exerted by of glucose to glucose-6-phosphate (Lenzen, 2014; Matschinsky
several neuropeptides, including neuropeptide-Y, gastrin- et al, 1998). Glucose-6-phosphate accumulates in the β cell, is
releasing peptide, and calcitonin gene–related peptide (CGRP). metabolized, and contributes to an increase in cellular adenos-
ine triphosphate (ATP). The β-cell membrane is rich in ATP-
SYNTHESIS AND STORAGE OF INSULIN dependent potassium channels, which subsequently undergo
closure in response to the surge in ATP, resulting in membrane
In 1923 Banting and McLeod, two Canadian surgeons, were depolarization (Rorsman et al, 2014; Yang et al, 2014). Depo-
awarded the Nobel Prize in Physiology or Medicine for the larization activates voltage-gated L-type calcium channels,
discovery of insulin (Banting, 1926). This 51 amino acid poly- leading to influx of calcium into the cell (Rorsman & Braun,
peptide is primarily responsible for maintaining serum glucose 2013; Rutter & Hodson, 2013) The increased intracellular
between 4 mM and 8 mM (70 to 140 mg/dL) during periods calcium concentration leads to margination of secretory gran-
of feeding and fasting (Rorsman & Braun, 2013). In addition, ules, their fusion with the cell membrane, and exocytosis of
insulin regulates lipid and protein metabolism. The gene their content, including insulin and its intermediate products
responsible for encoding insulin is located on the short arm of (e.g., C-peptide) (Yang et al, 2014). This process characterizes
chromosome 11 and leads to the translation of a preprohor- the “first phase” of insulin release, in which insulin is rapidly
mone protein known as preproinsulin within the β cell. Prepro- secreted within 3 to 5 minutes of glucose administration and
insulin consists of a leading sequence of 24 amino acids, terminates within 10 minutes. The loss of the first phase of
followed by three domains named “B,” “C,” and “A.” Succes- insulin secretion is one of the earliest metabolic defects identi-
sive cleavage processes take place starting at the time of transla- fied in type 2 diabetes mellitus (T2DM) (Nagamatsu et al,
tion until the final secretion. First, cleavage of the leading 2006). The “second phase” of insulin secretion is longer lasting
sequence in the endoplasmic reticulum leads to the formation (reaching a plateau in insulin secretion after 2 to 3 hours), and
of proinsulin. As the proinsulin is arranged into secretory gran- its regulation is not completely understood (Henquin, 2009;
ules in the trans-Golgi, additional proteases cleave the central Huang & Joseph, 2014). However, recent mathematic models
31 amino acid C-peptide. This leads to the formation of a suggest that the second phase is characterized by the recruit-
mature insulin peptide composed of an A-chain and B-chain ment and mobilization of intracellular granules containing
held together by two disulfide bonds, ready to be released in insulin (as opposed to predocked granules as in the first phase)
the secretory vesicle. Cleaved C-peptide and other intermediate in a dose-dependent glucose response (Stamper & Wang, 2013).
products, such as proinsulin, remain present in the secretory The amino acid arginine (L-arginine) is another well-known
granules and are eventually released together with the mature insulin secretagogue. Following uptake into the β cells through
insulin (Boron & Boulpaep, 2012). The mature insulin peptide a cationic amino-acid transporter (CAT), arginine leads to
has a plasma half-life of 4 minutes. It is rapidly internalized by depolarization of cell membrane, which triggers calcium influx
target organs expressing its receptor and degraded by the (Smith et al, 1997). Furthermore, L-arginine can stimulate the
kidneys and liver (Marques et al, 2004). Notably, C-peptide release of GLP-1, which acts at its receptor (GLP-1R) to
has a plasma half-life of 30 minutes and is excreted unchanged augment glucose-stimulated insulin secretion from pancreatic
by the kidneys, making it a clinically important marker of β cells (Clemmensen et al, 2013; Tolhurst et al, 2009).
endogenous insulin secretion (Jones & Hattersley, 2013). The secretion of insulin from pancreatic β cells is further
potentiated by the incretin effect through the enteroinsular axis
STIMULUS-SECRETION COUPLING (Diab & D’Alessio, 2010; Opinto et al, 2013). The incretin
FOR INSULIN SECRETION effect is the phenomenon whereby the presence of nutrients,
especially carbohydrates, in the duodenal lumen stimulates
The rise of islet cell transplantation has led to a renewed under- cells in the gut mucosa to release potent insulin secretagogues
standing of human β-cell regulation, building on earlier work (Drucker, 2013). The ingestion of nutrients stimulates duode-
completed in rodents. Insulin is released from β cells through nal and jejunal K cells to produce and release GIP, a well-
two mechanisms: unstimulated and stimulated secretion. studied incretin. In addition, GLP-1 is produced and release
Unstimulated secretion or basal insulin secretion occurs every by the L cells (also known as enteroglucagone cells), located in
6 to 8 minutes (Song et al, 2002). Stimulated secretion of the distal small bowel, colon, and rectum. It has been shown
insulin occurs in response to several stimuli, including glucose, that orally administered glucose is able to stimulate insulin
amino acids (e.g., arginine), acetylcholine (ACh), glutamate, secretion as much as 25% more than intravenously adminis-
and incretins such as gastric inhibitory peptide (GIP) and tered glucose, likely through the incretin effect (Ahrén, 2013).
glucagon-like peptide-1 (GLP-1). The change in extracellular GIP and GLP-1 play major roles in the enteroinsular axis,
glucose concentration, however, is the dominant factor control- mainly through activation of adenylate cyclase and subsequent
ling β-cell function. increase in intracellular cyclic adenosine monophosphate
Incretin
Acetylcholine GIP/GLP-1
Glucose
GLUT-1 M3
GLUT-2 M1 †Gαs
GLUT-3 Gαq*
Glucokinase PIP2
Glucose-6-P Phospholipase-C Adenylate
Cyclase
Pyruvate
L-Arginine CAT
IP3 DAG cAMP
KREBS
CYCLE Protein Kinase C Protein Kinase A
ATP-dependent Close Intracellular

↑ATP/ADP Ca2+ store
K+ channel Protein
Phosphorylation
Membrane
depolarization
Voltage-gated
Open ↑Cytosolic
Ca2+ channel
Ca2+ Insulin
Insulin
FIGURE 4.1. Stimulus-secretion coupling for insulin secretion. Insulin is secreted following food ingestion; glucose, acetylcholine (ACh), incretins,
and amino acids are the most important physiologic secretagogues. Glucose enters the human β cell mainly via the glucose transporters GLUT-1
and GLUT-3 and, to a lesser extent, via GLUT-2. Once in the β cell, glucose is phosphorylated to glucose-6-phosphate by an intracellular glucokinase
(GCK) and eventually is metabolized to produce adenosine triphosphate (ATP), leading to an elevation of the cytosolic ATP/diphosphate (ADP) ratio.
The increase in the intracellular ATP content is responsible for the closure of the ATP-dependent K+ channel. The resulting increased membrane
potential, caused by the closure of the ATP-dependent K+ channel, prompts the opening of voltage-dependent Ca2+ channels, leading to an increase
in intracellular Ca2+ levels. Arginine is a positively charged amino acid that leads to depolarization of the β cell following its uptake by a cationic
amino-acid transporter (CAT); the membrane depolarization leads to opening of voltage-dependent Ca2+ channels, allowing entry of Ca2+ into the
cell. The increase in intracellular Ca2+ triggers exocytosis of insulin-containing secretory granules. ACh, released from vagal efferent and α cell, binds
to the muscarinic ACh receptor M3, which is coupled with a phospholipase C, resulting in the production of inositol-1,4,5-triphosphate (IP3) and
diacylglycerol (DAG). Eventually, IP3 triggers the release of intracellular Ca2+, and DAG causes activation of protein kinase C (PKC). The incretins
(gastric inhibitory peptide [GIP] and glucagon-like peptide-1 [GLP-1]) bind to the extracellular domain of a G protein-coupled receptor to mediate
signal transduction by activation of adenylate cyclase, which acts to elevate intracellular cyclic adenosine monophosphate (cAMP) concentrations
and leads to activation of protein kinase A. Ultimately, the activation of protein kinase C and protein kinase A leads to protein phosporylation and
secretion of insulin. *Gαq, Membrane-associated heterotrimeric G protein that activates phospholipase C (PLC); †Gαs, membrane-associated het-
erotrimeric G protein that activates the cAMP-dependent pathway by activating adenylate cyclase; PIP2, phosphatidylinositol-4,5-bisphosphate.
(cAMP). The increase in cAMP leads to activation of protein increase in inositol-1,4,5-triphosphate (IP3) (Ruiz de Azua
kinase A, with subsequent phosphorylation and activation of et al, 2011). Additionally, ACh stimulates the somatostatin-
exocytosis-related proteins. Furthermore, cAMP activates secreting δ cell via M1 receptors. As somatostatin is known to
L-type calcium channels, culminating in insulin release (see inhibit insulin secretion, it appears that endogenous cholinergic
earlier). signaling provides a direct stimulatory and indirect inhibitory
Acetylcholine plays a pivotal role in glucose homeostasis. input to β cells, allowing further regulation of insulin secretion
Recently, it has been reported that in human islets, ACh acts (Molina et al, 2014).
primarily as a nonneuronal paracrine signal released from α cells
rather than as a neural signal, as previously described in rodent GLUCAGON AND OTHER ISLET HORMONES
islets (Rodriguez-Diaz, et al, 2011b). In fact, ACh stimulates
the insulin-secreting β cell via the muscarinic ACh receptors The islets of Langerhans secrete many additional hormones,
M3 and M5, causing release of insulin (Molina et al, 2014). including glucagon (α cells), somatostatin (δ cells), pancreatic
The activation of the M3 receptor leads to the release of calcium polypeptide (F cells), and ghrelin (ε cells) (Jain & Lammert,
from intracellular stores through a phospholipase C–mediated 2009). Furthermore, islets contain a variable, but small, number
of cells responsible for secreting pancreastatin, serotonin, and pancreatitis (Ito et al, 2007; Levy et al, 2006) (see Chapters 56
vasointestinal polypeptide (VIP) (Ohta et al, 2011; Sanlioglu to 58).
et al, 2012; Valicherla et al, 2013).Glucagon, a 29–amino-acid
peptide (molecular weight, 3.5 kDa), counteracts the effect of EXOCRINE PANCREAS
insulin by increasing blood glucose concentration through stim-
ulation of glycogenolysis, gluconeogenesis, and ketogenesis The exocrine pancreas constitutes 80% to 90% of the gland
(Cryer, 2012). Secretion from islet α cells directly into the mass and secretes the majority of digestive enzymes, as well as
portal system is largely in response to protein ingestion. Glu- approximately 2000 mL of colorless, odorless, and isosmotic
cagon at physiologic concentration exerts its function primarily alkaline protein-rich fluid (pH, 7.6-9.0) daily. It is mainly regu-
in liver tissue through activation of cAMP pathways, where lated by the neuroendocrine system, and it is integrated ana-
it promotes gluconeogenesis and indirectly ketogenesis tomically and physiologically with the endocrine pancreas,
(Ramnanan et al, 2011). In addition, glucagon indirectly stim- which helps modulate its function (Barreto et al, 2010).
ulates fatty oxidation through the carnitine acylcarnitine trans-
locase system (CAT). This leads to an increase in the ketone EXOCRINE PANCREAS STRUCTURE
bodies β-hydroxybutyric acid and acetoacetic acid, which exit
the liver to be used by other tissues as metabolic fuel. Glucagon The exocrine pancreas consists primarily of two distinct but
inhibition is caused by increased blood glucose concentration integrated units: the acinus and the ductal network.
of glucose as well as through paracrine effects of insulin and
somatostatin within the islet (Gylfe & Gilon, 2014). Acinus
Somatostatin, which acts primarily as an inhibitory hormone, The acinus is a functional unit mainly dedicated to the produc-
is secreted by the islet δ cell in addition to several other organs, tion and secretion of digestive enzymes (6 as much as 20 g
including the hypothalamus and the D cells of the gastrointesti- daily). It is composed of 15 to 100 pyramidal-shaped cells (as
nal tract. Among others, somatostatin inhibits insulin, glucagon, much as 30 µm apical base height) known as acinar cells
gastrin, and VIP. Its broad inhibition makes somatostatin and its (Cleveland et al, 2012). The acinar cells are polarized epithelial
pharmacologic analogues (e.g., octreotide) useful therapeutic cells rich in rough endoplasmic reticulum and characterized by
agents in the medical management of secreting pancreatic neu- an abundance of secretory zymogen granules within the apex.
roendocrine tumors (e.g., insulinoma) along with other medical The acinus is organized concentrically around a central lumen
diseases (e.g., Cushing disease, acromegaly, carcinoid, etc.) that is in continuity with the proximal end of an intercalated
(Lamberts et al, 1996) (see Chapter 65). Furthermore, soma- duct, where it drains its secretions. Several acini form a pancre-
tostatin analogues are used in treatment of some surgical com- atic lobule, which is separated from other lobules by thin layers
plications. Recently, pasireotide, a multisomatostatin receptor of connective tissue.
ligand, significantly decreased pancreatic leak complications fol-
lowing pancreatic surgery (Allen et al, 2014) (see Chapter 27). The Ductal Network
Pancreatic polypeptide (PP) is produced and released by The ductal network serves two critical functions: transport of
islet F cells; however, its physiologic role remains under inves- exocrine secretions from the acini to the duodenum and pro-
tigation (Holzer et al, 2012). Some studies suggest that absence duction of a solution rich in bicarbonate and electrolytes (Fig.
of PP secretion, resulting from removal of the uncinate process 4.2). The bicarbonate and water released in the ductal network
(rich in islet F cells) during pancreaticoduodenectomy, can lead facilitate the transport and flushing of acinar secretions through-
to pancreatogenic diabetes (Maeda & Hanazaki, 2011). out the pancreatic ducts and, most important, optimize the pH
Ghrelin, produced by the ε cells, also known as the “hunger of the solution in which pancreatic enzymes are secreted (Hegyi
hormone,” is a centrally active neuropeptide that participates & Petersen, 2013).
in metabolic regulation, growth hormone release, and energy The ductal network starts with small intercalated ducts
balance (Wierup et al, 2014; Heppner & Tong, 2014). originating from different acini, which then join to form an
intralobular duct. This serves to drain an individual pancreatic
PANCREATITIS CONSEQUENCES ON lobule. Intralobular ducts then drain into larger interlobular
ENDOCRINE PANCREAS FUNCTION ducts, which then empty into the main pancreatic duct, releas-
ing the pancreatic secretions into the duodenum, through the
Pancreatic endocrine insufficiency is a dreaded consequence of ampulla of Vater (Pandiri, 2014; Ashizawa et al, 2005).
acute pancreatitis. Recent studies suggest that following hospi- The ductal network is composed of highly specialized epi-
talization for the first episode of acute pancreatitis there is as thelial cells with varying morphologies and functions (Cleve-
much as a 40% risk of prediabetes and/or diabetes mellitus land et al, 2012). Cells of the intercalated duct are characterized
(Das et al, 2014). Approximately 15% of newly diagnosed dia- by minimal cytoplasm and by a squamous-shaped appearance.
betes mellitus occurs within 12 months from the first episodes On the contrary, cells in the main pancreatic duct have an
of acute pancreatitis, and the risk remains high, increasing abundance of cytoplasm rich in mitochondria and are charac-
significantly with time. It appears that development of endo- terized by a cuboidal shape.
crine insufficiency is independent from the severity of the Ductal cells contain substantial levels of cytoplasmic car-
episode of acute pancreatitis (Das et al, 2014). This suggests bonic anhydrase, an enzyme necessary for bicarbonate produc-
that the mechanisms that lead to endocrine insufficiency might tion (Ishiguro et al, 2012).
be partially unrelated to the amount of pancreatic necrosis.
Alternative mechanisms have been proposed and are currently Centroacinar Cells
under investigation. It is worth mentioning that the risk for Cuboidal-shaped centroacinar cells are present at the junction
diabetes mellitus can be as high as 80% in patients with chronic between the acinus and the ductal cells of the intercalated
A B C D
Centroacinar
cell
Intralobular
Intercalated duct
duct
Small
interlobular duct
Main/large
interlobular duct
Acinar Ductal
FIGURE 4.2. Anatomic organization of the pancreatic ductal network. A, Centroacinar cells are present at the junction between the intercalated
duct and the acinar cell (white arrowhead; scale bar: 10 µm). B, Intercalated duct (also known as terminal duct) originates at the level of the acini
and are composed of cell characterized by minimal cytoplasm and by squamous-shaped appearance. The intercalated ducts merge into intralobular
ducts (white arrowhead) that are lined by cuboidal epithelia and serves to drain an individual pancreatic lobule (scale bar: 10 µm). C, Intralobular
ducts join to form small interlobular ducts that eventually merge into (D) larger interlobular ducts that are lined by cuboidal epithelium (scale bar: 10
µm). Exocrine pancreatic secretions ultimately reach the main pancreatic duct and are released in the duodenum through the ampulla of Vater (not
shown). (Adapted from Reichert M, Rustgi AK: Pancreatic ductal cells in development, regeneration, and neoplasia, J Clin Invest 121:4572–
4578, 2011.)
duct. Recognized as morphologically distinct from acinar cells, compartment could contribute, at least in part, to the develop-
centroacinar cells are smaller than acinar cells (≈10 µm in ment of mucinous metaplasia and pancreatic intraepithelial
diameter), have a high nuclear-to-cytoplasm ratio, and have neoplasia (Strobel et al, 2010) (see Chapter 9B).
long cytoplasmic processes that allow contact with other cells
(i.e., centroacinar, acinar, and islet cells) (Cleveland et al,
2012; Leeson & Leeson, 1986; Pour, 1994). The role of the NEUROHORMONAL REGULATION OF EXOCRINE
centroacinar cells remains under investigation, but some PANCREATIC FUNCTION
authors suggest that centroacinar cells could represent progeni-
tor multipotent pancreatic cells and potentially be involved in
Digestive and Interdigestive Periods
malignant transformation (Seymour et al, 2007; Stanger et al, of Pancreatic Secretion
2005). Pancreatic exocrine secretion can be temporally categorized
into an interdigestive and digestive secretion period. Between
Ductal Epithelial Compartment meals, the intestinal migrating myoelectric complex (MMC) is
Ductal epithelial compartments (also known as pancreatic duct responsible for the cyclic stimulation of the exocrine pancreas
glands) are distributed along the pancreatic ducts and resemble (interdigestive period) (Zimmerman et al, 1992). The cholin-
blind outpouches or small branches originating from the pan- ergic stimuli that regulate the MMC cause cyclic pancreatic
creatic ducts. The cell lining of these outpouches consists of secretion of a fluid rich in bicarbonate (every 60 to 120 minutes).
columnar-shaped cells, characterized by abundant supranu- These secretions appear to facilitate the removal of bacteria and
clear cytoplasm and basally located nuclei (Cleveland et al, food debris from the small intestine; however, definitive agree-
2012). Their physiologic function remains under investigation; ment of their function has not been reached (Pandol, 2010).
however there is evidence to suggest that these cells undergo The digestive secretion period begins following the ingestion
selective expansion during chronic epithelial injuries. Some of a meal bolus and is characterized by three distinct phases:
authors have hypothesized that cells in the ductal epithelial cephalic, gastric, and intestinal. Each phase is determined by
the location of the meal bolus in the gastrointestinal tract, and Evidence suggests that muscarinic receptors (M1 and M3)
it is regulated by different secretory signals (Pandol, 2010). are predominantly expressed on acinar cells and are involved
The cephalic phase, elicited by the anticipation of food, the in regulating exocrine function, making ACh the principal neu-
smell, the taste, and the act of chewing, stimulates exocrine rotransmitter (Nakamura et al, 2013). This is supported by
secretion via the vagus nerve (Power & Schulkin, 2008). In this studies examining the mechanism of action of CCK. During
initial phase, pancreatic secretions are mainly composed of the intestinal phase of pancreatic secretion, the I cell of the
digestive enzymes and low levels of bicarbonate, indicating duodenum (also present in the jejunum), releases CCK primar-
principally an acinar-type secretion. ACh is the primary neu- ily in response to products of the digestion of fat, protein, and,
rotransmitter, although acinar cells have been shown to express to a lesser extent, starch. CCK then causes the release of diges-
G protein–coupled receptors (GPCRs) for gastrin-releasing tive enzymes and bile from the pancreas and gallbladder,
peptide (GRP) and VIP, suggesting a role for these peptides in respectively. Although only low levels of CCK receptor protein
the cephalic phase (Anagnostides et al, 1984; Holst et al, 1983). have been identified on human acinar cells, intrapancreatic
The gastric phase is initiated once the meal reaches the vagal nerve terminals express CCK receptors (Miyasaka et al,
stomach and is stimulated by gastric distention (Kreiss et al, 2002; Niebergall-Roth & Singer, 2006; Singer & Niebergall-
1996). A low volume of enzyme-rich secretion characterizes Roth, 2009). These receptors bind CCK and release ACh in
this phase, and it is accompanied by minimal water and bicar- the proximity of acinar cells. This suggests that the cholinergic
bonate secretion. pathway regulates exocrine function. Additional studies support
The intestinal phase begins with the entry of chyme and this hypothesis as the effect of CCK on pancreatic exocrine
gastric acid juice into the duodenum. In this phase, one of the secretion can be prevented and almost abolished by atropine
main regulatory mechanisms is the vasovagal enteropancreatic administration (Li et al, 1997; Mussa & Verberne, 2013;
reflex that is mediated through the dorsal vagus center. The Owyang, 1996; Soudah et al, 1992).
presence of chyme in the duodenum activates efferent fibers of Some ingested nutrients exert direct or indirect regulatory
enteric neurons that stimulate intrapancreatic postganglionic effects on pancreatic cells. Amino acids, especially phenylala-
neurons to release ACh. In addition, the presence of hydrogen nine, valine, methionine, and tryptophan are potent stimulants
ions (pH, ≤ 4.5) stimulates duodenal S cells to release secretin of exocrine pancreatic secretion. Furthermore, intraluminal
into the bloodstream. Secretin acts primarily on ductal cells fatty acids, monoglycerides, and, to a lesser extent, glucose
through its receptor GPCR, causing release of fluid and bicar- stimulate the secretion of digestive enzymes during the intesti-
bonate and, to a lesser extent, acinar secretion (Chey & Chang, nal phase (Pandol, 2010).
2014).
Feedback Inhibitory Regulation
Water, Bicarbonate, and Ion Secretion From the Feedback regulation to the exocrine pancreas is provided by
Ductal Network signals originating in the proximal and distal intestine. The
The role of the ductal network is of paramount importance for main factors identified are monitor peptide, luminal CCK-
the optimal function of the exocrine pancreas. The cells of the releasing factor (LCRF), secretin-releasing factors (e.g., phos-
ductal network are primarily under the control of secretin, pholipase A2), and peptide tyrosine tyrosine (PYY) (Barreto
which stimulates the centroacinar and ductal cells to release et al, 2010; Li et al, 2001; Liddle, 1995; Moran & Kinzig,
water and bicarbonate (Chey & Chang, 2014). These secre- 2004; Naruse et al, 2002). The action of these enzymes is
tions act as a vehicle for the transport of inactive digestive dependent on the presence or the absence of intraluminal
zymogens from the acinar cells to the duodenum. Furthermore, trypsin. Evidence suggests that trypsin inactivates monitor
their alkaline nature (pH, 7.6 to 9.0) helps neutralize the acidic peptide and LCRF, therefore preventing augmentation of CCK
chyme, in which nutrients are delivered from the stomach, release from the I cell. Once trypsin is occupied by the presence
resulting in an optimal neutral pH for the action of digestive of meal chyme, monitor peptide and LCRF are not digested
enzymes. The concentration of sodium bicarbonate (NaHCO3) and can augment CCK release from the I cell. Eventually, the
in pancreatic secretions can reach up to 140 to 150 mM, and excess of digestive protease in the duodenal lumen leads to
chloride secretion varies inversely to bicarbonate concentration, digestion of both monitor peptide and LCRF, preventing
keeping [HCO3−] + [Cl−] constant at approximately 160 mM further pancreatic enzyme secretion (Pandol, 2010).
(Ishiguro et al, 2012). Neuroendocrine L cells present in ileum and colon are stim-
Secretin also promotes an increase of blood flow to the ulated by the presence of intraluminal oleic acid to release PYY.
entire organ (Chey & Chang, 2014). During a meal, the blood This is a centrally active neuropeptide that exerts its action on
flow to the pancreas increases as much as 4 mL/min from a the area postrema of the brain, decreasing vagal cholinergic
baseline of 0.2 or 0.3 mL/min (Pandol, 2010). mediation of CCK-stimulated pancreatic secretion (Lin &
Taylor, 2004; Mussa & Verberne, 2013).
Regulation of Exocrine Secretion
Pancreatic secretions are tightly regulated through an intricate Digestive Enzymes
network of neural, humoral, and paracrine mediators. Many The exocrine pancreas releases proteolytic, amyolytic, lipolytic,
neurotransmitters, hormones, and growth factors have been and nuclease digestive enzymes. These enzymes are stored in
reported to influence pancreatic exocrine function. These zymogen granules either as proenzymes (i.e., trypsinogen, chy-
agents include, but are not limited to, ACh and catecholamine, motrypsinogen, procarboxypeptidase, prophospholipase, pro-
secretin (as earlier), nitric oxide (NO), VIP, GRP, neuropeptide elastase, mesotrypsin) or as active enzyme (i.e., a-amylase,
Y, galanin, substance P, CGRP, gastrin/cholecystokinin (CCK), lipase, DNase, RNase) (Whitcomb & Lowe, 2007). In addition,
and enkephalins (Barreto et al, 2010; Chandra & Liddle, 2012; the zymogen granules contain a trypsin inhibitor molecule
Chandra & Liddle, 2014). known as pancreatic secretory trypsin inhibitor (PSTI) (Kazal
ACINAR LUMEN
Gα q* CENTROACINAR CELL
ACh
IP3
CCK Gα q
Intracellular
PIP2
Phospholipase-C Ca2+ store ACINAR CELL
SUBSTANCE P Gα q
DAG
↑ Cytosolic
Ca2+
BOMBESIN Gαq
Secretion
Protein Kinase C
Protein
Phosphorylation
Protein Kinase A
SECRETIN †Gas
Adenylate cAMP
Cyclase
VIP Gα s
CENTROACINAR CELL
FIGURE 4.3. Stimulus-secretion coupling in pancreatic acinar cell. The primary secretagogues for acinar secretions bind to two types of cell
surface heterotrimeric G protein receptors. Acetylcholine (ACh; the primary secretagogue for acinar secretion), cholecystokinin (CCK), substance P,
and bombesin act through heterotrimeric G protein–coupled receptors (GPCRs) associated with phospholipase C. This results in the production of
two main messengers: inositol-1,4,5-triphosphate (IP3), which promotes the release of Ca2+ from intracellular stores; diacylglycerol (DAG), which
prompts the activation of protein kinase C (PKC). The increase in cytoplasmic Ca2+ concentration and activated PKC leads to protein phosphorylation
and digestive enzyme secretion. Secretin and vasoactive intestinal peptide (VIP) act through heterotrimeric GPCRs associated with adenylate cyclase.
Elevation in cytoplasmic cyclic adenosine monophosphate (cAMP) levels cause activation of protein kinase A (PKA), which ultimately leads to protein
phosphorylation and digestive enzyme secretion. The centroacinar cells are terminal ductal cells that are in close contact with acinar cells and could
represent progenitor multipotent pancreatic cells. *Gαq, Membrane-associated heterotrimeric G protein that activates phospholipase C (PLC);
†Gαs, membrane-associated heterotrimeric G protein that activates the cAMP-dependent pathway by activating adenylate cyclase; PIP2,
phosphatidylinositol-4,5-bisphosphate.
et al, 1948). PTSI forms a stable complex with trypsin near its adaptation remain under investigation, it is reasonable to
catalytic site, preventing its undesired activation (Pubols et al, believe that the regulation occurs at the level of gene
1974). transcription.
Following acinar stimulation, the zymogen granules fuse
with the apical acinar cell membrane, releasing their contents Stimulus-Secretion Coupling in Acinar Cell
in the pancreatic intercalated ducts that will eventually reach Increase in intracellular concentration of calcium is the funda-
the intestinal lumen. A brush-border glycoprotein peptidase mental event that stimulates the acinar cells to release the
present on the duodenal lumen, known as enterokinase, activates secretory granules containing the digestive enzymes (Fig. 4.3)
trypsinogen by removing its N-terminal hexapeptide fragment. (Low et al, 2010). Transmembrane heterotrimeric G proteins
The active form of trypsin is responsible for the catalytic activa- are the principal receptors for pancreatic acinar cell secreta-
tion of the remaining pancreatic proenzymes (Whitcomb & gogues and produce secondary messengers that ultimately act
Lowe, 2007). to release intracellular calcium (Messenger et al, 2014). Secre-
One characteristic of the acinar cell is its capacity to adapt tin and VIP bind to their specific GPCRs and cause activation
synthesis of digestive enzymes as a function of diet. Although of adenylate cyclase, generation of cAMP, and activation of
the mechanisms by which acinar cells are capable of this protein kinase A (Chey & Chang, 2014).
Pancreatic Ductal Lumen

+ – –
Na H2O Cl Cl HCO3–
SLC26A3/A6
CFTR
HCO3–
Cl–
Protein Kinase A CO2 + H2O
Carbonic Anhydrase
cAMP
H+ + HCO3–
Adenylate
Cyclase
H+ Na+ Na+ K+
Gαs*
ATP
Na+ H2O CO2 Interstitial Fluid

H+ Na+ Na+ K+
SECRETIN Capillary
FIGURE 4.4. Stimulus-secretion coupling in pancreatic ductal cell. Carbon dioxide (CO2) diffuses from the blood across the basolateral mem-
brane of the duct cell and is hydrated by carbonic anhydrase within the duct cell to form carbonic acid (H2CO3). Eventually, carbonic acid dissociates
into H+ and HCO3−. The extrusion of H+ via Na+-H+ exchanger, located across the basolateral membrane, leads to the accumulation of bicarbonate
(HCO3−) in the ductal cell. The intracellular bicarbonate (HCO3−) is then secreted into the ductal lumen by a chloride/bicarbonate (Cl−/HCO3−) exchanger
(SLC26A3/A6) in exchange for luminal Cl−. The cystic fibrosis transmembrane conductance regulator (CFTR) channel recycles Cl− back into the ductal
lumen, making it available for a new exchange. Secretin, the most important ductal cell secretagogue, binds to its heterotrimeric G protein–coupled
receptors, associated with adenylate cyclase located in the basolateral membrane. The resulting increase in cyclic adenosine monophosphate (cAMP)
leads to activation of protein kinase A (PKA) resulting in the activation of the CFTR channel. The activated CFTR channel accelerates the extrusion
of Cl− from the cell to the ductal lumen, causing the apical Cl−/ HCO3− to operate at a faster rate, leading to an increase in HCO3− secretion into the
ductal lumen. The net passage of bicarbonate across the duct cell generates an ionic and osmotic gradient, which favors paracellular passage of
sodium and water into the ductal space. ATP, Adenosine triphosphate; *Gαs, membrane-associated heterotrimeric G protein that activates the
cAMP-dependent pathway by activating adenylate cyclase.
CCK, ACh, bombesine/GRP, PAR-2–activating protease, the cellular receptor causing activation of protein kinase C and
and substance P bind to their respective GPCRs, which ulti- increase in intracellular calcium concentration. The end result is
mately act through the activation of phospholipase C and the the activation of cAMP-dependent Cl− channel that releases
production of IP3 (Singer & Niebergall-Roth, 2009; Weiss et al, intracellular Cl− into the ductal space, increasing the amount of
2008). IP3 binds act on its receptor in the endoplasmic reticu- Cl− available for the Cl−-HCO3− exchanger.
lum and stimulates Ca+2 release. Ultimately, activated protein The high concentration of ductal Cl− activates a Cl−-HCO3−
kinase A and protein kinase C phosphorylate specific intracel- antiport that results in an exchange of Cl− for HCO3−. Further-
lular proteins that lead to the secretion of pancreatic enzymes. more, Na+ and H2O are released into the ductal space drowned
by the ionic and osmotic gradient generated by the presence of
Stimulus-Secretion Coupling in the Ductal Cell bicarbonate in the duct lumen. Several K+ channels are present
The apical and the basolateral portion of the ductal cells are in the ductal cell membrane and appear to have an essential
involved in the stimulus-coupling secretion (Fig. 4.4). The apical role in generating and maintaining the electrochemical driving
membrane of ductal cells is equipped with cAMP-activated force for anion secretion. Ductal HCO3− secretion is not only
Cl−channels (also known as the cystic fibrosis transmembrane regulated by gastrointestinal hormones and cholinergic nerves
conductance regulator), Cl−-HCO3− exchanger (SLC26A3/A6), but is also influenced by luminal factors: intraductal pressure,
and water channels aquaporin 5 (AQP5). The basolateral mem- calcium concentration, and pathologic activation of protease
brane is rich in a Na+-H+ exchanger, N+,K+-ATPase, K+ conduc- and bile reflux (Ishiguro et al, 2012; Steward & Ishiguro 2009).
tance channels, and AQP1 (localized in both apical and
basolateral membrane). The interstitial portion of the ductal cells FUNCTIONAL ASSESSMENT
express receptors for the two major stimulants of ductal secre-
tion: secretin and ACh (Ishiguro et al, 2012; Steward & Ishiguro, Assessment of Endocrine Function
2009). Secretin binds to its cellular receptor, leading to the acti- The evaluation of the endocrine pancreas revolves around the
vation of adenylate cyclase and protein kinase A. ACh binds to assessment of β-cell function. These assessments should take
into account the capacity of the β cell to produce insulin, the

TABLE 4.1 Tests of Exocrine Pancreatic Function
response of the β cells to secretagogue stimuli, and the periph-
eral tissue resistance to insulin (characteristic of T2DM). Indirect Tests Direct Tests
Although each of these factors is important for accurate assess- Fecal fat quantification Noninvasive Tests
ment of β-cell function, their evaluation under physiologic con-
Qualitative (Sudan stain) Serum trypsinogen
ditions can be cumbersome. Among the challenges is the
Quantitative (72-hour stool Fecal chymotrypsin
nonlinear relationship between insulin secretion and insulin collection)
sensitivity, the nonconstant clearance of circulating insulin and
Pancreolauryl test Fecal elastase-1
hepatic extraction, and the dynamic insulin response to secre- 13
C-Mixed triglyceride
tagogues stimuli. Several test “models” have been developed in breath test
an attempt to account for the wide range of factors that influ-
Invasive Tests
ence β-cell function; however, many are difficult to execute and
Lundh test
lack standardization or accuracy (Cersosimo et al, 2014;
Antuna-Puente et al, 2011). Secretin test
Methods for evaluation of β-cell function include basal mea- CCK test
surement of plasma concentration of β-cell products, intrave- Secretin-CCK test
nous stimulation tests, and oral stimulation tests. Secretagogues and Imaging
The simplest methods of evaluation of β-cell function Secretin-enhanced MRI
require basal measurement of plasma concentration of fasting Secretin-enhanced MRCP
insulin, fasting C-peptide, fasting proinsulin/insulin ratio, and
CCK, cholecystokinin; MRCP, magnetic resonance cholangiopancreatography; MRI, magnetic
the homeostatic model assessment test. However, these tests resonance imaging.
often lack standardization and are hindered by their wide range
of sensitivity and specificity (Cersosimo et al, 2014).
The intravenous stimulation tests are mainly used for grade cholangiopancreatography [ERCP]) can easily identify
research purposes, and their clinical application is limited. advanced stages of pancreatic inflammation. The mild or
These tests include the intravenous tolerance test, the hyper- the initial forms of chronic pancreatitis are characterized by
glycemic glucose clamp, the graded glucose infusion, and the minimal anatomic changes and represent a diagnostic challenge
arginine stimulation test (Cersosimo et al, 2014). (Busireddy et al, 2014). It is in the latter scenario that pancre-
The oral stimulation tests are more commonly used in clini- atic function tests would find their best application.
cal practice and include the oral glucose tolerance test (OGTT) Pancreatic function tests (PFTs) can be grouped in two
and the mixed meal tolerance test (MTT). The OGTT and the major categories: indirect tests and direct tests (Table 4.1).
MTT provide a more physiologic stimulus to insulin secretion, Indirect PFTs are noninvasive tests that focus on the evalu-
as these tests elicit the full incretin effect (Elrick et al, 1964); ation of the consequences resulting from diminished or absent
furthermore, both tests take advantage of mathematical models digestive enzymes (e.g., steatorrhea).
that normalize insulin secretion levels to varying plasma glucose The aim of direct PFTs is to quantify and characterize pan-
concentrations (Cersosimo et al, 2014). The OGTT is widely creatic secretory content (i.e., digestive enzyme, bicarbonate,
used in clinical practice; after an overnight fast, subjects are and secretion volume); some tests require pancreatic stimula-
administered an oral glucose load (≈75 g) (Cobelli et al, 2007). tion through administration of a meal or hormonal secreta-
Blood samples are collected at baseline and at subsequent inter- gogues (e.g., secretin, CCK). Direct PFTs are further divided
vals to evaluate plasma concentration of insulin, glucose, into invasive, requiring gastrointestinal instrumentation via a
C-peptide, and other parameters of interest. double-lumen gastroduodenal tube, and not invasive (often
Impaired glucose tolerance is defined by a 2-hour glucose characterized by a lower sensitivity). Endoscopic versions of
value during an oral glucose tolerance test (OGTT) of direct pancreatic tests have been developed (ePFT) in which
7.8 mmol/L to 11.0 mmol/L, whereas overt T2DM presents pancreatic fluid secretions are collected directly from the second
values greater than or equal to 11.1 mmol/L (American Diabe- portion of the duodenum under direct visualization with the
tes Association, 2013). One of the major limitations of this test use of an endoscope (Law et al, 2012; Tayler & Parsi, 2011).
is its dependency on the unpredictable and variable intestinal In addition, the use of endoscopic ultrasound allows for simul-
absorption of glucose. taneous structural evaluation of the pancreatic parenchyma
(Albashir et al, 2010). Several different protocols for each test
Assessment of Exocrine Function have been developed, but agreement on an absolute gold-
Exocrine pancreatic insufficiency in adults is often the result of standard test has not been reached.
pancreatic inflammatory processes (e.g., chronic pancreatitis)
(see Chapters 57 and 58) and often a clinical problem Indirect Pancreatic Test
after pancreatic or gastric surgery, and it leads to maldigestion Indirect pancreatic tests are noninvasive tests that are well toler-
of fat, protein, and carbohydrates (Nakamura, et al, 2009b). ated by patients, although characterized by lower sensitivity
Ultimately, this will result in steatorrhea (>7 g of fecal fat in compared with the direct tests, especially in the earlier stages
24 hours), weight loss, and eventually malnutrition. The human of pancreatic exocrine insufficiency. One of the least invasive
pancreatic gland has a significant functional reserve; therefore tests is the quantification of fecal fat. It can be qualitative
more than 90% of pancreatic parenchyma has to be lost (Sudan stain) or quantitative (72-hour stool collection) and
before overt steatorrhea becomes evident (Levy et al, 2014). often of use during the response evaluation to pancreatic
Although imaging studies (e.g., CT, magnetic resonance enzyme therapy (Lindkvist, 2013). In the most common form,
cholangiopancreatography [MRCP], and endoscopic retro- patients are required to assume a diet of 100 g/day fat for 5
days and to collect the complete volume of feces for 3 days, spot fecal sample) are diagnostic for pancreatic insufficiencies
starting on day 3. Fecal content greater than 7 g/day is consid- in patients with chronic pancreatitis (Benini et al, 2013).
ered abnormal, and it is diagnostic for steatorrhea. This test However, fecal elastase-1 tests are often unreliable following
poses an unpleasant burden on laboratory personnel, and its pancreatic resection, where pancreatic insufficiency results
diagnostic utility is limited to the advanced stages of pancreatic from a combination of factors not solely depended on decreased
insufficiency. pancreatic function (e.g., abnormal hormonal stimulation,
Previously considered one of the most accurate of the non- abnormal mixing of food with digestive secretions, acidic intra-
invasive tests, the pacreolauryl test has limited value in patients luminal pH) (Benini et al, 2013). Furthermore, fecal elastase-1
with bile salt deficiency, renal failure, celiac disease, and post- levels are not influenced by the exogenous administration of
gastrectomy symptoms, and its use has been abandoned in pancreatic enzyme, and therefore this test cannot be utilized to
clinical practice (Elphick & Kapur, 2005; Friess & Michalski, evaluate the response to pancreatic enzyme replacement therapy
2009; Lankisch et al, 1983). (Leeds et al, 2011; Nøjgaard et al, 2012).
Another test available to quantify fat malabsorbtion is the Invasive direct PFTs require the use of a double-lumen col-
13
C–mixed triglyceride breath test. This test requires the oral lection tube, with one lumen terminating in the duodenum to
administration of a 13C–market substrate and relies on the pres- collect pancreatic secretions and one lumen resting in the
ence of intestinal pancreatic lipase activity. Ultimately, the 13C– gastric antrum to prevent gastric fluid from entering the duo-
market substrate is hydrolyzed by the intestinal lipase, yielding denum. Correct tube placement is confirmed via fluoroscopy
13
CO2 that is absorbed and eventually released across the pul- prior to administration of meal or segretagogues, and then
monary endothelium and quantified via mass spectrometry or pancreatic secretions are collected for 1 to 2 hours. Alterna-
infrared analysis. This test could represent a practical alternatively, an endoscope can be used, and pancreatic fluid can be
tive to the fecal fat test, although it is subject to similar limita- collected directly either through cannulation of the pancreatic
tions of all indirect tests with limited accuracy for the early duct or through the suction channel of the endoscope under
phases of pancreatic insufficiency (Keller et al, 2014; Naka- direct visualization from the duodenum. Of historic interest is
mura et al, 2009a). the Lundh test (Lundh,1962), where a physiologic stimulus to
pancreatic secretion is obtained through the administration of
Direct Pancreatic Function Test a standardized meal (composed of 300 mL of solution contain-
Noninvasive direct PFTs aim to quantify the fecal or serum ing 15% carbohydrate, 6% fat, and 5% protein). Endogenous
levels of pancreas-derived enzymes (e.g., serum trypsinogen, secretin and CCK, released in response to the standardized
fecal chymotrypsin, and fecal elastase). Although easy to meal, are necessary to stimulate pancreatic secretion, making
perform and well tolerated by patients, noninvasive PFTs are the Lundh test dependent on normal duodenal mucosal func-
hindered by their low sensitivity, often leading to inconclusive tion (Lundh, 1962; Pandol, 2010). The hormone-stimulated
results, especially in the early phases of chronic pancreatitis. tests are the cornerstone of the direct invasive pancreatic tests.
Furthermore, fecal measurement of pancreatic enzyme can lead These tests utilize the exogenous administration of secretin,
to false-positive results in the setting of nonpancreatic gastro- CCK, or a combination of secretin-CCK as a more reliable
intestinal disturbances and diarrhea. method of pancreatic stimulation (Jowell et al, 2000; Somogyi
Serum trypsinogen is considered a sensitive and specific test et al, 2003).
for advanced pancreatic insufficiency, although its accuracy for The secretin test requires intravenous administration of a
earlier stages of pancreatic insufficiency is low. A serum tryp- synthetic secretin bolus (0.2 µg/kg), followed by a continuous
sinogen concentration of less than 20 ng/mL is considered a collection of duodenal fluid in four aliquots of 15-minute inter-
reasonable cutoff for the diagnosis of pancreatic insufficiency vals each. The test measures bicarbonate concentration in each
(Jacobson et al, 1984). of the four aliquot samples. Exocrine insufficiency is evident
Fecal chymotrypsin concentration can be utilized for the when bicarbonate concentrations less than 80 mEq/L are
evaluation of pancreatic insufficiency. Chymotrypsin is specifi- recorded in each of the four aliquots, and severe exocrine insuf-
cally synthesized and secreted by the acinar cells of the pan- ficiency is diagnosed when bicarbonate concentration is less
creas. This test is influenced by exogenous pancreatic enzyme than 50 mEq/L (Chowdhury & Forsmark, 2003). Moreover,
administration and requires suspension of enzyme administra- the test can quantify the total volume output as well as the total
tion for at least 2 days before the test (Molinari et al, 2004). amount of bicarbonate secreted. These measurements are
Furthermore, chymotrypsin is not an ideal marker as it is employed as a secondary diagnostic test when bicarbonate con-
degraded during intestinal transit and can be diluted in the centrations are equivocal. Although theoretically useful, these
presence of diarrhea, leading to false-positive results. tests are not commonly employed because they are hindered by
Fecal elastase-1 appears to be a more reliable test compared the low accuracy caused by incomplete recovery of duodenal
to fecal chymotrypsin. Elastase-1 is not influenced by exoge- fluid.
nous pancreatic enzyme administration, and it is more stable Alternative approaches have been developed to overcome
than chymotrypsin during intestinal transit. Evidence suggests the challenges of gastroduodenal tube placement. One of these
that this study has a reasonably high sensitivity and specificity approaches is the purely endoscopic secretin test (ePFT). This
for pancreatic insufficiency (Leeds et al, 2011). Results obtained test has been shown to have similar accuracy to the classic
with the fecal elastase-1 test reliably correlate with the one secretin test. Duodenal aspirates are obtained through an endo-
obtained using imaging studies (ERCP, magnetic resonance scope at 0, 15, 30, 45, and 60 minutes after secretin stimulation
neurography [MRN]) (Bilgin et al, 2008), or the more sensitive and analyzed for bicarbonate concentration (Stevens et al,
direct invasive pancreatic tests (secretin test) (Löser et al, 1996; 2007; Tayler & Parsi, 2011).
Stein et al, 1996). Level of fecal elastase-1 less than or equal Another attempt made to overcome the limitations of duo-
to 15 µg/g of stool (enzyme-linked immunoabsorbent assay on denal fluid collection consists of the use of ERCP for the
analysis of pure pancreatic secretion collected directly from the CCK-stimulated gallbladder contraction, ultimately results in
pancreatic duct. However, the results obtained with this method dilution of the digestive enzymes. To avoid false-positive results,
have been unsatisfactory (Draganov et al, 2005). CCK tests some authors have advocated the use of perfusion markers,
using either CCK or a receptor agonist (e.g., cerulein) have although no definitive agreement exists.
been developed to measure pancreatic enzyme secretion. A Of recent interest are tests combining secretagogues with
simplified version of these tests measures lipase concentration imaging techniques. Examples of these tests are the secretin-
from duodenal fluid collected over an 80-minute period and enhanced MRI and the secretin-enhanced MRCP (Balci et al,
utilizes a cutoff lipase value of 780 IU/L (Conwell et al, 2002). 2010; Bian et al, 2013; Hansen et al, 2013). Secretin-enhanced
Acinar and ductal pancreatic exocrine function can be evalu- MRI utilizes diffusion-weighted MRI imaging to evaluate
ated simultaneously using the secretin-CCK test. This test increase in pancreatic capillary blood flow and pancreatic secre-
requires continuous collection of duodenal fluid and measure- tion following stimulation with secretin. In addition, secretin-
ment of total secretion output, bicarbonate concentration, and enhanced MRCP allows evaluation of duodenal filling as a
digestive enzyme concentration (Lieb & Draganov, 2008). function of pancreatic secretion (Sanyal et al, 2012). Although
Some patients may have a more pronounced deficit of one growing enthusiasm is developing around these tests, further
specific enzyme; therefore measuring more than one digestive studies are necessary to evaluate their performances with stan-
enzyme (i.e., amylase, lipase, and tryptase) in addition to bicar- dard invasive tests.
bonate can increase the sensitivity of this test. One pitfall of the
secretin-CCK test is that the large amount of fluid released by References are available at expertconsult.com.
the pancreas following secretin stimulation, combined with
Chapter 4 Pancreatic Physiology and Functional Assessment 76.e1
Conwell DL, et al: Analysis of duodenal drainage fluid after cholecys-

REFERENCES tokinin (CCK) stimulation in healthy volunteers, Pancreas 25(4):350–
354, 2002.
Ahrén B: Incretin dysfunction in type 2 diabetes: clinical impact and Cryer PE: Minireview: glucagon in the pathogenesis of hypoglycemia
future perspectives, Diabetes Metab 39(3):195–201, 2013. and hyperglycemia in diabetes, Endocrinology 153(3):1039–1048,
Albashir S, et al: Endoscopic ultrasound, secretin endoscopic pancre- 2012.
atic function test, and histology: correlation in chronic pancreatitis, Das SLM, et al: Newly diagnosed diabetes mellitus after acute pancre-
Am J Gastroenterol 105(11):2498–2503, 2010. atitis: a systematic review and meta-analysis, Gut 63(5):818–831,
Allen PJ, et al: Pasireotide for postoperative pancreatic fistula, N Engl 2014.
J Med 370(21):2014–2022, 2014. Diab DL, D’Alessio DA: The contribution of enteroinsular hormones
American Diabetes Association: Standards of medical care in to the pathogenesis of type 2 diabetes mellitus, Curr Diabet Rep
diabetes—2013, Diabetes Care 36(Suppl 1):S11–S66, 2013. 10(3):192–198, 2010.
Anagnostides A, et al: Sham feeding and pancreatic secretion. Evidence Draganov P, et al: Prospective evaluation of the accuracy of the intra-
for direct vagal stimulation of enzyme output, Gastroenterology ductal secretin stimulation test in the diagnosis of chronic pancre-
87(1):109–114, 1984. atitis, Clin Gastroenterol Hepatol 3(7):695–699, 2005.
Antuna-Puente B, et al: How can we measure insulin sensitivity/ Drucker DJ: Incretin action in the pancreas: potential promise, possible
resistance?, Diabetes Metab 37(3):179–188, 2011. perils, and pathological pitfalls, Diabetes 62(10):3316–3323, 2013.
Ashizawa N, et al: Three-dimensional structure of peripheral exocrine Elphick DA, Kapur K: Comparing the urinary pancreolauryl ratio and
gland in rat pancreas: reconstruction using transmission electron faecal elastase-1 as indicators of pancreatic insufficiency in clinical
microscopic examination of serial sections, Pancreas 31(4):401–404, practice, Pancreatology 5(2–3):196–200, 2005.
2005. Elrick H, et al: Plasma insulin response to oral and intravenous
Balci NC, et al: MRI and S-MRCP findings in patients with suspected glucose administration, J Clin Endocrinol Metab 24(10):1076–1082,
chronic pancreatitis: correlation with endoscopic pancreatic function 1964.
testing (ePFT), JMRI 31(3):601–606, 2010. Friess H, Michalski CW: Diagnosing exocrine pancreatic insufficiency
Banting FG: An address on diabetes and insulin: being the Nobel after surgery: when and which patients to treat, HPB (Oxford)
Lecture delivered at Stockholm on September 15th, 1925, Can Med 11(Suppl 3):7–10, 2009.
Assoc J 16(3):221–232, 1926. Gray H: Lewis WH, editor: Gray’s anatomy of the human body, ed 20,
Barker CJ, et al: The pancreatic islet as a signaling hub, Adv Biol Regul New York, 2000, Bartleby.com.
53(1):156–163, 2013. Gylfe E, Gilon P: Glucose regulation of glucagon secretion, Diabet Res
Barreto SG, et al: The islet-acinar axis of the pancreas: more than just Clin Pract 103(1):1–10, 2014.
insulin, Am J Physiol 299(1):G10–G22, 2010. Hansen TM, et al: Morphological and functional evaluation of chronic
Benini L, et al: Fecal elastase-1 is useful in the detection of steatorrhea pancreatitis with magnetic resonance imaging, World J Gastroenterol
in patients with pancreatic diseases but not after pancreatic resec- 19(42):7241–7246, 2013.
tion, Pancreatology 13(1):38–42, 2013. Hegyi P, Petersen OH: The exocrine pancreas: the acinar-ductal tango
Bian Y, et al: Quantification of pancreatic exocrine function of chronic in physiology and pathophysiology, Rev Physiol Biochem Pharmacol
pancreatitis with secretin-enhanced MRCP, World J Gastroenterol 165:1–30, 2013.
19(41):7177–7182, 2013. Henquin JC: Regulation of insulin secretion: a matter of phase control
Bilgin M, et al: Magnetic resonance imaging and magnetic resonance and amplitude modulation, Diabetologia 52(5):739–751, 2009.
cholangiopancreatography findings compared with fecal elastase 1 Heppner KM, Tong J: Mechanisms in endocrinology: regulation of
measurement for the diagnosis of chronic pancreatitis, Pancreas glucose metabolism by the ghrelin system: multiple players and
36(1):e33–e39, 2008. multiple actions, Eur J Endocriol 171(1):R21–R32, 2014.
Bonner-Weir S: The microvasculature of the pancreas, with emphasis Holst JJ, et al: Interrelation of nerves and hormones in stomach and
on that of the islets of Langerhans. In Go V, et al, editors: The pancreas, Scand J Gastroenterol Suppl 82:85–99, 1983.
pancreas: biology, pathobiology, and disease, ed 2, New York, 1993, Holzer P, et al: Neuropeptide Y, peptide YY and pancreatic polypep-
Raven. tide in the gut-brain axis, Neuropeptides 46(6):261–274, 2012.
Boron WF, Boulpaep EL: Medical physiology, ed 2, Philadelphia, 2012, Huang M, Joseph JW: Assessment of the metabolic pathways associated
Saunders/Elsevier. with glucose-stimulated biphasic insulin secretion, Endocrinology
Busireddy KK, et al: Pancreatitis-imaging approach, World J Gastroin- 155(5):1653–1666, 2014.
test Pathophysiol 5(3):252–270, 2014. Ishiguro H, et al: Physiology and pathophysiology of bicarbonate secre-
Cabrera O, et al: The unique cytoarchitecture of human pancreatic islets tion by pancreatic duct epithelium, Nagoya J Med Sci 74(1–2):1–18,
has implications for islet cell function, Proc Natl Acad Sci U S A 2012.
103(7):2334–2339, 2006. Ito T, et al: Pancreatic diabetes in a follow-up survey of chronic pan-
Cano DA, et al: Transcriptional control of mammalian pancreas organ- creatitis in Japan, J Gastroenterol 42(4):291–297, 2007.
ogenesis, Cell Mol Life Sci 71(13):2383–2402, 2014. Jacobson DG, et al: Trypsin-like immunoreactivity as a test for pancre-
Cersosimo E, et al: Assessment of pancreatic β-cell function: review of atic insufficiency, N Engl J Med 310(20):1307–1309, 1984.
methods and clinical applications, Curr Diabet Rev 10(1):2–42, Jain R, Lammert E: Cell-cell interactions in the endocrine pancreas,
2014. Diabetes Obes Metab 11(Suppl 4):159–167, 2009.
Chandra R, Liddle RA: Neurohormonal regulation of pancreatic secre- Jones AG, Hattersley AT: The clinical utility of C-peptide measure-
tion, Curr Opin Gastroenterol 28(5):483–487, 2012. ment in the care of patients with diabetes, Diabet Med 30(7):803–
Chandra R, Liddle RA: Recent advances in the regulation of pancreatic 817, 2013.
secretion, Curr Opin Gastroenterol 30(5):490–494, 2014. Jowell PS, et al: A double-blind, randomized, dose response study
Chey WY, Chang T-M: Secretin: historical perspective and current testing the pharmacological efficacy of synthetic porcine secretin,
status, Pancreas 43(2):162–182, 2014. Aliment Pharmacol Ther 14(12):1679–1684, 2000.
Chowdhury RS, Forsmark CE: Review article: pancreatic function Kazal LA, et al: Isolation of a crystalline trypsin inhibitor-anticoagulant
testing, Aliment Pharmacol Ther 17(6):733–750, 2003. protein from pancreas, J Am Chem Soc 70(9):3034–3040, 1948.
Clemmensen C, et al: Oral L-arginine stimulates GLP-1 secretion to Keller J, et al: 13C-mixed triglyceride breath test for evaluation of
improve glucose tolerance in male mice, Endocrinology 154(11):3978– pancreatic exocrine function in diabetes mellitus, Pancreas 43(6):842–
3983, 2013. 848, 2014.
Cleveland MH, et al: Exocrine ontogenies: on the development of Kreiss C, et al: Role of antrum in regulation of pancreaticobiliary secre-
pancreatic acinar, ductal and centroacinar cells, Sem Cell Dev Biol tion in humans, Am J Physiol 270(5 Pt 1):G844–G851, 1996.
23(6):711–719, 2012. Lamberts SW, et al: Octreotide, N Engl J Med 334(4):246–254, 1996.
Cobelli C, et al: Assessment of beta-cell function in humans, simulta- Lankisch PG, et al: Pancreolauryl test. Evaluation of a tubeless pan-
neously with insulin sensitivity and hepatic extraction, from intrave- creatic function test in comparison with other indirect and direct
nous and oral glucose tests, Am J Physiol Endocrinol Metab 293(1): tests for exocrine pancreatic function, Dig Dis Sci 28(6):490–493,
E1–E15, 2007. 1983.
Law R, et al: Endoscopic pancreatic function test using combined Nakamura H, et al: Usefulness of a 13C-labeled mixed triglyceride
secretin and cholecystokinin stimulation for the evaluation of chronic breath test for assessing pancreatic exocrine function after pancreatic
pancreatitis, Gastrointest Endosc 75(4):764–768, 2012. surgery, Surgery 145(2):168–175, 2009a.
Leeds JS, et al: The role of fecal elastase-1 in detecting exocrine Nakamura H, et al: Predictive factors for exocrine pancreatic insuffi-
pancreatic disease, Nat Rev Gastroenterol Hepatol 8(7):405–415, ciency after pancreatoduodenectomy with pancreaticogastrostomy,
2011. J Gastrointest Surg 13(7):1321–1327, 2009b.
Leeson TS, Leeson R: Close association of centroacinar/ductular and Nakamura K, et al: Distinct roles of M1 and M3 muscarinic acetylcho-
insular cells in the rat pancreas, Histol Histopathol 1(1):33–42, 1986. line receptors controlling oscillatory and non-oscillatory [Ca2+]i
Lenzen S: A fresh view of glycolysis and glucokinase regulation: history increase, Cell Calcium 54(2):111–119, 2013.
and current status, J Biol Chem 289(18):12189–12194, 2014. Naruse S, et al: Feedback regulation of pancreatic secretion by peptide
Lévy P, et al: Estimation of the prevalence and incidence of chronic YY, Peptides 23(2):359–365, 2002.
pancreatitis and its complications, Gastroentérol Clin Biol 30(6– Niebergall-Roth E, Singer MV: Enteropancreatic reflexes mediating the
7):838–844, 2006. pancreatic enzyme response to nutrients, Auton Neurosci 125(1–
Lévy P, et al: Epidemiology of chronic pancreatitis: burden of the 2):62–69, 2006.
disease and consequences, United European Gastroenterol J 2(5):345– Nøjgaard C, et al: Update of exocrine functional diagnostics in chronic
354, 2014. pancreatitis, Clin Physiol Funct Imaging 33(3):167–172, 2012.
Li Y, Hao Y, et al: High-affinity CCK-A receptors on the vagus nerve Ohta Y, et al: Convergence of the insulin and serotonin programs in
mediate CCK-stimulated pancreatic secretion in rats, Am J Physiol the pancreatic β-cell, Diabetes 60(12):3208–3216, 2011.
273(3 Pt 1):G679–G685, 1997. Opinto G, et al: Physiology of incretins and loss of incretin effect in
Li JP, et al: Pancreatic phospholipase A2 from the small intestine is a type 2 diabetes and obesity, Arch Physiol Biochem 119(4):170–178,
secretin-releasing factor in rats, Am J Physiol Gastrointest Liver Physiol 2013.
281(2):G526–G532, 2001. Owyang C: Physiological mechanisms of cholecystokinin action on
Liddle RA: Regulation of cholecystokinin secretion by intraluminal pancreatic secretion, Am J Physiol 271(1 Pt 1):G1–G7, 1996.
releasing factors, Am J Physiol 269(3 Pt 1):G319–G327, 1995. Pan FC, Wright C: Pancreas organogenesis: from bud to plexus to
Lieb JG, Draganov PV: Pancreatic function testing: here to stay for the gland, Dev Dyn 240(3):530–565, 2011.
21st century, World J Gastroenterol 14(20):3149–3158, 2008. Pandiri AR: Overview of exocrine pancreatic pathobiology, Toxicol
Lin HC, Taylor IL: Release of peptide YY by fat in the proximal but Pathol 42(1):207–216, 2014.
not distal gut depends on an atropine-sensitive cholinergic pathway, Pandol SJ: The exocrine pancreas, San Rafael, CA, 2010, Morgan &
Regul Pept 117(1):73–76, 2004. Claypool Life Sciences.
Lindkvist B: Diagnosis and treatment of pancreatic exocrine insuffi- Pour PM: Pancreatic centroacinar cells. The regulator of both exocrine
ciency, World J Gastroenterol 19(42):7258–7266, 2013. and endocrine function, Int J Pancreatol 15(1):51–64, 1994.
Löser C, et al: Faecal elastase 1: a novel, highly sensitive, and specific Power ML, Schulkin J: Anticipatory physiological regulation in feeding
tubeless pancreatic function test, Gut 39(4):580–586, 1996. biology: cephalic phase responses, Appetite 50(2–3):194–206, 2008.
Low JT, et al: Pancreatic acinar cell: new insights into the control of Pubols MH, et al: Trypsin inhibitor from human pancreas and pancre-
secretion, Int J Biochem Cell Biol 42(10):1586–1589, 2010. atic juice, J Biol Chem 249(7):2235–2242, 1974.
Lundh G: Pancreatic exocrine function in neoplastic and inflammatory Ramnanan CJ, et al: Physiologic action of glucagon on liver glucose
disease; a simple and reliable new test, Gastroenterology 42:275–280, metabolism, Diabetes Obes Metab 13(Suppl 1):118–125, 2011.
1962. Rodriguez-Diaz R, et al: Innervation patterns of autonomic axons in
Maeda H, Hanazaki K: Pancreatogenic diabetes after pancreatic resec- the human endocrine pancreas, Cell Metab 14(1):45–54, 2011a.
tion, Pancreatology 11(2):268–276, 2011. Rodriguez-Diaz R, et al: Alpha cells secrete acetylcholine as a non-
Marques RG, et al: C-peptide: much more than a byproduct of insulin neuronal paracrine signal priming beta cell function in humans, Nat
biosynthesis, Pancreas 29(3):231–238, 2004. Med 17(7):888–892, 2011b.
Matschinsky FM, et al: Pancreatic beta-cell glucokinase: closing the Rorsman P, Braun M: Regulation of insulin secretion in human pan-
gap between theoretical concepts and experimental realities, Diabetes creatic islets, Ann Rev Physiol 75(1):155–179, 2013.
47(3):307–315, 1998. Rorsman P, et al: ATP-regulated potassium channels and voltage-gated
McCulloch LJ, et al: GLUT2 (SLC2A2) is not the principal glucose calcium channels in pancreatic alpha and beta cells: similar functions
transporter in human pancreatic beta cells: implications for under- but reciprocal effects on secretion, Diabetologia 57(9):1749–1761,
standing genetic association signals at this locus, Mol Genet Metab 2014.
104(4):648–653, 2011. Ruiz de Azua I, et al: Novel insights into the function of β-cell M3
Merkwitz C, et al: The ductal origin of structural and functional het- muscarinic acetylcholine receptors: therapeutic implications, Trends
erogeneity between pancreatic islets, Prog Histochem Cytochem Endocrinol Metab 22(2):74–80, 2011.
48(3):103–140, 2013. Rutter GA, Hodson DJ: Minireview: intraislet regulation of insulin
Messenger SW, et al: Ca2+-regulated secretory granule exocytosis in secretion in humans, Mol Endocrinol 27(12):1984–1995, 2013.
pancreatic and parotid acinar cells, Cell Calcium 55(6):369–375, Sanlioglu AD, et al: Therapeutic potential of VIP vs PACAP in diabe-
2014. tes, J Mol Endocrinol 49(3):R157–R167, 2012.
Miyasaka K, et al: Amylase secretion from dispersed human pancreatic Sanyal R, et al: Secretin-enhanced MRCP: review of technique and
acini: neither cholecystokinin a nor cholecystokinin B receptors application with proposal for quantification of exocrine function,
mediate amylase secretion in vitro, Pancreas 25(2):161–165, 2002. AJR Am J Roentgenol 198(1):124–132, 2012.
Molina J, et al: Control of insulin secretion by cholinergic signaling in Seymour PA, et al: SOX9 is required for maintenance of the pancreatic
the human pancreatic islet, Diabetes 63(8):2714–2726, 2014. progenitor cell pool, Proc Natl Acad Sci U S A 104(6):1865–1870,
Molinari I, et al: Fecal chymotrypsin and elastase-1 determination on 2007.
one single stool collected at random: diagnostic value for exocrine Singer MV, Niebergall-Roth E: Secretion from acinar cells of the exo-
pancreatic status, Clin Biochem 37(9):758–763, 2004. crine pancreas: role of enteropancreatic reflexes and cholecystokinin,
Moore KL, et al: Abdomen. In Moore KL, editor: Clinically oriented Cell Biol Int 33(1):1–9, 2009.
anatomy, Baltimore, 2013, LWW, pp 265–268. Smith PA, et al: Electrogenic arginine transport mediates stimulus-
Moran TH, Kinzig KP: Gastrointestinal satiety signals II. Cholecysto- secretion coupling in mouse pancreatic beta-cells, J Physiol 499(Pt
kinin, Am J Physiol Gastrointest Liver Physiol 286(2):G183–G188, 3):625–635, 1997.
2004. Somogyi L, et al: Comparison of biologic porcine secretin, synthetic
Mussa BM, Verberne AJM: The dorsal motor nucleus of the vagus and porcine secretin, and synthetic human secretin in pancreatic function
regulation of pancreatic secretory function, Exp Physiol 98(1):25–37, testing, Pancreas 27(3):230–234, 2003.
2013. Song SH, et al: Pulsatile insulin secretion by human pancreatic islets,
Nagamatsu S, et al: Imaging docking and fusion of insulin granules J Clin Endocrinol Metab 87(1):213–221, 2002.
induced by antidiabetes agents: sulfonylurea and glinide drugs pref- Soudah HC, et al: Cholecystokinin at physiological levels evokes pan-
erentially mediate the fusion of newcomer, but not previously creatic enzyme secretion via a cholinergic pathway, Am J Physiol
docked, insulin granules, Diabetes 55(10):2819–2825, 2006. 263(1 Pt 1):G102–G107, 1992.
Chapter 4 Pancreatic Physiology and Functional Assessment 76.e3
Stamper IJ, Wang X: Mathematical modeling of insulin secretion and Tayler S, Parsi MA: Update on endoscopic pancreatic function testing,
the role of glucose-dependent mobilization, docking, priming and World J Gastroenterol 17(35):3957–3961, 2011.
fusion of insulin granules, J Theor Biol 318:210–225, 2013. Tolhurst G, et al: Nutritional regulation of glucagon-like peptide-1
Stanger BZ, et al: Pten constrains centroacinar cell expansion and secretion, J Physiol 587(Pt 1):27–32, 2009.
malignant transformation in the pancreas, Cancer Cell 8(3):185–195, Valicherla GR, et al: Pancreastatin is an endogenous peptide that
2005. regulates glucose homeostasis, Physiol Genomics 45(22):1060–1071,
Stein J, et al: Immunoreactive elastase I: clinical evaluation of a new 2013.
noninvasive test of pancreatic function, Clin Chem 42(2):222–226, van de Bunt M, Gloyn AL: A tale of two glucose transporters: how
1996. GLUT2 re-emerged as a contender for glucose transport into the
Stevens T, et al: The efficiency of endoscopic pancreatic function human beta cell, Diabetologia 55(9):2312–2315, 2012.
testing is optimized using duodenal aspirates at 30 and 45 minutes Weiss FU, et al: New advances in pancreatic cell physiology and patho-
after intravenous secretin, Am J Gastroenterol 102(2):297–301, 2007. physiology, Best Pract Res Clin Gastroenterol 22(1):3–15, 2008.
Steward MC, Ishiguro H: Molecular and cellular regulation of pancre- Whitcomb DC, Lowe ME: Human pancreatic digestive enzymes, Dig
atic duct cell function, Curr Opin Gastroenterol 25(5):447–453, 2009. Dis Sci 52(1):1–17, 2007.
Strobel O, et al: Pancreatic duct glands are distinct ductal compart- Wierup N, et al: The islet ghrelin cell, J Mol Endocrinol 52(1):R35–R49,
ments that react to chronic injury and mediate Shh-induced meta- 2014.
plasia, Gastroenterology 138(3):1166–1177, 2010. Yang S-N, et al: Ionic mechanisms in pancreatic β cell signaling, Cell
Syed A-B, et al: Pancreas size and volume on computed tomography Mol Life Sci 71(21):4149–4177, 2014.
in normal adults, Pancreas 41(4):589–595, 2012. Zimmerman DW, et al: Cyclic interdigestive pancreatic exocrine secre-
Tan GD: The pancreas, Anaesth Intensive Care Med 15(10):485–488, tion: is it mediated by neural or hormonal mechanisms? Gastroenter-
2014. ology 102(4 Pt 1):1378–1384, 1992.
CHAPTER 5
Liver blood flow: physiology, measurement,
and clinical relevance
Simon Turcotte
Circulation in the liver is unique because of a dual afferent because of the greater oxygen content found in arterial blood.
blood supply, derived from the hepatic artery (HA) and the In addition, the HA provides the exclusive blood supply to the
portal vein (PV) (see Chapter 2). The oxygen-rich arterial intrahepatic bile ducts through the peribiliary plexus (Fig. 5.2).
blood and the nutrient-rich portal venous blood merge in the In the hepatic lobules, the hepatic arterioles empty directly or
hepatic parenchymal microcirculation to sustain the complex via peribiliary plexi into the sinusoids. Direct artery-to-HV
functions of the liver, prior to returning to the heart through connections do not usually exist but may arise in liver disease.
the hepatic veins (HVs). This chapter outlines how liver blood Within the liver parenchyma, the pressure in the arterial
flow (LBF) is controlled to maintain the hepatic perfusion system is reduced toward that existing in the portal circulation
within an acceptable physiologic range, describes techniques and sinusoids. This is suggested to be achieved mainly by (1)
used for LBF measurement, and explores clinical situations in the presinusoidal arteriolar resistance in the peribiliary plexus
which blood flow is altered. and (2) the intermittent closure of the arterioles, which protect
the portal bloodstream from arterial pressure (Rappaport,
PHYSIOLOGY 1973). In the event of hepatic arterial occlusion, numerous
intrahepatic collaterals can provide a source of arterial blood.
Liver Blood Supply Additionally, extrahepatic collateral supply can develop after
The peculiar double afferent blood supply to the liver results hepatic arterial ligation and depends on the site of occlusion.
in 75% to 80% of the entering blood being partially deoxygen- If the common HA is interrupted, revascularization occurs
ated portal venous blood draining the stomach, intestine, through extrahepatic collaterals arising from (1) the inferior
spleen, and pancreas. The remainder is well-oxygenated blood phrenic arteries and (2) from the gastroduodenal arteries,
from the aorta, carried by the HA. Mixing of arterial and portal which derive blood flow from the superior mesenteric artery
blood occurs in terminal branches in the sinusoidal microcir- (Rappaport & Schneiderman, 1976). Ligation of the proper
culation around hepatocytes arranged into roughly polyhedral- hepatic arteries leads to revascularization mainly via a hyper-
shaped lobules, from their periphery toward their centrilobular trophied inferior phrenic circulation, which can develop con-
venule (Fig. 5.1). The centrilobular venules drain into the HVs nections with hepatic arteries within the liver (Jefferson et al,
and into the inferior vena cava (IVC). Although the liver mass 1956). If only the right or left HA is interrupted, intrahepatic
constitutes approximately 2.5% of the total body weight, the translobar anastomoses reestablish arterial flow in the ligated
liver receives nearly 25% of the cardiac output. The total LBF, system (Mays & Wheeler, 1974). Thus complete long-term
ranges between 800 to 1200 mL/min, which is equivalent to dearterialization of the liver by any form of arterial vascular
approximately 100 mL/min per 100 g liver wet weight. The occlusion is difficult to achieve.
liver blood volume is estimated to range from 25 to 30 mL per
100 g of liver wet weight, which accounts for 10% to 15% of Portal Vein
the total body blood volume. The sinusoids hold 60% of the The PV normally carries approximately 75% of the total blood
blood volume, whereas the remaining 40% lies in large vessels flow to the liver (90 mL/min per 100 g of liver weight) in a
(HA, PV, and HV) (Greenway & Stark, 1971; Lautt, 1977a). valveless, low-pressure/low-resistance venous system. The
Of note is the high compliance of the liver, calculated as the inferior and superior mesenteric veins join with the splenic vein
change in its blood volume per unit change in venous pressure. to form the PV, and jointly they collect the venous outflow from
The liver thus acts as an important blood reservoir because its the entire prehepatic splanchnic vascular bed (the intestinal
blood volume can expand considerably in cardiac failure or, in tract from the lower esophagus to the rectum plus the pancreas
case of bleeding episodes, can compensate as much as 25% of and spleen). The PV pressure ranges between 6 and 10 mm Hg
the hemorrhage by rapid expulsion of blood into the circulation in humans when measured by direct cannulation (Balfour et al,
(Lautt, 2007; Lautt & Greenway, 1976). 1954). Portal venous pressure depends primarily on the degree
of constriction of mesenteric and splanchnic arterioles, coupled
Hepatic Artery with the intrahepatic vascular resistance. Because portal blood
The HA normally supplies approximately 25% of the total is derived from postcapillary beds, it is partly deoxygenated.
blood flow to the liver (25 to 30 mL/min per 100 g of liver However, because of its large volume flow rate, it may supply
tissue) in a high-pressure/high-resistance system. The mean 50% to 70% of the liver’s normal oxygen requirement. During
pressure in the HA is similar to that in the aorta. The HA fasting states, the oxygen saturation in the portal blood
provides as much as 50% of the liver’s oxygen requirement approaches 85%, which is greater than other systemic veins.
77
Hepatic sinusoid
Centrilobular
venule
Hepatocytes Hepatic Bile canaliculi
lobule
Reticuloendothelial
Centrilobular cell
venule Hepatic sinusoid
Bile canaliculi
Hepatocyte
Portal triad
Branch of
A Hepatic lobules bile duct
Branch of
hepatic portal vein
Branch of
hepatic artery
Portal triad
Branch of
bile duct
Branch of
Hepatic hepatic
sinusoid portal vein
Hepatocytes
Branch of B Hepatocytes and sinusoids
hepatic artery
C Portal triad
FIGURE 5.1. The liver microcirculation. The liver is composed of thousands of roughly polyhedral structures called hepatic lobules, which are the
basic functional units of the organ. Hepatic lobule of some mammals, such as the pig, are delimited on all sides by connective tissue but have much
less connective tissue, and their boundaries are more difficult to distinguish in humans. A, A small central vein projecting through the center of each
hepatic lobule and several sets of blood vessels defining the periphery. The peripheral vessels are grouped primarily in connective tissue comprising
the portal tracts in the space of Mall, which include a branch of the portal vein and a branch of the hepatic artery, as well as a branch of the bile
duct. These comprise the portal triad. B, Both blood vessels to each lobule give off sinusoids, which run between plates of hepatocytes and drain
into the central vein. C, Micrograph showing components of the portal triad within the space of Mall (hematoxylin and eosin; ×220). (From Mescher
AL: Junqueira’s Basic Histology: Text and Atlas, 12th ed. New York, 2009, McGraw-Hill.)
Hepatic oxygen supply is diminished if portal blood flow is

significantly reduced, but the effect is minimized by an increase
in oxygen extraction from the hepatic arterial blood and not by
P
increasing flow (Lautt et al, 1987).
Hepatic Veins
The hepatic venous system is the systemic drainage tract of
the entire splanchnic circulation. A total LBF of 1.5 L/min is
PP
considered the normal value in the average male, but the range
can be quite wide (1 to 2 L/min). In normal conditions, the
A free pressure in the HVs and IVC is 1 to 2 mm Hg and is 1 to
5 mm Hg lower than the pressure measured in the sinusoids
and PV. The portal pressure gradient, defined as the difference
in pressure between the PV and the IVC, has been a useful
clinical indicator of the perfusion pressure of the liver with
portal blood (Berzigotti et al, 2013). The pressure gradient
in the liver is thus extremely low, in the range of 5 mm Hg
FIGURE 5.2. Arterial peribliary plexus. A cast of the portal vein (P), compared with all other organs, where it is in the range of
hepatic artery (A), and peribiliary arterial plexus (PP) of a rat, showing a
115 mm Hg (Lautt, 2009). Hepatic venous blood is normally
connection between a small artery and the plexus (arrow). The peribiliary
plexus forms a dense sheath around the bile duct. Bar = 100 µm. approximately two-thirds saturated with oxygen, but this may
(From Grisham JW, Nopanitaya W: Scanning electron microscopy of be markedly reduced during periods of low delivery of oxygen
casts of hepatic microvessels: review of methods and results. In Laut to the liver, when oxygen is extracted by hepatocytes. In resting
W: Hepatic Circulation in Health and Disease. New York, 1981, Raven states, the liver accounts for approximately 20% of the total
Press, Figure 4, p 98.) oxygen consumption of the body.
Chapter 5 Liver blood flow: physiology, measurement, and clinical relevance 79
A CLASSIC HEPATIC LOBULE B PORTAL LOBULE C PORTAL ACINUS

Drains blood from the portal Drains bile from Supplies oxygenated
vein and the hepatic artery to hepatocytes to the blood to hepatocytes
the hepatic or the central vein bile duct
Central vein
Hepatic arteriole
Zone III
Bile duct least
oxygenated
Portal vein
Zone II
Centrilobular Zone I
venule most
oxygenated
FIGURE 5.3. Structure-function conceptual liver units. To this day, there is no complete consensus on whether the microvascular unit of the liver
should be referred to as a lobule, centering on a hepatic vein, or an acinus, centering on a “portal triad” consisting of a terminal branch of the hepatic
artery, portal vein, and bile duct, encased within a limiting plate of cells defining the space of Mall. Three related conceptual units emphasizing different
aspects of hepatocyte activity have been proposed. A, The classic lobule emphasizes the endocrine function of hepatocytes as blood flows past
them toward the centrilobular venule. B, The portal lobule emphasizes the hepatocytes’ exocrine function and the flow of bile from regions of three
classic lobules toward the bile duct in the portal triad at the center. The area drained by each bile duct is roughly triangular. C, The liver acinus
concept proposed by Rappaport emphasizes the different oxygen and nutrient contents of blood at different distances along the sinusoids, with
blood from each portal area supplying cells in two or more classic lobules. Each hepatocyte’s major activity is determined by its location along the
oxygen/nutrient gradient: periportal cells of zone I get the most oxygen and nutrients and show metabolic activity generally different from the pericentral
hepatocytes of zone III, exposed to the lowest oxygen and nutrient concentrations. (From Boron WF, Boulpaep EL (eds): Medical Physiology: A
Cellular and Molecular Approach. Philadelphia, 2005, Saunders Elsevier.)
Hepatic Microcirculation with a diameter of 150 to 175 nm (Fig. 5.4). The sieve plates
Although the organization of the liver into morphologic and occupy as much as 8% of the endothelial surface and are not
functional units has been a matter of debate, the hexagonal uniform in size or distribution throughout the length of the
polyhedral-shaped hepatic lobule, encompassing hepatic micro- sinusoids. There is a decrease in diameter but an increase of
vascular subunits consisting of a portal triad of terminal frequency from periportal to centrilobular zones, which results
branches of the HA, PV, and bile duct; a network of sinusoids; in higher centrilobular porosity (Braet & Wisse, 2002; Wisse
and an efferent centrilobular venule is a widely accepted frame- et al, 1985). The fenestrae are dynamic structures that contract
work (Ekataksin & Kaneda, 1999; Malarkey et al, 2005) (see and dilate in response to alterations of sinusoidal blood flow
Fig. 5.1). The portal triads, surrounded by lymphatics and and perfusion pressure (Smedsrød et al, 1994). Red blood cells
autonomous nerves, all travel together in parallel in the space (RBCs) remain restricted within the sinusoids, whereas mol-
of Mall, through the liver parenchyma, and form portal tracts. ecules as large as albumin can pass through the fenestrations
Lymphatics transport proteins and other macromolecules that and enter the small space of Disse before making contact with
are trapped extravascularly due to hindrance of hepatocellular the microvilli of the hepatocytes (Lautt, 2009).
uptake, as in the case of cirrhosis, which will in turn contribute As represented in Fig. 5.5, other unique cellular compo-
to ascites formation (see Chapter 76). The hepatic sinusoids nents, such as the hepatic stellate cell (HSC) (Friedman, 2008)
correspond to the capillary bed of the liver and represent the and the Kupffer cell (KC), are found in the hepatic sinusoids
segment of the microcirculation in which supply of nutrients and may regulate the sinusoidal microcirculation in response to
and removal of metabolic products by hepatocytes takes place. various mediators (McCuskey, 2000). External to the endothe-
Bile canaliculi closely assemble around hepatocytes and collect lium cell lining, HSCs (also known as fat-storing cells, Ito cells,
bile flowing in an opposite direction from blood in the sinusoids. or hepatic perisinusoidal lipocytes) are contractile cells distrib-
As depicted in Fig. 5.3, histologic and physiologic studies of uted homogeneously around the exterior of the endothelial cells
the liver have given rise to three related ways to view the liver’s in the space of Disse, which is the space between the basal
microcirculation, emphasizing different functional aspects microvilli-rich surfaces of the hepatocytes and the sinusoidal
useful for the classification of various pathologic processes. lining cells. In addition to their well-known importance in
Apart from the absence of a basement membrane, the retinol metabolism and as key actors in the hepatic fibrogenic
structural peculiarity of hepatic sinusoids is their unique lining, response to injury (see Chapters 7 and 76), HSCs are capable
consisting of endothelial cells with flattened processes perfo- of compressing the sinusoidal diameter by squeezing the endo-
rated by small fenestrae. These open fenestrations are arranged thelial cells and therefore play a central role in the regulation
in clusters of 10 to 50 pores, forming so-called “sieve plates” of blood flow through hepatic sinusoids (Rockey, 2001; Zhang
et al, 1994). KCs are liver-specific macrophages, and, in con- direct regulation role of the sinusoidal microcirculation is
trast to HSCs, are anchored to the luminal side of the sinusoids. debated (Lautt, 2009; Vollmar & Menger, 2009).
They account for approximately 15% of the liver-cell popula-
tion and constitute approximately 80% of the total population Control of Liver Blood Flow
of macrophages in the body (Lautt, 2009). By their large bodies, The hepatic blood flow required to meet the physiologic func-
with cytoplasmic process that sometimes reach the opposite tion of the liver is mainly controlled by intrinsic physiologic
wall of a sinusoid, KCs represent a flow hindrance and can mechanisms that are independent of extrinsic innervation and
secrete large amounts of the vasodilator nitric oxide, but their vasoactive agents. Instead, the interrelationship of arterial and
portal inflow circuits is the major contributor to hepatic
perfusion.
PS Liver-Intrinsic Blood Flow Regulation

THE HEPATIC ARTERIAL BUFFER RESPONSE. Adequate and homoge-
H neous blood flow to the liver is necessary to sustain hepatic
functions and clearance of metabolites. Because the liver does
not control portal blood flow, which is simply the outflow of
the extrahepatic splanchnic organs, the main mechanism by
E which hepatic blood flow can remain constant relies on modula-
tion of the hepatic arterial flow (Fig. 5.6). Although this phe-
nomenon was observed in the late 19th and early 20th centuries
(Burton-Opitz, 1911; Gad, 1873), it has been characterized and
coined as the hepatic arterial buffer response (HABR) by Lautt
in 1981.
E HABR represents the ability of the HA to produce compen-
PS
H satory flow changes at the presinusoidal level in response to
changes in portal venous flow (see Fig. 5.6): If portal blood
FIGURE 5.4. Fenestrations in endothelial sinusoid lining. Electron flow is reduced, the HA dilates and increases its flow into the
microscopy of the luminal surface of the endothelium lining a sinusoid sinusoids, and the HA constricts when the portal flow is
in the liver shows grouped fenestrations. At the border are seen cut increased (Jakab et al, 1995; Lautt et al, 1990). In patients
edges of the endothelial cell (E) in this discontinuous sinusoid and undergoing abdominal surgery, a temporary occlusion of the
hepatocytes (H). Between these two cells is the thin perisinusoidal space PV resulted in a sharp increase in hepatic arterial flow of about
(PS), into which project microvilli from the hepatocytes surface. Blood 30%, whereas temporary occlusion of the HA did not have
plasma passes freely through the fenestrations into the perisinusoidal significant effect on portal venous flow. The HABR seems to
space, where the voluminous membrane of hepatocytes acts to remove operate under various physiologic and pathologic conditions
many high- and low-molecular-weight blood components and nutrients
and has even been suggested to operate prenatally (Ebbing
for storage and processing. Proteins synthesized and secreted from
hepatocytes, such as albumin, fibrinogen, and other blood proteins, are
et al, 2008). The HABR appears mainly regulated by the
released into the perisinusoidal space (×6500). (From Mescher AL: washout of adenosine, a potent vasodilator. Although adenosine
Junqueira’s Basic Histology: Text and Atlas, 12th ed. New York, 2009, is produced at a constant rate and secreted into the space of
McGraw-Hill; and Eddie Wisse, Electron Microscopy Unit, Department Mall (see Fig. 5.1), its concentration depends on the rate of
of Pathology, University of Maastricht, The Netherlands.) washout from the space of Mall into the sinusoids. When portal
Space of Disse
Hepatocyte Endothelial
Hepatic sinusoid cell
Hepatic Kupffer
stellate cell cell
FIGURE 5.5. Hepatic stellate and Kupffer cells relation to the liver microcirculation. Contractile hepatic stellate cells are distributed homogeneously
around the exterior of the endothelial cells in the space of Disse, the space between the basal microvilli-rich surfaces of the hepatocytes and the
sinusoidal lining cells. Kupffer cells, which are liver-specific macrophages, are anchored to the luminal side of the sinusoids. They can secrete vasoac-
tive mediators, and their large bodies may represent sinusoidal flow hindrance.
20 20
18 18
16 16
14 14
Blood flow (mL/min)
Blood flow (mL/min)

12 12
10 10
§
8 † 8
† †
6 * 6 *
4 4
2 2
0 0
A I II III IV V B I II III IV V
FIGURE 5.6. Hepatic arterial buffer response in cirrhotic (A) and control livers (B). Upon reduction of portal venous blood flow (open circles, black
line), there is a constant increase of hepatic arterial blood flow (solid circles, gray line). The portal flow is markedly diminish in cirrhotic liver, but the
buffer response is preserved. Values are means ± standard error of triplicate measurements per animal (n = 6). *P < .05 vs. I and II; †P < .05 vs. I,
II, and III. §P < .05 vs. I, II, II, and IV. (From Richter S, et al (eds): Impact of intrinsic blood flow regulation in cirrhosis: maintenance of hepatic arterial
buffer response. Am J Physiol Gastrointest Liver Physiol 279:G454-G462, 2000.)
blood flow decreases, less adenosine is washed away, and the arrangement precludes substances diffusing back from the
elevated adenosine concentration leads to dilation of the HA. hepatic parenchyma or venous blood into the hepatic arterial
Of importance, the source of the adenosine found in the space resistance vessels. Among other studies, this concept has been
of Mall remains to be elucidated. If portal blood flow is severely exemplified by isovolemic hemodilution or upregulation of
reduced, the buffer response results in the HA dilating maxi- hepatic enzymes, leading to oxygen deprivation to the liver
mally, as demonstrated by the inability to produce additional parenchyma, which does not result in HA dilation (Lautt &
dilation in response to intraarterial infusion of adenosine. Greenway, 1987). Therefore the hepatic metabolic demands do
Conversely, when portal flow is doubled, the HA constricts to not control the hepatic arterial flow, even if the liver parenchy-
a maximal extent, as demonstrated by the inability of intraarte- mal cells can release large quantities of potent vasoactive
rial norepinephrine to produce further constriction (Lautt et al, molecules during metabolic stress.
1990). Although the HABR is sufficiently powerful to regulate The term autoregulation refers to the tendency for local
the vascular tone in the HA over the full range from maximal arterial blood flow to remain constant in the face of pressure
vasodilation to maximal vasoconstriction, this mechanism is changes in the arteries that perfuse a given organ. Overall, the
capable of buffering 25% to 60% of the decreased portal flow degree of autoregulation is considered small in the liver, and
(Lautt et al, 1985). mixed results have been reported in animal models (Eipel et al,
Although adenosine appears to be the main mediator of 2010). In fact, the adenosine washout may well account for HA
the HABR, a possible contribution of the afferent sensory autoregulation because endogenous adenosine produced by the
nerves and neuropeptides is suggested by studies in fully HA tributaries can contribute to high presinusoidal adenosine
denervated animal livers (Mathie et al, 1980) and in trans- concentration in situations where hepatic clearance is impaired
planted human livers (Henderson et al, 1992; Payen et al, by a reduction in the portal flow, leading to hepatic arterial
1990), in which the HABR upon partial PV occlusion is par- vasodilation (Ezzat & Lautt, 1987).
tially impaired (Biernat et al, 2005; Ishikawa et al, 1995). It
has also recently been shown that hydrogen sulfide (H2S), a REGULATION OF INTRAHEPATIC RESISTANCE AT THE SINUSOIDAL
vasoactive gaseous mediator produced within the liver, appears LEVEL. Sinusoidal blood pressure and vascular resistance are so
to almost double the HABR by increasing the hepatic arterial low that a pressure gradient across the liver from the portal
conductance, and in turn, its inhibition has the opposite effect venous inflow to the hepatic venous outflow is only approxi-
(Siebert et al, 2008). mately 5 mm Hg. The low pressure gradient is remarkable,
considering that 30% of the inflow to the liver sinusoids is
THE HEPATIC INFLOW IS NOT CONTROLLED BY LIVER-INTRINSIC META- provided by the HA under arterial pressure. How the hepatic
BOLIC NEEDS. Until the mid-1970s, the hepatic arterial flow was pressure gradient is maintained has been studied in multiple
believed to be under metabolic control of the liver. As for most animal models. About 60 years ago, Knisley suggested the pres-
organs, the liver metabolism, estimated by its oxygen require- ence of sphincter-like structures at the entrance and exit of
ments, was postulated to participate in vascular inflow control. sinusoids that maintain the portal venous pressure gradients,
It has, however, been established that the liver normally receives but these proved to be species dependent (Lautt, 2009; Knisely
more oxygen than it requires and can extract more oxygen to et al, 1957; Vollmar & Menger, 2009). In humans, although
compensate for reduced delivery (Bredfeldt et al, 1985; Lautt, smooth muscle cells are found throughout all segments of
1976, 1977b). Additionally, the unique one-way sinusoidal flow the hepatic microvascular subunits, no such sphincter-like
TABLE 5.1 Liver-Intrinsic Vasoactive Molecules and Pathways Regulating the Microcirculation
Cellular Source and
Vasoactive Agent Function Enzyme System Distribution Target Cell Pathways
Thromboxane A2 Vasoconstriction, platelet COX-1, COX-2 SEC, KC SEC, platelet, TxA2R

activation and aggregation, leukocyte
leukocyte adhesion
Prostaglandin I2 Vasodilation, inhibition of COX-1, COX-2 SEC SEC, HSC PGI2R
platelet aggregation
Angiotensin II Vasoconstriction ACE HSC HSC AT1
Nitric oxide Vasodilation eNOS SEC VSMC, HSC sGC
Nitric oxide Vasodilation iNOS SEC, KC, VSMC, VSMC, HSC sGC
HSC, HC
Endothelin-1 Vasoconstriction SEC, HSC, KC VSMC, HSC, ETAR, ETB2R
SEC, KC
Endothelin-1 Vasodilation SEC, HSC, KC SEC ETB1R
Carbon monoxide Vasodilation HO-1 SEC, KC, VSMC, VSMC, HSC sGC
HSC, HC
Vasodilation HO-2 HC VSMC, HSC sGC
Hydrogen sulfide Vasodilation CSE (CBS) HSC, HC VSMC KATP channels
ACE, Angiotensin-converting enzyme; AT1, type 1 of angiotensin II; CBS, cystathionine-synthase; COX-1, COX-2, cyclooxygenase-1 and 2, respectively; CSE, cystathionine-lyase; eNOS, endothelial
constitutive nitric oxide synthase (type III); ETAR, endothelin type A receptor; ETB1R, endothelin type B1 receptor; HC, hepatocytes; HO-1, inducible heme oxygenase; HO-2, constitutive heme oxygenase;
HSC, hepatic stellate cells; iNOS, inducible nitric oxide synthase (type II); KATP, adenosine triphosphate (ATP)-sensitive potassium channel; KC, Kupffer cells; PGI2R, prostaglandin I2 receptor; SEC,
sinusoidal endothelial cells; sGC, soluble guanylate cyclase; TxA2R, thromboxane A2 receptor; VSMC, vascular smooth muscle cells.
From Vollmar B, Menger MD: The hepatic microcirculation: mechanistic contributions and therapeutic targets in liver injury and repair. Physiol Rev 89:1269-1339, 2009..
structures have been described for controlling the intrahepatic hypercarbia, whereas metabolic alkalosis has essentially no
blood flow. significant effect (Hughes et al, 1981). The sympathetic nervous
In contrast, there is a growing body of evidence that con- system is thought to be responsible for the hepatic arterial
tracting cells associated with the sinusoids, such as the HSCs vasoconstriction observed in hypercarbia and hypoxia (Mathie
and the sinusoidal endothelial cells, through a complex inter- & Blumgart, 1983).
play with vasoactive mediators, may dynamically regulate the Sympathetic Nervous System. The liver is a significant blood
hepatic microvascular blood flow. With such a low sinusoidal reservoir, and 50% of its blood volume may be mobilized by
perfusion pressure, local regulators at a single-cell level may nerve stimulation (Greenway & Lautt, 1989). Denervation
actively control the flow within the sinusoids toward the HVs. experiments have shown that the sympathetic nervous system
It seems plausible that the HSCs could dilate to pull outward is not involved in basal arterial tone in the liver (Mathie &
on the endothelial cells and enlarge the sinusoidal space (see Blumgart, 1983). Hepatic sympathetic nervous stimulation
Fig. 5.5); however, this does not seem to have been shown causes hepatic arterial vasoconstriction and reduced blood flow,
(Vollmar & Menger, 2009). Many endothelial mediators known which appears secondary to an autoregulatory response (Gre-
to control vascular tone by acting on HSC contractility have enway & Stark, 1971). Sensory denervated rats and pigs have
been described (Table 5.1), notably, (1) the endothelium- a diminished arterial buffer response on partial occlusion of the
derived relaxing factor nitric oxide (NO) (Palmer et al, 1987), PV (Ishikawa, 1995). Portal pressure increases as a result of an
(2) the endothelium-constricting factor endothelin-1 (ET-1) increase in portal venous resistance, but portal flow does not
(Hickey et al, 1985; Yanagisawa et al, 1988), and (3) two decrease unless there is a decrease in intestinal or splenic blood
vasodilatory gaseous molecules: carbon monoxide (CO) flow caused by simultaneous sympathetic stimulation of these
(Furchgott & Jothianandan, 1991; Lin & McGrath, 1988) and vascular beds. Although the HA contains both α-adrenergic and
H2S (Hosoki et al, 1997; Zhao et al, 2001). With the exception β-adrenergic receptors, the portal venous system is believed to
of H2S, a direct contribution of these vasoactive agents to the contain only α-receptors (Richardson & Withrington, 1981). At
HABR is not well established (Eipel et al, 2010). low doses, epinephrine causes hepatic and mesenteric arterial
vasodilation, whereas at high doses, vasoconstriction occurs in
Liver-Extrinsic Factors Affecting Liver Inflow the hepatic arterial and portal venous vascular beds and in the
ENDOGENOUS FACTORS mesenteric circulation (Greenway & Stark, 1971; Richardson
Blood Gas Tensions. Hypercarbia (partial pressure of carbon & Withrington, 1981).
dioxide in arterial blood [PaCO2] > 70 mm Hg) increases Other Endogenous Vasoactive Agents. Intraportal administra-
portal venous flow and decreases hepatic arterial flow in tion of exogenous vasoactive agents affects hepatic arterial
dogs (Hughes et al, 1979a), whereas hypocarbia (PaCO2 < resistance (Lautt et al, 1984). The mechanisms underlying this
30 mm Hg) decreases both (Hughes et al, 1979b). Systemic intrahepatic transvascular effect are not understood, but it is
hypoxia (partial pressure of oxygen in arterial blood [PaO2] < likely that the close anatomic association between arterioles and
70 mm Hg) causes a decrease in arterial flow but has no effect venules could permit this and may be a means by which hepatic
on the contribution from the PV (Hughes et al, 1979c). The arterial blood flow is finely controlled by endogenous agents,
response to metabolic acidosis is similar to that induced by such as gut hormones. Gastrin, secretin, cholecystokinin, and
vasoactive intestinal peptide cause vasodilation of the HA.

BOX 5.1 Summary of Methods Currently Used for Measuring
Hepatic blood flow is profoundly increased by glucagon as a
Liver Blood Flow
consequence of its strong vasodilatory action on the mesenteric
vasculature, but insulin has little hemodynamic effect on the Flow in Single Vessels
hepatic circulation. In addition, antagonists of calcitonin gene– Electromagnetic flowmeter*
related peptide and neurokinin significantly reduce hepatic Hepatic venous direct and wedge pressure measurement
*Doppler ultrasound
arterial blood flow, suggesting the presence of their receptors
Transient elastrography (indirect)
on the arterial vasculature (Biernat et al, 2005). Histamine
causes hepatic arterial dilation and, in the dog only, hepatic Total Liver Blood Flow
venous constriction. Bradykinin is a potent hepatic arterial Hepatocyte excretion
vasodilator that has little effect on the portal venous system. *Indocyanine green
The hepatic arterial vascular bed is dilated by most prostaglan- Galactose
dins; however, prostacyclin does not affect hepatic arterial flow Reticuloendothelial uptake
but increases portal blood flow through a vasodilator effect on *Technetium-99m–sulfur colloid
the prehepatic vascular bed. Vasopressin decreases portal flow Indicator dilution
and pressure by mesenteric arterial vasoconstriction but has Red blood cells (chromate-51)
Serum albumin (iodine-131, technetium-99m)
variable effects on the HA. Serotonin is believed to mediate
Indicator fractionation
vasoconstriction of portal radicles.
Microspheres (various labels)
Liver-extrinsic NO causes vasodilation in the hepatic arterial
and mesenteric vascular beds. Endothelin molecules can exert Hepatic Tissue Perfusion
a powerful and prolonged generalized systemic constriction Inert gas clearance (krypton-85, xenon-133, hydrogen)
(Miller et al, 1989; Zhang et al, 1994) that also has a direct Laser Doppler flowmetry
effect on the hepatic blood flow. Endothelins reduce hepatic In vivo fluorescent microscopy
perfusion (Kurihara et al, 1992), increase portal pressure Near-infrared spectroscopy
(Bauer et al, 1994; Isales et al, 1993; Tanaka et al, 1994; Tran-
*Common clinical use.
Thi et al, 1993), and reduce sinusoidal diameter (Bauer et al,
1994; Okumura et al, 1994; Zhang et al, 1994). Angiotensin
decreases hepatic arterial and portal blood flow and is one of a variety of indicator-clearance techniques were subsequently
the few substances to produce a significant vasoconstrictor developed, often confounded by the presence of liver disease.
effect on the HA. In contrast, H2S, either endogenously or Some of these techniques remain useful to estimate liver
exogenously, can reverse the norepinephrine-induced vasocon- function when planning major hepatectomies. By far today,
striction in an NO-independent fashion (Fiorucci et al, 2005). Doppler ultrasound (D-US) is the most common first-line
noninvasive technique used to assess liver vascularization and
EXOGENOUS FACTORS guide clinical management. The available methods are discussed
Anesthetic Agents. The effect of anesthetic agents on the LBF under three broad headings: (1) flow in single blood vessels,
has been mainly studied in animal models some 30 years ago. (2) total LBF, and (3) hepatic tissue perfusion. The techniques
Hepatic arterial and portal venous blood flow decreases pas- currently used for experimental and clinical use are listed in
sively in parallel with cardiac output during halothane inhala- Box 5.1.
tion, with little change in vascular resistance (Hughes et al,
1980; Thulin et al, 1975). Enflurane has been found to have
Flow in Single Vessels and Assessment
similar effects as those of halothane, although there is a decrease
in hepatic arterial vascular resistance as part of a generalized of Portal Hypertension
decrease in peripheral vascular resistance (Hughes et al, 1980). Invasive Techniques
NO concentrations of 30% to 70% reduce HA and PV flow, ELECTROMAGNETIC FLOWMETER. The direct and continuous
possibly as a result of a generalized stimulatory action on measurement of hepatic arterial and portal venous blood flow
α-adrenergic receptors (Thomson et al, 1982). Isoflurane with electromagnetic flowmeter probes remains the best avail-
seems to have minimal effects on hepatic arterial and portal able means of assessing individual vessel flow. Although the
venous flows, and the intravenous (IV) agent fentanyl may have technique has found widespread application in experiments
little effect on prehepatic splanchnic blood flow (Nagano et al, using large animals, its use in clinical situations has been limited
1990); thiopentone in low doses vasoconstricts the hepatic by the relatively extensive vascular dissection required for place-
arterial and mesenteric vascular beds (Thomson et al, 1986). ment of the probes and by the overestimation of true hepatic
tissue blood flow, which occurs in the presence of portosystemic
MEASUREMENT OF LIVER BLOOD FLOW shunts. In addition, relative movements of the probes can give
AND PERFUSION rise to errors and the need for repetitive calibrations. Using this
technique, total LBF in anesthetized subjects was determined
The earliest methods of measuring LBF involved direct invasive to be approximately 1 L/min, of which approximately 25% was
techniques, such as intravascular devices or venous outflow supplied by the HA (Schenk et al, 1962). Electromagnetic
collection (Burton-Opitz, 1910, 1911; MacLeod & Pearce, probes have also been used intraoperatively to assess the hemo-
1914). Currently, measurement of hepatic venous pressure dynamic status of the liver in cirrhosis (Ohnishi et al, 1987)
gradient (HVPG) remains the gold-standard technique to after liver surgery or transplantation (Takaoka et al, 1990). A
assess portal hypertension (Berzigotti et al, 2013) (see Chapters typical experimental preparation using electromagnetic flow-
76 and 87). Indirect determination of blood flow by the use of meter probes is illustrated in Figure 5.7. Because the US probes
frequent etiologies of chronic liver disease. The use of balloon-

tipped catheters is recommended to measure HVPG (Bosch
et al, 2009) because the volume of the liver circulation that is
measured is larger than that obtained by wedging the catheter
tip, which enhances the reliability and accuracy of the measure-
ment (Maleux et al, 2011).
Noninvasive Techniques
DOPPLER ULTRASOUND. Named after Christian Doppler for the
phenomenon he described, the principle of flow estimation by
Doppler is simple: Flow is a product of the average velocity of
the blood measured in the vessel of interest and the cross-
sectional area of the vessel. Two forms of D-US devices exist.
The first consists of a flowmeter with a US probe that is placed
directly on the vessel; measurement with such a device is
invasive. The second consists of a combined image scanner and
flowmeter (duplex) by which flow in a vessel can be measured
transcutaneously and noninvasively (see Chapter 15). In experi-
ments performed on anesthetized dogs, good correlation was
FIGURE 5.7. Experimental arrangement for measuring liver blood flow
found between portal venous flow measured by a transcutane-
in the dog with electromagnetic flow probes. Probes are placed around ous Doppler duplex system and electromagnetic flow probes
the portal vein (PV) and common hepatic artery (CHA). The gastroduo- fitted to the PV (Dauzat & Layrargues, 1989). Because the
denal vein (GDV) and gastroduodenal artery (GDA) are ligated as illus- development of color Doppler and better probe resolution with
trated to ensure that the flows measured by the probes are those that high-frequency transducers has improved accuracy, and because
actually perfuse the liver. the technique provides morphologic assessment of the liver, is
cheap, and can be performed at bedside, D-US has become a
widespread first-line technique. In patients with known cir-
are fitted directly to the vessel, such systems have been used rhosis, D-US has greater than 80% specificity diagnosing clini-
successfully in the investigational setting of intraoperative cally significant portal hypertension, but the sensitivity does not
(Henderson et al, 1992) and postoperative measurement of exceed 4% to 70%, particularly in compensated patients (Ber-
portal venous and hepatic arterial blood flow in liver transplant zigotti et al, 2013). Inversion of flow within the PV is 100%
patients (Payen et al, 1990). specific for clinically significant portal hypertension. D-US is
very accurate for detecting PV, HV, and HA thrombosis (Rossi
TRANSJUGULAR HEPATIC VENOUS PRESSURE MEASUREMENT. Because et al, 2006). Additionally, D-US is useful in the noninvasive
portal hypertension is responsible for most clinical consequences follow-up of transjugular intrahepatic portosystemic shunt
of cirrhosis, measurement of portal venous pressure is critical (TIPS) (Abraldes et al, 2005) (see Chapter 87).
to guide the clinical management of patients with chronic liver
diseases. Currently, the accuracy of invasive techniques has not Total Blood Flow
been surpassed by noninvasive measurement. Direct measure- Clearance Techniques
ments of portal pressure can be performed through transhepatic Substances reaching the liver via the HA or PV are equally well
or transvenous catheterization of the PV but are rarely used extracted (Lautt et al, 1984), and the rate of disappearance
because of the risk of intraperitoneal bleeding and visceral from the bloodstream of an indicator substance exclusively
perforation. Instead, measurement of hepatic venous direct and cleared by the liver is proportional to LBF. First applied to
wedge pressure by a transjugular approach has been developed humans by Bradley and colleagues in 1945, indirect clearance
as a safe and reproducible technique to assess portal hyperten- methods of LBF measurement rely on the Fick equation. The
sion (Fig. 5.8). As mentioned earlier, the pressure gradient flow measurement obtained by Bradley’s group depended on
between the PV and the IVC represents the liver portal perfu- the fact that intravenously injected bromosulfophthalein is
sion pressure, and its normal value is as high as 5 mm Hg. The removed from the bloodstream and excreted into the bile
HVPG is calculated with the difference between wedged hepatic entirely by hepatocytes. They derived a value for the rate of
venous pressure (WHVP) and free hepatic venous pressure hepatic bromosulfophthalein removal indirectly by determining
(FHVP) (see Chapter 87). It is based on the concept that when the rate of IV infusion of dye that maintained the arterial
the blood flow in a HV is blocked by a wedged catheter, concentration at a constant level, and by measuring the arte-
the static column of blood transmits the pressure from the riovenous concentration difference of bromosulfophthalein,
preceding communicated vascular territory, in this case, the they were able to calculate total hepatic blood flow. The mean
hepatic sinusoids. Thus WHVP is a measurement of the hepatic value obtained in a group of healthy individuals was 1.5 L/min.
sinusoidal pressure and not of portal pressure itself. As in cir- Indocyanine green (ICG) is another substance dependent
rhosis, the intersinusoidal communications are lost due to on hepatocyte extraction into bile. It was initially devised
fibrous sept and nodule formation, and the sinusoidal pressure for the measurement of blood flow and later used for the assess-
equilibrates with portal pressure. It has been well demonstrated ment of liver function by measuring functional hepatocyte
that WHVP adequately reflects portal pressure in alcoholic mass (see Chapter 3). HV canulation was initially performed
liver disease, hepatitis C–related (Perelló et al, 1999), and hepa- to calculate the true extraction efficiency of ICG because
titis B–related cirrhosis (Iwao et al, 1994), which are the most of its incomplete hepatic removal (Caesar et al, 1961). Many
2-4 cm
A B
mmHg
40
WHVP WHVP
20
FHVP FHVP
0
C D
FIGURE 5.8. Measurement of hepatic venous pressure. A, Free hepatic venous pressure (FHVP) is measured bymaintaining the tip of the catheter
free in the hepatic vein at 2 to 4 cm from its opening into the inferior vena cava. B, Wedged hepatic venous pressure (WHVP) is measured by
occluding the hepatic vein by inflating the angiographic balloon (arrow) at the tip of the catheter. Adequate occlusion of the hepatic vein is confirmed
by slowly injecting 5 mL of contrast dye into the vein with the balloon inflated. Please note the typical wedged pattern distal to the balloon. C, A
washout of contrast dye through communications with other hepatic veins (arrow) prevents a correct measurement of the hepatic venous pressure.
D, Typical tracing of pressures measured in the hepatic vein obtained using a multichannel recorder and adequately calibrated transducers. (From
Berzigotti et al: Assessing portal hypertension in liver diseases. Expert Rev Gastroenterol Hepatol 2013;7:141-155.)
investigators now use a simplified version of the original (Combes, 1960), or propranolol (George, 1979). The more
method, in which ICG is administered as a single IV bolus complete hepatic extraction of these substances overcomes the
instead of as an infusion, and hepatic extraction efficiency is need to cannulate a HV in patients with normal liver function.
determined from an analysis of the clearance curve derived A modification of the colloid extraction method developed in
from peripheral blood sampling or pulse dye densitometry, by the 1980s allows the derivation of the ratio of hepatic arterial
using an optical sensor placed on the finger (Akita et al, 2008; to total LBF, termed the “hepatic perfusion index.” The basis
Okochi et al, 2002). ICG plasma disappearance rate is the most of the technique is the ability to determine by dynamic scintig-
commonly used parameter, with a normal range between 16% raphy the temporal separation of accumulating hepatic activity
and 25% per minute and near-complete disappearance at 20 from the arterial and portal supplies after the IV administration
minutes (Sakka, 2007). It is now the most widely used quantita- of a bolus of technetium-99m–sulfur colloid (Fleming et al,
tive liver function test in the clinical setting (Clavien et al, 1981).
2007). Limitations of this technique include variations in
hepatic blood flow caused by intrahepatic and extrahepatic Other Techniques of Physiologic Interest
shunting, or portal thrombosis, which is common in liver INDICATOR DILUTION. The indicator dilution method relies on the
disease. application of the Stewart-Hamilton principle, also derived
Other hepatic clearance techniques have been investigated from the Fick equation (Stewart, 1897). In principle, the
in the past, such as colloidal clearance by the hepatic Kupffer hepatic blood flow is proportional to the amount of hepatic
cells (Dobson & Jones, 1952) and hepatocyte removal of galac- blood that has diluted an introduced indicator. This method
tose (Keiding, 1988), sorbitol (Zeeh et al, 1988), rose bengal involves the injection into the HA and PV of a labeled substance
that is not removed by the liver; changes in HV concentration γ-emissions of 133Xe after the injection of a saline solution of the
are measured by blood sampling or by monitoring the hepatic isotope into the PV.
isotope activity with an external detector. Such a method is
therefore independent of hepatocellular function and reliable, Laser Doppler Flowmetry
provided the indicator remains in the vascular space and is not Laser Doppler flowmetry (LDF) is a more recent but estab-
excreted prior to sampling. A modified thermal dilution tech- lished technique for the real-time measurement of microvascular
nique has been used to measure portal blood flow in humans RBC perfusion in the liver. By illuminating the tissue with
(Biber et al, 1983). Indicator dilution methods overestimate low-power laser light and capturing the backscattered light with
true blood flow to hepatic tissue when intrahepatic or extrahe- independent photodetectors, the Doppler shift of moving cells
patic shunts are present, although it is possible to measure can be transmitted as an electrical signal. Linearity of the LDF
azygos blood flow by thermal dilution in patients with cirrhosis signal from the liver with total-organ perfusion has been shown
(Bosch & Groszmann, 1984). (Shepherd et al, 1987), and the technique has been shown to
be sensitive to rapid changes in organ flow (Almond & Wheatley,
INDICATOR FRACTIONATION. The measurement of regional blood 1992). The technique has been applied successfully to measure
flow by fractional distribution of cardiac output was first LBF during liver transplantation in humans (Seifalian et al,
described by Sapirstein in 1956. Briefly, a known amount of 1997). A major drawback of the technique is that, owing to the
radioactive microspheres is injected into the left ventricle, and small volume of tissue interrogated by the laser, the LDF signal
a reference sample is withdrawn from a peripheral artery at a can only be used to measure arbitrary, instead of absolute blood
known rate. The microspheres are then extracted from the perfusion in a single area.
various vascular beds, where they have lodged in proportion
to the cardiac output. The hepatic arterial blood flow can be In Vivo Fluorescent Microscopy
determined directly by this method, but the portal flow con- Intravital microscopy was first described in the microvessels of
tribution is found indirectly by addition of the flow values in the frog tongue by Waller in 1846. Using this technique, indi-
the prehepatic splanchnic organs. Examination of the intrahe- vidual sinusoids and terminal venules can be visualized, and
patic distribution of microspheres has provided a means of changes in their diameters and the velocities with which eryth-
assessing the pattern of arterial flow in different liver regions rocytes pass through them can be seen (Menger & Messmer,
(Greenway & Oshiro, 1972). Because the microsphere method 1991). The introduction of fluorescent dyes has broadened the
requires the postmortem removal of the organs of interest for spectrum of in vivo microscopy in the liver from morphologic
radioactivity or colorimetric measurement, the additional analysis to the study of pathologic events. However, from a
determination of tissue weight enables flow per gram (i.e., hemodynamic point of view, intravital fluorescent microscopy
tissue perfusion) to be calculated. Microspheres may be used has problems of interpretation (Sherman et al, 1990). In per-
to determine the extent of portosystemic shunts. The frac- fused liver, a 2.5-fold increase in portal venous blood flow has
tional distribution in the liver may be measured with respect been found to be associated with only a 22% increase in sinu-
to systemic (lung) activity after PV injection, or it may be soidal RBC velocity, suggesting that changes in portal venous
estimated by injecting a second radioactive microsphere blood flow have only a minor effect on the capillary transit time
directly into the splenic or mesenteric venous system (Sherman et al, 1996).
(Groszmann et al, 1982).
Near-Infrared Spectroscopy
Hepatic Tissue Perfusion Near-infrared spectroscopy is a noninvasive technique that uses
Inert Gas Clearance light transmission and absorption to measure hemoglobin and
By exploiting the fact that radioactive gases such as krypton mitochondrial oxygenation. In contrast to visible light, which
(85Kr) and xenon (133Xe) distribute equally between tissue and can only penetrate a few millimeters, near-infrared light (700
blood according to a specific partition coefficient, the rate of to 1000 nm) can be detected through as much as 80 mm of
clearance of such gases can be measured after their injection tissue. The application of this technology to monitor liver
into the hepatic blood supply. After injection and rapid diffusion oxygenation has been validated in models of endotoxic shock
throughout the liver, the gas clears from the tissue into the blood in pigs (Nahum et al, 2006) and by intraoperative quantifica-
and is almost completely eliminated from the body after a single tion of congestion and mitochondrial redox during HV occlu-
passage through the lungs. The clearance rate is proportional to sion in living-donor transplantation (Ohdan et al, 2003).
hepatic tissue perfusion, which may be calculated by using a
standard formula (Leiberman et al, 1978). β-Emissions of 85Kr Investigational Techniques
are recorded by a Geiger-Müller tube or semiconductor (silicon) Orthogonal polarized spectral imaging has been used to obtain
detector placed on or immediately above the exposed liver images of the liver microcirculation comparable to those of fluo-
surface, whereas the γ-emissions of 133Xe are monitored trans- rescence microscopy but without the need for fluorescent dyes
cutaneously by a single scintillation crystal or a γ-camera; the (Langer et al, 2001). This system relies on the absorbance of
latter device allows simultaneous measurement of hepatic tissue hemoglobin and can discriminate a single RBC in an individual
perfusion in many regions of interest. Inert gas techniques capillary with a diameter of 5 µm. Studies with handheld devices
involve minimal trauma to the patient, and their accuracy is not using this technology have been applied to healthy living donors
markedly affected by the presence of hepatic cellular disease or to obtain sinusoidal RBC velocity and blood flow (Puhl et al,
nonperfusion shunts; however, some variability does occur, even 2003). A more recent successor to spectral imaging is sidestream
within the same subject, when multiple studies are performed. darkfield imaging, whereby a light guide surrounded by diodes
The first to use the inert gas method in the hepatic circulation can detect both red and white blood cells in the microcirculation
were Aronsen and colleagues (1968a), who recorded the without the surface reflections (Cerný et al, 2009).
Portal venous blood flow measured by magnetic resonance prioritized versus production of other hepatic proteins, such as
imaging (MRI) has been found to correlate well with flow albumin and transferrin (Sganga et al, 1985). More recently,
measured by D-US flow probes (Pelc et al, 1992). Subsequently, hemorrhagic shock has been recognized to result in generalized
the technique has been used to measure portal venous blood vascular endothelial dysfunction and impaired endothelial
flow in human liver transplantation candidates (Kuo et al, biosynthesis of NO. Endothelial NO that continues to be
1995). Further methodologic improvements in MRI with high- expressed by the liver during ischemia is believed to protect
resolution phase contrast and compressed sensing may lead to against the initial hepatic injury arising from severe hemor-
their increasing use in the study of hepatic arterial and venous rhage. By contrast, more prolonged hemorrhagic shock (>6
circulation in humans (Dyvorne et al, 2015). hours duration) induces greatly increased production of NO
Transient elastography (TE) is a well-validated technique owing to activation of an inducible NO synthase enzyme in
for the noninvasive assessment of liver fibrosis (Castera et al, hepatocytes and Kupffer cells (Peitzman et al, 1995).
2012) (see Chapter 3). Measurements are performed with a US In the transplantation era, the consequences of ischemia and
transducer built on the axis of a vibrator so that a vibration of reperfusion for the liver have been increasingly investigated
mild amplitude and low frequency is transmitted, inducing a and understood. The damage to the hepatic endothelium and
wave that propagates through the liver tissue, and pulse-echo parenchyma that results from postischemic reperfusion is
acquisitions are performed to measure the velocity of propaga- caused by numerous interrelated phenomena, including the
tion of the wave, which is directly correlated to tissue stiffness. action of locally liberated oxygen-derived free radicals and
Because fibrosis is the main determinant of tissue stiffness and excess formation of vasoconstrictor agents (Wendon, 1999).
of hepatic resistance to portal blood flow, TE has been tested Endogenous NO tends to protect the liver in the early reperfu-
in recent years as a novel way of obtaining numerical, objective, sion period after hepatic ischemia (Wang et al, 1995). Ischemia
and operator-independent noninvasive surrogate value of is also the primary signal for heat-shock protein production in
HVPG (Castera et al, 2012). In this emerging field, several liver tissue. Experimental studies of heat-shock protein precon-
publications support that high liver stiffness (≥21.1 kilopascal ditioning through intermittent portal clamping demonstrated
[kPa]) and low liver stiffness (<13.6 kPa) are accurate to respec- attenuated aminotransferase elevations and improved bile
tively rule in or out clinically significant portal hypertension, production after ischemia (Terajima et al, 1999).
but a considerable indeterminant zone (between 13.6 and 21.1
kPa) limits the use of TE in the precise assessment of HVPG Liver Atrophy
(Berzigotti et al, 2013). Liver atrophy results from a significant reduction of LBF
containing hepatotrophic substances (see Chapter 6). The
degree of atrophy depends on the degree of blood flow depriva-
CLINICAL RELEVANCE tion and may be distributed according to the source of depriva-
tion, including PV or HA blood flow or their combination.
Hemorrhagic Shock, Hypoperfusion, Atrophy and fatty degeneration of the canine liver after total
and Ischemia-Reperfusion Injury portal diversion through an Eck fistula initially was reported
Total LBF decreases approximately in relation to the degree of more than a century ago (Hahn et al, 1893). Partial or complete
the hemorrhage, and portal venous blood flow decreases in diversion of PV blood flow from the liver results in atrophy.
parallel to cardiac output; but similar to the coronary, pulmo- Complete portal venous flow diversion with interruption of all
nary, and cerebral circulations, hepatic arterial flow does not portal venous collaterals results in more profound liver atrophy
decrease until extremely low blood pressures are reached. As a than the partial deviation of portal venous flow resulting from
result, the hepatic oxygen supply tends to be maintained, side-to-side portacaval anastomoses (Bollman, 1961). A precise
although oxygen extraction greatly increases to preserve normal relationship between the degree of liver atrophy and the reduc-
total oxygen consumption (Smith et al, 1955). Hepatic blood tion of portal venous blood flow is still to be determined, pri-
volume can increase significantly in cardiac failure and can marily because of the relative composition and contribution of
compensate as much as 25% of hemorrhage from its large- portal flow and resultant compensatory hepatic arterial flow.
capacitance vessels (Lautt, 2007). Liver atrophy after portal diversion is not believed to be the
The clinical entity known as shock liver has long been recog- result of a decrease in absolute volume flow, but instead it is
nized, typically related to cardiogenic or hemorrhagic shock due to the effective loss of hepatotrophic substances in the
(Birgens et al, 1978). Hepatic dysfunction caused by hypoperfu- portal blood. Rats subjected to portal flow diversion with
sion is manifested pathologically by centrilobular necrosis and portacaval transposition had a decrease in relative liver weight
clinically by abdominal pain, cholestatic jaundice, and marked (Guest et al, 1977) despite the effective preservation of portal
elevation of serum aminotransferases. Three phases of liver perfusion from the IVC (Ryan et al, 1978). Dogs with “partial
injury attributed to ischemia were proposed by Champion and portacaval transposition” (Marchioro et al, 1967) or “splanch-
colleagues (1976), whereby the initial hepatic dysfunction nic flow division” (Starzl et al, 1973) with diversion of pancre-
would resolve as long as no additional insults (e.g., sepsis) were aticogastroduodenosplenic blood had atrophy in liver lobes,
incurred. Gottlieb and colleagues (1983) showed that hepatic although normal tissue perfusion was shown in all regions of
dysfunction in humans after trauma was related to a reduced the liver (Mathie et al, 1979).
hepatic blood flow rate as much as 70% of resting levels. Hepatic Histologically, arterial obstruction results in ischemic
blood flow was markedly reduced after injury, and although total changes, such as mitochondrial swelling, cell membrane disrup-
splanchnic oxygen delivery was decreased, oxygen consumption tion, platelet aggregation, and widening of the spaces of Disse
remained normal as a result of increased hepatic extraction. (Mallet-Guy et al, 1972). The fate of the liver after ligation of
In the ischemic state, upregulation of acute-phase proteins the HA depends largely on the extent of a functional collateral
(C-reactive protein, fibrinogen, ceruloplasmin, haptoglobin) is arterial circulation (Rappaport & Schneiderman, 1976). If
limited collaterals are present, liver infarction and necrosis may secondary biliary fibrosis (see Chapter 76) and shock after
occur after HA ligation, resulting in death. However, HA liga- biliary tract decompression. Approximately 20% of patients
tion results only in transient ischemic changes in the periphery with prolonged biliary obstruction experience clinically signifi-
of the hepatic acinus (zone III, see Fig. 5.3) in the presence of cant portal hypertension (Blumgart et al, 1984; Sedgwick et al,
adequate collaterals. Atrophy after HA ligation can occur 1966). The operative risk of biliary decompression in these
in liver segments that have compensatory collateral supply patients is significant. Technical difficulties of stricture repair—
to prevent necrosis. The effects of HA flow absence are magni- dense fibrous adhesions, hilar ductal involvement, and
fied by the presence of low portal venous blood flow, limited infection—are complicated by the risk of hemorrhage from
oxygen saturation, and superimposed infection (Rappaport & subhepatic and periductal varices and potential postoperative
Schneiderman, 1976). liver failure. In addition to the hemodynamic consequences of
chronic bile duct obstruction, sudden decompression of the
Impact of Acute and Chronic Bile Duct Obstruction on obstructed biliary tree also causes an abrupt decrease in
Liver Blood Flow wedged HV pressure, PV pressure, and arterial pressure within
Bile duct obstruction can affect hepatic hemodynamics signifi- 30 minutes of decompression in jaundiced dogs, leading to
cantly. In general, LBF is reduced in the presence of chronic hypotension and shock (Tamakuma et al, 1975). Similarly,
biliary obstruction, leading to hepatic dysfunction. Conversely, Steer and colleagues (1968) reported that rapid needle decom-
acute increases in bile duct pressure from early obstruction pression of an obstructed biliary tree in jaundiced dogs induced
result in a reflexive increase in LBF, which attempts to maintain a decreased arterial pressure, central venous pressure, and
adequate flow in the face of an increased pressure gradient portal venous pressure within 1 hour; they concluded that
opposing secretion and excretion of bile (see Chapter 8). Most sudden decompression of chronic biliary obstruction leads to
evidence suggests that the hemodynamic response of the liver sequestration of fluid within the liver, resulting in a decrease
to biliary obstruction is related, directly or indirectly, to changes in the effective circulating plasma volume and subsequent
in bile duct pressure. Given the limited space of Mall in the hypotension.
portal triad (see Fig. 5.1), it is conceivable that increased biliary
pressures may compress the portal capacitance vessels, leading Liver Resection and Regeneration
to increased arterial flow (Kanda et al, 1996). Acute serial The adult liver exhibits a remarkable potential to restore its
increases in bile duct pressure in dogs with complete bile duct cellular mass in response to injury through hepatocyte hyper-
obstruction increased hepatic arterial blood flow by 250% but plasia. Hepatic regeneration of the normal liver remnant pro-
did not affect portal venous blood flow. The precise mechanism ceeds rapidly after partial hepatic resection (Aronsen et al,
for reduction in LBF after chronic bile duct obstruction is 1970; Blumgart et al, 1971) (see Chapters 6 and 108D). Partial
unknown. Although increased portal vascular resistance is the liver resection without devascularization normally produces
accepted underlying cause, the primary site of this resistance little change in total blood flow to the liver. This occurs because
change has been considered to be presinusoidal (Reuter & the major contributor to total flow, the PV, is affected less by
Chuang, 1976), sinusoidal (Bosch et al, 1983), or postsinusoi- events occurring within the liver than by control mechanisms
dal (Tamakuma et al, 1975). The development of significant in the arterial resistance vessels of the prehepatic splanchnic
portosystemic shunting with an inverse correlation between bed. On the other hand, the failure of the liver to directly
shunting and portal venous blood flow can also contribute to control its portal venous flow may result in portal hyperperfu-
extrahepatic shunting (Bosch et al, 1983; Ohlsson, 1972). sion of a reduced parenchymal mass. Because essentially the
Relief of long-term obstruction does not result in the return same total blood flow is redistributed to a smaller mass of liver
of normal hemodynamics, suggesting irreversible intrahepatic tissue, a corresponding increase in tissue perfusion (mL/min
vascular damage (Aronsen et al, 1969). Furthermore, a 23% per unit tissue weight) would be anticipated in the in situ
reduction in effective LBF persisted for 1 to 5 years after remnant. Experimental studies support these expectations; an
operative decompression in patients with choledocholithiasis increase in hepatic tissue perfusion was observed in rats imme-
and jaundice for more than 2 weeks preoperatively (Aronsen, diately after two-thirds hepatectomy (Rice et al, 1977; Wheatley
1968b). Hunt (1979) serially measured LBF daily for 1 week et al, 1993; Wu et al, 1993). This increase in hepatic perfusion
after bile duct ligation in rats, using the 133Xe clearance tech- is due primarily to portal venous inflow, because hepatic arterial
nique to document the early hemodynamic response. Total LBF blood flow is low, and hepatic arterial resistance is high even
decreased steadily after the first postoperative day to a plateau 24 hours after partial hepatectomy in rats.
level of approximately 50% of the preoperative value 5 days In humans, an immediate increase in tissue perfusion of
after operation. Mathie and colleagues (1988) confirmed the approximately 120% occurs in the liver remnant (Mathie &
decrease in total LBF after bile duct ligation by measuring the Blumgart, 1982). A 60% partial hepatectomy results in a
individual portal venous and hepatic arterial components of doubling of the portal flow in the 40% of remnant liver tissue
LBF. Using electromagnetic flowmeters in dogs with complete (Troisi et al, 2005). Experimental evidence has suggested that
bile duct ligation, hepatic arterial and portal venous blood flow intrahepatic shear stress from increased portal flow is a regula-
were observed to decrease by 36% and 44%, respectively; they tor of liver regeneration (Nobuoka et al, 2006; Schoen et al,
also showed a 200% increase in intrahepatic portal resistance 2001). The significance of blood flow in relation to liver regen-
but a lesser increase in hepatic arterial resistance. Similarly, eration, however, continues to be debated since Mann (1944)
dogs with chronic bile duct ligation had decreased portal venous suggested that regenerative hyperplasia of the liver after partial
flow and had developed sinusoidal portal hypertension and resection was a function of portal blood flow and that the
extensive portosystemic shunting (Bosch et al, 1983). process could be prevented by portal flow diversion. However,
Chronic biliary obstruction can thus result in two hemody- regenerative hyperplasia normally occurs after partial liver
namic consequences: portal hypertension associated with resection in portacavally transposed animals, in which there is
no direct supply of portal blood nor the usual posthepatectomy et al, 2003). This phenomenon describes the impaired HA flow
increase in hepatic tissue perfusion (Fausto, 2000). by shifting of the main blood flow to the splenic or gastroduo-
denal artery in patients with hypersplenism. In whole-organ
Liver Blood Flow and Hemodynamic Studies liver recipients analyzed by D-US, HA vasoconstriction in
in Liver Transplantation response to portal hyperperfusion and exaggerated HABR
The HA buffer reponse is conserved following orthotopic liver produces a high resistive index with poor arterial perfusion
transplantation (OLT) despite denervation (Eipel et al, 2010; (Quintini et al, 2008). In a retrospective analysis of 650 OLTs,
Henderson et al, 1992). In a series of experiments by Payen a 5.1% incidence of splenic artery syndrome has been reported,
and colleagues (1990), serial clamping of the PV every 12 hours and prophylactic treatment with ligation of the splenic artery
for 7 days following OLT resulted in reciprocal increases in for patients at risk for development of splenic artery syndrome
hepatic arterial flow. OLT of a normal donor organ does not has been advocated (Mogl et al, 2010). A prospective study has
normalize the splanchnic and systemic hemodynamic altera- suggested that preoperative embolization of the splenic artery
tions of end-stage liver disease (Henderson, 1993). In fact, total leads to improved postoperative living-donor graft function
hepatic blood flow remains elevated 6 months after OLT (Umeda et al, 2007). Because splenic artery occlusion reduces
(Hadengue et al, 1993; Navasa et al, 1993) mainly because of the resistance to distal HA flow by reducing flow in the splenic
portal venous blood flow (Henderson et al, 1992; Paulsen & circulation, and consequently portal venous flow, it has been
Klintmalm, 1992). This suggests that baseline LBF may be suggested to revise the name of splenic artery steal syndrome
under direct sympathetic control, which is lost after OLT, to splenic artery syndrome, thereby underlining that the cause
leading to an unopposed rise. Azygos flow also remains elevated, is portal hyperperfusion and not arterial siphoning (Quintini
and other portosystemic shunts have been documented up to et al, 2008).
4 years after OLT (Navasa et al, 1993). Ligation of these por-
tosystemic collateral pathways has been shown to increase Portal Hypertension
portal venous blood flow (Fujimoto et al, 1995). Portal hypertension is a state of sustained increase in the
The hemodynamic consequences of OLT in human patients intraluminal pressure of the PV and its collaterals, associated
are difficult to interpret for several reasons: (1) The causes of with the most severe complications of cirrhosis, including
liver failure in end-stage transplant candidates are diverse; ascites, hepatic encephalopathy, and bleeding from gastro-
(2) immunosuppressive drugs are used, such as cyclosporine, esophageal varices (de Franchis et al, 2010) (see Chapters 76,
which causes arterial hypertension; and (3) to control systemic 81, and 82). A mean HVPG greater than 12 mm Hg has clas-
hypertension after OLT, patients may also be given vasodilators, sically been used to define portal hypertension, as variceal
which can cause persistently increased cardiac output. Cardiac bleeding does not occur at lower pressures than this (Garcia-
output data have been conflicting, with one group reporting Tsao et al, 1985). Measurement of HVPG is now considered
persistently elevated values (Hadengue et al, 1993) and others as one of the best surrogates of clinical events, using different
showing decreases 2 weeks and 2 months after OLT (Navasa thresholds to guide the management of cirrhotic patients, and
et al, 1993). Gadano and colleagues (1995) emphasized that a value of 10 mm Hg and greater is predictive of varices
factors such as anemia and sepsis may account for the deranged formation and liver decompensation (Table 5.2) (Berzigotti
hemodynamics after OLT. In a recent retrospective series of et al, 2013).
970 patients, new-onset heart failure developed in as many as
10% of patients following OLT after a median follow-up of 5.3 Hemodynamics
years, the majority of which were of nonischemic etiology Hemodynamic factors that influence portal hypertension are
(Qureshi et al, 2013). best understood by the flow-resistance principle that applies to
the portal venous system. Portal pressure depends on two basic
Small-for-Size Syndrome components: portal blood flow and hepatic portal vascular
The advent of living-donor partial liver transplantation and the resistance. Portal hypertension may result from a significant
enlargement of the resectable limit has introduced the phenom- increase in hepatic portal inflow from the prehepatic splanchnic
enon of small-for-size syndrome, whereby the pressure of the vasculature, an increase in intrahepatic portal resistance
full portal flow traveling through a small liver remnant leads to (IHPR), or both. Although simple in concept, multiple factors
a marked decreased arterial inflow, hypothesized to result from may influence both the components of the system and the
an intact HABR (Michalopoulos, 2010). In a porcine model, pathophysiology of portal hypertension.
portal flow to split grafts with a graft-to-recipient liver volume Increased portal pressure, diminished hepatic portal blood
ratio of 2 : 3 and 1 : 3 was inversely correlated to graft size flow, and an extensive extrahepatic collateral venous network
(Smyrniotis et al, 2002). In patients with right lobe living- supplied by a hyperdynamic splanchnic and systemic circula-
donor transplantation, the grafts are subjected to more than tion characterize the hemodynamics of portal hypertension in
double increases of portal blood flow, whereas the arterial flow cirrhotic patients. Extrahepatic shunts may account for at least
is strikingly decreased, likely to maintain total blood flow within 50% of the portal flow, whereas 80% of portal flow actually
an acceptable physiologic range (Marcos et al, 2000; Rocheleau reaching hepatocytes has been observed to bypass the sinusoidal
et al, 1999). The consequence of inadequate hepatic arterial vascular bed via intrahepatic shunts (Okuda et al, 1977). The
flow range from mild cholestasis and delayed synthetic function magnitude of extrahepatic shunt flow in patients with cirrhosis
to ischemic cholangitis and parenchymal infarct (Demetris was measured directly by thermal dilution assessment of azygos
et al, 2006) (see Chapter 120). blood flow; a value 300 mL/min greater than in patients without
Although arterial flow impairment appears to result from an portal hypertension was noted (Bosch & Groszmann, 1984).
active HABR, it has repeatedly been ascribed to the splenic The HA probably provides a greater relative contribution to the
artery steal syndrome in the past (Manner et al, 1991; Nüssler total LBF in patients with cirrhosis than in healthy individuals,
TABLE 5.2 Prognostic Significance of Hepatic Venous Pressure Gradient Thresholds According to the Compensated or Decompensated
Stage of Cirrhosis
Clinical Setting HVPG (mm Hg) Increased Risk of Threshold
Compensated cirrhosis 10 Presence (Garcia-Tsao et al, 1985) and development of gastroesophageal varices
(Groszmann, 2005)
First clinical decompensation in patients without varices (Ripoll, 2007)
Development of HCC (Ripoll, 2009)
Decompensation after surgery for HCC (Bruix, 1996; Llovet, 1999)
12 Variceal bleeding (Garcia-Tsao et al, 1985; Gluud, 1988; Lebrec, 1980; Merkel, 1992;
Vorobioff, 1996)
16 First clinical decompensation in patients with varices (Berzigotti, 2011) and mortality
(Merkel, 2000)
Decompensated cirrhosis 16 Variceal rebleeding and mortality (Stanley, 1998)
20 Failure to control variceal bleeding in patients actively bleeding from varices (Abraldes,
2008; Moitinho, 1999)
22 Mortality in patients with alcoholic cirrhosis and acute alcoholic hepatitis (Rincon, 2007)
30 Spontaneous bacterial peritonitis (Sersté, 2006)
HCC, Hepatocellular carcinoma.
From Berzigotti A, et al: Assessing portal hypertension in liver diseases. Expert Rev Gastroenterol Hepatol 7:141-155, 2013.
although it also was shown that 33% of the arterial blood may mediators, cytokines, growth factors, and endothelin (Lee et al,
flow through intrahepatic shunts to the systemic venous circula- 2015) (see Chapter 11). Capillarization of sinusoidal endothe-
tion (Groszmann et al, 1977). Vallance and Moncada proposed lial cells occurs by defenestration or loss of endothelial cell
the hypothesis that the low peripheral vascular resistance in pores and the appearance of a basement membrane (Varin &
portal hypertension may be caused by the stimulated produc- Huet, 1985). In the cirrhotic liver, the sites of vascular resistance
tion of NO induced by endotoxemia in 1991. The initial are still unclear. However, because portal and hepatic venules
experimental evidence was provided by Pizcueta and others can be found within fibrous septa, constriction or distortion of
(1992), who showed an increase in systemic and splanchnic portal venules, hepatic venules, or both may be involved (Kelty
vascular resistance in cirrhotic rats after the administration of et al, 1950). Disruption of hepatic architecture, with the devel-
an NO inhibitor. Although endotoxemia is likely responsible for opment of fibrous septa and abnormal nodules and circulation
the decompensation in end-stage cirrhosis, the hemodynamic leads to a sustained IHPR and portal pressure. While portal
alterations are due to NO in animal models (Lee et al, 1996; venous blood flow progressively decreases in cirrhosis, arterial
Niederberger et al, 1995). Constitutive NO synthase seems to resistance decreases and arterial flow increases, suggesting an
be upregulated in discrete anatomic locations, such as in the intact buffer response. Studies in cirrhotic rats have demon-
endothelium of the mesenteric artery and in the esophageal, strated higher hepatic arterial flows compared with normal
gastric, and jejunal mucosa (Mathie, 1999). control rats under baseline conditions (Richter et al, 2000).
The direction of portal blood flow has been proposed as a This finding was confirmed by using intraoperative measure-
contributor to the pathophysiology of portal hypertension. The ments in patients with end-stage cirrhosis undergoing living-
progression of intrahepatic disease and increasing sinusoidal donor liver transplantation (Aoki et al, 2005). Clinically, the
pressure has been postulated as contributing to reversal of flow vasodilaton of the splanchnic circulation likely serves to increase
in the PV, which further aggravates the injury by depriving the flow in the extrahepatic collateral circulation, leading to variceal
liver of nutrients (Warren & Muller, 1959). However, reversed, hemorrhage.
or hepatofugal, PV flow is relatively infrequent. Gaiani and
colleagues (1991) reported an incidence of 3.1% in 228 patients Treatments
assessed, whereas intrahepatic portal flow reversal has been Medical and surgical management strategies for portal hyper-
described in as many as 9% of patients and seen almost exclu- tension strive to improve patient survival by the reduction of
sively in those with Child-Turcotte-Pugh C cirrhosis (Tarantino pressure and flow in extrahepatic variceal vessels, mainly
et al, 1997). esophageal and gastric vessels, while preserving adequate portal
In portal hypertension, the increased mesenteric blood flow flow to the liver (see Chapters 81 to 87). Portosystemic shunting
in the hyperdynamic stage may be relatively less important and pharmacologic reduction of portal flow can provide effec-
than the elevated IHPR caused by the interplay of local regula- tive decompression, but both deprive the liver of portal flow.
tory mechanisms affecting sinusoidal hemodynamics (Grosz- Multiple pharmacologic agents have been investigated to
mann 1990). The traditional view of the source of increased reduce portal hypertension by diminishing hepatic portal inflow
portal pressure is fibrotic encroachment around portal radicles, from the mesenteric vascular bed. At a dose that decreases the
leading to increased IHPR. The pathogenesis of cirrhosis heart rate by 25%, the β-blocker propranolol significantly
involves initial hepatocyte necrosis and inflammation with reduced the risk of rebleeding in cirrhotic patients who were
subsequent transformation of HSCs into myofibroblasts. HSC otherwise in good condition (Lebrec et al, 1981) (see Chapter
activation results in the collagenization of the space of Disse 82). Propranolol exerts its action by two mechanisms: decreased
(Martinez-Hernandez, 1985) (see Chapter 7). Several factors cardiac output as a result of β1-adrenergic cardiac receptor
have been implicated in HSC activation, such as inflammatory blockade and antagonism of β2-adrenoceptors in the splanchnic
vasculature, which leaves the vasoconstrictive influence of PREVENTION OF VARICEAL BLEEDING. All patients with cirrhosis
α-adrenergic receptors unopposed, resulting in a decrease should be screened by endoscopy for varices at diagnosis. The
portal flow and pressure. Vasopressin causes generalized periph- treatment of underlying liver disease, when possible, may
eral vasoconstriction (Bosch et al, 1988), whereas the effect of reduce portal hypertension and prevent its clinical complica-
somatostatin is specific to the splanchnic vascular bed (Kravetz tions. There is no pharmaceutical agent proven effective to
et al, 1984) and resultsg from glucagon-release inhibition and prevent the formation of varices. Patients with small varices
direct vasoconstriction. Serotonin may play a significant role in with red marks or Child-Turcotte-Pugh C cirrhosis have an
maintaining increased portal pressure, and smooth muscle increased risk of bleeding and should thus be treated with
serotonin-receptor antagonists have been shown to lower the nonselective β-blockers. Patients with medium to large varices
pressure in cirrhosis (Hadengue et al, 1987; Mastaï et al, 1989). also can benefit from endoscopic band ligation for the preven-
Although it was originally thought that IHVR in cirrhosis tion of the first variceal bleeding episode (primary prophylaxis)
was irreversible, evidence has supported that it can be reduced (see Chapter 82). In centers where adequate resources and
pharmacologically (Bhathal & Grossman, 1985). The nitrova- expertise are available, HVPG measurements can routinely be
sodilators isosorbide dinitrate and isosorbide mononitrate were used for prognostic and therapeutic indications (see Table 5.2).
observed to lower the portal pressure in portal hypertensive A decrease in HVPG of at least 20% from baseline or to
animals (Blei & Gottstein, 1986) and to increase hepatic (but 12 mm Hg or less after chronic treatment with a nonselective
not azygos) blood flow in patients with cirrhosis (Navasa et al, β-blocker has been demonstrated clinically relevant in the
1989), suggesting that they may act by reducing intrahepatic setting of primary prophylaxis of variceal bleeding.
portal vascular resistance. Application of nitroglycerin by
transdermal tape to patients with cirrhosis resulted in a reduc- TREATMENT OF ACUTE BLEEDING (SEE CHAPTER 83). Critical initial
tion in portal pressure without affecting hepatic blood flow steps include airway protection, particularly in patients with
(Iwao et al, 1991), and IV nitroglycerin caused a 24% decrease altered mental status or those with hemodynamic instability,
in IHVR in patients with cirrhosis (Groszmann, 1990). In resuscitation with fluid and blood products, and correction of
animal models, IHVR can be reduced by prostaglandin E2, the coagulopathy and thrombocytopenia. Patients with variceal
endothelin receptor antagonist isoprenaline, nitroprusside, hemorrhage are often bacteremic as a result of a concomitant
papaverine, and verapamil. (Ballet, 1991; Bhathal & Grossman, infectious process (spontaneous bacterial peritonitis, urinary
1985; Reichen & Le, 1986; Reichen et al, 1998). tract infection, or pneumonia), and clinical trials have shown
The use of TIPS, first reported by Rössle and colleagues better outcomes when empiric antibiotic therapy is initiated
(1989), now largely replaces shunt surgery. TIPS is currently early. Vasoconstrictive drugs that reduce portal pressure (soma-
the treatment of choice for recurrent variceal bleeding in tostatin and vasopressin analogues) and endoscopic variceal
patients who are refractory to conservative medical manage- ablation (ligation and sclerotherapy) are the mainstay of initial
ment (Parker, 2014) (see Chapter 87). management (see Chapters 82 and 83). As shown in single trials
Surgical treatment of portal hypertension may be performed and by meta-analysis, this initial strategy controls 80% to 85%
by one of the many portosystemic shunt procedures; the initial of bleeding episodes (Gonzalez et al, 2008). Early assessment
clinical application of the portacaval shunt was reported 50 of prognosis is important to guide further management, because
years after its description by Eck (Whipple, 1945) (see Chapter patients at high risk of treatment failure benefit from early TIPS
85 and 86). The hemodynamic consequences of shunt surgery placement (within 72 hours) (García-Pagán et al, 2010) (see
depend on the particular shunt performed, the nature and Chapter 87). An HVPG of 20 mm Hg or higher, Child-Pugh
severity of the disease, and the hemodynamic condition of the class C, and active bleeding at endoscopy are the variables most
patient. End-to-side portacaval shunts divert all portal blood consistently found to predict 5 day rebleeding treatment failure
flow away from the liver, whereas less complete diversions (Bambha et al, 2008). TIPS may not be an option in some cases,
reduce portal flow in proportion to the degree to which the for example, in face of portal thrombosis, in which situation a
shunt reduces portal pressure. The HA flow may increase by surgical shunt or a devascularization procedure (Dagenais et al,
100%, but even a maximal flow increase can usually only partly 1994) is indicated (see Chapters 84, 85, and 86). Emergency
compensate for loss of portal flow (Mathie & Blumgart, 1983b). portacaval shunt has a success rate of 95% in stopping bleeding
Hepatic oxygen consumption tends to be maintained by in this context. The death rate of the operation is, however, not
increased oxygen extraction from the available arterial supply. insignificant, but generally related to the status of the patient’s
Total portacaval shunts are very effective in reducing portal liver function. Approximately 40% of patients experience
pressure and preventing bleeding from esophageal varices. encephalopathy after portacaval surgical shunting. Hepatic
However, because of bypass of the hepatic circulation, liver insufficiency is accelerated, and liver failure is the cause of death
failure and encephalopathy are common complications of the in approximately two-thirds of those who die after an emergency
operation. Therefore partial shunts, such as the side-to-side, portacaval shunt. Balloon tamponade should only be used in
mesocaval, and proximal or distal splenorenal shunts, are pre- massive bleeding as a temporary measure until definitive treat-
ferred when technically feasible. Selective shunts, such as the ment is instituted (Garcia-Tsao & Bosch, 2010).
distal splenorenal (Warren) shunt, in which the gastrosplenic
collaterals are decompressed via the splenic vein into the left Blood Flow in Hepatic Tumors
renal vein, leaving the PV intact, are affective but usually too It has been demonstrated that tumors of the liver, whether
time consuming for use in emergency operations. primary or secondary, are perfused almost exclusively with
Current consensus on treatment approaches are here sum- arterial blood (Breedis & Young, 1954). Hepatocellular carci-
marized for prevention and treatment of variceal bleeding in noma and liver metastases from neuroendocrine tumors are,
patients with cirrhosis (de Franchis et al, 2010; Garcia-Tsao & however, more arterialized and less necrotic than most other
Bosch, 2010). primary or metastatic liver tumors. The arterial uptake of
hepatic solid lesions thus provides important differential diag- can also be effectively used to deliver high doses of chemo-
nosis information when assessed by computed tomography therapy directly to the liver for the treatment of patients with
scan and MRI performed with IV contrast, provided that colorectal cancer liver metastasis and unresectable intrahepatic
images are acquired at time of arterial enhancement in addition cholangiocarcinoma (Goéré et al, 2010; Jarnagin et al, 2009;
to venous enhancement (Forner et al, 2008) (see Chapters 18, Kemeny, 2013; Kemeny et al, 2009) (see Chapter 99).
19, 91, and 93). The neovasculature of tumor tissue lacks
smooth muscle and therefore does not respond to vasocon Effect of Laparoscopy on Liver Blood Flow
strictor agents, enabling increased delivery and retention of The use of a CO2 pneumoperitoneum in laparoscopic surgery
chemotherapeutic drugs. The vasoconstrictors epinephrine, has been demonstrated to substantially reduce portal venous
phenylephrine, and angiotensin have improved cytotoxic drug flow in parallel with the intraperitoneal pressure (Jakimowicz
delivery to liver tumors in human and animal models, but this et al, 1998; Schilling et al, 1997). Conflicting data exist in
has not translated into common clinical practice (Bloom et al, support of the maintenance of the HABR effect in high-pressure
1987; Goldberg et al, 1991; Hemingway et al, 1991). pneumoperitoneum. Rat models using fluorescent microspheres
A variety of transarterial techniques have been used to supported preserved hepatic arterial flow during decreased
selectively embolize and deliver chemotherapy to liver tumors, portal venous flow (Yokoyama et al, 2002), but this was not
taking advantage of the fact that they derive disproportionately supported by others (Richter et al, 2001), who saw a parallel
greater blood supply from the hepatic arterial circulation com- decrease in arterial and portal venous flows during laparoscopy.
pared with the surrounding liver (see Chapter 96). Emboliza- Most surgical teams use a 10 to 14 mm Hg pneumoperitoneum
tion is often performed with Gelfoam, which dissolves after a during laparoscopic liver resection, which allows good control
few weeks, but other inert agents are also used and are probably of bleeding (Tranchart et al, 2015) (see Chapter 105). A case-
more effective for occluding vessels. Some centers use inert matched analysis suggested that the positive pressure of pneu-
particles without chemotherapy (i.e., bland embolization), but moperitoneum was probably the main factor explaining the
most combine the procedure with chemotherapy (i.e., transar- decreased blood loss during laparoscopy when compared with
terial chemoembolization, or TACE). Doxorubicin, mitomycin, open liver surgery (Tranchart et al, 2013). Some groups have
and cisplatin in various combinations are the drugs most often suggested the avoidance of head-up positioning and pressures
given. The embolized material causes temporary blood flow greater than 15 mm Hg during laparoscopy to preserve LBF
interruption and potentially improves the uptake of chemo- (Junghans et al, 1997; Klopfenstein et al, 1998). So far, no
therapeutic agents in tumor tissue and, consequently, reduces major accident related to the use of a CO2 pneumoperitoneum
systemic toxicity (Liapi & Geschwind, 2010). Lipiodol and during laparoscopic liver resection has been reported in
drug-eluting beads, which lodge in the tumor, have also been approximately 6000 cases (Dagher et al, 2014). Furthermore,
used as a carrier for chemotherapy. More recently, radioactive a swine model has been used to prospectively demonstrate that
microspheres emitting yttrium-90 have been used. These pro- multiple gas embolisms frequently occur during laparaoscopic
cedures are mainly performed in the palliative setting for liver resection without significant modification of hemodynam-
patients with hepatocellular carcinoma not amenable to liver ics (Jayaraman et al, 2009; Schmandra et al, 2004).
resection or transplantation. It remains unclear if the addition
of chemotherapeutic agents provides much benefit beyond the Acknowledgements
necrosis produced by occlusion of the hepatic arterial supply Thank you to Drs. Blumgart, Wheatley, Mathie, and Rocha for
alone (Lammer et al, 2010; Llovet et al, 2002; Lo et al, 2002; their previous contributions to this chapter.
Malagari et al, 2010; Maluccio et al, 2008; Salem et al, 2011).
Hepatic arterial infusion of chemotherapy delivered by References are available at expertconsult.com.
catheters connected to subcutaneously placed ports or pumps
Chapter 5 Liver blood flow: physiology, measurement, and clinical relevance 92.e1
REFERENCES through gastroesophageal collaterals in cirrhosis, Hepatology 4:424–

429, 1984.
Abraldes JG, et al: Utility of color Doppler ultrasonography predicting Bradley SE, et al: The estimation of hepatic blood flow in man, J Clin
tips dysfunction, Am J Gastroenterol 100:2696–2701, 2005. Invest 24:890–897, 1945.
Abraldes JG, et al: Hepatic venous pressure gradient and prognosis in Braet F, Wisse E: Structural and functional aspects of liver sinusoidal
patients with acute variceal bleeding treated with pharmacologic and endothelial cell fenestrae: a review, Comp Hepatol 1:1, 2002.
endoscopic therapy, J Hepatol 48(2):229–236, 2008. Bredfeldt JE, et al: Compensatory mechanisms in response to an ele-
Akita H, et al: Real-time intraoperative assessment of residual liver vated hepatic oxygen consumption in chronically ethanol-fed rats,
functional reserve using pulse dye densitometry, World J Surg Am J Physiol 248:G507–G511, 1985.
32:2668–2674, 2008. Breedis C, Young G: The blood supply of neoplasms in the liver, Am J
Almond NE, Wheatley AM: Measurement of hepatic perfusion in the Pathol 30:969–985, 1954.
rat by laser Doppler flowmetry, Am J Physiol 262:G203–G209, 1992. Bruix J, et al: Surgical resection of hepatocellular carcinoma in cirrhotic
Aoki T, et al: Intraoperative direct measurement of hepatic arterial patients: prognostic value of preoperative portal pressure, Gastroen-
buffer response in patients with or without cirrhosis, Liver Transpl terology 111(4):1018–1022, 1996.
11:684–691, 2005. Burton-Opitz R: The vascularity of the liver: I. The flow of blood in the
Aronsen KF: Late effects of biliary stasis on the effective liver blood hepatic artery, Q J Exp Physiol 3:297–313, 1910.
flow, Acta Chir Scand 134:278–281, 1968b. Burton-Opitz R: The vascularity of the liver: the influence of the portal
Aronsen KF, et al: Disappearance rate of 133Xe from dog liver with blood flow upon the flow in the hepatic artery, Q J Exp Physiol
different routes of administration, Nucl Med (Stuttg) 7:119–124, 4:93–102, 1911.
1968a. Caesar J, et al: The use of indocyanine green in the measurement of
Aronsen KF, et al: Liver blood flow studies during and after various hepatic blood flow and as a test of hepatic function, Clin Sci 21:43–57,
periods of total biliary obstruction in the dog, Acta Chir Scand 1961.
135:55–59, 1969. Castera L, et al: Noninvasive evaluation of portal hypertension using
Aronsen KF, et al: Evaluation of hepatic regeneration by scintillation transient elastography, J Hepatol 56:696–703, 2012.
scanning, cholangiography and angiography in man, Ann Surg 171: Cerný V, et al: In situ assessment of the liver microcirculation in
567–574, 1970. mechanically ventilated rats using sidestream dark-field (SDF)
Balfour DC Jr, et al: Hepatic vein pressure studies for evaluation of imaging, Physiol Res 58:49–55, 2009.
intrahepatic portal hypertension, AMA Arch Surg 68:442–447, 1954. Champion HR, et al: A clinicopathologic study of hepatic dysfunction
Ballet F: Hepatic resistance in isolated perfused normal and cirrhotic following shock, Surg Gynecol Obstet 142:657–663, 1976.
liver. In Ballet F, Thurman RG, editors: Research in perfused liver, Clavien PA, et al: Strategies for safer liver surgery and partial liver
London, 1991, INSERM/John Libbey, pp 339–360. transplantation, N Engl J Med 356:1545–1559, 2007.
Bambha K, et al: Predictors of early re-bleeding and mortality after Combes B: Estimation of hepatic blood flow in man and dogs by
acute variceal haemorrhage in patients with cirrhosis, Gut 57:814– I131-labelled rose bengal, J Lab Clin Med 56:537–543, 1960.
820, 2008. Dagenais M, et al: Experience with radical esophagogastric devascular-
Bauer M, et al: ET-1 induced alterations of hepatic microcirculation: ization procedures (Sugiura) for variceal bleeding outside Japan,
sinusoidal and extrasinusoidal sites of action, Am J Physiol 267: World J Surg 18:222–228, 1994.
G143–G149, 1994. Dagher I, et al: International experience for laparoscopic major liver
Berzigotti A, et al: Prognostic value of a single HVPG measurement resection, J Hepatobiliary Pancreat Sci 21:732–736, 2014.
and Doppler-ultrasound evaluation in patients with cirrhosis and Dauzat M, Layrargues GP: Portal vein blood flow measurements using
portal hypertension, J Gastroenterol 46(5):687–695, 2011. pulsed Doppler and electromagnetic flowmetry in dogs: a compara-
Berzigotti A, et al: Assessing portal hypertension in liver diseases, Expert tive study, Gastroenterology 96:913–919, 1989.
Rev Gastroenterol Hepatol 7(2):141–155, 2013. de Franchis R, Baveno V Faculty: Revising consensus in portal hyper-
Bhathal PS, Grossman HJ: Reduction of the increased portal vascular tension: report of the Baveno V consensus workshop on methodology
resistance of the isolated perfused cirrhotic rat liver by vasodilators, of diagnosis and therapy in portal hypertension, J Hepatol 53:762–
J Hepatol 1:325–337, 1985. 768, 2010.
Biber B, et al: Portal blood flow in man during surgery, measured by Demetris AJ, et al: Pathophysiologic observations and histopathologic
a modification of the continuous thermodilution method, Scand J recognition of the portal hyperperfusion or small-for-size syndrome,
Gastroenterol 18:233–239, 1983. Am J Surg Pathol 30:986–993, 2006.
Biernat J, et al: Role of afferent nerves and sensory peptides in the Dobson EL, Jones HB: The behaviour of intravenously injected particu-
mediation of hepatic artery buffer response, J Physiol Pharmacol late material: its rate of disappearance from the blood stream as a
56:133–145, 2005. measure of liver blood flow, Acta Med Scand Suppl 273:1–71, 1952.
Birgens HS, et al: The shock liver. Clinical and biochemical findings in Dyvorne HA, et al: Quantification of hepatic blood flow using a high-
patients with centrilobular liver necrosis following cardiogenic shock, resolution phase-contrast MRI sequence with compressed sensing
Acta Med Scand 204:417–421, 1978. acceleration, AJR Am J Roentgenol 204:510–558, 2015.
Blei AT, Gottstein J: Isosorbide dinitrate in experimental portal hyper- Ebbing C, et al: Hepatic artery hemodynamics suggest operation of a
tension: a study of factors that modulate the hemodynamic response, buffer response in the human fetus, Reprod Sci 15:166–178, 2008.
Hepatology 6:107–111, 1986. Eipel C, et al: Regulation of hepatic blood flow: the hepatic arterial
Bloom ND, et al: Enhancement of tumor blood flow and tumoricidal buffer response revisited, World J Gastroenterol 16:6046–6057, 2010.
effect of doxorubicin by intraportal epinephrine in experimental liver Ekataksin W, Kaneda K: Liver microvascular architecture: an insight
metastasis, Arch Surg 122:1269–1272, 1987. into the pathophysiology of portal hypertension, Semin Liver Dis
Blumgart LH, et al: Observations on liver regeneration after right 19:359–382, 1999.
hepatic lobectomy, Gut 12:922–928, 1971. Ezzat WR, Lautt WW: Hepatic arterial pressure-flow autoregulation is
Blumgart LH, et al: Benign bile duct stricture following cholecystec- adenosine mediated, Am J Physiol 252:H836–H845, 1987.
tomy: critical factors in management, Br J Surg 71:836–843, 1984. Fausto N: Liver regeneration, J Hepatol 32:19–31, 2000.
Bollman JL: The animal with an Eck fistula, Physiol Rev 41:607–621, Fiorucci S, et al: The third gas: H2S regulates perfusion pressure in
1961. both the isolated and perfused normal rat liver and in cirrhosis,
Bosch J, et al: Chronic bile duct ligation in the dog: hemodynamic Hepatology 42:539–548, 2005.
characterization of a portal hypertensive model, Hepatology 3:1002– Fleming JS, et al: In vivo assessment of hepatic-arterial and portal-
1007, 1983. venous components of liver perfusion: concise communication,
Bosch J, et al: Effects of vasopressin on the intravariceal pressure in J Nucl Med 22:18–21, 1981.
patients with cirrhosis: comparison with the effects on portal pres- Forner A, et al: Diagnosis of hepatic nodules 20 mm or smaller in
sure, Hepatology 8:861–865, 1988. cirrhosis: prospective validation of the noninvasive diagnostic criteria
Bosch J, et al: The clinical use of HVPG measurements in chronic liver for hepatocellular carcinoma, Hepatology 47:97–104, 2008. Erratum
disease, Nat Rev Gastroenterol Hepatol 6:573–582, 2009. in Hepatology 2008;47:769.
Bosch J, Groszmann RJ: Measurement of azygous venous blood flow Friedman SL: Hepatic stellate cells: protean, multifunctional, and
by a continuous thermal dilution technique: an index of blood flow enigmatic cells of the liver, Physiol Rev 88:125–172, 2008.
Fujimoto M, et al: Influence of spontaneous portosystemic collateral Henderson JM: Abnormal splanchnic and systemic hemodynamics of
pathways on portal hemodynamics in living-related liver transplanta- end-stage liver disease: what happens after liver transplantation?,
tion in children. Doppler ultrasonographic study, Transplantation Hepatology 17:514–516, 1993.
60:41–45, 1995. Henderson JM, et al: High cardiac output of advanced liver disease
Furchgott RF, Jothianandan D: Endothelium-dependent and -indepen- persists after orthotopic liver transplantation, Hepatology 15:258–262,
dent vasodilation involving cyclic GMP: relaxation induced by nitric 1992.
oxide, carbon monoxide and light, Blood Vessels 28:52–61, 1991. Hickey KA, et al: Am J Physiol Cell Physiol 248:C550–C556, 1985.
Gad J: Studies on the relations of the blood stream of the portal vein to the Hosoki R, et al: The possible role of hydrogen sulfide as an endogenous
blood stream in the hepatic artery. Dissertation, Berlin, 1873, G. Schade. smooth muscle relaxant in synergy with nitric oxide, Biochem Biophys
Gadano A, et al: Hemodynamics after orthotopic liver transplantation: Res Commun 237:527–531, 1997.
study of associated factors and long-term effects, Hepatology 22: Hughes RL, et al: The effect of hypercarbia on hepatic blood flow and
458–465, 1995. oxygen consumption in the greyhound, Br J Anaesth 51:289–296,
Gaiani S, et al: Prevalence of spontaneous hepatofugal portal flow in 1979a.
liver cirrhosis. Clinical and endoscopic correlation in 228 patients, Hughes RL, et al: Liver blood flow and oxygen consumption during
Gastroenterology 100:160–167, 1991. hypocapnia and IPPV in the greyhound, J Appl Physiol 47:290–295,
García-Pagán JC, et al; Early TIPS (Transjugular Intrahepatic Porto- 1979b.
systemic Shunt) Cooperative Study Group: Early use of TIPS in Hughes RL, et al: Systemic hypoxia and hyperoxia, and liver blood flow
patients with cirrhosis and variceal bleeding, N Engl J Med 362: and oxygen consumption in talc greyhound, Pflugers Arch 381:151–
2370–2379, 2010. 157, 1979c.
Garcia-Tsao G, Bosch J: Management of varices and variceal hemor- Hughes RL, et al: Effects of enflurane and halothane on liver blood
rhage in cirrhosis, N Engl J Med 362:823, 2010. flow and oxygen consumption in the greyhound, Br J Anaesth
Garcia-Tsao G, et al: Portal pressure, presence of gastroesophageal 52:1079–1086, 1980.
varices and variceal bleeding, Hepatology 5:419–424, 1985. Hughes RL, et al: Liver blood flow and oxygen consumption during
George CF: Drug kinetics and hepatic blood flow, Clin Pharmacokinet metabolic acidosis and alkalosis in the greyhound, Clin Sci 60:355–
4:433–448, 1979. 361, 1981.
Gluud C, et al: Prognostic indicators in alcoholic cirrhotic men, Hepa- Hunt DR: Changes in liver blood flow with development of biliary
tology 8(2):222–227, 1988. obstruction in the rat, Aust N Z J Surg 49:733–737, 1979.
Goéré D, et al: Prolonged survival of initially unresectable hepatic Isales CM, et al: Endothelin-1 induces cholestasis, which is mediated
colorectal cancer patients treated with hepatic arterial infusion of by an increase in portal pressure, Biochem Biophys Res Commun
oxaliplatin followed by radical surgery of metastases, Ann Surg 191:1244–1251, 1993.
251:686–691, 2010. Ishikawa M, et al: Hemodynamic changes in blood flow through the
Goldberg JA, et al: The use of angiotensin II as a potential method denervated liver in pigs, J Invest Surg 8:95–100, 1995.
of targeting cytotoxic microspheres in patients with intrahepatic Iwao T, et al: Wedged hepatic venous pressure reflects portal venous
tumour, Br J Cancer 63:308–310, 1991. pressure during vasoactive drug administration in nonalcoholic cir-
Gonzalez R, et al: Meta-analysis: Combination endoscopic and drug rhosis, Dig Dis Sci 39:2439–2444, 1994.
therapy to prevent variceal rebleeding in cirrhosis, Ann Intern Med Iwao T, et al: Hemodynamic study during transdermal application of
149:109–122, 2008. nitroglycerin tape in patients with cirrhosis, Hepatology 13:124–128,
Gottlieb ME, et al: Hepatic perfusion and splanchnic oxygen consump- 1991.
tion in patients postinjury, J Trauma 23:836–843, 1983. Jakab F, et al: The interaction between hepatic arterial and portal
Greenway CV, Lautt WW: Hepatic circulation. In Schultz SG, et al, venous blood flows; simultaneous measurement by transit time
editors: Handbook of physiology: the gastrointestinal system, New York, ultrasonic volume flowmetry, Hepatogastroenterology 42:18–21,
1989, American Physiological Society, Oxford University Press, pp 1995.
1519–1564. Jakimowicz J, et al: Laparoscopic insufflation of the abdomen reduces
Greenway CV, Oshiro G: Intrahepatic distribution of portal and hepatic portal venous flow, Surg Endosc 12:129–132, 1998.
arterial blood flows in anaesthetized cats and dogs and the effects of Jarnagin WR, et al: Regional chemotherapy for unresectable primary
portal occlusion, raised venous pressure and histamine, J Physiol liver cancer: results of a phase II clinical trial and assessment of
(Lond) 227:473–485, 1972. DCE-MRI as a biomarker of survival, Ann Oncol 20:1589–1595,
Greenway CV, Stark RD: Hepatic vascular bed, Physiol Rev 51:23–65, 2009.
1971. Jayaraman S, et al: The association between central venous pressure,
Groszmann RJ: Pathophysiology of portal hypertension. In Gentilini P, pneumoperitoneum, and venous carbon dioxide embolism in lapa-
et al, editors: Liver diseases and renal complications, New York, 1990, roscopic hepatectomy, Surg Endosc 23:2369–2373, 2009.
Raven Press, pp 165–173. Jefferson NC, et al: Formation of effective collateral circulation
Groszmann RJ, et al: Intrahepatic arteriovenous shunting in cirrhosis following excision of hepatic artery, Am J Physiol 184:589–592,
of the liver, Gastroenterology 73:201–204, 1977. 1956.
Groszmann RJ, et al: Splanchnic hemodynamics in portal-hypertensive Junghans T, et al: Does pneumoperitoneum with different gases, body
rats: measurement with gamma-labeled microspheres, Am J Physiol positions, and intraperitoneal pressures influence renal and hepatic
242:G156–G160, 1982. blood flow?, Surgery 121:206–211, 1997.
Groszmann RJ, et al: β-blockers to prevent gastroesophageal varices in Kanda H, et al: Hepatic blood flow after acute biliary obstruction
patients with cirrhosis, N Engl J Med 353(21):2254–2261, 2005. and drainage in conscious dogs, Hepatogastroenterology 43:235–240,
Guest J, et al: Portacaval transposition and subsequent partial hepatec- 1996.
tomy in the rat: effects on liver atrophy, hypertrophy and regenerative Keiding S: Galactose clearance measurements and liver blood flow,
hyperplasia, Br J Exp Pathol 58:140–146, 1977. Gastroenterology 94:477–481, 1988.
Hadengue A, et al: Persistence of systemic and splanchnic hyperkinetic Kelty RH, et al: The relation of the regenerated liver nodule to the
circulation in liver transplant patients, Hepatology 17:175–178, vascular bed in cirrhosis, Gastroenterology 15:285–295, 1950.
1993. Kemeny NE: The re-birth of hepatic arterial infusion for colorectal liver
Hadengue A, et al: Beneficial hemodynamic effects of ketanserin in metastases, J Gastrointest Oncol 4:118–120, 2013.
patients with cirrhosis: possible role of serotonergic mechanisms in Kemeny NE, et al: Conversion to resectability using hepatic artery
portal hypertension, Hepatology 7:644–647, 1987. infusion plus systemic chemotherapy for the treatment of unresect-
Hahn M, et al: Die eckische Fistel zwischen der Unteren Hohlvene und able liver metastases from colorectal carcinoma, J Clin Oncol
der Pfortader und ihre Folgen für den Organismus, Arch Exp Pathol 27:3465–3471, 2009.
Pharmacol 32:161–210, 1893. Klopfenstein CE, et al: Effects of abdominal CO2 insufflation and
Hemingway DM, et al: The effects of intra-arterial vasoconstrictors on changes of position on hepatic blood flow in anesthetized pigs, Am
the distribution of a radiolabelled low molecular weight marker in J Physiol 275:H900–H905, 1998.
an experimental model of liver tumour, Br J Cancer 63:495–498, Knisely MH, et al: Hepatic sphincters: brief summary of present-day
1991. knowledge, Science 125:1023–1026, 1957.
Kravetz D, et al: Comparison of intravenous somatostatin and vaso- MacLeod JJR, Pearce RG: The outflow of blood from the liver as
pressin infusions in treatment of acute variceal hemorrhage, Hepatol- affected by variations in the condition of the portal vein and hepatic
ogy 4:442–446, 1984. artery, Am J Physiol 35:87–105, 1914.
Kuo PC, et al: Magnetic resonance imaging and hepatic hemodynam- Malagari K, et al: Prospective randomized comparison of chemoembo-
ics: correlation with metabolic function in liver transplantation lization with doxorubicin-eluting beads and bland embolization with
candidates, Surgery 117:373–379, 1995. BeadBlock for hepatocellular carcinoma, Cardiovasc Intervent Radiol
Kurihara T, et al: Relationship between endothelin and thromboxane 33:541–551, 2010.
A2 in rat liver microcirculation, Life Sci 51:PL281–PL285, Malarkey DE, et al: New insights into functional aspects of liver mor-
1992. phology, Toxicol Pathol 33:27–34, 2005.
Lammer J, et al; PRECISION V Investigators: 2010: Prospective ran- Maleux G, et al: Prospective study comparing different indirect
domized study of doxorubicin-eluting-bead embolization in the methods to measure portal pressure, J Vasc Interv Radiol 22:1553–
treatment of hepatocellular carcinoma: results of the PRECISION V 1558, 2011.
study, Cardiovasc Intervent Radiol 33:41–52, 2010. Mallet-Guy Y, et al: Note on immediate ultrastructural lesions of the
Langer S, et al: Orthogonal polarization spectral imaging as a tool for liver following experimental ligation of the hepatic artery, Lyon Chir
the assessment of hepatic microcirculation: a validation study, 68:170–175, 1972.
Transplantation 71:1249–1256, 2001. Maluccio MA, et al: Transcatheter arterial embolization with only
Lautt WW: Method for measuring hepatic uptake of oxygen or other particles for the treatment of unresectable hepatocellular carcinoma,
blood-borne substances in situ, J Appl Physiol 40:269–274, 1976. J Vasc Interv Radiol 19:862–869, 2008.
Lautt WW: Hepatic vasculature: a conceptual review, Gastroenterology Mann FC: The William Henry Welch Lectures. II. Restoration
73:1163–1169, 1977a. and pathologic reactions of the liver, J Mt Sinai Hosp 11:65–74,
Lautt WW: Control of hepatic and intestinal blood flow: effect of iso- 1944.
volemic haemodilution on blood flow and oxygen uptake in the intact Manner M, et al: Arterial steal: an unusual cause for hepatic hypoperfu-
liver and intestines, J Physiol Lond 265:313–326, 1977b. sion after liver transplantation, Transpl Int 4:122–124, 1991.
Lautt WW: Role and control of the hepatic artery. In Lautt WW, editor: Marchioro TL, et al: The effect of partial portacaval transposition on
Hepatic circulation in health and disease, New York, 1981, Raven Press, the canine liver, Surgery 61:723–732, 1967.
pp 203–226. Marcos A, et al: The interrelationship between portal and arterial blood
Lautt WW: Regulatory processes interacting to maintain hepatic blood flow after adult to adult living donor liver transplantation, Transplan-
flow constancy: vascular compliance, hepatic arterial buffer response, tation 70:1697–1703, 2000.
hepatorenal reflex, liver regeneration, escape from vasoconstriction, Martinez-Hernandez A: The hepatic extracellular matrix: II. Electron
Hepatol Res 37:891–903, 2007. immunohistochemical studies in rats with CCl4-induced cirrhosis,
Lautt WW: Hepatic circulation: physiology and pathophysiology. In collo- Lab Invest 53:166–186, 1985.
quium series: integrated systems physiology: from molecule to function to Mastaï R, et al: Serotonin blockade in conscious, unrestrained cirrhotic
disease, San Rafael, CA, 2009, Morgan & Claypool Life Sciences. dogs with portal hypertension, Hepatology 9:265–268, 1989.
Free full text available at: http://www.ncbi.nlm.nih.gov/books/ Mathie RT: The hepatic haemodynamic effects of nitric oxide. In
NBK53073/. Mathie RT, Griffith TM, editors: The haemodynamic effects of nitric
Lautt WW, Greenway CV: Hepatic venous compliance and role of liver oxide, London, 1999, Imperial College Press, pp 292–317.
as a blood reservoir, Am J Physiol 231:292–295, 1976. Mathie RT, Blumgart LH: Hepatic tissue perfusion studies during
Lautt WW, Greenway CV: Conceptual review of the hepatic vascular partial hepatectomy in man, Surg Gastroenterol 1:297–302, 1982.
bed, Hepatology 7:952–963, 1987. Mathie RT, Blumgart LH: Effect of denervation on the hepatic haemo-
Lautt WW, et al: The comparative effects of substances via the hepatic dynamic response to hypercapnia and hypoxia in the dog, Pflugers
artery or portal vein on hepatic arterial resistance, liver blood volume Arch 397:152–157, 1983.
and hepatic extraction in cats, Hepatology 4:927–932, 1984. Mathie RT, Blumgart LH: The hepatic haemodynamic response to
Lautt WW, et al: Adenosine as putative regulator of hepatic arterial flow acute portal venous blood flow reductions in the dog, Pflugers Arch
(the buffer response), Am J Physiol 248:H331–H338, 1985. 399:223–227, 1983b.
Lautt WW, et al: Quantitation of the hepatic arterial buffer response to Mathie RT, et al: The hepatic arterial blood flow response to portal vein
graded changes in portal blood flow, Gastroenterology 98:1024–1028, occlusion in the dog, Pflugers Arch 386:77–83, 1980.
1990. Mathie RT, et al: The role of blood flow in the control of liver size,
Lebrec D, et al: Portal hypertension, size of esophageal varices, and risk J Surg Res 27:139–144, 1979.
of gastrointestinal bleeding in alcoholic cirrhosis, Gastroenterology Mathie RT, et al: Hepatic hemodynamics after chronic obstruction
79(6):1139–1144, 1980. of the biliary tract in the dog, Surg Gynecol Obstet 166:125–130,
Lebrec D, et al: Propranolol for prevention of recurrent gastrointestinal 1988.
bleeding in patients with cirrhosis: a controlled study, N Engl J Med Mays ET, Wheeler CS: Demonstration of collateral arterial flow after
305:1371–1374, 1981. interruption of hepatic arteries in man, N Engl J Med 290:993–996,
Lee YA, et al: Pathobiology of liver fibrosis: a translational success story, 1974.
Gut 64:830–841, 2015. Michalopoulos GK: Liver regeneration after partial hepatectomy: criti-
Lee FY, et al: Role of endotoxaemia in hyperdynamic circulation in rats cal analysis of mechanistic dilemmas, Am J Pathol 176:2–13, 2010.
with extrahepatic or intrahepatic portal hypertension, J Gastroenterol McCuskey RS: Morphological mechanisms for regulating blood flow
Hepatol 11:152–158, 1996. through hepatic sinusoids, Liver 20:3–7, 2000.
Leiberman DP, et al: The hepatic arterial and portal venous circulations Menger MD, Messmer K: In vivo fluorescent microscopy for quantita-
of the liver studied with a krypton-85 clearance technique, J Surg tive analysis of the hepatic microcirculation in hamsters and rats, Eur
Res 25:154–162, 1978. Surg Res 23:158–169, 1991.
Liapi E, Geschwind JF: Chemoembolization for primary and metastatic Merkel C, et al: Prognostic usefulness of hepatic vein catheterization in
liver cancer, Cancer J 16(2):156–162, 2010. patients with cirrhosis and esophageal varices, Gastroenterology
Lin H, McGrath JJ: Vasodilating effects of carbon monoxide, Drug 102(3):973–979, 1992.
Chem Toxicol 11:371–385, 1988. Merkel C, et al: Long-term results of a clinical trial of nadolol with or
Llovet JM, et al: Intention-to-treat analysis of surgical treatment for without isosorbide mononitrate for primary prophylaxis of variceal
early hepatocellular carcinoma: resection versus transplantation, bleeding in cirrhosis, Hepatology 31(2):324–329, 2000.
Hepatology 30(6):1434–1440, 1999. Miller WL, et al: Integrated cardiac, renal, and endocrine actions of
Llovet JM, et al: Arterial embolisation or chemoembolisation versus endothelin, J Clin Invest 83:317–320, 1989.
symptomatic treatment in patients with unresectable hepatocellular Mogl MT, et al: Evolving experience with prevention and treatment of
carcinoma: a randomised controlled trial, Lancet 359:1734–1739, splenic artery syndrome after orthotopic liver transplantation, Transpl
2002. Int 23:831–841, 2010.
Lo CM, et al: Randomized controlled trial of transarterial lipiodol Moitinho E, et al: Prognostic value of early measurements of portal
chemoembolization for unresectable hepatocellular carcinoma, pressure in acute variceal bleeding, Gastroenterology 117(3):626–631,
Hepatology 35:1164–1171, 2002. 1999.
Nagano K, et al: Hepatic oxygen supply-uptake relationship and Reuter SR, Chuang VP: The location of increased resistance to portal
metabolism during anesthesia in miniature pigs, Anesthesiology 75: blood flow in obstructive jaundice, Invest Radiol 11:54–59, 1976.
902–910, 1990. Rice GC, et al: Liver tissue blood flow measured by 85Kr clearance in
Nahum E, et al: Correlation of transcutaneous hepatic near-infrared the anaesthetized rat before and after partial hepatectomy, Br J Exp
spectroscopy readings with liver surface readings and perfusion Pathol 58:243–250, 1977.
parameters in a piglet endotoxemic shock model, Liver Int 26:1277– Richardson PDI, Withrington PG: Liver blood flow: II. Effects of drugs
1282, 2006. and hormones on liver blood flow, Gastroenterology 81:356–375,
Navasa M, et al: Reduction of portal pressure by isosorbide-5-mono- 1981.
nitrate in patients with cirrhosis. Effects on splanchnic and systemic Richter S, et al: Impact of intrinsic blood flow regulation in cirrhosis:
hemodynamics and liver function, Gastroenterology 96:1110–1118, maintenance of hepatic arterial buffer response, Am J Physiol Gas-
1989. trointest Liver Physiol 279(2):G454–G462, 2000.
Navasa M, et al: Hemodynamic and humoral changes after liver trans- Richter S, et al: Loss of physiologic hepatic blood flow control (“hepatic
plantation in patients with cirrhosis, Hepatology 17:355–360, 1993. arterial buffer response”) during CO2-pneumoperitoneum in the rat,
Niederberger M, et al: Normalization of nitric oxide production cor- Anesth Analg 93:872–877, 2001.
rects arterial vasodilation and hyperdynamic circulation in cirrhotic Rincon D, et al: Prognostic value of hepatic venous pressure gradient
rats, Gastroenterology 109:1624–1630, 1995. for in-hospital mortality of patients with severe acute alcoholic hepa-
Nobuoka T, et al: Portal blood flow regulates volume recovery of the titis, Aliment Pharmacol Ther 25(7):841–848, 2007.
rat liver after partial hepatectomy: molecular evaluation, Eur Surg Res Ripoll C, et al: Hepatic venous pressure gradient predicts clinical
38:522–532, 2006. decompensation in patients with compensated cirrhosis, Gastroenter-
Nüssler NC, et al: Diagnosis and treatment of arterial steal syndromes ology 133(2):481–488, 2007.
in liver transplant recipients, Liver Transpl 9:596–602, 2003. Ripoll C, et al: Hepatic venous pressure gradient predicts development
Ohdan H, et al: Intraoperative near-infrared spectroscopy for evaluat- of hepatocellular carcinoma independently of severity of cirrhosis, J
ing hepatic venous outflow in living-donor right lobe liver, Transplan- Hepatol 50(5):923–928, 2009.
tation 76:791–797, 2003. Rocheleau B, et al: Hepatic artery buffer response following left portal
Ohlsson EG: The arterial circulation in the liver after total biliary vein ligation: its role in liver tissue homeostasis, Am J Physiol
obstruction in dogs, Acta Chir Scand 138:1–58, 1972. 277:G1000–G1007, 1999.
Ohnishi K, et al: Portal hemodynamics in idiopathic portal hyperten- Rockey DC: Hepatic blood flow regulation by stellate cells in normal
sion (Banti’s syndrome). Comparison with chronic persistent hepa- and injured liver, Semin Liver Dis 21:337–349, 2001.
titis and normal subjects, Gastroenterology 92:751–758, 1987. Rössle M, et al: New non-operative treatment for variceal haemorrhage,
Okochi O, et al: ICG pulse spectrophotometry for perioperative liver Lancet 2:153, 1989.
function in hepatectomy, J Surg Res 103:109–113, 2002. Rossi S, et al: Contrast-enhanced versus conventional and color
Okuda K, et al: Percutaneous transhepatic catheterization of the portal Doppler sonography for the detection of thrombosis of the portal
vein for the study of portal hemodynamics and shunts. A preliminary and hepatic venous systems, AJR Am J Roentgenol 186:763–773,
report, Gastroenterology 73:279–284, 1977. 2006.
Okumura S, et al: Vasoactive effect of endothelin-1 on rat liver in vivo, Ryan CJ, et al: Liver blood flow measurements in the portacavally
Hepatology 19:155–161, 1994. transposed rat before and after partial hepatectomy, Br J Exp Pathol
Palmer RM, et al: Nitric oxide release accounts for the biological activ- 59:111–115, 1978.
ity of endothelium-derived relaxing factor, Nature 327:524–526, Sakka SG: Assessing liver function, Curr Opin Crit Care 13:207–214,
1987. 2007.
Parker R: Role of transjugular intrahepatic portosystemic shunt in the Salem R, et al: Radioembolization results in longer time-to-progression
management of portal hypertension, Clin Liver Dis 18:319–334, 2014. and reduced toxicity compared with chemoembolization in patients
Paulsen AW, Klintmalm GBG: Direct measurement of hepatic blood with hepatocellular carcinoma, Gastroenterology 140:497–507, 2011.
flow in native and transplanted organs, with accompanying systemic Sapirstein LA: Fractionation of the cardiac output of rats with isotopic
hemodynamics, Hepatology 16:100–111, 1992. potassium, Circ Res 4:689–692, 1956.
Payen DM, et al: Portal and hepatic arterial blood flow measurements Schenk WG Jr, et al: Direct measurement of hepatic blood flow in
of human transplanted liver by implanted Doppler probes: interest surgical patients: with related observations on hepatic flow dynamics
for early complications and nutrition, Surgery 107:417–427, 1990. in experimental animals, Ann Surg 156:463–469, 1962.
Peitzman AB, et al: Hemorrhagic shock, Curr Probl Surg 32:925–1002, Schilling MK, et al: Splanchnic microcirculatory changes during CO2
1995. laparoscopy, J Am Coll Surg 184:378–382, 1997. Erratum in J Am
Pelc LR, et al: Arterial and venous blood flow: noninvasive quantitation Coll Surg 1997;185:423.
with MR imaging, Radiology 185:809–812, 1992. Schmandra TC, et al: Risk of gas embolism in hand-assisted versus
Perelló A, et al: Wedged hepatic venous pressure adequately reflects total laparoscopic hepatic resection, Surg Technol Int 12:137–143,
portal pressure in hepatitis C virus-related cirrhosis, Hepatology 30: 2004.
1393–1397, 1999. Schoen JM, et al: Shear stress-induced nitric oxide release triggers the
Pizcueta P, et al: Modulation of the hyperdynamic circulation of cir- liver regeneration cascade, Nitric Oxide 5:453–464, 2001.
rhotic rats by nitric oxide inhibition, Gastroenterology 103:1909–1915, Sedgwick CE, et al: Management of portal hypertension secondary to
1992. bile duct strictures: review of 18 cases with splenorenal shunt, Ann
Puhl G, et al: Noninvasive in vivo analysis of the human hepatic Surg 163:949–953, 1966.
microcirculation using orthogonal polarization spectral imaging, Seifalian AM, et al: Hepatic microcirculation during human orthotopic
Transplantation 75:756–761, 2003. liver transplantation, Br J Surg 84:1391–1395, 1997.
Quintini C, et al: “Splenic artery steal syndrome” is a misnomer: the Sersté T, et al: Relationship between the degree of portal hypertension
cause is portal hyperperfusion, not arterial siphon, Liver Transpl and the onset of spontaneous bacterial peritonitis in patients with
14:374–379, 2008. cirrhosis, Acta Gastroenterol Belg 69(4):355–360, 2006.
Qureshi W, et al: Clinical predictors of post-liver transplant new-onset Sganga G, et al: Reprioritization of hepatic plasma protein release in
heart failure, Liver Transpl 19:701–710, 2013. trauma and sepsis, Arch Surg 120:187–199, 1985.
Rappaport AM: The microcirculatory hepatic unit, Microvasc Res Shepherd AP, et al: Evaluation of an infrared laser-Doppler blood
6:212–228, 1973. flowmeter, Am J Physiol 252:G832–G839, 1987.
Rappaport AM, Schneiderman JH: The function of the hepatic artery, Sherman IA, et al: Dynamics of arterial and portal venous flow interac-
Rev Physiol Biochem Pharmacol 76:129–175, 1976. tions in perfused rat liver: an intravital microscopic study, Am J
Reichen J, et al: The effect of endothelin and its antagonist Bosentan Physiol 271:G201–G210, 1996.
on hemodynamics and microvascular exchange in cirrhotic rat liver, Sherman IA, et al: Hepatic microvascular changes associated with
J Hepatol 28:1020–1030, 1998. development of liver fibrosis and cirrhosis, Am J Physiol 258:H460–
Reichen J, Le M: Verapamil favorably influences hepatic microvascular H465, 1990.
exchange and function in rats with cirrhosis of the liver, J Clin Invest Siebert N, et al: H2S contributes to the hepatic arterial buffer response
78:448–455, 1986. and mediates vasorelaxation of the hepatic artery via activation of
K(ATP) channels, Am J Physiol Gastrointest Liver Physiol 295:G1266– Troisi R, et al: Effects of hemiportocaval shunts for inflow modulation
G1273, 2008. on the outcome of small-for-size grafts in living donor liver trans-
Smedsrød B, et al: Cell biology of liver endothelial and Kupffer cells, plantation, Am J Transplant 5:1397–1404, 2005.
Gut 35:1509–1516, 1994. Umeda Y, et al: Preoperative proximal splenic artery embolization: a
Smith JJ, et al: Oxygen saturation of hepatic blood during hemorrhagic safe and efficacious portal decompression technique that improves
shock in the dog, Circ Res 3:311–314, 1955. the outcome of live donor liver transplantation, Transpl Int 20:947–
Smyrniotis V, et al: Hemodynamic interaction between portal vein and 955, 2007.
hepatic artery flow in small-for-size split liver transplantation, Transpl Vallance P, Moncada S: Hyperdynamic circulation in cirrhosis: a role
Int 15:355–360, 2002. for nitric oxide?, Lancet 337:776–778, 1991.
Stanley AJ, et al: Haemodynamic parameters predicting variceal haem- Varin F, Huet PM: Hepatic microcirculation in the perfused cirrhotic
orrhage and survival in alcoholic cirrhosis, QJM 91(1):19–25, 1998. rat liver, J Clin Invest 76:1904–1912, 1985.
Starzl TE, et al: The origin, hormonal nature, and action of hepatotro- Vollmar B, Menger MD: The Hepatic microcirculation: mechanistic
phic substances in portal venous blood, Surg Gynecol Obstet contributions and therapeutic targets in liver injury and repair,
137:179–199, 1973. Physiol Rev 89:1269–1339, 2009.
Steer ML, et al: Chronic biliary obstruction: hemodynamic effects of Vorobioff J, et al: Prognostic value of hepatic venous pressure gradient
decompression, Surg Forum 19:342–344, 1968. measurements in alcoholic cirrhosis: a 10-year prospective study,
Stewart GN: Researches on the circulation time and on the influences Gastroenterology 111(3):701–709, 1996.
which affect it. IV. The output of the heart, J Physiol 22:159, Waller A: Microscopic examination of some of the principal tissues of
1897. the animal tongue, as observed in the tongue of the living frog and
Takaoka F, et al: Intraoperative evaluation of Eurocollins and Univer- toad, Philos Mag J Sci 29:397–405, 1846.
sity of Wisconsin preservation solutions in patients undergoing Wang Y, et al: Inhibition of nitric oxide synthesis aggravates reperfusion
hepatic transplantation, Transplantation 49:544–547, 1990. injury after hepatic ischemia and endotoxemia, Shock 4:282–288,
Tamakuma S, et al: Relationship between hepatic hemodynamics and 1995.
biliary pressure in dogs: its significance in clinical shock following Warren WD, Muller WH: A classification of some hemodynamic
biliary decompression, Jpn J Surg 5:255–568, 1975. changes in cirrhosis and their surgical significance, Ann Surg
Tanaka A, et al: Endothelin-1 stimulates bile acid secretion and vesicu- 150:413–427, 1959.
lar transport in the isolated perfused rat liver, Am J Physiol 266: Wendon JA: Local consequences of reperfusion in the liver. In Grace
G324–G329, 1994. PA, Mathie RT, editors: Ischaemia-reperfusion injury, Oxford, 1999,
Tarantino L, et al: Reverse flow in intrahepatic portal vessels and liver Blackwell Science, pp 56–64.
function impairment in cirrhosis, Eur J Ultrasound 6:171–177, 1997. Wheatley AM, et al: Effect of orthotopic transplantation and chemical
Terajima H, et al: Reduction of hepatic microcirculatory failure caused denervation of the liver on hepatic hemodynamics in the rat,
by normothermic ischemia/reperfusion-induced injury by means of J Hepatol 19:442–450, 1993.
heat shock preconditioning, Shock 12:329–334, 1999. Whipple AO: The problem of portal hypertension in relation to the
Thomson IA, et al: Effects of certain IV anaesthetics on liver blood flow hepatosplenopathies, Ann Surg 122:449–475, 1945.
and hepatic oxygen consumption in the greyhound, Br J Anaesth Wisse E, et al: The liver sieve: considerations concerning the structure
58:69–80, 1986. and function of endothelial fenestrae, the sinusoidal wall and the
Thomson IA, et al: Effects of nitrous oxide on liver haemodynamics space of Disse, Hepatology 5:683–692, 1985.
and oxygen consumption in the greyhound, Anaesthesia 37:548–553, Wu Y, et al: Hormonal and splanchnic hemodynamic alterations follow-
1982. ing hepatic resection, J Surg Res 55:44–48, 1993.
Thulin L, et al: Effect of controlled halothane anaesthesia on splanch- Yanagisawa M, et al: Nature 332:411–115, 1988.
nic blood flow and cardiac output in the dog, Acta Anaesth Scand Yokoyama Y, et al: Hepatic vascular response to elevated intraperitoneal
19:146–153, 1975. pressure in the rat, J Surg Res 105:86–94, 2002.
Tranchart H, et al: Laparoscopic liver resection with selective prior Zeeh J, et al: Steady-state extrarenal sorbitol clearance as a measure of
vascular control, Am J Surg 205:8–14, 2013. hepatic plasma flow, Gastroenterology 95:749–759, 1988.
Tranchart H, et al: Bleeding control during laparoscopic liver resection: Zhao W, et al: The vasorelaxant effect of H(2)S as a novel endogenous
a review of literature, J Hepatobiliary Pancreat Sci 22:371–378, 2015. gaseous K(ATP) channel opener, EMBO J 20:6008–6016, 2001.
Tran-Thi TA, et al: Regulation of endothelin-1 action on the perfused Zhang JX, et al: Endothelin-1 induces direct constriction of hepatic
rat liver, FEBS Lett 318:353–357, 1993. sinusoids, Am J Physiol 266:G624–G632, 1994.
CHAPTER 6
Liver regeneration: mechanisms and clinical
relevance
Jeroen de Jonge and Kim M. Olthoff
And now the last recess of the Black Sea opened up and they regain most of its functional mass within 2 weeks (Humar et al,
[the Argonauts] caught sight of the high crags of the Kaukasos, 2004; Kamel et al, 2003; Marcos et al, 2000; Pomfret et al,
where Prometheus stood chained by every limb to the hard 2003). Factors that limit the achievement of curative tumor
rock with fetters of bronze, and fed an eagle on his liver. The resection and small-for-size (SFS) transplantations are the high
bird kept eagerly returning to its feed. They saw it in the morbidity and mortality rates associated with insufficient
afternoon flying high above the ship with a strident whirr. It volume of the liver remnant or graft (Fig. 6.1). While hepatobili-
was near the clouds, yet it made all their canvas quiver to its ary and transplant surgeons continue to expand the magnitude
wings as it beat by. For its form was not that of an ordinary and complexity of liver resection, understanding liver regenera-
bird: the long quill-feathers of each wing rose and fell like a tion becomes increasingly relevant. Many tumors that were
bank of polished oars. Soon after the eagle had passed, they previously considered to be unresectable are now amenable to
heard Prometheus shriek in agony as it pecked at his liver. The complete resection through induction chemotherapy and inno-
air rang with his screams till at length they saw the flesh- vative treatment strategies to increase liver remnant volume
devouring bird fly back from the mountain by the same way (Clavien et al, 2007). Portal vein embolization (PVE), portal
as it came. vein ligation, and to the extreme, in Associating Liver Partition
Argonautica 2.1238f and Portal vein ligation for Staged hepatectomy (ALPPS),
causes atrophy of the ipsilateral hemiliver and hypertrophy of
The ability of the liver to regenerate was recognized by the the contralateral side and are particularly valuable in patients
Greeks in the ancient myth of Prometheus, the Titan god of who have underlying liver disease. Because of the increased
forethought, who gave fire to the mortals and angered Zeus. success of liver surgery, more patients with underlying liver
Prometheus was chained to the Caucasus Mountains, and each disease are now considered suitable candidates for liver resec-
day he would be tormented by Zeus’ eagle Ethon as it devoured tion, even with Child-Pugh grade A cirrhosis and some degree
his liver. Each night, the damaged liver would heal so the eagle of portal hypertension (Boleslawski et al, 2012; Chen et al,
could begin anew, illustrating the incredible capacity of the liver 2012; Zhong et al, 2014). Patients with the metabolic syn-
to regenerate. drome, morbid obesity with concomitant steatosis, or nonalco-
In contrast to other solid organs, the human liver has the holic steatohepatitis (NASH) have, however, increased risk of
unique ability to regenerate after toxic injury, chronic inflam- liver failure after liver surgery (Belghiti et al, 2000; McCormack
mation, and surgical resection. The terms regeneration, hyper- et al, 2007), as are elderly patients older than 70 years with
plasia, and hypertrophy are used synonymously in literature, restricted hepatic reserve capacity.
but hyperplasia is the most precise from a cellular standpoint. In colorectal malignant disease, approximately 50% of the
The damaged hepatic lobes do not grow back in the same way patients will experience liver metastasis in the course of their
a lizard’s tail regrows, but rather, there is a hyperplastic response disease. With the substantially improved chemotherapy for
(defined as increasing cell number) in the remnant liver, leading colorectal cancer during the last decade, the resectability rate
to its hypertrophic appearance (defined as enlarging liver size) has gradually increased to 20% to 30% (Chua et al, 2010;
(Lory et al, 1994). This process is highly regulated, involving Robinson et al, 2011; Uetake et al, 2015). As a consequence,
multiple cell types, extrahepatic signals, complex molecular more patients are having liver surgery after neoadjuvant or
pathways, and cellular interactions. A delicate balance in initia- induction chemotherapy, which has considerable effects on liver
tion of regeneration and exerting an intracellular growth parenchyma and can inhibit liver regeneration.
response to restore lost liver mass, while maintaining normal Regeneration also is crucial in liver transplantation. In
metabolic function, is necessary for survival. The inability to deceased-donor transplantation, hepatocyte loss occurs in the
maintain this process leads to poor liver function and ultimately form of ischemia/reperfusion (I/R) injury owing to the neces-
liver failure after surgery. sary preservation period from procurement to implantation
and damage that may have occurred in the donor. Due to the
CLINICAL RELEVANCE OF LIVER scarcity of organs, more “marginal” organs are accepted for
REGENERATION transplantation, which increases the need for regeneration and
recovery in the environment of I/R injury and use of immuno-
Today, liver surgery is routinely and safely accomplished for suppressive drugs. One of the landmark advances in liver
malignant and benign disease. This substantial progress in both transplantation is the ability to use segmental liver grafts
resection and transplantation relies on the ability of the liver to obtained from either a deceased donor or a living donor. In the
93
regenerative signaling observed in a rat liver transplant model

of I/R injury is similar to that observed after PHx.
General Features of Liver Regeneration

Liver regeneration after surgical resection is carried out by
proliferation of existing mature cellular populations. These
include hepatocytes, biliary and fenestrated endothelial cells,
Kupffer cells, and cells of Ito (stellate cells) (Gressner, 1995)
(Fig. 6.2). The kinetics of cell proliferation and the growth
factors produced by proliferating hepatocytes suggest that
hepatocytes provide the mitogenic stimuli leading to prolifera-
tion of the other cells. The degree of hepatocyte proliferation is
directly proportional to the degree of injury (Bucher & Swaf-
field, 1964). Immediately after liver resection, the rate of DNA
synthesis in hepatocytes begins to increase while they exit the
resting state of the cell cycle (G0) and enter G1, traverse to DNA
synthesis (S phase), and ultimately undergo mitosis (M phase).
The induction of DNA synthesis occurs later in the nonparen-
FIGURE 6.1 Remnant liver or liver graft volume limits tumor resection chymal cells (at approximately 48 hours for Kupffer and biliary
and small-for-size transplantation. Here is an example of an “extreme” epithelial cells and at approximately 96 hours for endothelial
liver resection, only leaving segment IV of the liver in case of colorectal cells (Grisham, 1969; Widmann & Fahimi, 1975). The first
liver metastases. In this situation patient’s recovery solely depends on
peak of DNA synthesis occurs at 40 hours after resection in
robust liver regeneration. (Courtesy of Professor Peter Lodge, The Leeds
Teaching Hospitals, Leeds, UK; and Cees Verhoef, Erasmus MC, Rot-
rodents and at 7 to 10 days in primates. In small animals, the
terdam, Netherlands.) regenerative response returns the liver to the preresection mass
in 1 week to 10 days. Clinical studies from living-donor trans-
plantation suggest that a significant amount of regeneration
latter situation, success of the procedure relies on relatively occurs in human within 2 weeks following resection and is
rapid hepatic regeneration in both donor and recipient. The nearly complete at 3 months following resection (Olthoff, 2015;
minimal amount of functional liver necessary for successful Everson et al, 2013; Emond et al, 2015).
transplantation or for safe recovery in the donor is a major Hepatocyte proliferation starts in the periportal areas of the
concern. Donor graft size, recipient weight, portal hyperten- lobules (Rabes et al, 1976) and then proceeds to the pericentral
sion, I/R injury, and the recipient’s disease severity all contribute areas by 36 to 48 hours. Liver histology at day 3 to 4 after PHx
to the amount of posttransplant regeneration and recovery is characterized by clumps of small hepatocytes surrounding
needed (Lee, 2015). Efforts to decrease the amount of liver capillaries, which change into true hepatic sinusoids. The
removed from the donor to minimize donor risk results in hepatic matrix composition also changes from high laminin
smaller grafts for the recipients and real challenges in postop- content to primarily containing fibronectin and collagen types
erative recovery. IV and I. After a 70% hepatectomy, restoration of the original
number of hepatocytes theoretically requires 1.66 proliferative
BASIC CHARACTERISTICS cycles per residual hepatocyte. In fact, most of the hepatocytes
OF LIVER REGENERATION (95% in young and 75% in very old rats) in the residual lobes
participate in one or two proliferative events (Stocker & Heine,
Models of Liver Regeneration 1971). Hepatocytes, however, have an almost unlimited capac-
Liver regeneration is most clearly shown in the experimental ity to regenerate because transplantation of several hundreds of
model that was pioneered in 1931 by Higgins and Anderson healthy hepatocytes can repopulate a whole damaged liver in a
(1931). In this model, a simple two-thirds partial hepatectomy calculated minimum of 69 doublings (Rhim et al, 1994).
(PHx) is performed, without damage to the lobes left behind.
This leads to enlargement of the residual lobes to make up for Liver Stem Cells
the mass of the removed lobes in 5 to 7 days. Other well-known In contrast to other regenerating tissues (bone marrow, skin),
models of liver regeneration are associated with extensive tissue primary liver regeneration after surgical trauma is not depen-
injury and inflammation and include the use of hepatic toxins, dent on a small group of progenitor cells (stem cells). However,
such as ethanol (EtOH) (Kaplowitz, 2002), carbon tetrachlo- in response to liver damage inflicted by agents such as galac-
ride (CCl4) (Reynolds, 1963) and galactosamine (GalN) tosamine, hepatocytes are unable to replicate. In this situation
(Dabeva & Sahafritz, 1993), bile duct ligation (Kountouras a population of cells known as “oval cells” proliferates to replace
et al, 1984) or portal vein ligation (Bilodeau et al, 1999), and the hepatic parenchyma (Fausto & Campbell et al, 2003),
I/R injury (Jaeschke, 1998). Newer models include transgenic although this paradigm has been questioned in mice. In distinct
albumin promoter urokinase-type plasminogen activator (uPA) approaches to determine whether cells other than hepatocytes
fusion constructs (Heckel et al, 1990), FAH/RAG2 knockout themselves could be the source of new hepatocytes in oval cell
mice (Azuma et al, 2007; Strom et al, 2010) and PHx in injury, two groups found no evidence of such liver stem/
zebrafish (Kan et al, 2009). In each model, the different toxic progenitor cells (Schaub et al, 2014; Yanger et al, 2014).
agents injure specific liver-cell subpopulations, and therefore In the human situation, hepatic progenitor cells (HPCs)
PHx is the preferred in vivo model to study the regenerative participate in repopulation of the liver after acute massive
response. Debonera and colleagues (2001) demonstrated that necrosis and have also been identified in chronic liver disease
Chapter 6 Liver regeneration: mechanisms and clinical relevance 95
Vascular endothelial
Stellate cell
growth factor Hepatocyte
Platelets
Inactive hepatocyte Transforming

growth factor growth factor-β
S phase
Lipopoly- Extracellular
saccharides proteases Inhibition
Serotonin
Hepatocyte
growth factor
Tumor Farnesoid X
Kupffer cell
necrosis
receptor
factor-α
G1 phase
Leukocyte
Interleukin-6
Nucleus
G0 or G1 Release
phase Matrix
Sinusoidal metalloproteases
endothelial cell Transforming
Epidermal growth factor-α
Bile acids growth factor
FIGURE 6.2 Pathways of liver regeneration initiated by major hepatectomy. After hepatectomy, nonparenchymal cells, such as stellate cells, Kupffer
cells, leukocytes, and platelets, are activated by soluble factors. As a result, Kupffer cells release tumor necrosis factor-α and interleukin-6. The
cytokines cause a priming of the remnant hepatocytes, and concurrently, extracellular proteases such as urokinase-type plasminogen activator convert
inactive hepatocyte growth factor to its active form. The cytokines and the growth factors act in concert to initiate the reentry of quiescent hepatocytes
(in the G0 phase) into the cell cycle from the G1 phase to the S phase, resulting in DNA synthesis and hepatocyte proliferation. The metabolic burden
is indicated by the accumulation of bile acids in the blood, which enter the hepatocytes and drive increased protein and DNA synthesis. To signal
the end of proliferation, transforming growth factor-β blocks further replication. (From Clavien, et al: Strategies for safer liver surgery and partial liver
transplantation. N Engl J Med 356:1545-1559, 2007.)
(Roskams et al, 1998). The human HPC originates from the flow after PHx resulted in hepatocyte proliferation and hyper-
canals of Hering (Carpentier et al, 2011; Carpino et al, 2011; plasia (Starzl et al, 1977). From a recent experimental partial
Rodrigo-Torres et al, 2014) and play an important role in liver transplant model, we know that increased portal flow
acetaminophen-induced injury (Kofman et al, 2005). Huch (from a universal 100 mL/100 g liver tissue/min to 200 mL/
and colleagues (2015) described conditions allowing long-term 100 g/min) is essential for liver regeneration. However, if portal
expansion of these adult bile duct–derived bipotent progenitor flow exceeds 250 mL/100 g/min (portal hyperperfusion), the
cells from human liver, which enables disease modeling, toxi- process is completely abolished (Hessheimer et al, 2011).
cology studies, and regenerative medicine. Experiments in parasymbiotic rats demonstrated the existence
In common with other tissues, there is persuasive evidence of humoral factors in the induction of liver growth after PHx
that, in the liver, stem cells can be the founder cells of primary (Moolten & Bucher, 1967).
hepatic malignancies, such as hepatocellular carcinoma (HCC). Interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α)
Prominin-1 (CD133) and Sal-like protein 4 (SALL4) have have since then been identified as the earliest factors triggering
been identified as cancer stem cell markers in HCC (Ma, 2012; activation of several transcription factors during regeneration
Oikawa et al, 2013; Song et al, 2008; Tang et al, 2008, 2012; (Cressman et al, 1996; Yamada et al, 1997) (Fig. 6.3). In
You et al, 2010). IL-6–deficient or TNF-α receptor–deficient mice, liver regen-
eration after hepatectomy is delayed (Streetz et al, 2000) but
Induction of Proliferation: Priming and not completely abolished (Fujita et al, 2001; Michalopoulos
Cell-Cycle Progression et al, 1984). Therefore other blood-derived mitogens, such as
Within minutes after PHx, specific immediate early genes are hepatocyte growth factor (HGF), were identified as putative
activated in remnant hepatocytes (Haber et al, 1993). These hepatic growth factors during liver regeneration (Fausto, 2000).
genes include protooncogenes that play an important role in Hepatocytes in normal liver are not ready to respond to mito-
normal cell-cycle progression, such as JUN, FOS, MYC, and genic signals without a set of “priming” events that switch them
KRAS (Morello et al, 1990; Taub, 2004; Thompson et al, 1986) into a responsive state. This has been described by Webber and
and the transcriptional factors nuclear factor (NF)-κB, signal colleagues (1994), who identified priming factors involved in
transducer and activator of transcription 3 (STAT3), activator initiating and triggering of the hepatic response to injury and
protein-1 (AP-1), and CCAAT enhancer–binding protein-β (C/ concomitant growth factors and their receptors, allowing com-
EBPβ) (Cressman et al, 1995; FitzGerald et al, 1995). petent hepatocytes to progress through the cell cycle. Priming
Historically, the onset of liver regeneration has been attrib- is accomplished by the release of preformed cytokines that
uted to a flow-dependent response by which increased relative subsequently activate transcription factor complexes and allow
Hepatocyte
STAT3
gp130 JAK P
Kupffer cell
C3a/C5a IL-6
TNF-α TLR
IL-6R JAK P –VE
P
P SOCS3
Activated
NF-κB STAT3 homodimer
IL-6
SOCS3
TNF IL-6 mRNA P mRNA
mRNA P
TNF-α
Multiple biological responses:
Cell survival/proliferation
FIGURE 6.3 Stimulated by components of the innate immune system, Kupffer cells produce and secrete interleukin-6 (IL-6) and tumor necrosis
factor-α (TNF-α) to kick-start the regenerative response. IL-6 helps stimulate hepatocyte proliferation via signal transducer and activator of transcription
3 (STAT3) activation; in turn, this response is negatively regulated by suppressor of cytokine signaling 3 (SOCS3). gp130, Glycoprotein 130; IL-6R,
interleukin-6 receptor; JAK, Janus activating kinase; mRNA, messenger RNA; NF-κB, nuclear factor kappa B; P, phosphate; TLR, Toll-like receptor;
VE, negative feedback. (From Alison et al: Stem cells in liver regeneration, fibrosis and cancer: the good, the bad and the ugly. J Pathol 217:282-
298, 2009.)
the cell to exit G0 into G1 of the cell cycle. This group includes poor activation of nuclear factor kappaB (NF-κB) and STAT3.
TNF-α and IL-6 (Aldeguer et al, 2002). Growth factors include In the initiation of growth factors, uPA appears to play an
the potent hepatocyte mitogens HGF, transforming growth important role. uPA and its downstream effector plasminogen
factor-α (TGF-α), and heparin-binding epidermal growth increase within 1 to 5 minutes after PHx and rapidly cleave the
factor (HB-EGF). This process is further controlled by HGF precursor pro-HGF. Blocking uPA delays the appearance
co-mitogens, such as insulin, glucagon, steroid hormones, par- of HGF and thereby delays liver regeneration, whereas blocking
ticularly estradiol, and epinephrine, which facilitate activity of plasminogen-activator inhibitor (PAI) accelerates the release of
the mitogens, and by downregulation of growth factor inhibi- HGF and liver regeneration (Currier et al, 2003).
tors such as activin A and TGF-β. The nuclear receptor (NR) superfamily is a set of transcrip-
So far, a few factors have been identified to be possibly tion factors, acting as conductors of differentiated liver func-
responsible for the release of these priming cytokines and tions. NRs are master transcriptional regulators of different
growth factors in the onset of liver regeneration. The first is homeostatic processes (e.g., development, cell differentiation,
endotoxin lipopolysaccharide (LPS), produced in the gut by metabolism, proliferation, and apoptosis), and can be modu-
gram-negative bacteria. Circulating LPS is an extremely strong lated by different signals (e.g., hormones, vitamins, lipids)
signal for Kupffer cells to start the cascade resulting in hepato- (Karpen & Trauner, 2010; Wagner et al, 2011). NRs (48 in
cyte replication. Rats treated with antibiotics and germ-free humans, 49 in rodents) are key players in the modulation of
rodents have a delayed peak of DNA replication after PHx, liver physiology and development, being also involved in cell
confirming the importance of LPS (Cornell et al, 1990). This growth and differentiation (Mangelsdorf et al, 1995). Some
signal is MyD88 (myeloid differentiation primary response NRs are regulated by small lipophilic ligands (e.g., hormones,
gene 88) dependent, but the LPS receptor Toll-like receptor vitamins, dietary lipids, bile acids (BAs), and xenobiotics),
(TLR) 2, TLR2, and CD14 do not play roles in regulating whereas other NRs, namely ‘‘true orphans,’’ regulate trans
cytokine production or DNA replication after PHx (Campbell cription independently from binding to specific ligands (Man-
et al, 2006). Another major finding is the demonstration that gelsdorf & Evans, 1995). NRs are also implicated in liver
cytokine activation and DNA replication are severely impaired regeneration modulation (Huang et al, 2006; Vacca et al, 2014)
in mice lacking the complement components C3a and C5a and in the pathophysiology of liver disease.
(Strey et al, 2003). In particular, mice lacking both C3a and NRs are suitable targets for pharmacologic approaches
C5a have impaired production of TNF and IL-6 after PHx and aimed to the control of hepatocyte proliferation (Vacca et al,
2013) because they may modulate a number of early changes several mechanisms, including binding by CDK-inhibitory
essential for the liver regeneration and HCC, such as the activa- proteins, such as p21. Feedback signals to this process are
tion of transcription factors (AP-1; NF-κB; STAT3; and C/ provided by suppressor of cytokine signaling 3 (SOCS 3) and
EBPβ, and the expression of immediate early genes (FBJ TGF-β, also regulated by IL-6. Other studies confirmed the
murine osteosarcoma viral oncogene homolog FOS, JUN, importance of NF-κB (Arai et al, 2003; Fukuhara et al, 2003;
MYC avian myelocytomatosis viral oncogene homolog [c-Myc], Togo et al, 2004) and showed immediate upregulation of apop-
liver regenerating factor-1 [LRF-1], early growth response-1, totic genes (Fas and caspases) in livers that failed owing to
[EGR-1]), cytokines, and growth factors (Huang et al, 2006; excessive resection (Morita et al, 2002) (Fig. 6.4).
Turmelle et al, 2006). In addition, many NR ligands can induce As shown above, the initiation of liver regeneration is associ-
hepatocyte proliferation, also in the absence of liver injury (i.e., ated with strong inflammatory cell signals. It is necessary for
‘‘direct hyperplasia’’) (Columbano & Shinozuka, 1996; Colum- this inflammation to be controlled to allow progression of the
bano et al, 2005; Locker et al, 2003). This is the case of fibrates regenerative pathways. The NF-κB inhibitory and ubiquitin-
(agonists of PPAR-α), thyroid hormones, and the halogenated editing A20 protein (TNFAIP3) plays a key role in the liver’s
hydrocarbon TCPOBOP (agonist of the constitutive androstane protective response to injury, particularly to its antiinflamma-
receptor, CAR) (Columbano & Ledda-Columbano, 2003; tory armamentarium (da Silva et al, 2014).
Ohmura et al, 1996). In a model of 70% mouse hepatectomy, A20 is significantly upregulated in the liver following PHx
the NRs FXRβ, PPAR-δ, and THRA mRNA expression levels and protects hepatocytes from apoptosis and ongoing inflam-
represented the identity card of proliferating cells, thus high- mation by inhibiting NF-κB( da Silva et al, 2013; Longo et al,
lighting these NRs as candidate biomarkers of liver prolifera- 2005). A20 also allows proliferation and optimizes metabolic
tion, and as potential targets for novel pharmacological control and energy production following LR (liver regeneration,
approaches (Vacca et al, 2014). liver resection), as demonstrated by increased enzymatic activ-
Another humoral factor that triggers the concerted regen- ity of cytochrome c oxidase or mitochondrial complex IV
erative response in hepatocytes was discovered. Extracellular (Damrauer et al, 2011). A20-based therapies could be benefi-
adenosine triphosphate (ATP) has emerged as a rapidly acting cial in future prevention and treatment of hepatic failure after
signaling molecule that, after PHx, leads to rapid and transient liver resection.
activation of c-Jun–amino-terminal kinase (JNK) signaling, The cytokine-induced form of nitric oxide synthase (iNOS)
induction of immediate early genes FOS and JUN, and also seems to play an important role in scavenging oxygen radi-
AP-1 DNA-binding activity (Gonzales et al, 2010; Tackett et al, cals and protection from apoptosis, caused by an uncontrolled
2014). inflammatory reaction mediated through IL-6 and TNF-α (Rai
et al, 1998).
Distinct Intracellular Pathways in Liver Regeneration The HGF/c-Met pathway is important for sustaining DNA
The importance of liver regeneration in sustaining life is dem- synthesis after injury and activates various downstream path-
onstrated by the plethora of well-characterized pathways ways that involve phosphoinositide 3-kinase (PI3K), ERK, and
involved and the fact that liver regeneration is usually only serine-threonine kinase Akt/protein kinase B (AKT) (Huh et al,
blunted or delayed in experimental models that lack a “critical” 2004). This pathway cross-talks with the Wnt/β-catenin signal-
pathway. The immediate early genes that are activated in regen- ing pathway, which has come to the forefront in liver biology
erating liver were first identified by Mohn and colleagues (1990) during the last several years (Monga, 2011). Increased levels of
and Haber and colleagues (1993), who described induction HGF results in β-catenin dissociation along with nuclear trans-
patterns of more than 70 genes activated in the first week after location (Monga et al, 2002) and upregulation of downstream
two-thirds PHx in a rat model, normally not expressed in the targets of this pathway, such as cyclin D1, MYC, uPAR, matrix
quiescent liver. A mouse complementary DNA (cDNA) array metalloproteinases (MMPs), and EGF receptor (Michalopou-
has also been used to study the crucial role of IL-6 in liver los et al, 1997). Vascular endothelial growth factor (VEGF)
regeneration in IL-6–deficient mice posthepatectomy, showing interacts with endothelial cells in the liver to increase HGF
a 36% reduction in the early gene expression compared with production from nonparenchymal cells (LeCouter et al, 2003).
wild-type mice. In the absence of IL-6, activation of genes such In the transplant setting, microarray analysis of SFS rat liver
as those encoding the uPA receptor (uPAR) and PAI-1, crucial grafts showed upregulation of vasoconstrictive and adhesion
for HGF activation and the MAPK pathway, were remarkably molecule genes at early time points postreperfusion, with later
delayed (Li et al, 2002). IL-6 activates the JAK/STAT3 and increases in genes associated with inflammation and cell death
MAPK signaling pathways via the gp130/IL-6R complex. This and downregulation of genes related to energy metabolism
leads to activation of an array of immediate and delayed early (Man et al, 2003), which was confirmed in the situation of
genes required for normal liver-specific metabolic functions, clinical deceased-donor and living-donor liver transplantation
repair and hepatoprotection from injury (Taub, 2003; Wueste- (LDLT) (de Jonge et al, 2009).
feld et al, 2003; Zimmers et al, 2003). STAT3 is crucial for cells
to progress from G1 to S phase and for activating the MYC gene, Remodeling of the Liver
a gene required for cell cycle progression. Other intracellular Remodeling of the newly regenerated liver tissue begins with
signaling pathways that involve the receptor tyrosine kinases the repopulation and maturation of nonparenchymal cells, such
p38, mitogen-activated protein kinase (MAPK), phosphorylated as endothelial cells, stellate cells, and biliary epithelial cells.
extracellular signal-regulated kinases (pERKs), and JNK are Newly formed hepatocytes form clusters into which replicating
also rapidly activated. Progression through the cell cycle is regu- endothelial cells invade to form new sinusoids. To restore
lated by cyclins and cyclin-dependent kinases (CDKs). Various normal architecture, stellate cells, which are located between
combinations bind to form kinase complexes that are active at endothelial cells and hepatocytes, synthesize extracellular
distinct points within the cell cycle and tightly controlled by matrix (ECM) proteins and TGF-β1, which can regulate the
Chemokines,
Hormones,
Transmitters
Survival factors Growth factors Extracellular
(e.g., interleukins,
(e.g., IGF-1) (e.g., TGF- α, EGF ) matrix
serotonin)
GPCR Integrins
RTK RTK
cdc42 Wnt
Grb2/SOS Fyn/Shc
PLC
Frizzled
P13K G-Protein Ras FAK Dishevelled
Src
AKT Raf GSK-3 β
PKC Adenylate
cyclase MEK Hedgehog
Akk α NF- κB APC
Cytokines PKA MEKK MAPK MKK

IκB β-catenin
JAKs
Patched
(e.g., EPC)
STAT3, 5
TCF
Cytokine receptor
Myc: Mad: ERK JNKs

Bcl-xL Max Max β-catenin: TCF
Fos Jun
SMO
Cytochrome C CREB CyclD
p16 Gli
Rb CDK4
Caspase-9 p15
E2F
Gene regulation
CyclE p27
Caspase-8 Apoptosis
ARF CDK2
p21
Cell
FADD mdm2
Bcl-2 proliferation
p53
Bad
Mt Bax
abnormality
FasR sensor Bim
Death factors
(e.g. FasL,TNF)
FIGURE 6.4 Intracellular pathways of liver regeneration. Multiple intracellular signaling pathways are rapidly activated. Progression through the cell
cycle is regulated by cyclins and cyclin-dependent kinases (CDKs), tightly controlled by several mechanisms, including binding by CDK-inhibitory
proteins, such as p21. APC, Activated protein C; ARF, ADP-ribosylation factor; Bcl, B-cell lymphoma 2; CREB, cAMP response element-binding
protein; EGF, epidermal growth factor; EPC, endothelial progenitor cell; ERK, extracellular signal-regulated kinase; FADD, Fas-associated death
domain; FAK, focal adhesion kinase; FasR, Fas receptor; GPCR, G-protein–coupled receptors; GSK, glycogen synthetase kinase; IGF-1, insulin-like
growth factor-1; JAK, Janus activating kinase; JNK, c-Jun–amino-terminal kinase; MAPK, mitogen-activated protein kinase; MEK, MAP/ERK kinase;
MEKK, MAP/ERK kinase kinase; Mt, mitochondrion; NF-κB, nuclear factor kappa B; PI3K, phosphoinositide 3-kinase; PKA, PKC, protein kinases A
and C, respectively; PLC, phospholipase C; RTK, receptor tyrosine kinase; SMO, Smoothened (protein); SOS, Son of Sevenless protein; Src, sarcoma;
STAT, signal transducer and activator of transcription; TCF, T-cell factor; TGF, transforming growth factor; TNF, tumor necrosis factor. (Original from
http://en.wikipedia.org/wiki/File:Signal_transduction_pathways.png.)
production of hepatic ECM. VEGF, angiopoietins 1 and 2, inhibitory effect on proliferating hepatocytes, potentially signal-
TGF-α, fibroblast growth factor (FGF)-1 and FGF-2, and ing the end of rapid regeneration (Kim et al, 1998).
HGF all are likely involved in the angiogenic process. Angio-
statin, an inhibitor of angiogenesis, causes delayed and sup- Maintaining Liver Function During Regeneration
pressed liver regeneration in mice (Drixler et al, 2002; Vogten After volume loss, hepatocytes must adapt rapidly and seek a
et al, 2004). Remodeling of the ECM is associated with the compromise between maintenance of continued differentiated
activation of the urokinase/plasminogen pathway and the function and cellular replication to permit survival. After toxic
membrane-bound MMP pathway. MMP-9 is one of the most injury, resection, or transplantation, the balance is dramatically
active proteins during liver remodeling and regeneration (Kim shifted to the crucial tasks of recovery and regeneration at the
et al, 2000). MMPs not only remodel the ECM but also expense of normal hepatic metabolism. The success of restoring
regulate immune responses (Duarte et al, 2015). Compared lost liver mass, repairing tissue injury, and resolving inflamma-
with fatty controls, MMP-9(−/−) steatotic livers showed signifi- tion determines the ability of the liver to support normal meta-
cantly reduced leukocyte infiltration, proinflammatory cytokine bolic function and determines the ability of the liver to recover
expression, and liver necrosis. Loss of MMP-9 activity preserved (Huang & Rudnick, 2014; Strey et al, 2004) (Fig. 6.5). Several
platelet endothelial cell adhesion molecule-1 expression, a of the expressed immediate early genes encode enzymes and
modulator of vascular integrity at the endothelial cell-cell junc- proteins that are involved in regulating gluconeogenesis, a very
tions in steatotic livers after I/R injury (Kato et al, 2014). important process after PHx to compensate for the lost glyco-
The MMPs in combination with HGF, EGF, and TGF-β1 gen content and to produce sufficient glucose for the whole
act to remodel the ECM, changing the levels of several ECM organism (Haber et al, 1995; Rosa et al, 1992). There is rapid
proteins, such as collagen, fibronectin, laminin, and entactin. increased expression of genes involved in glucose homeostasis
The maturation and thickening of the ECM seems to have an after PHx. Most notably, these include phosphoenolpyruvate
Metabolism: Recovery from

Synthesis injury:
Storage Regeneration
Degradation Apoptosis
Homeostasis Inflammation
Detoxification Restructuring
Liver function
Increased Increased need

metabolic for recovery and
stress regeneration
Recovery Metabolic
regeneration function
Patient factors: Liver factors:

• Critical illness • Age
• Medical comorbidities • Steatosis
• Infection • Remnant size
B • Complex surgery • Ischemia
FIGURE 6.5 Metabolic balance between regeneration and maintaining liver function. A, In times of relative quiescence, there is a balance within
the liver of metabolic function and continuous liver cell replacement or restructuring as needed. B, In times of stress or after injury or resection, there
is an increased need for metabolic function or regeneration and recovery. If metabolic need is great as a result of conditions within the patient, there
may not be sufficient energy balance within the liver to regenerate sufficiently, and the liver may not recover. If the requirements for regeneration and
repair are overwhelming, there may be decreased metabolic function, affecting patient outcome.
carboxykinase (PEPCK), glucose-6-phosphatase (G6Pase), after PHx require mitochondrial fatty-acid β-oxidation. PPAR-
and insulin-like growth factor binding protein-1 (IGFBP-1), α may be a crucial modulator controlling energy flux important
controlled at the level of transcription by insulin (downregula- for repair of liver damage and regeneration (Anderson et al,
tion), glucagon/cyclic adenosine monophosphate (upregula- 2002).
tion), and glucocorticoids (upregulation) (Diamond et al, 1993; In a microarray analysis of gene expression profiles after
Mohn et al, 1990). Livers lacking IGFBP-1 display abnormal living-donor liver transplantation, we demonstrated that C/
liver regeneration (Leu et al, 2003). Decreased expression of EBPα was downregulated, as well as HNF-4α and PPAR-α (de
genes that oppose gluconeogenesis also occurs rapidly pos- Jonge et al, 2009). Expression of many other liver-specific
thepatectomy. Insulin itself can be a potent growth factor genes, such as IGFBP-1 and G6Pase, are regulated in the basal
mediated through the insulin receptor, and insulin and gluca- state by HNF-1. The transcriptional activity of HNF-1 is
gon have long been established as important “gut-derived” upregulated during liver regeneration by binding of HNF-1 to
growth factors (Starzl et al, 1975). Liver-specific transcription the growth-induced transcription factors STAT3 and AP-1
factors (hepatocyte nuclear factors [HNFs]), have an important (Leu et al, 2001).
role in determining the level of glucose production, fatty-acid New insights into how the liver fulfills the adaptive response
metabolism, and liver-specific secreted proteins. C/EBPα regu- to metabolic needs during regeneration, may come from the
lates expression of genes involved in hepatic glucose and lipid tight regulation of lipid, glucose, and bile acid (BA) metabolism
homeostasis, has antiproliferative proprieties, and is downregu- through the class III NAD1-dependent histone deacetylase
lated during liver regeneration after hepatectomy (Costa et al, SIRT1 (Ruderman et al, 2010). SIRT1 is activated in situations
2003; Mischoulon et al, 1992). of low energy availability and links nutritional status with meta-
During early regeneration, the liver accumulates fat. Neither bolic homeostasis. It regulates adenosine monophosphate
the mechanisms responsible for nor the functional significance activated protein kinase (AMPK). Contrary to SIRT1, mam-
of this transient steatosis has been determined. Suppression of malian target of rapamycin (mTOR) is activated in high-energy
hepatocellular fat accumulation is associated with impaired conditions and controls cell growth and proliferation (Sengupta
hepatocellular proliferation after PHx, indicating that hepato- et al, 2010). mTOR promotes protein synthesis and this axis is
cellular fat accumulation is specifically regulated during, and essential to regulate the cell cycle during liver regeneration after
may be essential for, normal liver regeneration (Shteyer et al, PHx. BA is also essential for the regeneration of the liver after
2004). Some data show that decreasing lipid peroxidation levels PHx (Huang et al, 2006), although, when present in excess,
by vitamin treatment after PHx produces an attenuation of BA can be toxic and promote hepatocyte death. Therefore a
cellular apoptosis and a marked increase in the proliferation fine regulation of BA metabolism comes from the tight regula-
process, suggesting that the modulation of lipid peroxidation tion of lipid, glucose, and BA metabolism through the class
has a role in the liver regeneration process (Ronco et al, 2002). III nicotinamide adenine dinucleotide-1–dependent histone
Hepatocytes in the periportal regions that divide and replicate deacetylase sirtuin 1 (SIRT1) (Ruderman et al, 2010). SIRT1
is activated in situations of low-energy availability and links the role of C/EBPα, a key regulator of liver proliferation, in the
nutritional status with metabolic homeostasis. It regulates termination of regeneration was demonstrated. Complex forma-
adenosine monophosphate–activated protein kinase. Contrary tion of C/EBPα and the chromatin remodeling protein HDAC1
to SIRT1, mammalian target of rapamycin (mTOR) is activated represses other key regulators of liver proliferation: C/EBPα,
in high-energy conditions and controls cell growth and prolif- p53, FXR, SIRT1, PGC1α (peroxisome proliferator-activated
eration (Sengupta et al, 2010). mTORC1 promotes protein [PPAR]-γ coactivator-1), and telomerase reverse transcriptase
synthesis, and this is essential to preserve liver homeostasis and (TERT). The C/EBPβ-HDAC1 complexes also repress promot-
a proper response to injury. The farnesoid X receptor FXR ers of enzymes of glucose synthesis PEPCK and G6Pase. Proper
(NR1H4) is the master regulator of BA, lipid, and glucose cooperation of C/EBP and chromatin remodeling proteins
metabolism. Recently, the role of SIRT1 as a key regulator of seems essential for the termination of liver regeneration after
the regenerative response of the liver, controlling BA homeo- surgery and for maintenance of liver functions (Jin et al, 2015).
stasis, protein synthesis, and cell proliferation through deacety- Additional work strongly implicates the detection of blood
lation of FXR and histones, and regulation of mTOR was BA levels by nuclear receptors, as a regulator of liver growth
established (Garcia-Rodriguez et al, 2014). (Huang et al, 2006). In addition, Mst1 and Mst2, mammalian
genes comparable to those in the Drosophila Hippo kinase
Termination of Proliferation signaling cascade, which regulates wing mass during develop-
The size of the liver is highly regulated and is controlled by the ment, can also control hepatocyte proliferation (Dong et al,
functional needs of the organism. This observation infers the 2007). Overexpression of YES-associated protein (YAP), the
existence of a master regulator of the liver/body-mass ratio, that mammalian counterpart to Yorki—the last gene in the Dro-
is a “hepatostat.” sophila Hippo kinase cascade—in a transgenic mouse model
From LDLT, we know that differences are present between leads to massive liver hyperplasia, reaching 25% of body weight.
donors and recipients in the percentage reconstitution of the YAP transcriptionally activates cell-cycle proteins such as Ki-67
standard liver volume (LV) (80% vs. 93% at 3 months), prob- and c-Myc and also inhibitors of apoptosis, and its phosphory-
ably caused by the need for functional liver mass to compensate lation by way of activation of the Hippo pathway blocks its
for long-standing liver disease (Olthoff et al, 2015). ability to shuttle to the nucleus (Reddy & Irvine, 2008). It is
The most well-known antiproliferative factors within the therefore possible that the Hippo kinase pathway has a decisive
liver are TGF-β and related family members such as activin role in determining overall liver size (Hong et al, 2014; Plouffe
(Derynck & Zhang, 2003). TGF-β is produced mainly by et al, 2015; Wu et al, 2015; Zheng et al, 2011).
hepatic stellate cells, but in the early phase, it forms inhibitory
complexes with SKI proto-oncogene (SKI) and SKI like proto- LIVER ATROPHY
oncogene (SnoN) (Macías-Silva et al, 2002), rendering hepato-
cytes initially resistant to TGF-β (Koniaris et al, 2003). Later Classically, atrophy is triggered by an obstruction of portal
on, TGF-β acts through a heteromeric receptor complex, which venous blood flow or is the result of chronic obstruction of the
subsequently phosphorylate proteins of the SMAD family bile duct. When atrophy occurs unilaterally, the opposite lobe
(protein homologs of both the Drosophila protein mothers of the liver responds with a hypertrophic response, and the
against decapentaplegic (MAD) and the Caenorhabditis elegans anatomic result is a rotation of the liver around the hilar axis.
protein SMA), particularly SMAD2 and SMAD3 (Lonn et al, The liver frequently is grossly distorted, and anatomic land-
2009). The downregulation of microRNA 23b (miR23b) may marks are markedly changed, most commonly seen accompa-
further contribute to activation of the TGF-β1/SMAD3 signal- nied by a rotation of the portal triad structures (Fig. 6.6). In
ing pathway during the termination stage (Yuan et al, 2011). recent years, surgeons have sometimes used preoperative embo-
Upon activation, Smad2, Smad3 and Smad4 assemble in a lization of the right portal vein to induce hypertrophy when the
common complex, translocate into the nucleus and activate. future liver remnant (FLR) is small.
The Smad2/3-Smad 4 complex translocates into the nucleus
and activates target genes that negatively regulate the cell cycle Mechanisms of Liver Atrophy
(Nakao et al, 1997). Reactive oxygen species (ROS) enhance The death of liver cells in atrophy generally is divided into
synthesis and activation of TGF-β (Koli et al, 2008), which may necrosis and apoptosis. The distinction is important because
account for the reduced regeneration after ischemia and reper- necrosis is a nonregulated traumatic disruption of a cell occur-
fusion. Interacting with the TGF-β/Smad signaling could restore ring when it encounters overwhelming injury, whereas apoptosis
regeneration in a model of SFS liver grafts (Zhong et al, 2010). is an inducible, highly orchestrated cascade of events that is
Similarly, activin A blocks hepatocyte mitogenesis and shows physiologic. Necrotic cells lose membrane integrity, leak lyso-
diminished signaling during liver regeneration when its cellular- somal enzymes, and induce a large inflammatory response.
receptor level is reduced. Its receptor level is restored once liver Apoptosis is energy dependent and allows cells to shrink and
regeneration is terminated (Date et al, 2000). The level of die without inducing inflammation. Apoptotic cells are charac-
activin receptor messenger RNA (mRNA) expression was terized by plasma membrane blebbing, chromosomal conden-
shown to be an important determinant in the magnitude of sation, and nuclear DNA fragmentation (Steller, 1995).
regeneration in portal vein ligation and PHx (Takamura et al,
2005a, 2005b). SOCSs are important negative regulators of Portal Vein–Induced Hepatic Atrophy
cytokine signaling that prevent the tyrosine phosphorylation of The most common clinical scenario of hepatic atrophy is portal
STAT proteins. It has been shown that IL-6 signaling in the liver vein–induced atrophy; occlusion of the portal vein leads to
causes the rapid upregulation of SOCS3, which correlates with ischemia in areas of the liver, and atrophy in this setting is a
the subsequent downregulation of phosphorylated STAT3, result of liver cell death secondary to necrosis and apoptosis
thereby terminating the IL-6 signal (Campbell et al, 2001). Also (Ikeda et al, 1995; Shibayama et al, 1991). Ischemic necrosis
A B
C
FIGURE 6.6 Hepatic atrophy. A, Computed tomography appearance of the liver in a patient with papillary hilar cholangiocarcinoma involving the
left hepatic duct. Note the atrophic left lobe and intrahepatic ductal dilation predominantly on the left. B, Gross appearance of the liver in the same
patient. Inset shows intraluminal view of the common bile duct with tumor extruding from the left hepatic duct (arrow).
of centrilobular areas of the liver predominates in the first 3 impaired apoptosis and less injury and fibrosis compared with
days of cell death. Areas peripheral to the necrotic liver cells wild-type mice (Canbay et al, 2002; Gujral et al, 2004; Miyoshi
predominantly undergo apoptotic cell death, and apoptosis et al, 1999). More recent data, however, suggest a nonischemic
persists long after necrosis subsides. Oxygen levels and mito- model of necrosis/oncosis as the predominant process leading
chondrial function help determine which cells will undergo to cell death after common bile duct ligation, with cell swelling
necrosis or apoptosis (Nishimura et al, 1998). Models of portal and without apoptotic caspase-3 activation (Fickert et al, 2005).
vein ischemia in rats have confirmed a caspase-dependent
apoptosis and have indicated that Kupffer cells are involved in Clinical Causes of Atrophy
generating reactive oxygen substrates and other acute-phase Hilar cholangiocarcinoma is a frequent cause of biliary atrophy,
reactants, culminating in mitochondrial dysfunction and apop- induced by occlusion of a major bile duct. Biliary occlusion,
tosis (Kohli et al, 1999). often accompanied by portal vein compromise leading to
atrophy of the liver, occurs approximately 20% of the time with
Biliary-Induced Hepatic Atrophy this disease and has significant surgical implications. Also
The molecular mechanisms involved in biliary obstruction postcholecystectomy bile duct strictures lead to hepatic atrophy
leading to hepatic atrophy are much more centered on apopto- 10% to 15% of the time (Hadjis & Blumgart, 1987; Pottakkat
sis, with little or no involvement of acute necrosis. Cholestasis et al, 2009). Choledocholithiasis and hepatolithiasis are infre-
results in the accumulation of toxic bile salts, which induce quent causes of atrophy, as are benign tumors, such as papil-
apoptosis through the Fas-mediated pathway. In this case, lomas, cystadenomas, and granular cell tumors (Blumgart &
TNF-α and Fas ligand bind to the Fas death receptors, leading Kelly, 1984). Strictures caused by parasitic biliary infections,
to a cascade of intracellular events, including cytochrome c such as Clonorchis sinensis and Ascaris lumbricoides, also have
release from mitochondria and activation of apoptosis-mediating been known to cause biliary obstruction and associated atrophy
caspases. In Fas-deficient mice, bile duct ligation resulted in of the liver. Occlusion of the hepatic artery alone would not
induce atrophy, although arterial radioembolization techniques due to high mitochondrial susceptability to oxidative stress
are currently used to increase resectability in the treatment (Lenaz et al, 2000).
of HCCs (Toso et al, 2010). Lobar radioembolization with
yttrium-90 or holmium-166 can cause atrophy of a lobe through Biliary Obstruction
β-emission, causing necrosis. This way “radiohepatectomy” Cholestasis results in the accumulation of toxic bile salts, which
allows growth of the contralateral lobe with a median increase induce apoptosis through the Fas-mediated pathway. In this
of 24% after 3 months (Garlipp et al, 2014; Vouche et al, 2013), pathway, TNF-α and Fas ligand bind to the Fas death recep-
which may enable resection. tors, leading to cytochrome c release from mitochondria and
activation of apoptosis-mediating caspases. In Fas-deficient
Compensatory Regeneration Triggered by Atrophy mice, bile duct ligation resulted in impaired apoptosis and less
In the embolized portion of liver, Kupffer cells release TNF-α, injury and fibrosis compared with wild-type mice (Canbay et al,
inducing hepatocyte swelling and hyperplasia in the contralateral 2002; Gujral et al, 2004; Miyoshi et al, 1999). Other molecular
liver. Additionally, increased portal flow in the nonembolized mechanisms involved in reduces regenerative capacity with
lobe induces proliferation and activates several cytoplasmic biliary obstruction are much suppressed expression of c-Myc
growth-promoting signal transduction pathways involved in (Tracy et al, 1991), C/EBP, and cyclin E (Nakano et al, 2001).
hepatic regeneration, including c-Jun and the MAPK pathways. Production of HGF (Hu et al, 2003), EGF (Bissig et al, 2000),
Kupffer cells in the contralateral liver begin to produce TNF-α and IL-6 (Fujiwara et al, 2001) is also altered. Finally, biliary
as well, stimulated either by increased portal flow or through obstruction also impairs enterohepatic circulation and thus
the direct effect of circulating growth factors (Kim et al, 2001). negatively affects the regenerative capacity. Clinically relevant
Stellate cells in the injured portion of the liver produce HGF is that external biliary drainage for obstructive jaundice mark-
(Parekkadan et al, 2007; Skrtic et al, 1999). mRNA levels of edly suppresses liver regeneration after PHx (Suzuki et al,
HGF in the ischemic portion of liver were noted to be 10 to 1994), whereas internal biliary drainage preserves this capacity
20 times that of normal, with most upregulation occurring in (Saiki et al, 1999). The mechanism was demonstrated in a rat
Kupffer cells (Hamanoue et al, 1992). model in which oral BAs were given before PHx. Liver regen-
eration was significantly increased through activation of the
FACTORS INFLUENCING LIVER REGENERATION FXR signaling pathway (Ding et al, 2015).
Patient-Related Factors Diabetes Mellitus

Age Insulin is one of the most important hepatotrophic factors in
Liver age is a significant factor in hepatocellular regeneration. portal venous blood (Starzl et al, 1975). The binding protein of
Older livers do not regenerate as quickly as younger livers and IGF, a molecule similar to insulin, also rises substantially after
show delayed regeneration after acute injury and impaired PHx (Demori et al, 2000; Ghahary et al, 1992). Thus impaired
function after liver transplantation (Burroughs et al, 2006; Feng secretion of insulin and IGF in diabetic patients prevents liver
et al, 2006). Rodent models have shown reduced and delayed regeneration following PHx, as reflected by decreased synthesis
thymidine kinase uptake in older animals after PHx, and there of RNA, DNA, and protein on the first postoperative day
is a striking difference in the magnitude of DNA synthesis and (Yamada et al, 1977). Enhancement of mitochondrial phos-
timing of hepatocyte replication between young and old livers phorylation activity in the remnant liver following PHx is
(Taguchi et al, 2001). The CDK inhibitor p21 is expressed at inhibited in proportion to the severity of impaired insulin secre-
high levels in old mouse livers, as is cyclin B1, a regulator of tion. Insulin gene transfer via the spleen enhances liver regen-
G2/M phase of the cell cycle (Ledda-Columbano et al, 2004). eration without causing liver damage and improves nutritional
Aging has been shown to be associated with a progressive status after hepatectomy in diabetic rats (Matsumoto et al,
decline in growth hormone secretion and Foxm1B expression 2003). Multiple regression analysis in clinical PVE has shown
(Wang et al, 2002). Treatment of old mice with growth hormone that diabetes mellitus is a risk factor for reduced hypertrophy
can restore hepatocyte proliferation, with increased Foxm1B in the nonembolized lobe (Imamura et al, 1999). These results
and cyclin B1 expression and significant reduction in p27 demonstrate the importance of insulin in hepatic regeneration
protein levels (Krupczak-Hollis et al, 2003). Recently, a critical and strict glucose control should be aimed for in liver surgery
role of the glycogen synthase kinase 3β-cyclin D3 pathway in and PVE (Yokoyama et al, 2007).
the loss of the regenerative capacity in old livers was shown,
which could be overcome by exogenous growth hormone Nutritional Status
substitution (Jin et al, 2009). Also, aging switches the C/EBP-α Hepatic regeneration is metabolically intensive and requires a
pathway of growth arrest in liver from CDK inhibition to large amount of energy. Liver regeneration after hepatectomy
repression of E2F transcription. This blocks the activation of is associated with a derangement in energy metabolism, mea-
the c-Myc promoter in old livers after PHx and in tissue culture sured by a decrease in the ratio of ATP to its hydrolysis product
models (Iakova et al, 2003; Timchenko, 2009). inorganic phosphate. This depleted energy status is mirrored in
On a clinical level, several publications on living-donor biochemical indices of liver function and restitution parallels
transplantation report worse graft survival as donor age the course of restoration of hepatic cell mass (Mann et al,
increases above 45 years (Abt et al, 2004), although it is unclear 2002). Nutritional support is undoubtedly the most physiologic
what the mechanisms behind this involve. Little is known manner to enhance liver regeneration, but there still is little, if
regarding differential gene expression in old versus young livers, any, clear information regarding the effect of specific nutrients
but results from the United Network for Organ Sharing on liver regeneration in humans (Holecek, 1999).
(UNOS) database show reduced resistance in old donor livers From animal models we know that malnutrition is associ-
to prolonged cold ischemia times (Cassuto et al, 2008), possibly ated with higher postoperative mortality and reduced
regeneration following PHx (Skullman et al, 1990). When Steatohepatitis, or acute inflammation in the setting of fatty
improving nutritional status, enteral feeding should be pre- infiltration, carries an even higher risk and eventually results in
ferred, because rats given enteral nutrition showed much better fibrosis and cirrhosis.
weight gain after 70% hepatectomy than those given isocaloric In transplantation, liver grafts with moderate to severe ste-
nutrition parenterally (Delany et al, 1994). atosis (>30% to ≥60%) have higher rates of primary nonfunc-
Recently, a small trial with supplementation with branched- tion, higher aminotransferases, and poorer graft survival in the
chain amino acids–enriched nutrients showed improved nutri- transplant setting (Verran et al, 2003), possibly as a result of
tional state in LDLT recipients in the early posttransplant increased inflammation and the inability to initiate repair and
period and shortened the posttransplant catabolic phase regeneration mechanisms.
(Yoshida et al, 2012). Supplementation of glutamine, one of Progressive scarring and hepatic fibrosis leads to accumula-
the sources of DNA and protein synthesis, has been shown tion of abnormal collagen secreted into the ECM by stellate
to promote liver regeneration (Ito & Higashiguchi et al, 1999). cells (Kim et al, 1998). Impaired regeneration is thought to
Essential fatty acids, components of the cell membrane be the result of the decreased diffusion of nutrients and
and precursors of several functional mediators, also play an hepatotrophic factors to the hepatocytes, and ultimately the
important role in hepatic regeneration. Interestingly, dextrose architectural liver abnormalities created by scar contracture
supplementation has an inhibitory effect of on liver regenera- form a physical barrier, preventing hepatocytes from proliferat-
tion, associated with increased expression of C/EBP-α, p21, ing (Poynard et al, 1997).
and p27 (Weymann et al, 2009), although this effect was not
found in a previous study (Nishizaki et al, 1996). Pharmacologic Therapy
Many exogenous agents can affect liver regeneration, including
Gender frequently prescribed drugs and neoadjuvant chemotherapy.
Sex steroids are known to induce transient hepatocellular pro- Numerous medications associated with impaired liver regenera-
liferation and to improve fatty-acid metabolism (Repa et al, tion, secondary to the induction of steatosis, are certain antiar-
2000). Estrogen receptors are found on hepatocytes, and serum rhythmic agents, antibiotics, antiviral agents, anticonvulsants,
estradiol is increased substantially after PHx in rodents and steroids, calcium channel blockers, statins, and antiglycemic
humans (Francavilla et al, 1989). Pretreatment of rats with medications. β-Blockers and nonsteroidal antiinflammatory
17β-estradiol induces hepatocyte DNA synthesis in vitro and drugs (NSAIDs) exert a more direct negative influence on liver
accelerates liver regeneration in vivo; administration of tamoxi- regeneration. β-Blockers impair liver regeneration by decreas-
fen, a mixed-estrogen agonist/antagonist, slows regeneration ing portal blood flow to the liver and blocking the trophic
when given soon after PHx (Francavilla et al, 1989). In contrast, effects of epinephrine. NSAIDs directly inhibit cyclooxygenase,
testosterone levels decline in men and in male rats after PHx. part of the C/EBP-β–mediated liver regenerative pathway. Liver
However, there is no clinical evidence that shows significant regeneration after PHx is delayed in rodents treated with either
differences between the sexes in humans after liver resection. β-blockers or NSAIDs (Hong et al, 1997; Rudnick et al, 2001).
Although this effect has not been reflected in increased morbid-
Intrinsic Liver Disease: Steatohepatitis ity or mortality after liver resection in patients taking β-blockers
The regenerative response of the liver can be seriously affected or NSAIDs, the potential harmful effect of any medication on
by preexisting intrinsic liver conditions, such as steatohepatitis, hepatic regeneration must be considered before resection.
fibrosis, and cirrhosis. Steatohepatitis is a condition met more There are an increasing number of patients having liver
frequently in the general population (Adams et al, 2005; Bhala surgery after neoadjuvant or induction chemotherapy. Major
et al, 2011) and poses challenges in the management of patients drawbacks of hepatotoxic chemotherapy are the sinusoidal
undergoing liver resection. Hepatic steatosis affects regenera- obstruction syndrome (SOS), associated with oxaliplatin
tion on several molecular levels. Lipid accumulation has been (Eloxatin, Sanofi Aventis, Bridgewater, NJ) (Rubbia-Brandt
associated with hepatocyte mitochondrial damage caused by et al, 2004) and the chemotherapy-associated steatohepatitis
free radical injury. Steatotic livers in rats show delayed mitosis (CASH) (Vauthey et al, 2006), which is associated with irino-
and increased mortality after PHx, which may be due to abnor- tecan (Campto, Pfizer, New York). This chemotherapy-
mal TNF and IL-6 signaling (Selzner et al, 2000). The coordi- associated steatohepatitis increases postoperative mortality and
nated induction of JNKs and ERKs is disrupted after PHx in specifically deaths from postoperative liver failure (Fernandez
fatty livers of ob/ob mice (a model for steatohepatitis), with et al, 2005). Sinusoidal obstruction also impairs liver regenera-
enhanced AKT and inhibition of PEPCK (Yang et al, 2001). tion after extensive liver resections and increases postoperative
Cyclin D1 induction is abolished along with STAT3 and morbidity but may be prevented by concomitant administration
reduced ATP levels, which may arrest cell-cycle progression. of bevacizumab (Avastin, Hoffmann-LaRoche, Indianapolis,
Hepatocyte mitochondrial damage associated with lipid accu- IN) (Rubbia-Brandt et al, 2011; Vigano et al, 2013). Bevaci-
mulation is caused by free radical injury from fatty-acid oxida- zumab, a monoclonal antibody targeting VEGF, is given in
tion. Abnormalities in induction of cytochrome P-450 may be combination with cytotoxic chemotherapy to improve resect-
one mechanism in the pathophysiology of these findings in fatty ability (Gruenberger et al, 2015) and survival in patients with
livers and may contribute to poor regeneration (Farrell, 1999; metastatic colorectal cancer (Hurwitz et al, 2004). Well-
Kurumiya et al, 2000; Neuschwander-Tetri et al, 2003). The designed preclinical studies have demonstrated that inhibition
presence of underlying steatosis has a considerable impact on of angiogenesis can inhibit wound healing (Scappaticci et al,
operative morbidity and mortality after major hepatic resection 2005). Bevacizumab does not appear to adversely affect the
(Behrns et al, 1998; de Meijer et al, 2010; Vetelainen et al, results of PHx in humans (D’Angelica et al, 2007). Animal
2007), with a significantly higher rate of complications if studies have demonstrated that liver regeneration depends on
marked steatosis (≥30%) is present (Kooby et al, 2003). VEGF and angiogenesis (Drixler et al, 2002). In a murine
model, anti–VEGF receptor therapy slightly impaired liver and activation of cytokines, transcription factors, and immedi-
regeneration and cell proliferation after PHx compared with ate early genes and a greater magnitude of hepatocellular rep-
control individuals (Van Buren et al, 2008). Clinically, no dif- lication (Debonera et al, 2001). The liver can tolerate ischemic
ferences in postoperative liver insufficiency were seen if stopped injury only to a certain “point of no return;” however, after
at least 6 to 8 weeks before hepatic resection (Reddy SK et al, which the damage is too extensive, and the liver or allograft is
2008; Zorzi et al, 2008). unable to maintain functional homeostasis and regenerative
Regarding the effect of sorafenib (Nexavar, Bayer), a multi- capabilities, which results in liver dysfunction and graft failure
kinase inhibitor used for treatment of HCC, conflicting results (Debonera et al, 2002).
have been reported in experimental research. Sorafenib did In addition, recent studies have demonstrated that lack of
not impact on liver regeneration when ceased before surgery; blood flow occurring during cold preservation for transplanta-
however, administration after hepatectomy affected late liver tion markedly deteriorates the liver sinusoidal endothelial cells
regeneration in a mouse study, and reduced phospho-ERK (LSEC) protective phenotype by downregulating the expres-
levels and wound-healing complications were observed (Hora sion of the transcription factor Kruppel-like factor 2 (KLF2),
et al, 2011). In another rat experiment, however, no significant which orchestrates the transcription of a variety of protective
change in liver regeneration related to sorafenib exposure was genes, including the endothelial synthase of NO (eNOS), the
found (Shi et al, 2012). antithrombotic molecule thrombomodulin, or the antioxidant
transcription factor NRF2 ( Gracia-Sancho et al, 2010; Peralta
Liver Transplantation et al, 2013; Russo et al, 2012).
Regeneration also is crucial in liver transplantation. In cadaveric
transplantation, hepatocyte loss occurs in the form of I/R injury Minimal Liver Mass
owing to the necessary preservation period from procurement The amount of liver mass transplanted has been shown to be
to implantation and damage that may have occurred in the an important factor after transplantation. Early experimental
donor. Regenerative mechanisms are actively engaged after studies addressing regeneration after transplantation showed
transplantation, depending on the length and degree of preser- that a SFS graft adapts to its environment and achieves a size
vation injury (Debonera et al, 2001; Olthoff, 2002). Similarly, equal to the original native liver (Kam et al, 1987). It became
hepatocytes lost to the alloimmune response require replace- apparent that the graft size-to-recipient ratio was crucial in that
ment. Regeneration also is necessary in the setting of trans- grafts that were too small had decreased survival (Francavilla
planting a SFS graft into a larger recipient. This is the case in et al, 1994). These findings correlated with early clinical experi-
the setting of adult-to-adult living-donor liver transplantation ence in living-donor transplantation in that some SFS segmental
(AALDLT) (Olthoff, 2003). Ischemic injury is minimized in grafts developed a “SFS syndrome” that was associated with
AALDLT, in that the preservation period is short; however, this significant functional impairment, shown by prolonged cho-
technique supplies a graft that is by definition too small, requir- lestasis and histologic changes consistent with ischemic injury
ing vigorous immediate hepatocyte proliferation. By transplant- and associated with poor outcome. Liver grafts with a graft
ing only 50% to 60% of what is the expected LV in adults, volume of less than 40% of calculated standard LV were associ-
recipients (and donors) must rely on the rapid regeneration of ated with poor graft survival and prolonged hyperbilirubinemia
a partial liver in addition to maintaining the basic metabolic (Emond et al, 1996; Kiuchi et al, 1999; Lo et al, 1999; Zhong
functions required of the liver. The National Institutes of Health et al, 2006). The development of segmental graft dysfunction
(NIH)-sponsored study in adult-to-adult living-donor trans- in the A2ALL study was also associated with worse patient
plantation (A2ALL) has investigated the role of numerous outcome (Olthoff et al, 2015). Animal models of partial liver
donor and recipient factors in regeneration. The size of the graft transplantation have studied the interplay between the
remnant liver or graft had the greatest impact on rate and regenerative response and ischemic injury in the setting of 50%
quantity of regeneration (Olthoff et al, 2015). and 30% size grafts. The partial grafts showed a robust regen-
In a pilot study investigating molecular mechanisms associ- erative response if the ischemic injury was minimal. It became
ated with human liver regeneration, we noted differences in apparent, however, that when these partial grafts were subjected
hepatic gene expression between donors with complete regen- to ischemic injury of moderate to prolonged time periods, there
eration compared with those with less successful regeneration— was a significant effect on survival, with extensive hepatic
genes mainly related to cell proliferation, inflammation, necrosis, the inability to initiate or maintain the regenerative
metabolism (aminoacyl transfer RNA synthesis), and stress response, and decreased survival. These findings show the
pathways (acute-phase response) were among the most signifi- diminished tolerance of SFS grafts for additional injury beyond
cantly regulated pathways. In contrast, the poor regeneration transplantation itself ( Debonera et al, 2004; Selzner et al,
group demonstrated very little change in expression before and 2002).
after resection. The lack of significant change in genomic profile Although ischemic injury in AALDLT is minimized, the
in the poorly regenerating livers suggests a possible inhibition amount of critical liver mass required for transplantation in
or delay in initiation of recovery and regeneration molecular living donation remains in question. Most centers have defined
pathways (Hashmi et al, 2015). liver mass as graft-to-recipient body-weight ratio or as a per-
centage of the standard LV. No uniform method of measuring
Ischemic Injury or reporting graft volume in relation to the recipient has been
Warm and cold ischemic injury is an unavoidable component established. Clinical experience with living-donor and split
of transplantation. After prolonged cold ischemia of whole-liver grafts has led to an accepted lower limit of 0.8% graft-to-
grafts, there is initiation of the cell-cycle pathways with upregu- recipient body-weight ratio, or 40% of the standard LV. Donor
lation of markers of liver regeneration as described earlier. and recipient characteristics, and graft factors, significantly
When ischemic injury is significant, there is a greater expression influence these minimal accepted standard volumes. Patients
with fulminant hepatic failure, severe portal hypertension, and 2004; Tralhao et al, 2002), have been reported to inhibit
significant disease severity, as manifested by a high Model for hepatic regeneration. The precise mechanism responsible for
End-Stage Liver Disease (MELD) score, and patients with viral infection–related suppression is unclear, although it may
significant metabolic stress may require more LV than stable be partly mediated by the inhibition of cell-cycle–dependent
patients transplanted under elective conditions (Marcos et al, molecules. Because most transplant patients in the United
2000). Although AALDLT in acutely ill patients with fulminant States have hepatitis C, it is important to know whether this
failure is often performed successfully, usually in parts of the vigorous regenerative response in the transplanted graft has a
world with no access to deceased donor transplantation, many significant effect on the kinetics of viral replication in hepatitic
centers in the West are not performing them because of the C virus (HCV)–positive individuals. Investigators have noted
uncertainty of knowing if a partial graft has enough volume to that HCV RNA replication is enhanced in proliferating cells,
support the recovery of such a recipient (Campsen et al, 2008; suggesting that viral replication is regulated by cell-cycle–
Olthoff et al, 2011). The accumulation of additional stressful dependent factors (Erhardt et al, 2002). Clinically, progressive
stimuli, such as sepsis or renal failure, may push a relatively liver disease has been reported to be associated with increased
small graft into failure. hepatocyte proliferation (Poynard et al, 1997). Alternatively,
HCV core proteins have been shown to have significant inter-
Effect of Immunosuppression action with cellular promoters and regulators of cell growth,
Within the graft environment, the host immune response needs which may affect liver regeneration (Blindenbacher et al, 2003;
to be inhibited to avoid acute allograft rejection, and inhibition Hayashi et al, 2000; Lee et al, 2002; Ray et al, 1998; Shrivas-
of this response may interfere with the recovery of liver grafts tava et al, 1998). Early single-center studies suggested
requiring active regeneration of hepatocytes. Glucocorticoids, that recipients of living-donor liver transplants may have an
routinely used in immunosuppression protocols, have been earlier and more severe recurrence of HCV compared with
shown to inhibit cell-cycle progression markedly in PHx recipients of whole deceased-donor grafts; however, this was
models and in transplant models with ischemic injury not validated in any of the prospective controlled studies
(Debonera et al, 2003; Nagy et al, 1998; Tamasi et al, 2001). (Berenguer, 2006; Ghobrial et al, 2002; Shiffman et al, 2004;
Cyclosporine and tacrolimus may have differential effects on Terrault et al, 2007).
regeneration in a dose-dependent fashion (Francavilla et al, Concomitant bacterial infections also alter liver regenera-
1991). Sirolimus, with its antiproliferative action, interferes tion. Earlier reports showed enhanced liver regeneration in rats
with hepatocyte replication (Francavilla et al, 1992; Palmes after inflammation prior to hepatic resection, due to stimula-
et al, 2008). The rapid hepatocyte replication and smaller liver tion of LPS; upregulation of proinflammatory cytokines, such
mass also may interfere with metabolism and pharmacokinet- as IL-6 and TNF-α; and HGF ( Sekine et al, 2004; Weiss et al,
ics of certain drugs. Preliminary studies have shown that 2001), all of which are chief mediators of hepatic regeneration.
AALDLT recipients require lower doses of tacrolimus in the A recent study showed—more in line with clinical observations—
early postoperative period than patients receiving whole grafts a significantly delayed regeneration kinetic in a rat model
(Trotter et al, 2002). The ability to measure functional recov- of combined liver resection and intraperitoneal sepsis, with
ery of these recipients would help in the assessment of hepa- hyperinflammation and increased liberation of proinflamma-
tocellular function and metabolic demands in these regenerating tory cytokines (Seehofer et al, 2007). The relation between
partial liver grafts. inflammation and poor liver regeneration was confirmed in
patients with early allograft dysfunction (EAD) occurring in
Donor Age the first week post liver transplantation. EAD was associated
As with liver regeneration in the nontransplant setting; old with an inflammatory response in the perioperative period and
grafts do not regenerate as quickly as young livers. A clinical a specific pattern of 25 cytokines, chemokines, and immuno-
study of living donors showed a greater graft/standard LV receptors. Patients with EAD showed higher MCP-1 (mono-
in the young donor livers posttransplant compared with middle- cyte chemoattractant protein-1 [CCL2]), IL-8 (CXCL8), and
aged and old donor grafts. The old livers also had a higher RANTES ((regulated on activation, normal T-cell expressed
prothrombin time in the early period postoperatively (Ikegami and secreted; CCL5) chemokine levels in the early postopera-
et al, 2000). Statistics from the UNOS database show that the tive period, suggesting upregulation of the NF-κB pathway,
graft survival of older living-donor grafts is inferior to younger in addition to higher levels of chemokines and cytokines
grafts (Abt et al, 2004) and increasing donor and recipient age associated with T-cell immunity, including MIG (monokine
affected both short- and long-term survival in the NIH A2ALL induced by interferon-γ; CXCL9), IP-10 (interferon-γ–induc-
study (Olthoff et al, 2005, 2015). In the deceased-donor trans- ible protein-10; CXCL10) and IL-2R (interleukin-2 receptor)
plant setting, grafts older than 55 to 60 years have poorer (Friedman et al, 2012).
long-term survival combined with longer cold ischemic times
(Cassuto et al, 2008). Age may affect the regeneration and Other Factors
recovery of the living donor and the recipient. Many groups Several other factors may influence the regenerative response
limit the upper age limit of the donor in the 50- to 60-year after transplantation. Increased portal venous flow has been
range, although no definite age has been specified. implicated in more rapid regeneration (Jiang et al, 2009). In
addition to the reduced cell mass of the graft itself, portal
Inflammation: Viral Hepatitis and Bacterial Infections blood flow dynamics are altered, and the graft is subjected to
Very little is known about how rapid hepatocellular prolifera- increased portal blood flow and pressure. Increased portal
tion affects other processes. Inflammation in general, and some flow to about two times baseline level is associated with
viral infections, such as hepatitis B and C and murine cyto- increased recovery (Eguchi et al, 2003); however, significant
megalovirus specifically (Marshall et al, 2005; Sun & Gao, portal hyperperfusion is, on the other hand, regarded as
central to the problem in SFS grafts, leading to decreased to patients with cirrhosis and are only rough estimates of
regeneration and increased mortality (Hessheimer et al, 2011; functional reserve. Bilirubin, albumin, international normal-
Tracy et al, 1991). Larger grafts and younger donors have ized ratio, and platelet count become abnormal only in
been recommended in patients with severe portal hyperten- advanced cirrhosis. Magnetic resonance imaging or computed
sion, as well as modulation of portal inflow. Poor hepatic tomography (CT) can measure residual LV accurately, but
venous drainage has been shown to inhibit regeneration quantitative functional testing is not as precise. Traditional
(Scatton et al, 2008), and segments with poor venous drain- techniques to measure the LV to be resected before hepatec-
age become atrophied with time. tomy can lead to inaccurate estimates of functional residual
As in liver resection, female gender may have a positive effect liver (FRL) volumes because of the presence of dilated bile
on regeneration of partial grafts in murine models (Yokoyama ducts, multiple tumors, undetected lesions, compromised LV
et al, 2007). Estrogen may be responsible for the better toler- due to cholestasis or previous chemotherapy, cholangitis, vas-
ance to various stresses due to a reduced inflammatory response cular obstruction, steatosis or cirrhosis, or segmental atrophy
and a reduced oxygen radical production, leading to improved and/or hypertrophy from tumor growth (Azoulay et al, 2000;
hepatic regeneration. However, also in transplantation, there is Kubota et al, 1997). Accurate preoperative assessment by CT
no clinical evidence that supports the results of these experi- volumetry is necessary as significant interpatient variation
mental studies. exists in hepatic volumes. In living-donor liver transplantation,
total liver volume (TLV) and FRL can be used to predict
EXPERIMENTAL STRATEGIES TO PROMOTE postresection function because the donor liver is normal.
LIVER REGENERATION However, the TLV of the recipient’s diseased liver is not a
useful index of function. Values calculated from graft weight-
Although there has yet to be developed any reliable interven- to-recipient body weight ratio (GRBWR), or standardized LV
tion to improve liver regeneration in the clinical realm, based on recipient BSA are used to predict minimum adequate
numerous approaches have been successful in the experimen- graft volume (Higashiyama et al, 1993; Vauthey et al, 2000).
tal setting. After major hepatectomy or in the situation of a In segmental graft liver transplantation, a GRBWR greater
SFS liver graft in liver transplantation, usually a robust than 0.8% or a graft-to-weight ratio (graft weight divided by
priming reaction by highly elevated levels of IL-6 and TNF-α standard liver weight of recipient) greater than 40% is recom-
is present, but induction of cell-cycle progression fails (Debo- mended to achieve graft and patient survival of greater than
nera et al, 2004). In an animal model, this problem could be 90% (Lo et al, 1999). In comparison, extended resection of
overcome by a single injection of HB-EGF (Mitchell et al, 80% of functional parenchyma can be performed in the
2005). Another strategy that may be useful in a SFS situation absence of chronic liver disease for hepatobiliary malignancies
is blockage of the receptor for advanced glycation end prod- (Abdalla et al, 2002). Recommended minimal functional
ucts (RAGE). Blockage of RAGE greatly improved survival remnant LV following extended hepatectomy is greater than
after an 85% hepatectomy in a rat model by increasing TNF 25% in a normal liver and greater than 40% in an “injured”
and IL-6 production and enhanced expression of the antiin- liver, with moderate to severe steatosis, cholestasis, fibrosis,
flammatory cytokine IL-10 (Cataldegirmen et al, 2005). cirrhosis, or following chemotherapy (Shoup et al, 2003).
Recently, the role of TGF-β as the most potent growth inhibi- Quantitative liver function tests measure the liver’s ability to
tory polypeptide currently known was assessed in a model of metabolize or extract test compounds and can identify patients
SFS liver transplantation. After SFS liver transplantation, a with impaired function at earlier stages of disease but have
sharp rise in TGF-β1 was found, forming a heteromeric limited application in predicting a liver’s ability to regenerate
receptor complex. Phosphorylation of this complex activates after major resection. Indocyanine green clearance (ICG) is
proteins of the Smad family, particularly Smad2 and Smad3 regarded as an accurate assessment of functional reserve and
(Lonn et al, 2009). Activated Smad2 and Smad3 form a can help predict mortality (Lau et al, 1997), but more is being
complex with Smad4, move into the nucleus and activate learned about measuring hepatic function in diseased livers
target genes expressing regulatory proteins and exerting its using quantitative functional testing, such as methionine
inhibitory effect on hepatocyte proliferation. Smad7, an inhibi- breath tests, cholate clearance, liver single-photon emission
tory Smad, associates stably with the TGF-β receptor complex CT scans, and liver scintigraphy and phosphorus-31 magnetic
and was shown to inhibit the TGF-β–dependent cell-cycle resonance spectroscopy (Corbin et al, 2002). When compared
arrest (Zhong et al, 2010). Therefore anti–TGF-β therapy with ICG and CT volumetric data, hepatobiliary scintigraphy
holds promise as a new strategy to improve regeneration of is a reproducible accurate tool to assess functional liver uptake
SFS grafts in clinical practice. and excretion, preoperative liver function reserve, and remnant
liver function, and it allows monitoring of postoperative liver
CLINICAL IMPLICATIONS function regeneration (Dinant et al, 2007). Assessment of
future remnant volume distinguishes those who will most
When to Stimulate Liver Regeneration Preoperatively likely benefit from preoperative liver enhancement techniques
The assessment of hepatic function before resection is diffi- PVE or hepatic artery embolization.
cult, but we do know that the risk for perioperative complica- For living donors, a remnant of at least 30% is recom-
tions increases when the remnant LV is too small, particularly mended to maintain the highest level of donor safety. The
in diseased livers (Yigitler et al, 2003). Hepatic function A2ALL study investigated quantitative liver function in detail
seems to recover quickly after resection but is difficult to in a series of living donors and found significant alterations in
measure. Conventional clinical blood tests and liver biopsy are function in the early days after donation and some subsequent
inadequate measures of hepatic function. The Child-Pugh clinical findings of elevated portal pressure, such as low platelets
classification classification and MELD score apply primarily and higher spleen size (Emond et al, 2015; Everson et al, 2013;
Trotter et al, 2011). If this will affect donors long term is still (Selzner et al, 1999). The first clinical attempt to minimize
unknown. ischemic injury during liver resection was performed by inter-
rupting long ischemic intervals with multiple short periods of
The Use of Portal Vein Embolization reperfusion (Makuuchi et al, 1987). The protective effect of
to Promote Regeneration IPC involves many different mechanisms, including inhibition
The selective embolization techniques increase tolerance to of apoptosis and preservation of cellular ATP content in patients
major hepatic resection by reducing the LV requiring resection undergoing major liver resection. (Clavien et al, 2003; Rudiger
and inducing hypertrophy of the FLR to approximate target et al, 2002). A recent cDNA microarray study in humans
limits in patients with large tumors or abnormal liver function. demonstrated that IPC triggers the overexpression of IL-1Ra,
Criteria for selection of patients for PVE before major hepatec- iNOS, and Bcl-2, which counteracts the ischemia-induced
tomy are FLR size; factors compromising liver function, includ- proinflammatory and proapoptotic activation (Barrier et al,
ing previous chemotherapy, hepatitis, and cirrhosis; 2005). IPC has also been described as promoting liver regen-
and the planned complexity of the procedure (Abdalla et al, eration via the upregulation of cytokines such as TNF-α and
2001; Yigitler et al, 2003). It is recommended when predicted IL-6, various heat-shock proteins, and the downregulation of
FLR is less than 20% to 25% in a normal liver and less than TGF-β (Gomez et al, 2007). Ischemic preconditioning by 10
40% in a liver with compromised function (Hemming et al, minutes of portal triad inflow occlusion and 10 minutes of
2003). Stimulation by PVE increases circulating IL-6 and reperfusion was shown to be effective both in liver resection,
TNF-α (Feingold et al, 1988; Koga & Ogasawara, 1991), with particularly in patients with mild to moderate steatosis (Clavien
activation of the mitogenic cascade, similar to PHx. In fact, et al, 2003), and in liver transplantation (Franchello et al, 2009;
marginal contralateral regeneration less than 5% after PVE is Jassem et al, 2009). There was no difference in protective
a strong predictor of liver failure after subsequent liver resection potential between intermittent clamping (15 minutes ischemia
(Ribero et al, 2007). A significant increase in DNA synthesis and 5 minutes reperfusion) and ischemic precondition with
and mRNA expression of HGF has been observed in the subsequent inflow occlusion for a maximum of 75 minutes,
nonligated lobe (Uemura et al, 2000), whereas HGF expression except for patients older than 65 years, who benefited more
is only slightly elevated, and negative regulators of hepatocyte from intermittent clamping to attenuate liver injury. Pharma-
proliferation, such as TGF-β and IL-1β, are strongly expressed cologic induction of heat-shock proteins could play a beneficial
in the shrinking ligated lobe. It is important to keep in mind role in the recovery of liver function after hepatectomy, but
that these factors also may promote tumor outgrowth in the clinical trials have to be awaited.
FLR, and continuation of chemotherapy during PVE for malig-
nant conditions should be considered (Covey et al, 2008;
Regenerative Potential of the Liver
Pamecha et al, 2009). After Chemotherapy
As discussed before, an increasing number of patients with
Associating Liver Partition With Portal Vein Ligation tumors undergo extensive chemotherapy with multiple drugs
for Staged Hepatectomy (ALPPS) before surgery. The complication rate and mortality after major
In 2012, a radical new surgical procedure was introduced to liver resection is increased in those patients, compared with
stimulate liver regeneration in patients with a small future patients not receiving these drugs (Fernandez et al, 2005;
remnant liver (Schnitzbauer et al, 2012). In this two-step Vauthey et al, 2006). The deleterious effect of chemotherapy
approach, the liver parenchyma is transected between segments on regeneration seems to increase with the total number of
II/III and IV, with concomitant ligation of the right portal vein. cycles given and shows a sharp rise after 5 courses (Karoui
One week to 10 days later, impressive hypertrophy of the left et al, 2006). Therefore we advocate no more than six cycles of
lateral segments has occurred with a median volume increase FOLFOX/FOLFIRI (5-FU, leukovorin, oxaliplatin/irinotecan)–
of 74% (range, 21% to 192%) (Schlegel et al, 2014), and in a containing chemotherapy before liver resection, with a 3-week
second stage, an extended right hemihepatectomy can be per- interval in between. The anti-VEGF monoclonal bevacizumab
formed. The procedure was initially hampered by high mortality has a long half-life and theoretically should be stopped 6 to 8
(11% up to 19%) and high morbidity (as high as 40%) (de weeks before surgery, necessitating close collaboration with the
Santibanes & Clavien, 2012; Schadde et al, 2015), but with medical oncologist in the timing of surgery. Results from PVE
increasing experience and better patient selection, better results before major liver resection under continuous bevacizumab,
are gained without perioperative mortality (Hernandez- however, showed no deterioration in increase in FLR volume
Alejandro et al, 2014). 4 weeks after PVE (Zorzi et al, 2008). Recently, significantly
Although ALPPS seems a promising technique to boost impaired hypertrophy was reported in the same situation (Aus-
regeneration in selected patients, many recommend consider- silhou et al, 2009), but this may also be attributed to extensive
ation of alternative approaches, such as adequate portal and concomitant chemotherapy (Vauthey & Zorzi, 2009). The
hepatic venous embolizations, before an all-operative approach optimal window between the completion of bevacizumab and
is attempted (Vauthey & Mise, 2014). surgery therefore remains uncertain (Clavien et al, 2007).
Ischemic Preconditioning (IPC)
NEW HORIZONS AND FUTURE PERSPECTIVES
to Stimulate Regeneration
Liver resection, transplantation, and trauma can result in pro- Therapeutic Use of Stem Cells
longed deprivation of tissue oxygen, converting cellular metabo- In addition to hepatocyte proliferation and hepatic progenitor
lism to anaerobic pathways. Reperfusion and, consequently, the cells in liver regeneration, bone marrow (BM)-derived cells
restoration of oxygen delivery lead to liver injury. This phenom- have the ability to engraft as hepatocytes (Alison et al, 2009;
enon is known as I/R injury, which impairs liver regeneration Duncan et al, 2009; Friedman & Krause, 2009). Adult BM
includes two well-defined populations of stem cells: hematopoi- In the context of disease, infusion of (BM) MSC has shown
etic stem cells (HSC), which give rise to all mature lineages of some beneficial effects in patients with liver failure (Peng et al,
blood and mesenchymal stem cells (MSC), which can differen- 2011). In chronic liver failure, some Phase I trials involving the
tiate into bone, cartilage, muscle, and fat. injection of autologous BM cells to cirrhotic patients have
Evidence for BM-to-hepatocyte transition has been demon- reported modest improvements in clinical scores (reviewed in
strated by analyzing liver samples after male-to-female BM Houlihan & Newsome, 2008; Lin et al, 2008). However, the
transplantation in rodents and humans (Alison et al, 2000; BM may also harbor cells with fibrogenic potential, which
Lagasse et al, 2000; Oertel & Shafritz, 2008; Theise et al, contribute significantly to end-stage liver fibrosis (Forbes et al,
2000). However, the extent to which this occurs and the mecha- 2004; Russo et al, 2006). Therefore MSC therapy is sometimes
nisms are still under debate (Fausto, 2004; Oertel et al, 2008). referred to as a two-edged sword, and concerns were raised
Estimates of repopulation by hematopoietic stem cells vary about MSCs promoting liver cirrhosis by differentiation into
from 0.01% to as high as 40% but are often overestimated (Hall myofibroblasts, which are involved in the fibrotic response (di
et al, 2012). Bonzo et al, 2008). The latest publications in this field indicate
In the discussion about the mechanism whereby the hema- that the use of MSCs is safe and has a beneficial effect on liver
topoietic stem cells acquire a hepatocyte phenotype, both fibrosis (Chang et al, 2009; Li et al, 2012; Pan et al, 2014;
fusion of stem cells and hepatocytes (Vassilopoulos et al, Roderfeld et al, 2010; Zhao et al, 2009). Thus MSC-mediated
2003; Wang et al, 2003) and transdifferentiation of stem cells liver therapy is a promising field, which significantly improves
into hepatocytes have been proven (Harris et al, 2004; Jang liver function and increases animal survival in experimental
et al, 2004). The highest levels of BM-derived hepatocytes liver injury models. Its role in liver regeneration after surgery
were found in humans with severe liver disease, suggesting remains to be determined.
that tissue damage may promote engraftment as hepatocytes.
The major fraction of mobilized BM stem cells expresses the Decellularized Hepatic Matrix and Hepatic
chemokine receptor CXCR4. At the same time, the mRNA Tissue Engineering
level of its ligand (SDF-1) is increased in the damaged liver
tissue. These results provide a clue that the CXCR4/SDF-1 Due to the shortage of organs for transplant, research on
interaction may be important for the mobilization of progeni- alternate modalities, such as hepatic tissue engineering, has
tor stem cells from BM to the damaged liver (Khurana & gained momentum. The applications of such engineered organs
Mukhopadhyay, 2007). Clinically, the release of adult stem could not just be seen in the setting of transplantation but also
cells from BM was demonstrated after partial liver resection as support of failing liver function after large resections.
for benign and malignant conditions (De Silvestro et al, 2004; An innovative advance in this field has been the realization
Gehling et al, 2005), and BM-derived stem cells increased the of an important role of ECM in maintenance of differentiated
regeneration of contralateral liver after clinical PVE 2.5 fold hepatocyte phenotype. Recently, strategies were developed to
(am Esch et al, 2005). derive intact ECM from a liver by using a decellularization
Additional studies report that MSCs also promote liver process (Fig. 6.7).
repair in cases of liver damage (Duncan et al, 2009). MSCs, This strategy is based upon removal of cells from an organ,
isolated from human umbilical cords (Yan et al, 2009), adipose leaving a complex mixture of structural and functional proteins
tissue (Banas et al, 2008), bone marrow (Kuo et al, 2008), or that constitute the ECM (Caralt et al, 2014), which is then
rat bone marrow (Abdel Aziz et al, 2007) improved liver func-
tion of rodents undergoing acute liver damage (e.g., CCl4
injections).
The therapeutic effects of MSCs or MSC-derived hepato-
cytes in liver injury can be explained by three primary mecha-
nisms. First, MSCs generate cells that function as normal
hepatocytes, after fusing with metabolically defective hepato-
cytes (Kallis et al, 2007; Lagasse et al, 2000; Wang et al, 2003).
The second mechanism is soluble factors secreted by MSCs in
response to acute damage. Infusion of human MSC-conditioned
medium into rats treated with D-galactosamine (i.e., acute liver
damage) improved liver function after 24 hours (Parekkadan
et al, 2007; van Poll et al, 2008), with a 90% decrease in
apoptosis and a threefold increase in the number of proliferat-
ing hepatocytes. The same was shown after 70% hepatectomy,
with upregulated hepatic gene expression of cytokines and
growth factors relevant for cell proliferation, angiogenesis, and
antiinflammatory responses (Fouraschen et al, 2012).
Finally, the paracrine effects of MSCs may be exerted
by sharing of shed microvesicles (MVs) (Herrera et al, 2010).
Intercellular exchange of protein and RNA-containing mic- FIGURE 6.7 Decellularized liver scaffold. A pig liver was treated with
roparticles is an increasingly important mode of cell-cell 4% Triton X-100 and 0.1% ammonia for approximately 16 hours at a
communication, and MSCs may redirect the behavior of dif- low flow rate of 60 mL/min. Vascular structures such as the vena cava
ferentiated hepatic cells by horizontal transfer of mRNA shuttled retain their strength, whereas the liver extracellular matrix is totally dis-
by MVs (Deregibus et al, 2007; Ratajczak et al, 2006). posed of all cell types.
reseeded with an appropriate population of cells (Sabetkish significant reduction of hepatic steatosis, which was associated
et al, 2014; Zhou et al, 2014) and connected to the bloodstream with reduced cholesterol synthesis rates and stimulation of
and biliary system. Using the whole-organ acellular matrix as hepatic fatty-acid oxidation. The clinical relevance of miRNAs
a three-dimensional scaffold for seeding hepatocyte-like cells, a was shown in differences in spontaneous recovery from acute
fully functional transplantable bioengineered liver graft may liver failure. Patients with spontaneous recovery from acute
become a reality. liver failure showed significantly higher serum levels of miR-122
One of the major remaining obstacles toward clinical appli- and liver tissue levels compared with nonrecovered patients,
cation is now to choose a cell source for liver repopulation. So with strong downregulation of miRNA target genes that impair
far, adult primary hepatocytes have been the primary choice, liver regeneration, including heme oxygenase-1; programmed
but scarcity of high-quality human hepatocytes limits tissue cell death 4; and the CDK inhibitors p21, p27, and p57 (John
engineering applications. et al, 2014). Besides upregulation of miR-122 and miR-21 after
With the advent of technologies enabling reprogramming of partial liver resection, other miRs are downregulated: miR-22a,
adult somatic cells to a pluripotent state (induced pluripotent miR-26a, miR-30b, miR378, Let-7f, and Let-7g. Inhibition of
stem cell, iPS), it may now become possible to generate the miR33 improves liver regeneration after PHx in mice, indicat-
large numbers of inducible human hepatocytes (iHeps), needed ing miRNAs as critical regulators of hepatocyte proliferation
to recellularize a liver bioscaffold (Berger et al, 2015; Hay et al, during liver regeneration (Bandiera et al, 2015; Chen et al,
2008a, 2008b; Huang et al, 2011; Sekiya & Suzuki, 2011; 2011; Cirera-Salinas et al, 2012; Song et al, 2010). Also in liver
Sullivan et al, 2009; Yu et al, 2013). transplantation, distinct patterns of successful and failed regen-
There is, however, concern about the plasticity of these cells eration could be discerned, with inhibition of miRNAs 150,
to form bile ducts, a main hurdle to usefulness in clinically 663, and 503 being associated with successful regeneration
transplantable liver matrix engineering. A solution to this (Salehi et al, 2013).
problem may come from recently discovered bipotential liver In a murine model of hepatectomy in increasing age,
organoids (Huch et al, 2015). These cells are positive for Lgr5, upregulated miRNAs may contribute to the aging-related
the receptor for the Wnt agonist R-spondin. They are able decline in oxidative defense by targeting various classes of
to differentiate into both hepatocytes and cholangiocytes, glutathione-S-transferases (Maes et al, 2008). Antagonizing
as precursors of bile ducts, depending on culture media these miRNAs may reverse the decline of regeneration and
composition. oxidative defense mechanisms in aging liver.
The intricate spatiotemporal environment of a decellularized
liver matrix, with additional use of nonparenchymal cells of the SUMMARY
liver, may provide the ideal niche for functional differentiation
of such organoids. This is truly an evolving and timely field with In the last century, knowledge about liver regeneration has
much ongoing research in which knowledge of liver regenera- rapidly evolved from a truly mythical black box event into
tion is essential. growing understanding of the pathways involved in this amazing
complex multistep process. Much was learned about the
The Role of miRNA in Liver Regeneration dynamics and redundant intracellular signaling pathways of
A family of tiny regulatory RNAs, known as microRNAs liver regeneration, but less is still known about the exact signals
(miRNAs), was found to have profound roles in the control of that initiate and stop liver regeneration. Our advanced knowl-
diverse aspects of hepatic function and dysfunction, including edge on liver regeneration and prevention of liver failure led to
hepatocyte growth, stress response, metabolism, viral infection safer extreme liver resections for benign and malignant diseases
and proliferation, gene expression, and maintenance of hepatic and the use of living-liver donors in liver transplantation.
phenotype (Ambros, 2004; Elmen et al, 2008). miRNAs are Despite our better understanding, there has been little
small endogenous noncoding RNAs that posttranscriptionally structured advancement in therapeutic options in case of liver
repress the expression of protein-coding genes by base-pairing failure due to insufficient liver regeneration. New challenges lie
with the 39 untranslated regions of the target mRNAs (Bartel, ahead in the use of therapeutic strategies to enhance liver
2004; Grimson et al, 2007). regeneration in patients in whom normal regeneration fails and
In 2002, miR-122 was identified as an abundant miRNA in thus push the possibilities of liver resection to the next level.
the liver (Lagos-Quintana et al, 2002) and characterized as the Furthermore, while promoting liver cell proliferation, we must
most frequent miRNA isolated in the adult liver (Chang et al, be very cautious not to stimulate tumor growth in patients with
2004). Using distinct protocols to silence miR-122, evidence primary or metastatic liver tumors as a consequence of our
for the overall importance of miR-122 in the regulation of liver therapy.
metabolism was found (Esau et al, 2006; Krutzfeldt et al,
2005). Silencing miR-122 in high-fat–fed mice resulted in a References are available at expertconsult.com.
Chapter 6 Liver regeneration: mechanisms and clinical relevance 109.e1
REFERENCES Boleslawski E, et al: Hepatic venous pressure gradient in the assessment

of portal hypertension before liver resection in patients with cirrhosis,
Abdalla EK, et al: Portal vein embolization: rationale, technique and Br J Surg 99(6):855–863, 2012.
future prospects, Br J Surg 88(2):165–175, 2001. Bucher NLR, Swaffield MN: The rate of incorporation of labeled
Abdalla EK, et al: Extended hepatectomy in patients with hepatobiliary thymidine into the deoxyribonucleic acid of regenerating rat liver in
malignancies with and without preoperative portal vein embolization, relation to the amount of liver excised, Cancer Res 24:1611–1625,
Arch Surg 137(6):675–680, discussion 680–671, 2002. 1964.
Abdel Aziz MT, et al: Therapeutic potential of bone marrow-derived Burroughs AK, et al: 3-month and 12-month mortality after first liver
mesenchymal stem cells on experimental liver fibrosis, Clin Biochem transplant in adults in Europe: predictive models for outcome, Lancet
40(12):893–899, 2007. 367(9506):225–232, 2006.
Abt PL, et al: Allograft survival following adult-to-adult living donor Campbell JS, et al: Expression of suppressors of cytokine signaling
liver transplantation, Am J Transplant 4(8):1302–1307, 2004. during liver regeneration, J Clin Invest 107(10):1285–1292, 2001.
Adams LA, et al: The natural history of nonalcoholic fatty liver disease: Campbell JS, et al: Proinflammatory cytokine production in liver
a population-based cohort study, Gastroenterology 129(1):113–121, regeneration is Myd88-dependent, but independent of Cd14, Tlr2,
2005. and Tlr4, J Immunol 176(4):2522–2528, 2006.
Aldeguer X, et al: Interleukin-6 from intrahepatic cells of bone marrow Campsen J, et al: Outcomes of living donor liver transplantation for
origin is required for normal murine liver regeneration, Hepatology acute liver failure: the adult-to-adult living donor liver transplanta-
35(1):40–48, 2002. tion cohort study, Liver Transpl 14(9):1273–1280, 2008.
Alison MR, et al: Hepatocytes from non-hepatic adult stem cells, Canbay A, et al: Fas enhances fibrogenesis in the bile duct ligated
Nature 406(6793):257, 2000. mouse: a link between apoptosis and fibrosis, Gastroenterology
Alison MR, et al: Stem cells in liver regeneration, fibrosis and cancer: 123(4):1323–1330, 2002.
the good, the bad and the ugly, J Pathol 217(2):282–298, 2009. Caralt M, et al: Optimization and critical evaluation of decellularization
Ambros V: The functions of animal microRNAs, Nature 431(7006):350– strategies to develop renal extracellular matrix scaffolds as biological
355, 2004. templates for organ engineering and transplantation, Am J Transplant
am Esch JS 2nd, et al: Portal application of autologous CD133+ bone 15(1):64–75, 2014.
marrow cells to the liver: a novel concept to support hepatic regen- Carpentier R, et al: Embryonic ductal plate cells give rise to cholangio-
eration, Stem Cells 23(4):463–470, 2005. cytes, periportal hepatocytes, and adult liver progenitor cells, Gas-
Anderson SP, et al: Delayed liver regeneration in peroxisome prolifer- troenterology 141(4):1432–1438, 1438.e1431–1434, 2011.
ator-activated receptor-alpha-null mice, Hepatology 36(3):544–554, Carpino G, et al: Biliary tree stem/progenitor cells in glands of extra-
2002. hepatic and intraheptic bile ducts: an anatomical in situ study
Arai M, et al: Gene expression profiling reveals the mechanism and yielding evidence of maturational lineages, J Anat 220(2):186–199,
pathophysiology of mouse liver regeneration, J Biol Chem 278(32): 2011.
29813–29818, 2003. Cassuto JR, et al: The cumulative effects of cold ischemic time and older
Aussilhou B, et al: Preoperative liver hypertrophy induced by portal donor age on liver graft survival, J Surg Res 148(1):38–44, 2008.
flow occlusion before major hepatic resection for colorectal metas- Cataldegirmen G, et al: RAGE limits regeneration after massive liver
tases can be impaired by bevacizumab, Ann Surg Oncol 16(6):1553– injury by coordinated suppression of TNF-alpha and NF-kappaB,
1559, 2009. J Exp Med 201(3):473–484, 2005.
Azoulay D, et al: Resection of nonresectable liver metastases from Chang YJ, et al: Mesenchymal stem cells facilitate recovery from chemi-
colorectal cancer after percutaneous portal vein embolization, Ann cally induced liver damage and decrease liver fibrosis, Life Sci
Surg 231(4):480–486, 2000. 85(13–14):517–525, 2009.
Azuma H, et al: Robust expansion of human hepatocytes in Fah−/−/ Chang J, et al: miR-122, a mammalian liver-specific microRNA, is
Rag2−/−/Il2rg−/− mice, Nat Biotechnol 25(8):903–910, 2007. processed from hcr mRNA and may downregulate the high affinity
Banas A, et al: IFATS collection: in vivo therapeutic potential of human cationic amino acid transporter CAT-1, RNA Biol 1(2):106–113,
adipose tissue mesenchymal stem cells after transplantation into 2004.
mice with liver injury, Stem Cells 26(10):2705–2712, 2008. Chen X, et al: miRNA regulation of liver growth after 50% partial
Bandiera S, et al: miR-122 - A key factor and therapeutic target in liver hepatectomy and small size grafts in rats, Transplantation 91(3):293–
disease, J Hepatol 62(2):448–457, 2015. 299, 2011.
Barrier A, et al: Ischemic preconditioning modulates the expression of Chen X, et al: Severity of portal hypertension and prediction of post-
several genes, leading to the overproduction of IL-1Ra, iNOS, and operative liver failure after liver resection in patients with Child-Pugh
Bcl-2 in a human model of liver ischemia-reperfusion, FASEB J grade A cirrhosis, Br J Surg 99(12):1701–1710, 2012.
19(12):1617–1626, 2005. Chua TC, et al: Systematic review of randomized and nonrandomized
Bartel DP: MicroRNAs: genomics, biogenesis, mechanism, and func- trials of the clinical response and outcomes of neoadjuvant systemic
tion, Cell 116(2):281–297, 2004. chemotherapy for resectable colorectal liver metastases, Ann Surg
Behrns KE, et al: Hepatic steatosis as a potential risk factor for major Oncol 17(2):492–501, 2010.
hepatic resection, J Gastrointest Surg 2(3):292–298, 1998. Cirera-Salinas D, et al: Mir-33 regulates cell proliferation and cell cycle
Belghiti J, et al: Seven hundred forty-seven hepatectomies in the 1990s: progression, Cell Cycle 11(5):922–933, 2012.
an update to evaluate the actual risk of liver resection, J Am Coll Surg Clavien PA, et al: Strategies for safer liver surgery and partial liver
191(1):38–46, 2000. transplantation, N Engl J Med 356(15):1545–1559, 2007.
Berenguer M: Live donor liver transplantation for hepatitis C: new data, Clavien PA, et al: A prospective randomized study in 100 consecutive
old story, Liver Transpl 12(4):516–519, 2006. patients undergoing major liver resection with versus without ische-
Berger DR, et al: Enhancing the functional maturity of iPSC-derived mic preconditioning, Ann Surg 238(6):843–850, discussion 851–852,
human hepatocytes via controlled presentation of cell-cell interac- 2003.
tions in vitro, Hepatology 61(4):1370–1381, 2015. Columbano A, Ledda-Columbano GM: Mitogenesis by ligands of
Bhala N, et al: The natural history of nonalcoholic fatty liver disease nuclear receptors: an attractive model for the study of the molecular
with advanced fibrosis or cirrhosis: an international collaborative mechanisms implicated in liver growth, Cell Death Differ 10(Suppl
study, Hepatology 54(4):1208–1216, 2011. 1):S19–S21, 2003.
Bilodeau M, et al: Evaluation of hepatocyte injury following partial Columbano A, et al: Gadd45beta is induced through a CAR-dependent,
ligation of the left portal vein, J Hepatol 30(1):29–37, 1999. TNF-independent pathway in murine liver hyperplasia, Hepatology
Bissig KD, et al: Epidermal growth factor is decreased in liver of rats 42(5):1118–1126, 2005.
with biliary cirrhosis but does not act as paracrine growth factor Columbano A, Shinozuka H: Liver regeneration versus direct hyperpla-
immediately after hepatectomy, J Hepatol 33(2):275–281, 2000. sia, FASEB J 10(10):1118–1128, 1996.
Blindenbacher A, et al: Expression of hepatitis c virus proteins inhibits Corbin IR, et al: Regenerative activity and liver function following
interferon alpha signaling in the liver of transgenic mice, Gastroen- partial hepatectomy in the rat using (31)P-MR spectroscopy, Hepa-
terology 124(5):1465–1475, 2003. tology 36(2):345–353, 2002.
Blumgart LH, Kelley CJ: Hepaticojejunostomy in benign and malignant Cornell RP, et al: Depressed liver regeneration after partial hepatectomy
high bile duct stricture: approaches to the left hepatic ducts, Br J of germ-free, athymic and lipopolysaccharide-resistant mice, Hepa-
Surg 71(4):257–261, 1984. tology 11(6):916–922, 1990.
Costa RH, et al: Transcription factors in liver development, differentia- Ding L, et al: Bile acid promotes liver regeneration via farnesoid X
tion, and regeneration, Hepatology 38(6):1331–1347, 2003. receptor signaling pathways in rats, Mol Med Rep 11(6):4431–4437,
Covey AM, et al: Combined portal vein embolization and neoadjuvant 2015.
chemotherapy as a treatment strategy for resectable hepatic colorec- Dong J, et al: Elucidation of a universal size-control mechanism in
tal metastases, Ann Surg 247(3):451–455, 2008. Drosophila and mammals, Cell 130(6):1120–1133, 2007.
Cressman DE, et al: Rapid activation of the Stat3 transcription complex Drixler TA, et al: Liver regeneration is an angiogenesis- associated
in liver regeneration, Hepatology 21(5):1443–1449, 1995. phenomenon, Ann Surg 236(6):703–711, discussion 711-712, 2002.
Cressman DE, et al: Liver failure and defective hepatocyte regeneration Duarte S, et al: Matrix metalloproteinases in liver injury, repair and
in interleukin-6- deficient mice, Science 274(5291):1379–1383, fibrosis, Matrix Biol 44-46:147–156, 2015.
1996. Duncan AW, et al: Stem cells and liver regeneration, Gastroenterology
Currier AR, et al: Plasminogen directs the pleiotropic effects of uPA in 137(2):466–481, 2009.
liver injury and repair, Am J Physiol Gastrointest Liver Physiol Eguchi S, et al: Relationship between portal venous flow and liver
284(3):G508–G515, 2003. regeneration in patients after living donor right-lobe liver transplan-
Dabeva MD, Shafritz DA: Activation, proliferation, and differentiation tation, Liver Transpl 9(6):547–551, 2003.
of progenitor cells into hepatocytes in the D-galactosamine model of Elmen J, et al: LNA-mediated microRNA silencing in non-human
liver regeneration, Am J Pathol 143(6):1606–1620, 1993. primates, Nature 452(7189):896–899, 2008.
Damrauer SM, et al: A20 modulates lipid metabolism and energy Emond JC, et al: Functional analysis of grafts from living donors:
production to promote liver regeneration, PLoS ONE 6(3):e17715, implications for the treatment of older recipients, Ann Surg
2011. 224(4):544–554, 1996.
D’Angelica M, et al: Lack of evidence for increased operative morbidity Emond JC, et al: Changes in liver and spleen volumes after living liver
after hepatectomy with perioperative use of bevacizumab: a matched donation: a report from the adult-to-adult living donor liver trans-
case-control study, Ann Surg Oncol 14(2):759–765, 2007. plantation cohort study (A2ALL), Liver Transpl 21(2):151–161,
da Silva CG, et al: A20 promotes liver regeneration by decreasing 2015.
SOCS3 expression to enhance IL-6/STAT3 proliferative signals, Erhardt A, et al: Hepatitis C virus core protein induces cell proliferation
Hepatology 57(5):2014–2025, 2013. and activates ERK, JNK, and p38 MAP kinases together with the
da Silva CG, et al: A20–an omnipotent protein in the liver: prometheus MAP kinase phosphatase MKP-1 in a HepG2 Tet-Off cell line,
myth resolved? Adv Exp Med Biol 809:117–139, 2014. Virology 292(2):272–284, 2002.
Date M, et al: Differential regulation of activin A for hepatocyte growth Esau C, et al: miR-122 regulation of lipid metabolism revealed by in
and fibronectin synthesis in rat liver injury, J Hepatol 32(2):251–260, vivo antisense targeting, Cell Metab 3(2):87–98, 2006.
2000. Everson GT, et al: Functional elements associated with hepatic regen-
Debonera F, et al: Activation of interleukin-6/STAT3 and liver regeneration in living donors after right hepatic lobectomy, Liver Transpl
eration following transplantation, J Surg Res 96(2):289–295, 2001. 19(3):292–304, 2013.
Debonera F, et al: Dexamethasone inhibits early regenerative response Farrell G: Effects of disease on expression and regulation of CYPs, Mol
of rat liver after cold preservation and transplantation, Hepatology Aspects Med 20(1–2):55–70, 137, 1999.
38(6):1563–1572, 2003. Fausto N: Liver regeneration, J Hepatol 32(Suppl 1):19–31, 2000.
Debonera F, et al: Partial liver grafts with prolonged cold preservation Fausto N: Liver regeneration and repair: hepatocytes, progenitor cells,
initiate blunted regenerative responses, Am J Transplant 2(Suppl and stem cells, Hepatology 39(6):1477–1487, 2004.
3):156, 2002. Fausto N, Campbell JS: The role of hepatocytes and oval cells in liver
Debonera F, et al: Severe preservation injury induces Il-6/STAT3 regeneration and repopulation, Mech Dev 120(1):117–130, 2003.
activation with lack of cell cycle progression after partial liver graft Feingold KR, et al: Tumor necrosis factor stimulates DNA synthesis in
transplantation, Am J Transplant 4(12):1964–1971, 2004. the liver of intact rats, Biochem Biophys Res Commun 153(2):576–582,
de Jonge J, et al: Unique early gene expression patterns in human 1988.
adult-to-adult living donor liver grafts compared to deceased donor Feng S, et al: Characteristics associated with liver graft failure: the
grafts, Am J Transplant 9(4):758–772, 2009. concept of a donor risk index, Am J Transplant 6(4):783–790, 2006.
Delany HM, et al: Contrasting effects of identical nutrients given Fernandez FG, et al: Effect of steatohepatitis associated with irinotecan
parenterally or enterally after 70% hepatectomy, Am J Surg or oxaliplatin pretreatment on resectability of hepatic colorectal
167(1):135–143, discussion 143-144, 1994. metastases, J Am Coll Surg 200(6):845–853, 2005.
de Meijer VE, et al: Systematic review and meta-analysis of steatosis as Fickert P, et al: Oncosis represents the main type of cell death in mouse
a risk factor in major hepatic resection, Br J Surg 97(9):1331–1339, models of cholestasis, J Hepatol 42(3):378–385, 2005.
2010. FitzGerald MJ, et al: Rapid DNA binding by nuclear factor kappa B
Demori I, et al: Increased insulin-like growth factor binding protein-4 in hepatocytes at the start of liver regeneration, Cell Growth Differ
expression after partial hepatectomy in the rat, Am J Physiol Gastro- 6(4):417–427, 1995.
intest Liver Physiol 278(3):G384–G389, 2000. Forbes SJ, et al: A significant proportion of myofibroblasts are of bone
Deregibus MC, et al: Endothelial progenitor cell derived microvesicles marrow origin in human liver fibrosis, Gastroenterology 126(4):955–
activate an angiogenic program in endothelial cells by a horizontal 963, 2004.
transfer of mRNA, Blood 110(7):2440–2448, 2007. Fouraschen SM, et al: Secreted factors of human liver-derived mesen-
Derynck R, Zhang YE: Smad-dependent and Smad-independent chymal stem cells promote liver regeneration early after partial hepa-
pathways in TGF-beta family signalling, Nature 425(6958):577–584, tectomy, Stem Cells Dev 21(13):2410–2419, 2012.
2003. Francavilla A, et al: Effects of rapamycin on cultured hepatocyte pro-
de Santibanes E, Clavien PA: Playing Play-Doh to prevent postopera- liferation and gene expression, Hepatology 15(5):871–877, 1992.
tive liver failure: the “ALPPS” approach, Ann Surg 255(3):415–417, Francavilla A, et al: Estradiol and testosterone levels in patients under-
2012. going partial hepatectomy. A possible signal for hepatic regeneration?
De Silvestro G, et al: Mobilization of peripheral blood hematopoietic Dig Dis Sci 34(6):818–822, 1989.
stem cells following liver resection surgery, Hepatogastroenterology Francavilla A, et al: The effect of estrogen and tamoxifen on hepatocyte
51(57):805–810, 2004. proliferation in vivo and in vitro, Hepatology 9(4):614–620, 1989.
Diamond RH, et al: Novel delayed-early and highly insulin-induced Francavilla A, et al: Studies on mechanisms of augmentation of liver
growth response genes. Identification of HRS, a potential regulator regeneration by cyclosporine and FK 506, Hepatology 14(1):140–143,
of alternative pre-mRNA splicing, J Biol Chem 268(20):15185–15192, 1991.
1993. Francavilla A, et al: Small-for-size liver transplanted into larger recipi-
di Bonzo LV, et al: Human mesenchymal stem cells as a two-edged ent: a model of hepatic regeneration, Hepatology 19(1):210–216,
sword in hepatic regenerative medicine: engraftment and hepatocyte 1994.
differentiation versus profibrogenic potential, Gut 57(2):223–231, Franchello A, et al: Ischemic preconditioning (IP) of the liver as a safe
2008. and protective technique against ischemia/reperfusion injury (IRI),
Dinant S, et al: Risk assessment of posthepatectomy liver failure using Am J Transplant 9(7):1629–1639, 2009.
hepatobiliary scintigraphy and CT volumetry, J Nucl Med 48(5): Friedman RS, Krause DS: Regeneration and repair: new findings in
685–692, 2007. stem cell research and aging, Ann N Y Acad Sci 1172:88–94, 2009.
Friedman BH, et al: Serum cytokine profiles associated with early Heckel JL, et al: Neonatal bleeding in transgenic mice expressing
allograft dysfunction in patients undergoing liver transplantation, urokinase-type plasminogen activator, Cell 62(3):447–456, 1990.
Liver Transpl 18(2):166–176, 2012. Hemming AW, et al: Preoperative portal vein embolization for extended
Fujita J, et al: Effect of TNF gene depletion on liver regeneration after hepatectomy, Ann Surg 237(5):686–691, discussion 691–693, 2003.
partial hepatectomy in mice, Surgery 129(1):48–54, 2001. Hernandez-Alejandro R, et al: Can we improve the morbidity and mor-
Fujiwara Y, et al: Cytokine characteristics of jaundice in mouse liver, tality associated with the associating liver partition with portal vein
Cytokine 13(3):188–191, 2001. ligation for staged hepatectomy (ALPPS) procedure in the manage-
Fukuhara Y, et al: Gene expression profile in the regenerating rat liver ment of colorectal liver metastases? Surgery 157(2):194–201, 2014.
after partial hepatectomy, J Hepatol 38(6):784–792, 2003. Herrera MB, et al: Human liver stem cell-derived microvesicles acceler-
Garcia-Rodriguez JL, et al: SIRT1 controls liver regeneration by regu- ate hepatic regeneration in hepatectomized rats, J Cell Mol Med
lating bile acid metabolism through farnesoid X receptor and mam- 14(6B):1605–1618, 2010.
malian target of rapamycin signaling, Hepatology 59(5):1972–1983, Hessheimer AJ, et al: Decompression of the portal bed and twice-
2014. baseline portal inflow are necessary for the functional recovery of a
Garlipp B, et al: Left-liver hypertrophy after therapeutic right-liver “small-for-size” graft, Ann Surg 253(6):1201–1210, 2011.
radioembolization is substantial but less than after portal vein embo- Higashiyama H, et al: Graft size assessment by preoperative computed
lization, Hepatology 59(5):1864–1873, 2014. tomography in living related partial liver transplantation, Br J Surg
Gehling UM, et al: Partial hepatectomy induces mobilization of a 80(4):489–492, 1993.
unique population of haematopoietic progenitor cells in human Higgins GM, Anderson RM: Experimental pathology of the liver. I.
healthy liver donors, J Hepatol 43(5):845–853, 2005. Restoration of the liver of the white rat following partial surgical
Ghahary A, et al: Effects of partial hepatectomy on hepatic insulinlike removal, Arch Pathol 12:186–202, 1931.
growth factor binding protein-1 expression, Hepatology 15(6):1125– Holecek M: Nutritional modulation of liver regeneration by carbohy-
1131, 1992. drates, lipids, and amino acids: a review, Nutrition 15(10):784–788,
Ghobrial RM, et al: Rapid and severe early HCV recurrence following 1999.
adult living donor liver transplantation, Am J Transplant 2(Suppl Hong L, et al: The Hippo signaling pathway in liver regeneration and
3):163A, 2002. tumorigenesis, Acta Biochim Biophys Sin (Shanghai) 47(1):46–52,
Gomez L, et al: Inhibition of mitochondrial permeability transition 2014.
improves functional recovery and reduces mortality following acute Hong JH, et al: dl-propranolol negatively regulates the transcription of
myocardial infarction in mice, Am J Physiol Heart Circ Physiol proliferating cell nuclear antigen (PCNA)-gene and thereby sup-
293(3):H1654–H1661, 2007. presses DNA synthesis in regenerating rat liver, Biochem Mol Biol Int
Gonzales E, et al: ATP release after partial hepatectomy regulates liver 42(1):103–112, 1997.
regeneration in the rat, J Hepatol 52(1):54–62, 2010. Hora C, et al: Effect of sorafenib on murine liver regeneration, Hepatol-
Gracia-Sancho J, et al: Flow cessation triggers endothelial dysfunction ogy 53(2):577–586, 2011.
during organ cold storage conditions: strategies for pharmacologic Houlihan DD, Newsome PN: Critical review of clinical trials of bone
intervention, Transplantation 90(2):142–149, 2010. marrow stem cells in liver disease, Gastroenterology 135(2):438–450,
Gressner AM: Cytokines and cellular crosstalk involved in the activa- 2008.
tion of fat-storing cells, J Hepatol 22(Suppl 2):28–36, 1995. Hu RH, et al: Profile of hepatocyte growth factor in patients with
Grimson A, et al: MicroRNA targeting specificity in mammals: deter- obstructive jaundice, Hepatogastroenterology 50(54):1987–1990,
minants beyond seed pairing, Mol Cell 27(1):91–105, 2007. 2003.
Grisham JW: Cellular proliferation in the liver, Recent Results Cancer Huang P, et al: Induction of functional hepatocyte-like cells from
Res 17:28–43, 1969. mouse fibroblasts by defined factors, Nature 475(7356):386–389,
Gruenberger T, et al: Bevacizumab plus mFOLFOX-6 or FOLFOXIRI 2011.
in patients with initially unresectable liver metastases from colorectal Huang W, et al: Nuclear receptor-dependent bile acid signaling is
cancer: the OLIVIA multinational randomised phase II trial, Ann required for normal liver regeneration, Science 312(5771):233–236,
Oncol 26(4):702–708, 2015. 2006.
Gujral JS, et al: Reduced oncotic necrosis in Fas receptor-deficient Huang J, Rudnick DA: Elucidating the metabolic regulation of liver
C57BL/6J-lpr mice after bile duct ligation, Hepatology 40(4):998– regeneration, Am J Pathol 184(2):309–321, 2014.
1007, 2004. Huch M, et al: Long-term culture of genome-stable bipotent stem cells
Haber BA, et al: High levels of glucose-6-phosphatase gene and protein from adult human liver, Cell 160(1–2):299–312, 2015.
expression reflect an adaptive response in proliferating liver and Huh CG, et al: Hepatocyte growth factor/c-met signaling pathway is
diabetes, J Clin Invest 95(2):832–841, 1995. required for efficient liver regeneration and repair, Proc Natl Acad Sci
Haber BA, et al: Induction patterns of 70 genes during nine days after USA 101(13):4477–4482, 2004.
hepatectomy define the temporal course of liver regeneration, J Clin Humar A, et al: Liver regeneration after adult living donor and deceased
Invest 91(4):1319–1326, 1993. donor split-liver transplants, Liver Transpl 10(3):374–378, 2004.
Hadjis NS, Blumgart LH: Role of liver atrophy, hepatic resection and Hurwitz H, et al: Bevacizumab plus irinotecan, fluorouracil, and leu-
hepatocyte hyperplasia in the development of portal hypertension in covorin for metastatic colorectal cancer, N Engl J Med 350(23):2335–
biliary disease, Gut 28(8):1022–1028, 1987. 2342, 2004.
Hall SR, et al: Overestimation of hematopoietic stem cell frequencies Iakova P, et al: Aging reduces proliferative capacities of liver by switch-
in human liver grafts, Hepatology 57(6):2547–2549, 2012. ing pathways of C/EBPalpha growth arrest, Cell 113(4):495–506,
Hamanoue M, et al: Rapid and marked induction of hepatocyte growth 2003.
factor during liver regeneration after ischemic or crush injury, Hepa- Ikeda K, et al: The ultrastructure, kinetics and intralobular distribution
tology 16(6):1485–1492, 1992. of apoptotic hepatocytes after portal branch ligation with special
Harris RG, et al: Lack of a fusion requirement for development reference to their relationship to necrotic hepatocytes, Arch Histol
of bone marrow-derived epithelia, Science 305(5680):90–93, Cytol 58(2):171–184, 1995.
2004. Ikegami T, et al: The impact of donor age on living donor liver trans-
Hashmi SK, et al: Genomics of liver transplant injury and regeneration, plantation, Transplantation 70(12):1703–1707, 2000.
Transplant Rev (Orlando) 29(1):23–32, 2015. Imamura H, et al: Preoperative portal vein embolization: an audit of
Hay DC, et al: Highly efficient differentiation of hESCs to functional 84 patients, Hepatology 29(4):1099–1105, 1999.
hepatic endoderm requires ActivinA and Wnt3a signaling, Proc Natl Ito A, Higashiguchi T: Effects of glutamine administration on liver
Acad Sci USA 105(34):12301–12306, 2008a. regeneration following hepatectomy, Nutrition 15(1):23–28, 1999.
Hay DC, et al: Efficient differentiation of hepatocytes from human Jaeschke H: Mechanisms of reperfusion injury after warm ischemia of
embryonic stem cells exhibiting markers recapitulating liver develop- the liver, J Hepatobiliary Pancreat Surg 5(4):402–408, 1998.
ment in vivo, Stem Cells 26(4):894–902, 2008b. Jang YY, et al: Hematopoietic stem cells convert into liver cells within
Hayashi J, et al: Hepatitis C virus core protein activates the MAPK/ days without fusion, Nat Cell Biol 6(6):532–539, 2004.
ERK cascade synergistically with tumor promoter TPA, but not with Jassem W, et al: Effect of ischemic preconditioning on the genomic
epidermal growth factor or transforming growth factor alpha, Hepa- response to reperfusion injury in deceased donor liver transplanta-
tology 32(5):958–961, 2000. tion, Liver Transpl 15(12):1750–1765, 2009.
Jiang SM, et al: Role of splanchnic hemodynamics in liver regeneration Kurumiya Y, et al: Differential suppression of liver-specific genes in
after living donor liver transplantation, Liver Transpl 15(9):1043– regenerating rat liver induced by extended hepatectomy, J Hepatol
1049, 2009. 32(4):636–644, 2000.
Jin J, et al: Cooperation of C/EBP family proteins and chromatin Lagasse E, et al: Purified hematopoietic stem cells can differentiate into
remodeling proteins is essential for termination of liver regeneration, hepatocytes in vivo, Nat Med 6(11):1229–1234, 2000.
Hepatology 61(1):315–325, 2015. Lagos-Quintana M, et al: Identification of tissue-specific microRNAs
Jin J, et al: The age-associated decline of glycogen synthase kinase 3beta from mouse, Curr Biol 12(9):735–739, 2002.
plays a critical role in the inhibition of liver regeneration, Mol Cell Lau H, et al: Evaluation of preoperative hepatic function in patients
Biol 29(14):3867–3880, 2009. with hepatocellular carcinoma undergoing hepatectomy, Br J Surg
John K, et al: MicroRNAs play a role in spontaneous recovery from 84(9):1255–1259, 1997.
acute liver failure, Hepatology 60(4):1346–1355, 2014. LeCouter J, et al: Angiogenesis-independent endothelial protection of
Kallis YN, et al: Bone marrow stem cells and liver disease, Gut liver: role of VEGFR-1, Science 299(5608):890–893, 2003.
56(5):716–724, 2007. Ledda-Columbano GM, et al: Aging does not reduce the hepatocyte
Kam I, et al: Evidence that host size determines liver size: Studies in proliferative response of mice to the primary mitogen TCPOBOP,
dogs receiving orthotopic liver transplants, Hepatology 7(2):362–366, Hepatology 40(4):981–988, 2004.
1987. Lee SG: A complete treatment of adult living donor liver transplanta-
Kamel IR, et al: Liver regeneration after living adult right lobe trans- tion: a review of surgical technique and current challenges to expand
plantation, Abdom Imaging 28(1):53–57, 2003. indication of patients, Am J Transplant 15(1):17–38, 2015.
Kan NG, et al: Compensatory growth mechanisms regulated by BMP Lee MN, et al: Hepatitis C virus core protein represses the p21 pro-
and FGF signaling mediate liver regeneration in zebrafish after moter through inhibition of a TGF-beta pathway, J Gen Virol 83(Pt
partial hepatectomy, FASEB J 23(10):3516–3525, 2009. 9):2145–2151, 2002.
Kaplowitz N: Biochemical and cellular mechanisms of toxic liver injury, Lenaz G, et al: Mitochondrial bioenergetics in aging, Biochim Biophys
Semin Liver Dis 22(2):137–144, 2002. Acta 1459(2–3):397–404, 2000.
Karoui M, et al: Influence of preoperative chemotherapy on the risk of Leu JI, et al: Interleukin-6-induced STAT3 and AP-1 amplify hepato-
major hepatectomy for colorectal liver metastases, Ann Surg 243(1):1– cyte nuclear factor 1-mediated transactivation of hepatic genes,
7, 2006. an adaptive response to liver injury, Mol Cell Biol 21(2):414–424,
Karpen SJ, Trauner M: The new therapeutic frontier—nuclear recep- 2001.
tors and the liver, J Hepatol 52(3):455–462, 2010. Leu JI, et al: Impaired hepatocyte DNA synthetic response posthepa-
Kato H, et al: MMP-9 deficiency shelters endothelial PECAM-1 tectomy in insulin-like growth factor binding protein 1-deficient
expression and enhances regeneration of steatotic livers after ische- mice with defects in C/EBP beta and mitogen-activated protein
mia and reperfusion injury, J Hepatol 60(5):1032–1039, 2014. kinase/extracellular signal-regulated kinase regulation, Mol Cell Biol
Khurana S, Mukhopadhyay A: Characterization of the potential sub- 23(4):1251–1259, 2003.
population of bone marrow cells involved in the repair of injured Li W, et al: STAT3 contributes to the mitogenic response of hepatocytes
liver tissue, Stem Cells 25(6):1439–1447, 2007. during liver regeneration, J Biol Chem 277(32):28411–28417, 2002.
Kim TH, et al: Differential expression and distribution of focal adhe- Li T, et al: Exosomes derived from human umbilical cord mesenchymal
sion and cell adhesion molecules in rat hepatocyte differentiation, stem cells alleviate liver fibrosis, Stem Cells Dev 22(6):845–854, 2012.
Exp Cell Res 244(1):93–104, 1998. Lin WR, et al: The role of bone marrow-derived cells in fibrosis, Cells
Kim TH, et al: Expression and activation of pro-MMP-2 and pro- Tissues Organs 188(1–2):178–188, 2008.
MMP-9 during rat liver regeneration, Hepatology 31(1):75–82, Lo CM, et al: Minimum graft size for successful living donor liver
2000. transplantation, Transplantation 68(8):1112–1116, 1999.
Kim RD, et al: Impact of cell swelling on proliferative signal transduc- Locker J, et al: A common set of immediate-early response genes in
tion in the liver, J Cell Biochem 83(1):56–69, 2001. liver regeneration and hyperplasia, Hepatology 38(2):314–325, 2003.
Kiuchi T, et al: Impact of graft size mismatching on graft prognosis in Longo CR, et al: A20 protects mice from lethal radical hepatectomy by
liver transplantation from living donors, Transplantation 67(2):321– promoting hepatocyte proliferation via a p21waf1-dependent mecha-
327, 1999. nism, Hepatology 42(1):156–164, 2005.
Kofman AV, et al: Dose- and time-dependent oval cell reaction in Lonn P, et al: Regulating the stability of TGFbeta receptors and Smads,
acetaminophen-induced murine liver injury, Hepatology 41(6):1252– Cell Res 19(1):21–35, 2009.
1261, 2005. Lory J, et al: The pathology of the atrophy/hypertrophy complex (AHC)
Koga M, Ogasawara H: Induction of hepatocyte mitosis in intact adult of the liver. A light microscopic and immunohistochemical study,
rat by interleukin-1 alpha and interleukin-6, Life Sci 49(17):1263– Histol Histopathol 9(3):541–554, 1994.
1270, 1991. Ma S: Biology and clinical implications of CD133(+) liver cancer stem
Kohli V, et al: Endothelial cell and hepatocyte deaths occur by apoptosis cells, Exp Cell Res 319(2):126–132, 2012.
after ischemia-reperfusion injury in the rat liver, Transplantation Macías-Silva M, et al: Up-regulated transcriptional repressors SnoN
67(8):1099–1105, 1999. and Ski bind smad proteins to antagonize transforming growth
Koli K, et al: Transforming growth factor-beta activation in the lung: factor-? signals during liver regeneration, J Biol Chem 277(32):28483–
focus on fibrosis and reactive oxygen species, Antioxid Redox Signal 28490, 2002.
10(2):333–342, 2008. Maes OC, et al: Murine microRNAs implicated in liver functions and
Koniaris LG, et al: Liver regeneration, J Am Coll Surg 197(4):634–659, aging process, Mech Ageing Dev 129(9):534–541, 2008.
2003. Makuuchi M, et al: Safety of hemihepatic vascular occlusion during
Kooby DA, et al: Impact of steatosis on perioperative outcome follow- resection of the liver, Surg Gynecol Obstet 164(2):155–158, 1987.
ing hepatic resection, J Gastrointest Surg 7(8):1034–1044, 2003. Man K, et al: Intragraft gene expression profiles by cDNA microarray
Kountouras J, et al: Prolonged bile duct obstruction: a new experimen- in small-for-size liver grafts, Liver Transpl 9(4):425–432, 2003.
tal model for cirrhosis in the rat, Br J Exp Pathol 65(3):305–311, Mangelsdorf DJ, Evans RM: The RXR heterodimers and orphan recep-
1984. tors, Cell 83(6):841–850, 1995.
Krupczak-Hollis K, et al: Growth hormone stimulates proliferation of Mangelsdorf DJ, et al: The nuclear receptor superfamily: the second
old-aged regenerating liver through forkhead box m1b, Hepatology decade, Cell 83(6):835–839, 1995.
38(6):1552–1562, 2003. Mann DV, et al: Metabolic control patterns in acute phase and regen-
Krutzfeldt J, et al: Silencing of microRNAs in vivo with “antagomirs, erating human liver determined in vivo by 31-phosphorus magnetic
Nature 438(7068):685–689, 2005. resonance spectroscopy, Ann Surg 235(3):408–416, 2002.
Kubota K, et al: Measurement of liver volume and hepatic functional Marcos A, et al: Liver regeneration and function in donor and recipient
reserve as a guide to decision-making in resectional surgery for after right lobe adult to adult living donor liver transplantation,
hepatic tumors, Hepatology 26(5):1176–1181, 1997. Transplantation 69(7):1375–1379, 2000.
Kuo TK, et al: Stem cell therapy for liver disease: parameters governing Marcos A, et al: Selection and outcome of living donors for adult to
the success of using bone marrow mesenchymal stem cells, Gastro- adult right lobe transplantation, Transplantation 69(11):2410–2415,
enterology 134(7):2111–2121, 2121.e2111–2113, 2008. 2000.
Marshall A, et al: Relation between hepatocyte G1 arrest, impaired Olthoff KM, et al: Liver regeneration after living donor transplantation:
hepatic regeneration, and fibrosis in chronic hepatitis C virus infec- adult-to-adult living donor liver transplantation cohort study, Liver
tion, Gastroenterology 128(1):33–42, 2005. Transpl 21(1):79–88, 2015.
Matsumoto T, et al: Insulin gene transfer with adenovirus vector via the Palmes D, et al: Impact of rapamycin on liver regeneration, Virchows
spleen safely and effectively improves posthepatectomized conditions Arch 452(5):545–557, 2008.
in diabetic rats, J Surg Res 110(1):228–234, 2003. Pamecha V, et al: Effect of portal vein embolisation on the growth
McCormack L, et al: Hepatic steatosis is a risk factor for postoperative rate of colorectal liver metastases, Br J Cancer 100(4):617–622,
complications after major hepatectomy: a matched case-control 2009.
study, Ann Surg 245(6):923–930, 2007. Pan XN, et al: Bone marrow-derived mesenchymal stem cell therapy
Michalopoulos GK, DeFrances MC: Liver regeneration, Science for decompensated liver cirrhosis: a meta-analysis, World J Gastroen-
276(5309):60–66, 1997. terol 20(38):14051–14057, 2014.
Michalopoulos G, et al: Control of hepatocyte replication by two serum Parekkadan B, et al: Immunomodulation of activated hepatic stellate
factors, Cancer Res 44(10):4414–4419, 1984. cells by mesenchymal stem cells, Biochem Biophys Res Commun
Mischoulon D, et al: Growth-dependent inhibition of CCAAT 363(2):247–252, 2007.
enhancer-binding protein (C/EBP alpha) gene expression during Parekkadan B, et al: Mesenchymal stem cell-derived molecules reverse
hepatocyte proliferation in the regenerating liver and in culture, Mol fulminant hepatic failure, PLoS ONE 2(9):e941, 2007.
Cell Biol 12(6):2553–2560, 1992. Peng L, et al: Autologous bone marrow mesenchymal stem cell
Mitchell C, et al: Heparin-binding epidermal growth factor-like growth transplantation in liver failure patients caused by hepatitis B:
factor links hepatocyte priming with cell cycle progression during short-term and long-term outcomes, Hepatology 54(3):820–828,
liver regeneration, J Biol Chem 280(4):2562–2568, 2005. 2011.
Miyoshi H, et al: Hepatocyte apoptosis after bile duct ligation in the Peralta C, et al: Hepatic ischemia and reperfusion injury: effects on the
mouse involves Fas, Gastroenterology 117(3):669–677, 1999. liver sinusoidal milieu, J Hepatol 59(5):1094–1106, 2013.
Mohn KL, et al: Immediate-early gene expression differs between Plouffe SW, et al: Disease implications of the Hippo/YAP pathway,
regenerating liver, insulin-stimulated H-35 cells, and mitogen-stim- Trends Mol Med 21(4):212–222, 2015.
ulated Balb/c 3T3 cells. Liver-specific induction patterns of gene 33, Pomfret EA, et al: Liver regeneration and surgical outcome in donors
phosphoenolpyruvate carboxykinase, and the jun, fos, and egr fami- of right-lobe liver grafts, Transplantation 76(1):5–10, 2003.
lies, J Biol Chem 265(35):21914–21921, 1990. Pottakkat B, et al: Surgical management of patients with post-cholecys-
Monga SP: Role of Wnt/beta-catenin signaling in liver metabolism and tectomy benign biliary stricture complicated by atrophy-hypertrophy
cancer, Int J Biochem Cell Biol 43(7):1021–1029, 2011. complex of the liver, HPB (Oxford) 11(2):125–129, 2009.
Monga SP, et al: Hepatocyte growth factor induces Wnt-independent Poynard T, et al: Natural history of liver fibrosis progression in patients
nuclear translocation of beta-catenin after Met-beta-catenin dissocia- with chronic hepatitis C, Lancet 349(9055):825–832, 1997.
tion in hepatocytes, Cancer Res 62(7):2064–2071, 2002. Rabes HM, et al: Analysis of cell cycle compartments of hepatocytes
Moolten FL, Bucher NL: Regeneration of rat liver: transfer of humoral after partial hepatecomy, Cell Tissue Kinet 9(6):517–532, 1976.
agent by cross circulation, Science 158(798):272–274, 1967. Rai RM, et al: Impaired liver regeneration in inducible nitric oxide
Morello D, et al: c-myc, c-fos, and c-jun regulation in the regenerating synthasedeficient mice, Proc Natl Acad Sci USA 95(23):13829–13834,
livers of normal and H-2K/c-myc transgenic mice, Mol Cell Biol 1998.
10(6):3185–3193, 1990. Ratajczak J, et al: Membrane-derived microvesicles: important and
Morita T, et al: Mechanism of postoperative liver failure after excessive underappreciated mediators of cell-to-cell communication, Leukemia
hepatectomy investigated using a cDNA microarray, J Hepatobiliary 20(9):1487–1495, 2006.
Pancreat Surg 9(3):352–359, 2002. Ray RB, et al: Hepatitis C virus core protein represses p21WAF1/Cip1/
Nagy P, et al: Dexamethasone inhibits the proliferation of hepatocytes Sid1 promoter activity, Gene 208(2):331–336, 1998.
and oval cells but not bile duct cells in rat liver, Hepatology Reddy BV, Irvine KD: The Fat and Warts signaling pathways: new
28(2):423–429, 1998. insights into their regulation, mechanism and conservation, Develop-
Nakano K, et al: Lower activity of CCAAT/enhancer-binding protein ment 135(17):2827–2838, 2008.
and expression of cyclin E, but not cyclin D1, activating protein-1 Reddy SK, et al: Addition of bevacizumab to irinotecan- and oxaliplatin-
and p21(WAF1), after partial hepatectomy in obstructive jaundice, based preoperative chemotherapy regimens does not increase mor-
Biochem Biophys Res Commun 280(3):640–645, 2001. bidity after resection of colorectal liver metastases, J Am Coll Surg
Nakao A, et al: TGF-beta receptor-mediated signalling through Smad2, 206(1):96–106, 2008.
Smad3 and Smad4, EMBO J 16(17):5353–5362, 1997. Repa JJ, et al: Disruption of the sterol 27-hydroxylase gene in mice
Neuschwander-Tetri BA, Caldwell SH: Nonalcoholic steatohepatitis: results in hepatomegaly and hypertriglyceridemia. Reversal by cholic
summary of an AASLD Single Topic Conference, Hepatology acid feeding, J Biol Chem 275(50):39685–39692, 2000.
37(5):1202–1219, 2003. Reynolds ES: Liver parenchymal cell injury. I. Initial alterations of the
Nishimura Y, et al: Mitochondrial dysfunction and cytoskeletal disrup- cell following poisoning with carbon tetrachloride, J Cell Biol
tion during chemical hypoxia to cultured rat hepatic sinusoidal 19:139–157, 1963.
endothelial cells: the pH paradox and cytoprotection by glucose, Rhim JA, et al: Replacement of diseased mouse liver by hepatic cell
acidotic pH, and glycine, Hepatology 27(4):1039–1049, 1998. transplantation, Science 263(5150):1149–1152, 1994.
Nishizaki T, et al: Nutritional support after hepatic resection: a random- Ribero D, et al: Portal vein embolization before major hepatectomy and
ized prospective study, Hepatogastroenterology 43(9):608–613, 1996. its effects on regeneration, resectability and outcome, Br J Surg
Oertel M, Shafritz DA: Stem cells, cell transplantation and liver repopu- 94(11):1386–1394, 2007.
lation, Biochim Biophys Acta 1782(2):61–74, 2008. Robinson S, et al: Systemic chemotherapy and its implications for resec-
Ohmura T, et al: Hepatocyte proliferation induced by a single dose of tion of colorectal liver metastasis, Surg Oncol 20(2):57–72, 2011.
a peroxisome proliferator, Am J Pathol 148(3):815–824, 1996. Roderfeld M, et al: Bone marrow transplantation demonstrates medul-
Oikawa T, et al: Sal-like protein 4 (SALL4), a stem cell biomarker in lar origin of CD34(+) fibrocytes and ameliorates hepatic fibrosis in
liver cancers, Hepatology 57(4):1469–1483, 2013. Abcb4(−/−) mice, Hepatology 51(1):267–276, 2010.
Olthoff KM: Molecular pathways of regeneration and repair after liver Rodrigo-Torres D, et al: The biliary epithelium gives rise to liver pro-
transplantation, World J Surg 26(7):831–837, 2002. genitor cells, Hepatology 60(4):1367–1377, 2014.
Olthoff KM: Hepatic regeneration in living donor liver transplantation, Ronco MT, et al: Modulation of balance between apoptosis and pro-
Liver Transpl 9(10C):S35–S41, 2003. liferation by lipid peroxidation (LPO) during rat liver regeneration,
Olthoff KM, et al: Outcomes of adult living donor liver transplantation: Mol Med 8(12):808–817, 2002.
comparison of the Adult-to-adult Living Donor Liver Transplanta- Rosa JL, et al: Gene expression of regulatory enzymes of glycolysis/
tion Cohort Study and the national experience, Liver Transpl gluconeogenesis in regenerating rat liver, Biochem J 287(Pt 1):113–
17(7):789–797, 2011. 116, 1992.
Olthoff KM, et al: Outcomes of 385 adult-to-adult living donor liver Roskams T, et al: Hepatic OV-6 expression in human liver disease and
transplant recipients: a report from the A2ALL Consortium, Ann rat experiments: Evidence for hepatic progenitor cells in man,
Surg 242(3):314–323, discussion 323–325, 2005. J Hepatol 29(3):455–463, 1998.
Rubbia-Brandt L, et al: Severe hepatic sinusoidal obstruction associ- Shrivastava A, et al: Ectopic expression of hepatitis C virus core protein
ated with oxaliplatin-based chemotherapy in patients with metastatic differentially regulates nuclear transcription factors, J Virol 72(12):
colorectal cancer, Ann Oncol 15(3):460–466, 2004. 9722–9728, 1998.
Rubbia-Brandt L, et al: Gene expression profiling provides insights into Shteyer E, et al: Disruption of hepatic adipogenesis is associated with
pathways of oxaliplatin-related sinusoidal obstruction syndrome in impaired liver regeneration in mice, Hepatology 40(6):1322–1332,
humans, Mol Cancer Ther 10(4):687–696, 2011. 2004.
Ruderman NB, et al: AMPK and SIRT1: a long-standing partnership? Skrtic S, et al: Hepatocyte-stimulated expression of hepatocyte growth
Am J Physiol Endocrinol Metab 298(4):E751–E760, 2010. factor (HGF) in cultured rat hepatic stellate cells, J Hepatol
Rudiger HA, et al: Comparison of ischemic preconditioning and 30(1):115–124, 1999.
intermittent and continuous inflow occlusion in the murine liver, Skullman S, et al: Influence of malnutrition on regeneration and
Ann Surg 235(3):400–407, 2002. composition of the liver in rats, Acta Chir Scand 156(10):717–722,
Rudnick DA, et al: Prostaglandins are required for CREB activation 1990.
and cellular proliferation during liver regeneration, Proc Natl Acad Song W, et al: Expression and clinical significance of the stem cell
Sci USA 98(15):8885–8890, 2001. marker CD133 in hepatocellular carcinoma, Int J Clin Pract 62(8):
Russo FP, et al: The bone marrow functionally contributes to liver 1212–1218, 2008.
fibrosis, Gastroenterology 130(6):1807–1821, 2006. Song G, et al: MicroRNAs control hepatocyte proliferation during liver
Russo L, et al: Addition of simvastatin to cold storage solution prevents regeneration, Hepatology 51(5):1735–1743, 2010.
endothelial dysfunction in explanted rat livers, Hepatology 55(3): Starzl TE, et al: A hundred years of the hepatotrophic controversy, Ciba
921–930, 2012. Found Symp 55:111–129, 1977.
Sabetkish S, et al: Whole-organ tissue engineering: Decellularization Starzl TE, et al: Portal hepatotrophic factors, diabetes mellitus and
and recellularization of three-dimensional matrix liver scaffolds, acute liver atrophy, hypertrophy and regeneration, Surg Gynecol
J Biomed Mater Res A 103(4):1498–1508, 2014. Obstet 141(6):843–858, 1975.
Saiki S, et al: Preoperative internal biliary drainage is superior to Steller H: Mechanisms and genes of cellular suicide, Science 267(5203):
external biliary drainage in liver regeneration and function after 1445–1449, 1995.
hepatectomy in obstructive jaundiced rats, Ann Surg 230(5):655–662, Stocker E, Heine WD: Regeneration of liver parenchyma under normal
1999. and pathological conditions, Beitr Pathol 144(4):400–408, 1971.
Salehi S, et al: Human liver regeneration is characterized by the coor- Streetz KL, et al: Interleukin 6 and liver regeneration, Gut 47(2):309–
dinated expression of distinct microRNA governing cell cycle fate, 312, 2000.
Am J Transplant 13(5):1282–1295, 2013. Strey CW, et al: The proinflammatory mediators C3a and C5a are
Scappaticci FA, et al: Surgical wound healing complications in meta- essential for liver regeneration, J Exp Med 198(6):913–923, 2003.
static colorectal cancer patients treated with bevacizumab, J Surg Strey CW, et al: Partial hepatectomy induced liver proteome changes
Oncol 91(3):173–180, 2005. in mice, Proteomics 4:1–8, 2004.
Scatton O, et al: Impact of localized congestion related to venous Strom SC, et al: Chimeric mice with humanized liver: tools for the
deprivation after hepatectomy, Surgery 143(4):483–489, 2008. study of drug metabolism, excretion, and toxicity, Methods Mol Biol
Schadde E, et al: Systematic review and meta-analysis of feasibility, 640:491–509, 2010.
safety, and efficacy of a novel procedure: associating liver partition Sullivan GJ, et al: Generation of functional human hepatic endoderm
and portal vein ligation for staged hepatectomy, Ann Surg Oncol from human induced pluripotent stem cells, Hepatology 51(1):329–
22(9):3109–3120, 2015. 335, 2009.
Schaub JR, et al: Evidence against a stem cell origin of new hepatocytes Sun R, Gao B: Negative regulation of liver regeneration by innate
in a common mouse model of chronic liver injury, Cell Rep 8(4): immunity (natural killer cells/interferon-gamma), Gastroenterology
933–939, 2014. 127(5):1525–1539, 2004.
Schlegel A, et al: ALPPS: from human to mice highlighting accelerated Suzuki H, et al: Internal biliary drainage, unlike external drainage, does
and novel mechanisms of liver regeneration, Ann Surg 260(5):839– not suppress the regeneration of cholestatic rat liver after partial
846, discussion 846–847, 2014. hepatectomy, Hepatology 20(5):1318–1322, 1994.
Schnitzbauer AA, et al: Right portal vein ligation combined with in situ Tackett BC, et al: P2Y2 purinergic receptor activation is essential for
splitting induces rapid left lateral liver lobe hypertrophy enabling efficient hepatocyte proliferation in response to partial hepatectomy,
2-staged extended right hepatic resection in small-for-size settings, Am J Physiol Gastrointest Liver Physiol 307(11):G1073–G1087, 2014.
Ann Surg 255(3):405–414, 2012. Taguchi T, et al: Differences in DNA synthesis in vitro using isolated
Seehofer D, et al: Intraabdominal bacterial infections significantly alter nuclei from regenerating livers of young and aged rats, Mech Ageing
regeneration and function of the liver in a rat model of major hepa- Dev 122(2):141–155, 2001.
tectomy, Langenbecks Arch Surg 392(3):273–284, 2007. Takamura K, et al: Activin and activin receptor expression changes in
Sekine K, et al: Plasma hepatocyte growth factor is increased in early- liver regeneration in rat, J Surg Res 126(1):3–11, 2005a.
phase sepsis, J Infect Chemother 10(2):110–114, 2004. Takamura K, et al: Possible endocrine control by follistatin 315 during
Sekiya S, Suzuki A: Direct conversion of mouse fibroblasts to hepato- liver regeneration based on changes in the activin receptor after a partial
cyte-like cells by defined factors, Nature 475(7356):390–393, 2011. hepatectomy in rats, Hepatogastroenterology 52(61):60–66, 2005b.
Selzner M, et al: Ischemia impairs liver regeneration after major tissue Tamasi V, et al: The effect of dexamethasone on P450 activities in regen-
loss in rodents: Protective effects of interleukin-6, Hepatology erating rat liver, Biochem Biophys Res Commun 286(2):239–242,
30(2):469–475, 1999. 2001.
Selzner M, et al: Mechanisms of ischemic injury are different in the Tang Y, et al: Progenitor/stem cells give rise to liver cancer due to
steatotic and normal rat liver, Hepatology 32(6):1280–1288, 2000. aberrant TGF-beta and IL-6 signaling, Proc Natl Acad Sci USA
Selzner N, et al: Cold ischemia decreases liver regeneration after partial 105(7):2445–2450, 2008.
liver transplantation in the rat: a TNF-?/IL-6-dependent mechanism, Tang KH, et al: CD133(+) liver tumor-initiating cells promote tumor
Hepatology 36(4 Pt 1):812–818, 2002. angiogenesis, growth, and self-renewal through neurotensin/interleu-
Sengupta S, et al: Regulation of the mTOR complex 1 pathway by kin-8/CXCL1 signaling, Hepatology 55(3):807–820, 2012.
nutrients, growth factors, and stress, Mol Cell 40(2):310–322, 2010. Taub R: Hepatoprotection via the IL-6/Stat3 pathway, J Clin Invest
Shi JH, et al: RAF-targeted therapy for hepatocellular carcinoma in the 112(7):978–980, 2003.
regenerating liver, J Surg Oncol 107(4):393–401, 2012. Taub R: Liver regeneration: from myth to mechanism, Nat Rev Mol Cell
Shibayama Y, et al: Recovery from hepatic necrosis following acute Biol 5(10):836–847, 2004.
portal vein embolism with special reference to reconstruction of Terrault NA, et al: Outcomes in hepatitis C virus-infected recipients of
occluded vessels, J Pathol 165(3):255–261, 1991. living donor vs. deceased donor liver transplantation, Liver Transpl
Shiffman ML, et al: Histologic recurrence of chronic hepatitis C virus 13(1):122–129, 2007.
in patients after living donor and deceased donor liver transplanta- Theise ND, et al: Derivation of hepatocytes from bone marrow cells in
tion, Liver Transpl 10(10):1248–1255, 2004. mice after radiation-induced myeloablation, Hepatology 31(1):235–
Shoup M, et al: Volumetric analysis predicts hepatic dysfunction in 240, 2000.
patients undergoing major liver resection, J Gastrointest Surg Thompson NL, et al: Sequential protooncogene expression during rat
7(3):325–330, 2003. liver regeneration, Cancer Res 46(6):3111–3117, 1986.
Timchenko NA: Aging and liver regeneration, Trends Endocrinol Metab Webber EM, et al: In vivo response of hepatocytes to growth factors
20(4):171–176, 2009. requires an initial priming stimulus, Hepatology 19(2):489–497,
Togo S, et al: Mechanism of liver regeneration after partial hepatectomy 1994.
using mouse cDNA microarray, J Hepatol 40(3):464–471, 2004. Weiss YG, et al: Compensatory hepatic regeneration after mild, but not
Toso C, et al: The place of downstaging for hepatocellular carcinoma, fulminant, intraperitoneal sepsis in rats, Am J Physiol Gastrointest
J Hepatol 52(6):930–936, 2010. Liver Physiol 280(5):G968–G973, 2001.
Tracy TF Jr, et al: Cholestasis without cirrhosis alters regulatory liver Weymann A, et al: p21 is required for dextrose-mediated inhibition of
gene expression and inhibits hepatic regeneration, Surgery 110(2): mouse liver regeneration, Hepatology 50(1):207–215, 2009.
176–182, discussion 182–183, 1991. Widmann JJ, Fahimi HD: Proliferation of mononuclear phagocytes
Tralhao JG, et al: Paracrine in vivo inhibitory effects of hepatitis B virus (Kupffer cells) and endothelial cells in regenerating rat liver. A light
X protein (HBx) on liver cell proliferation: an alternative mechanism and electron microscopic cytochemical study, Am J Pathol 80(3):
of HBx-related pathogenesis, Proc Natl Acad Sci USA 99(10):6991– 349–366, 1975.
6996, 2002. Wu H, et al: Integration of Hippo signalling and the unfolded protein
Trotter JF, et al: Laboratory test results after living liver donation in the response to restrain liver overgrowth and tumorigenesis, Nat Commun
adult-to-adult living donor liver transplantation cohort study, Liver 6:6239, 2015.
Transpl 17(4):409–417, 2011. Wuestefeld T, et al: Interleukin-6/glycoprotein 130-dependent pathways
Trotter JF, et al: Living donor liver transplant recipients achieve rela- are protective during liver regeneration, J Biol Chem 278(13):11281–
tively higher immunosuppressant blood levels than cadaveric recipi- 11288, 2003.
ents, Liver Transpl 8(3):212–218, 2002. Yamada Y, et al: Initiation of liver growth by tumor necrosis factor:
Turmelle YP, et al: Rosiglitazone inhibits mouse liver regeneration, Deficient liver regeneration in mice lacking type I tumor necrosis
FASEB J 20(14):2609–2611, 2006. factor receptor, Proc Natl Acad Sci USA 94(4):1441–1446, 1997.
Uemura T, et al: Different expression of positive and negative regulators Yamada T, et al: Insulin requirements for hepatic regeneration following
of hepatocyte growth in growing and shrinking hepatic lobes after hepatectomy, Ann Surg 185(1):35–42, 1977.
portal vein branch ligation in rats, Int J Mol Med 5(2):173–179, Yan Y, et al: Mesenchymal stem cells from human umbilical cords
2000. ameliorate mouse hepatic injury in vivo, Liver Int 29(3):356–365,
Uetake H, et al: A multicenter phase II trial of mFOLFOX6 plus beva- 2009.
cizumab to treat liver-only metastases of colorectal cancer that are Yang SQ, et al: Disrupted signaling and inhibited regeneration in obese
unsuitable for upfront resection (TRICC0808), Ann Surg Oncol mice with fatty livers: implications for nonalcoholic fatty liver disease
22(3):908–915, 2015. pathophysiology, Hepatology 34(4 Pt 1):694–706, 2001.
Vacca M, et al: Clustering nuclear receptors in liver regeneration identi- Yanger K, et al: Adult hepatocytes are generated by self-duplication
fies candidate modulators of hepatocyte proliferation and hepatocar- rather than stem cell differentiation, Cell Stem Cell 15(3):340–349,
cinoma, PLoS ONE 9(8):e104449, 2014. 2014.
Vacca M, et al: Nuclear receptors in regenerating liver and hepatocel- Yigitler C, et al: The small remnant liver after major liver resection:
lular carcinoma, Mol Cell Endocrinol 368(1–2):108–119, 2013. how common and how relevant? Liver Transpl 9(9):S18–S25, 2003.
Van Buren G 2nd, et al: Effect of molecular therapeutics on liver Yokoyama Y, et al: Mechanisms of hepatic regeneration following portal
regeneration in a murine model, J Clin Oncol 26(11):1836–1842, vein embolization and partial hepatectomy: a review, World J Surg
2008. 31(2):367–374, 2007.
van Poll D, et al: Mesenchymal stem cell-derived molecules directly Yokoyama Y, et al: Which gender is better positioned in the process of
modulate hepatocellular death and regeneration in vitro and in vivo, liver surgery? Male or female? Surg Today 37(10):823–830, 2007.
Hepatology 47(5):1634–1643, 2008. Yoshida R, et al: Branched-chain amino acid-enriched nutrients
Vassilopoulos G, et al: Transplanted bone marrow regenerates liver by improve nutritional and metabolic abnormalities in the early post-
cell fusion, Nature 422(6934):901–904, 2003. transplant period after living donor liver transplantation, J Hepatobili-
Vauthey JN, et al: Standardized measurement of the future liver ary Pancreat Sci 19(4):438–448, 2012.
remnant prior to extended liver resection: methodology and clinical You H, et al: Epigenetic regulation of cancer stem cell marker CD133
associations, Surgery 127(5):512–519, 2000. by transforming growth factor-beta, Hepatology 51(5):1635–1644,
Vauthey JN, et al: Chemotherapy regimen predicts steatohepatitis and 2010.
an increase in 90-day mortality after surgery for hepatic colorectal Yu B, et al: Reprogramming fibroblasts into bipotential hepatic stem
metastases, J Clin Oncol 24(13):2065–2072, 2006. cells by defined factors, Cell Stem Cell 13(3):328–340, 2013.
Vauthey JN, Mise Y: Commentary on “Can we improve the morbidity Yuan B, et al: Down-regulation of miR-23b may contribute to activa-
and mortality associated with the associating liver partition with tion of the TGF-beta1/Smad3 signalling pathway during the termi-
portal vein ligation for staged hepatectomy (ALPPS) procedure in nation stage of liver regeneration, FEBS Lett 585(6):927–934,
the management of colorectal liver metastases? Surgery 157(2):207– 2011.
210, 2014. Zhao Q, et al: Stem/progenitor cells in liver injury repair and regenera-
Vauthey JN, Zorzi D: In search of the black sheep: is it bevacizumab tion, Biol Cell 101(10):557–571, 2009.
or extended chemotherapy? Ann Surg Oncol 16(6):1463–1464, 2009. Zheng T, et al: Hippo signaling in oval cells and hepatocarcinogenesis,
Verran D, et al: Clinical experience gained from the use of 120 steatotic Cancer Lett 302(2):91–99, 2011.
donor livers for orthotopic liver transplantation, Liver Transpl Zhong JH, et al: Hepatic resection is safe and effective for patients with
9(5):500–505, 2003. hepatocellular carcinoma and portal hypertension, PLoS ONE
Vetelainen R, et al: Steatosis as a risk factor in liver surgery, Ann Surg 9(9):e108755, 2014.
245(1):20–30, 2007. Zhong Z, et al: Liver regeneration is suppressed in small-for-size liver
Vigano L, et al: Liver resection for colorectal metastases after chemo- grafts after transplantation: involvement of c-Jun N-terminal kinase,
therapy: impact of chemotherapy-related liver injuries, pathological cyclin D1, and defective energy supply, Transplantation 82(2):241–
tumor response, and micrometastases on long-term survival, Ann 250, 2006.
Surg 258(5):731–740, discussion 741–742, 2013. Zhong Z, et al: Inhibition of TGF-ß/Smad signaling improves regenera-
Vogten JM, et al: Angiostatin inhibits experimental liver fibrosis in tion of small-for-size rat liver grafts, Liver Transpl 16(2):181–190,
mice, Int J Colorectal Dis 19(4):387–394, 2004. 2010.
Vouche M, et al: Radiation lobectomy: time-dependent analysis of Zhou Q, et al: Stem cells with decellularized liver scaffolds in liver
future liver remnant volume in unresectable liver cancer as a bridge regeneration and their potential clinical applications, Liver Int
to resection, J Hepatol 59(5):1029–1036, 2013. 35(3):687–694, 2014.
Wagner M, et al: Nuclear receptors in liver disease, Hepatology Zimmers TA, et al: Massive liver growth in mice induced by systemic
53(3):1023–1034, 2011. interleukin 6 administration, Hepatology 38(2):326–334, 2003.
Wang X, et al: The Forkhead Box m1b transcription factor is essential Zorzi D, et al: Chemotherapy with bevacizumab does not affect liver
for hepatocyte DNA replication and mitosis during mouse liver regeneration after portal vein embolization in the treatment of
regeneration, Proc Natl Acad Sci USA 99(26):16881–16886, 2002. colorectal liver metastases, Ann Surg Oncol 15(10):2765–2772, 2008.
Wang X, et al: Cell fusion is the principal source of bone-marrow-
derived hepatocytes, Nature 422(6934):897–901, 2003.
CHAPTER 7
Liver fibrogenesis: mechanisms and
clinical relevance
Diana Vetter and Scott L. Friedman
Liver fibrosis represents a scarring response to either acute or Remarkably, as many as 45% of all deaths are related to some
chronic liver injury. Following acute liver injury, parenchymal kind of fibrosis (Wynn, 2008), stressing the importance of this
cells regenerate to successfully preserve hepatocellular mass response and explaining the growing interest in this field of
and function. This acute process is associated with an inflam- research. For decades, fibrosis was considered an irreversible
matory and fibrogenic response but with limited deposition of disease that progresses to cirrhosis with a greater risk for hepa-
extracellular matrix (ECM). In contrast, prolonged liver injury tocellular carcinoma and with development of liver failure. This
leads to sustained production of growth factors, proteolytic meant that the only potential treatment for liver fibrosis was
enzymes, angiogenic factors, and fibrogenic cytokines. These liver transplantation once cirrhosis was present.
events culminate in the accumulation of extracellular matrix, Research during the past 35 years has yielded increasing
forming septa that coalesce into broad bands of scar tissue insight into the cellular and molecular mechanisms of this
encircling nodules of hepatocytes and leading to altered micro- disease, uncovering an orchestrated pathophysiology, and iden-
vascular structure (Friedman, 2004; Lee et al, 2015) (Fig. 7.1). tifying the hepatic stellate cell (HSC) as the central cell type in
This late stage of fibrosis, termed cirrhosis, ultimately impairs fibrogenesis (Friedman et al, 1985) and, most important,
liver function and leads to portal hypertension and its revealing the potential reversibility of the disease and discovery
complications. of potential therapeutic targets.
Typically, progression of fibrosis to cirrhosis evolves for
decades before clinical events ensue, but disease may progress MOLECULAR AND CELLULAR MECHANISMS
more rapidly following repeated episodes of severe acute alco- OF FIBROSIS
holic hepatitis, subfulminant hepatitis (especially due to drug
toxicity), and fibrosing cholestasis in patients with hepatitis C The anatomic arrangement of the parenchymal and nonparen-
virus (HCV) reinfection following liver transplantation. In chymal cells of the liver contributes to its unique role as an
addition, there have been reports of rapidly progressive acute immune organ and helps explain how the liver responds to an
HCV with fibrosis in men coinfected with human immunode- insult. The liver is composed primarily of epithelial cells (hepa-
ficiency virus (HIV) (Fierer et al, 2008). tocytes and cholangiocytes), as well as resident nonparenchy-
Genetic and environmental factors also influence the course mal cells that include hepatic macrophages (Kupffer cells),
of the disease. For example, polymorphisms in a number of sinusoidal endothelium, and HSCs. In addition to Kupffer
candidate genes involving the inflammatory (e.g., Toll-like cells, a growing list of specialized immune cells have been
receptor 4 [TLR4] [Guo et al, 2009]) or the immune (Powell characterized, including dendritic cells, natural killer (NK)
et al, 2000) response may influence the progression of liver cells, and natural killer T (NKT) cells, emphasizing that the
fibrosis in humans. In HCV, the development of a seven-gene liver represents a key organ in the regulation of innate immunity
signature has good prognostic value in assessing the risk (Crispe, 2014; Gao & Radaeva, 2013) (see Chapter 10).
for cirrhosis development (Huang et al, 2007) and has been The liver capsule extends as septa into the liver, delineating
further validated in two separate cohorts (Pradat et al, 2010; hepatic lobules that form the structural units of the liver. The
Trepo et al, 2011). However, with the development of highly lobule forms a hexagonal structure with portal triads (including
effective and well-tolerated therapies for HCV, this kind of branches of the hepatic portal vein, the hepatic artery, and the
genetic test has diminishing utility for HCV disease, because bile duct) localized in the periphery of the lobule, and with a
fibrosis progression risk is unlikely to influence the decision portal vein branch in the center (see Fig. 7.1). Hepatocyte plates
whether to treat the infection, as eventually most patients will radiate outward from the central vein and are separated from
probably be treated. Unfortunately, the development of a each other by sinusoids. The latter form the connecting element
similar genetic risk score for nonalcoholic fatty liver disease between the branches of the hepatic portal veins and hepatic
(NAFLD) has been elusive, possibly because the disease is arteries with the central vein. Kupffer cells, NK cells, NKT
more heterogeneous. cells, and dendritic cells, all of which are important components
The main etiologies of liver fibrosis in Western countries are of the innate immune system, reside in the hepatic sinusoids.
chonic HCV and hepatitis B virus (HBV) infection, alcohol The subendothelial space between the sinusoidal endothelium
abuse, and nonalcoholic steatohepatitis (NASH) (see Chapters and hepatocytes is also termed the space of Disse. Thus the
70 and 71). As a generalized tissue response to chronic injury, HSCs, which lie in the space of Disse, have direct contact with
fibrosis also occurs in many other organs (heart, lung, kidneys) endothelial cells and hepatocytes. Sinusoidal endothelial cells
and typically represents the result of an ongoing inflammation. are highly fenestrated, which allows unimpeded flow of plasma
110
Chapter 7 Liver fibrogenesis: mechanisms and clinical relevance 111
Normal liver
Portal triad
Bile duct
Hepatocytes
HSC
Sinusoidal space of Disse
Portal vein Terminal

Sinusoidal endothelial cells KC hepatic
vein
Hepatic
arteriole
Fibrotic liver
HSC activation Loss of hepatocyte

and proliferation microvilli
Loss of endothelial fenestrations Distortion

of veins
Increase in fibril-forming
collagen in space of Disse
Fibril-forming collagens (types I, III,V)

Basement membrane collagens (types IV,VI)
B Glycoconjugates (laminin, fibronectin, glycosaminoglycans, tenascin)
FIGURE 7.1. Matrix and cellular alteration in hepatic fibrosis. Normal liver parenchyma contains epithelial cells (hepatocytes) and nonparen-
chymal cells: fenestrated sinusoidal endothelium, hepatic stellate cells (HSCs), and Kupffer cells (KCs). A, Sinusoids are separated from hepatocytes
by a low-density basement membrane–like matrix confined to the space of Disse, which ensures metabolic exchange. Upon injury, the HSCs become
activated and secrete large amounts of extracellular matrix (ECM), resulting in progressive thickening of the septa. B, Deposition of ECM in the space
of Disse leads to the loss of both endothelial fenestrations and hepatocyte microvilli, which results in both the impairment of normal bidirectional
metabolic exchange between portal venous flow and hepatocytes and the development of portal hypertension. (From Hernandez-Gea V, Friedman
SL: Pathogenesis of liver fibrosis, Annu Rev Pathol 6:425–456, 2011.)
from sinusoidal blood into the space of Disse. Through this factor kappa B (NF-κB) in Kupffer cells, which leads to
arrangement, hepatocytes and HSCs are exposed directly to increased tumor necrosis factor-α (TNF-α) production (see
plasma derived largely from venous blood draining the Fig. 7.2). TNF-α in turn induces neutrophil infiltration and
intestine. stimulates mitochondrial oxidant production in hepatocytes,
which are then sensitized to undergo apoptosis. Furthermore,
Common Triggers of Hepatic Fibrogenesis ROS and acetaldehyde, the main degradation product of
Ongoing insult to the liver will lead to an increased inflamma- alcohol, both activate HSCs and stimulate inflammatory signals
tory state with activation of hepatic stellate cells, which ulti- (Maher et al, 1994). Interestingly, many of the same gut defects
mately tilts the profibrotic and antifibrotic balance toward in alcoholic liver disease are now also implicated in NASH, with
fibrosis. Viral infection, reactive oxygen species (ROS), and bile additional focus on the nature of the microbiome as well as the
acids are among the most common stress signals for the liver integrity of the gut mucosa as determinants of fatty liver disease
(Fig. 7.2). An in vitro study further suggests that free fatty acids (De Minicis et al, 2014; Dumas et al, 2014; Fouts et al, 2012).
promote fibrogenesis by indirect activation of HSCs (Wobser In HCV infection, the virus escapes the immune response
et al, 2009), which may be relevant to the pathogenesis of fatty and infects hepatocytes (Wang et al, 2013). This causes oxida-
liver disease. tive stress, again leading to recruitment of inflammatory cells
In alcoholic liver disease, ethanol decreases gut motility, and HSC activation. HSCs can also be directly activated either
increases epithelial permeability, and promotes overgrowth by HCV through membrane receptors (Mazzocca et al, 2005;
of gram-negative bacteria. Consequently, lipopolysaccharide Schulze-Krebs et al, 2005), or by HBV (Martin-Vilchez et al,
(LPS) concentration is elevated in portal blood, through the 2008) (see Chapter 70).
TLR4 signaling complex, to generate ROS via reduced nico- Bile acids are hepatotoxic agents and typically target hepa-
tinamide adenine dinucleotide phosphate (NADPH) oxidase tocytes but may also injure biliary epithelium (Higuchi &
(Paik et al, 2003; Roh & Seki, 2013; Vazquez-Torres et al, Gores, 2003). In addition to their potential role in provoking
2004; Zhang et al, 2012). Oxidants then upregulate nuclear damage, there is an increasing appreciation for the role of bile
Small intestine
Gram-neg.
bacteria↑
Kupffer cells (macrophages)
LPS in portal blood ↑

T cells
Inflammatory TNF-α
cytokines Neutrophils
ROS
Alcohol Mito. oxidant Apoptosis qHSCs
production
NASH
ROS
HCV Oxidative Sensitized to
stress↑ apoptosis NADPH
HBV aHSCs Proliferation
FXR Contractility
Free fatty acids Fibrogenesis
Bile acids via EGFR Altered matrix
degradation
HSC chemotaxis
Infammatory
signaling
Myofibroblasts
FIGURE 7.2. Pathways of cellular injury and fibrosis. This diagram depicts the key pathways of cellular injury and fibrosis. The main causes of
chronic liver injury are alcohol, nonalcoholic steatohepatitis (NASH), viral infection, and injury from bile acids in cholestatic conditions. All factors
activate hepatic stellate cells (HSCs), which is a key event in liver fibrogenesis. Alcohol can promote gram-negative bacterial overgrowth of the small
intestine and/or reduced gut integrity, thereby increasing lipopolysaccharide (LPS) in the portal blood. LPS activates Kupffer cells (hepatic macro-
phages), which increase the mitochondrial oxidant production in hepatocytes by way of tumor necrosis factor-α (TNF-α), thereby sensitizing them
to apoptosis. Kupffer cells also promote local accumulation of T cells and neutrophils, which, along with apoptotic hepatocytes, stimulate the activa-
tion of HSCs. Damage to hepatocytes by NASH or infection with hepatitis B or C viruses (HBV, HCV) promotes oxidative stress, further sensitizing
hepatocytes to apoptosis. Free fatty acids also increase the intracellular oxidative stress of hepatocytes. Bile acids inhibit activation of HSCs via a
farnesoid X receptor (FXR) pathway. aHSCs, Activated HSCs; EGFR, endothelial growth factor receptor; ERK-1, extracellular signal-regulated kinase-
1; mito., mitochondrial; NADPH, reduced nicotinamide adenine dinucleotide phosphate; qHSCs, quiescent HSCs; ROS, reactive oxygen species;
TNF-α, tumor necrosis factor-α.
acids as ligands for nuclear receptors, in particular the farnesoid lung, bone marrow, and pancreas, among others (Iwaisako
X receptor (FXR), which drives an entire cellular program that et al, 2014; Lua et al, 2014). The relative importance of each
can alter hepatocellular metabolism and bile secretion and fibrogenic cell type in liver fibrogenesis may depend on the
composition (Adorini et al, 2012). Remarkably, the therapeutic origin of the liver injury. Recent fate-tracing studies using
benefit of vertical sleeve gastrectomy has also been ascribed to genetically altered reporter mice implicate stellate cells as the
FXR signaling in an animal model, raising the possibility that dominant source of MFBs (Mederacke et al, 2013) in paren-
intestinal FXR alone may be sufficient to drive weight loss and chymal liver disease; however, some contribution from biliary
improved metabolic parameters in NASH (Ryan et al, 2014). portal fibroblasts is important in cholestatic liver disease
Oxidant stress, mediated by ROS, is a common mediator of (Perepelyuk et al, 2013; Wells, 2014).
injury in many liver diseases in which damaged hepatocytes HSC activation can be divided conceptually into two phases.
become apoptotic or necrotic, thereby releasing ROS (Nieto First there is initiation, with early changes in gene expression
et al, 2002) and NADPH oxidase, which both activate HSCs and phenotype, resulting from paracrine stimulation, primarily
(Canbay et al, 2004). Injured hepatocytes also release inflam- due to changes in surrounding ECM, as well as exposure to
matory cytokines and soluble factors that activate Kupffer cells lipid peroxides and products of damaged hepatocytes. Next
and stimulate the recruitment of activated T cells. This inflam- there is perpetuation, which results from the effects of these
matory milieu further stimulates the activation of resident stimuli on maintaining the activated phenotype and generating
HSCs. Bile acids also directly stimulate proliferation of HSCs fibrosis. Within the nucleus, a growing number of transcription
by activating the epidermal growth factor receptor (Svegliati- factors regulate HSC activation, including peroxisome prolif-
Baroni et al, 2005). In contrast to hepatocytes, HSCs are pro- erator–activated receptors (PPARs), retinoid receptors, liver X
tected from bile acid–induced apoptosis by excluding bile acids receptor, REV-ERBα, NF-κB, FXR, GATA4, vitamin D recep-
(Svegliati-Baroni et al, 2005). tor, JunD, Kruppel-like factor 6, and FOXF1 (Lee et al, 2015;
NAFLD is increasingly prevalent due to increased rates of Mann & Mann, 2009). A number of general and cell type–
childhood and adult obesity in the United States and Western specific membrane receptors and signaling pathways also
Europe (see Chapter 71). In fact, the percentage of liver trans- control HSC biology, including receptor tyrosine kinases, che-
plantations performed for this indication is rapidly rising, which mokine receptors, and integrins (Lee et al, 2015). In addition
is likely to accelerate further as more viral hepatitis patients are to these novel pathways, there has been substantial progress in
effectively cured with antivirals, combined with a growing prev- understanding how epigenetics contributes to stellate cell
alence of obesity (Charlton, 2013). NAFLD can progress to biology and fibrosis through both histone modifications and the
NASH, with consequent fibrosis and cirrhosis (Loomba & contribution of microRNAs (Lee et al, 2015; Mann, 2014).
Sanyal, 2013; Michelotti et al, 2013; Noureddin et al, 2013; As noted above, portal fibroblasts (Beaussier et al, 2007;
Singh et al, 2014). The pathogenesis is not fully understood. Ramadori & Saile, 2004), and bone marrow–derived MFBs
However, multiple convergent pathologies are increasingly (Russo et al, 2006) have also been identified as collagen-
implicated, including insulin resistance, oxidant stress, altered producing cells in the injured liver. Although earlier studies
adipokine balance, lipotoxicity, effects of the microbiome, and implicated epithelial-mesenchymal transition as a source of
enhanced inflammation, among others (De Minicis et al, 2014; fibrogenic cells (Rygiel et al, 2008), more recent findings
Tacke & Yoneyama, 2013; Wree et al, 2013). strongly refute its importance in liver (Chu et al, 2011).
Hepatic Stellate Cell Activation: Hepatic Myofibroblasts Functions of Hepatic Myofibroblasts

The HSC has emerged as a central regulator of the liver’s Hepatic MFBs have functions that are distinct from their qui-
fibrotic and repair responses (Fig. 7.3) (Hernandez-Gea & escent cells of origin. They are profibrogenic and promitotic,
Friedman, 2011). In a healthy liver, the HSC is a quiescent cell they have a chemotactic and vasoregulatory role, and they
type that contains cytoplasmic retinoid droplets, representing control the degradation of ECM. They also have important
the major storage site for vitamin A in the body, and expresses immune and phagocytic functions (Friedman, 2008b, Gao &
the markers desmin and glial fibrillary acidic protein Radaeva, 2013, Gao & Xu, 2014). The regulation of ECM
(Friedman, 2008b). During liver injury, HSCs undergo activa- accumulation and degradation by HSCs is reviewed in the next
tion in response to a range of inflammatory and injury signals section.
produced by damaged hepatocytes and biliary cells, by changes
in the composition of the ECM, by proangiogenic growth Fibrogenesis
factors such as vascular endothelial growth factor (VEGF) and The major profibrogenic signal in liver is the cytokine TGF-β1
angiopoietin, and by fibrogenic cytokines that include trans- (Bachem et al, 1989). TGF-β1 is secreted mainly by MFBs
forming growth factor-β (TGF-β1), connective tissue growth (Bissell et al, 1995) but also by platelets (Bachem et al, 1989)
factor (CTGF), angiotensin II, and leptin, among others (Lee and Kupffer cells (Bilzer et al, 2006). It functions by activating
et al, 2015; Lemoinne et al, 2013). the type II TGF-β receptor, which recruits the type I TGF-β
Activation of HSCs is accompanied by loss of retinoid drop- receptor. SMAD2 and SMAD3 then associate with the TGF-βI
lets and accumulation of α-smooth muscle actin, a myogenic receptor (Dooley et al, 2001), are phosphorylated and recruit
filament that confers increased cellular contractility. Activated SMAD4. This tri-heteromeric complex then translocates to the
HSCs are characteristically α–smooth muscle actin—and nucleus, where it activates profibrogenic transcription factors.
desmin-positive cells, but there is heterogeneity in the patterns TGF-β also activates the mitogen-activated protein kinase
of intracellular filament expression among HSCs. Highly acti- (MAPK) p38 pathway (Cao et al, 2002), which stimulates
vated subsets of HSCs are also called hepatic myofibroblasts additional SMAD-independent collagen type 1 synthesis and,
(MFB) (Friedman, 2008a), a cell type that is also characteristic in contrast to the SMAD-dependent collagen type 1 synthesis,
of wound healing in a range of tissues, including skin, kidney, also leads to a posttranscriptionally regulated stabilization of
Senescent/Inactivated/
Quiescent Activated
Apoptotic
Functions Features Features
• Matabolic homeostasis • Metabolic reprogramming • Increased NK-cell–mediated

Molecular Phenotype
cell death of senescent

• ECM homeostasis Liver Injury • Autophagy fuels activation
HSCs
• Cell injury amplification
• Vasoregulation • Infectious
• ‘‘Classic’’ activated changes: • Inactivated HSCs ‘‘primed’’
• Retinoid metabolism • Metabolic for reactivation
• Retinoid loss •↑ Chemotaxis
• Congestive • ↑ Fibrogenesis •↑
Contractility
• Autoimmune • ↑ ECM turnover • ↑ Inflammatory
• ↑ Proliferation signaling
• Alcohol
Senescent
• Toxic
• Drugs
Cellular Phenotype
Inactivated
Apoptotic
Organ Phenotype
Normal Fibrosis/Cirrhosis Regression
FIGURE 7.3. Functions, features and phenotypes of hepatic stellate cells in normal and diseased liver. Hepatic stellate cells may exist as
several different phenotypes with distinct molecular and cellular functions and features, each of which contributes significantly to liver homeostasis
and disease. Quiescent stellate cells are critical to the normal metabolic functioning of the liver. Liver injury provokes transdifferentiation of quiescent
stellate cells to their activated phenotype, leading to metabolic reprogramming, increased autophagy to fuel the metabolic demands, amplification
of parenchymal injury, and the development of “classic” phenotypic features of activated hepatic stellate cells/myofibroblasts. Through these changes,
activated stellate cells drive the fibrotic response to injury and the development of cirrhosis. As liver injury subsides, activated stellate cells can be
eliminated by one of three pathways: apoptosis, senescence, or reversion to an inactivated phenotype. Senescent stellate cells are more likely to be
cleared by natural killer (NK)-cell–mediated cell death, whereas inactivated stellate cells remain “primed” to respond to further liver injury. This reduc-
tion in the number of activated stellate cells contributes to the regression of fibrosis or cirrhosis and repair of the liver in most, but not all, patients.
The relative contribution of these three pathways of stellate cell clearance to fibrosis regression is not yet clear. ECM, Extracellular matrix; HSCs,
hepatic stellate cells. (From Lee YM, et al: Pathobiology of liver fibrosis—a translational success story, Gut 64:830–841, 2015.)
the collagen type 1 messenger RNA (mRNA) (Tsukada et al, Marra, 2010). All pathways downstream of the βPDGF-β
2005). Local activation of TGF-β1 at the cell surface by inte- receptor, the key receptor isoform in HSCs, promote prolifera-
grins has led to the prospect of antagonizing integrins as an tion. First, c-Jun N-terminal kinase is stimulated through
antifibrotic therapy (Henderson et al, 2013). In addition to MAPK (Schwabe et al, 2001); second, PDGF receptor stimu-
TGF-β1, CTGF (Lipson et al, 2012), and Hedgehog signaling lates the RAS/RAF complex, followed by mitogen-induced
have also been implicated as important fibrogenic mediators in extracellular kinase and extracellular signal-regulated kinase
liver injury and repair (Michelotti et al, 2013; Xie et al, 2013). engagement (Schwabe et al, 2001); third, activation of the
PI3K pathway, leading to AKT (Protein kinase B) activation
Proliferation and phosphorylation of the 70S6 kinase (Reif et al, 2003).
The predominant stimulus to MFB proliferation is the mitogen
platelet-derived growth factor (PDGF) (Borkham-Kamphorst Immunoregulation
et al, 2004), in addition to other mitogens, including epidermal The liver is a microenvironment of diminished immunogenicity,
growth factor, VEGF, and fibroblast growth factor (Mann & which is necessary to cope with the high exposure of antigens
from the portal vein (Crispe, 2003) (see Chapter 10). This
feature also accounts for the tolerance of liver transplantation Vasoregulation
across ABO barriers and may contribute to the chronic nature MFBs play an important role in the regulation of sinusoidal
of HBV or HCV, in which the virus persists despite the develop- blood flow and may contribute to portal hypertension that is
ment of an immune response. Upon entry of the antigen to the characteristic of advanced liver disease (see Chapters 76 and
sinusoid, classic antigen-presenting cells (Kupffer cells, den- 79,). The release of endothelin-1 (ET-1) can stimulate their
dritic cells) are first encountered. Subsequently, HSCs in the contraction through the endothelin type A (ETA) receptor
space of Disse may contact antigens. Indeed, HSCs display a (Khimji & Rockey, 2011), thereby promoting tissue contrac-
wide range of immunoregulatory functions and are an essential tion, increasing portal resistance, and generating portal hyper-
part of the local immunogenicity (Crispe, 2014; Gao & Radaeva, tension. On the other hand, MFBs and endothelial cells also
2013; Gao & Xu, 2014; Jiang et al, 2013; Kubes & Mehal 2012; secrete nitric oxide (NO), which is the physiologic antagonist
Watanabe et al, 2009). of ET-1.
Hepatic MFBs produce a range of proinflammatory and
antiinflammatory cytokines (see Chapter 11) and recruit Structural Features of Hepatic Fibrogenesis
lymphocytes through secretion of chemokines (monocyte che- In hepatic fibrosis, the total amount of collagen is increased up
moattractant protein-1, interleukin-8 [IL-8], C-C chemokine to sixfold, whereas the parenchymal mass (e.g., hepatocytes) is
21 [CCL21], regulated on activation, normal T-cell expressed progressively diminished. The composition of the ECM changes
and secreted, C-C chemokine receptor 5 [CCR5]) (Gao & Xu, with progression of disease (see Fig. 7.1). Collagen type IV in
2014; Marra & Tacke, 2014), thus amplifying the inflammatory the space of Disse is replaced by interstitial collagens, type I and
response. However, upon activation they exert a profound III. Additionally, the discontinuous basal membrane beneath
immunosuppressive activity by inducing T-cell apoptosis (Yu the sinusoidal endothelial cells is replaced by a continuous base-
et al, 2004). In the setting of liver transplantation, MFB- ment membrane, and sinusoidal fenestrations are reduced. This
induced T-cell apoptosis via programmed death ligand-1 (Yu decreased porosity (also known as “capillarization”), combined
et al, 2004) may enable local immunotolerance of the liver. In with perisinusoidal fibrosis, scar contraction, and formation of
liver fibrosis, MFBs may further regulate the contribution of intrahepatic shunts, contribute to increased hepatic venous
lymphocytes to the course of hepatic fibrosis by ingesting dis- pressure and portal hypertension. Fibrillar collagens that are
ease-associated lymphocytes (Muhanna et al, 2008). produced by MFBs also interact with MFBs via discoidin
The interaction between HSCs and immune cells is bidirec- domain receptors and integrins (Olaso et al, 2001), thereby
tional. T cells activate HSCs by interferon-γ (IFN-γ), which inhibiting apoptosis and increasing MFB proliferation.
upregulates both stimulatory (CD80, CD86, CD54) and inhib- With the maturation of the fibrotic scar, not only is the
itory (B7-H1) surface molecules, and enhances both inflamma- amount of collagen increased, but the scar also becomes
tory and suppressive cytokines. However, the inhibitory increasingly insoluble through chemical cross-linking by lysyl
molecules are thought to override the stimulatory counterparts, oxidase 2 (LOXL2), tissue transglutaminase, and a disintegrin
resulting in immunosuppression. Lymphocytes can also mediate and metalloproteinase with thrombospondin-type repeats
hepatic fibrosis by activating HSCs. CD8-positive T lympho- metalloproteinase with thrombosponin type I motif (ADAMTS2)
cytes are more fibrogenic toward stellate cells than CD4 T (Kesteloot et al, 2007). Indeed, stellate cells are an important
lymphocytes (Safadi et al, 2004). This may explain, in part, source of these cross-linking enzymes (Perepelyuk et al, 2013).
why patients coinfected with HIV and HCV have accelerated Cross-linking makes the fibrous septa progressively resistant to
fibrosis, as their CD4:CD8 cell ratios are reduced. Of the CD4- proteolysis by matrix metalloproteinases (MMPs). The long-
positive T lymphocytes, previously called T-helper cells, the standing clinical dogma that the slower the pace of injury, the
humoral immunity mediated by T-helper 2 cells (Th2) is pro- less reversible the scar, is supported by animal studies in which
fibrogenic in liver injury, whereas the cell-mediated immunity even advanced fibrosis of short duration is reversible. Thus the
by the Th1 cells via IFN-γ, TNF-α, and IL-2 is antifibrogenic reversibility of a scar may be limited primarily by the extent of
(Shi et al, 1997). collagen cross-linking. Clinically, increased septal thickness and
HSCs can also function as antigen-presenting cells (Winau smaller nodule size, both of which reflect more advanced stages
et al, 2007). They can interact with bacterial LPSs directly via of fibrosis, are significant predictors of worse clinical outcomes
TLR4, which amplifies their activation. TLR4 signaling leads (Nagula et al, 2006). Efforts to therapeutically increase the
to downregulation of a TGF-β pseudoreceptor, BMP (bone solubility of collagen have shown that an antibody to LOXL2
morphogenic protein) and activin membrane-bound inhibitor, has marked antifibrotic effects in animal models of fibrosis in
which thereby amplifies fibrogenic activity of MFBs (Seki & liver and other organs (Barry-Hamilton et al, 2010), which has
Brenner, 2008). Signaling through TLR4 may be elicited not led to its use in ongoing clinical trials in fibrosis and cancer
only by exogenous ligands, including LPS, but also by endog- (Nishioka et al, 2012).
enous ligands, including high-mobility group box 1 protein
(Pradere et al, 2010; Roh & Seki, 2013; Wang et al, 2013). The Regulation of Collagen Deposition and Degradation
discovery of endogenous ligands for TLR4 has been part of a The deposition and degradation of collagen are tightly regulated.
larger appreciation that many cells, including HSCs, possess MMPs are the key enzymes that degrade fibrillar collagens
intracellular machinery known as the inflammasome, which (collagen types I and III) and noncollagenous ECM substrates
transduces signals arising from cellular damage (Henao-Mejia (Iredale et al, 2013). The tissue inhibitor metalloproteinases
et al, 2012; Kubes & Mehal, 2012; Vandanmagsar et al, 2011; (TIMPs) are their major antagonists by inactivating proteases
Wree et al, 2014). The role of the inflammasome is especially and by inhibiting MFB apoptosis (Murphy et al, 2002).
pertinent to understanding the pathogenesis of inflammation Both decreased levels of interstitial collagenases and increased
and fibrosis in NAFLD and NASH. levels of MMP inhibitors in liver injury create an imbalance that
favors reduced degradation of fibrillar collagens in hepatic fibro- including two stages of cirrhosis (“incomplete”and “complete”
sis.The interstitial collagenases MMP-1, MMP-8, and MMP-13 cirrhosis) (Ishak et al, 1995). The METAVIR score (an acronym
in humans and MMP-13 in rodents unwind the triple-helical derived from a French Investigator group) is a simple, widely
collagen type I, which is the principal collagen in the fibrotic applied five-stage scoring system (METAVIR Cooperative
liver, so that each α-chain is presented to the active site of the Study Group, 1994; Poynard et al, 1997) that is the most com-
enzyme that cleaves the collagen (Iredale et al, 2013). Other monly used worldwide. It incorporates the fibrosis scores F0 to
MMPs (e.g., MMP-2) cannot unwind the triple-helical collagen F4, and the activity scores A0 to A3, which assess the amount
and thus cannot degrade intact collagen type I alone. In early of necroinflammation. Specifically, F0 = no fibrosis, F1 = fibro-
liver injury, MMP-2 degrades the low-density basement mem- sis without septa, F2 = few septa, F3 = numerous septa without
brane present in the subendothelial space (Zhou et al, 2004). cirrhosis, and F4 = cirrhosis; further, A0 = no necroinflamma-
Its replacement with fibril-forming matrix impairs hepatocyte tory activity, A1 = mild necroinflammatory activity, A2 =
differentiation and function. During progressive fibrosis, expres- moderate necroinflammatory activity, and A3 = severe necro-
sion of MMP-1 (humans) or MMP-13 (rodents) is decreased, inflammatory activity.
and MMP-2 expression increases (Milani et al, 1994; Preaux For NAFLD, a separate scoring system for grading inflam-
et al, 1999). In parallel, the expression of tissue inhibitor mation and staging fibrosis has been widely adopted (Brunt
metalloproteinases (TIMP)-1 and TIMP-2, which inhibit the et al, 2011; Kleiner et al, 2005; Sanyal et al, 2011), which
collagen-degrading MMPs, is increased (Murphy et al, 2002; captures key histologic features of disease progression that are
Ramachandran & Iredale, 2009). distinct from viral liver disease (see Chapter 71). Specifically,
Hepatic macrophages are increasingly recognized as the key NAFLD and NASH, as well as alcoholic liver disease, are pri-
cellular determinant of matrix degradation, with some contri- marily centrilolobular rather than the periportal distribution
bution by dendritic and other inflammatory cells (Jiao et al, typical of viral liver diseases. The NAFLD Kleiner stages are
2012; Mitchell et al, 2009; Tacke & Zimmermann, 2014). In stage 0 = no fibrosis, stage 1 = perisinusoidal or periportal
mouse models, macrophages augment fibrogenesis during pro- fibrosis (1a: mild, zone 3; 1b: moderate, zone 3; 1c: portal/
gression of liver fibrosis, whereas during resolution, they hasten periportal), stage 2 = periportal and perisinusoidal fibrosis,
matrix degradation through increased production of MMP-13 stage 3 = bridging fibrosis, and stage 4 = cirrhosis.
(Fallowfield et al, 2007). More important, there is substantial There are other scoring systems specifically developed for
heterogeneity of macrophages in liver injury and resolution to different etiologies of liver fibrosis. For example, Ludwig and
account for these divergent activities, with a subset known as coworkers proposed a four-stage system to describe fibrosis for
Ly6c-lo cells implicated in degrading matrix during fibrosis both primary biliary cirrhosis (Ludwig et al, 1978) and scleros-
regression (Pellicoro et al, 2014; Ramachandran & Iredale, ing cholangitis (Ludwig et al, 1981).
2012; Ramachandran et al, 2012). Despite these major For fibrosis staging, there is increasing reliance on absolute
advances in identifying the key fibrolytic cell in liver fibrosis quantification of collagen in liver biopsy samples as assessed by
regression, it is not clear which is the major interstitial collage- computerized morphometry, instead of using a discontinuous
nase in fibrosis regression, because MMP-1 is only expressed scoring system composed of discrete stages. Indeed, collagen
at low levels in liver. proportionate area assessment is far more predictive of clinical
outcomes, even in NASH, and therefore its use is rapidly
DIAGNOSIS AND CLINICAL MONITORING gaining popularity, especially as an end point in clinical trials
OF HEPATIC FIBROSIS (Calvaruso et al, 2009; Huang et al, 2014; Manousou et al,
2013).
Many patients with chronic liver disease may initially be seen There is a great need for reliable, quantitative noninvasive
with late-stage fibrosis, because earlier stages are often asymp- diagnostics for fibrosis, and recent studies indicate steady
tomatic. Thus clinicians must have a high index of suspicion progress. Cross-sectional imaging studies such as computed
for occult fibrosis, especially in patients with unexplained eleva- tomography (CT) and magnetic resonance imaging (MRI) can
tions of liver enzymes, splenic enlargement, stigmata of liver demonstrate features of advanced liver disease, such as nodu-
disease, and/or laboratory or imaging findings suggestive of larity and signs of portal hypertension (splenomegaly, enlarged
portal hypertension (see Chapter 79). caudate lobe, esophageal varices). Diffusion-weighted MRI
When chronic liver disease is suspected, the liver biopsy measures the apparent diffusion coefficient of water, a param-
remains the gold standard for diagnosing and staging of liver eter that depends upon tissue structure. It has compared favor-
fibrosis. However, it is an invasive procedure with risk of ably to other noninvasive measures for determining advanced
adverse events and, equally important, a high likelihood of fibrosis (Murphy et al, 2015). More recent advances in MR
sampling (Bedossa et al, 2003; Ratziu et al, 2005) and inter- methods include MR elastography (Loomba et al, 2014), MR
pathologist and intrapathologist variability (Bedossa et al, fat fraction, and other related technologies (Banerjee et al,
2003). At least one third of biopsies may differ by one fibrosis 2014; Noureddin et al, 2013b & 2013a; Tang et al, 2014).
stage between the right side and left side hepatic lobes in HCV There are also efforts to develop newer MR probes that will
(Regev et al, 2002) and NAFLD (Ratziu et al, 2005). Shorter enable quantification of total liver collagen or elastin content
biopsies are associated with an increase in reported diagnoses (Ehling et al, 2013; Fuchs et al, 2013).
of mild and moderate fibrosis at the cost of more severe fibrosis, Biochemical parameters associated with hepatocyte injury
representing an understaging of fibrosis (Bedossa et al, 2003). (aspartate aminotransferase [AST], alanine aminotransferase
There are several commonly used histologic staging systems [ALT]), cholestatic liver injury (bilirubin, alkaline phospha-
for fibrosis (see Chapter 76). The Histology Activity Index tase), impaired liver synthetic function (apolipoproteins, cho-
score reported by Knodell includes three stages (Knodell et al, lesterol, coagulation factors, α2-macroglobulin, hyaluronic
1981), whereas the Ishak score differentiates six stages, acid, albumin, globulins), or impaired hepatic clearance of
endogenous or exogenous substances from the circulation (bili- to liver biopsies that were staged according to the Kleiner score
rubin, bile acid, caffeine, lidocaine metabolites, bromsulpha- (Adams & Angulo, 2007; Angulo et al, 2007). If ELF was used
lein, methacetin, indocyanine green, cholate, or ammonia) to delineate any fibrosis (using thresholds with a sensitivity and
(Everson et al, 2012) can provide information on the presence specificity of 90%, respectively), the authors concluded that
and cause of the disease. These tests all assess impaired liver liver biopsy could have been avoided in 48% of the patients.
function in fibrosis and may prove to be more quantitative and
sensitive than tests of liver injury or morphology. Indeed, pul-
Cytokines and Chemokines Associated
monary function tests (e.g., spirometry) have been a mainstay with Hepatic Fibrosis
of clinical assessment in lung disease rather than tissue analysis Of the cytokines and chemokines that are associated with
of lung for decades. Continued progress in using functional hepatic fibrosis, TGF-β1 is the dominant stimulus to the pro-
rather than structural assessment of liver disease is expected. duction of ECM by HSCs. Hepatic mRNA levels of TGF-β1
are increased in chronic liver disease in association with
Biochemical Tests increases in mRNA levels of type I collagen (Houglum et al,
A more direct approach incorporates serum molecules associ- 1994). In HCV-related chronic liver disease, serum TGF-β
ated with fibrosis, including those involved in deposition or levels correlate with the Knodell fibrosis score, but not with
degradation of ECM, as well as specific fibrogenic cytokines clinical, biochemical, or virologic parameters (Nelson et al,
associated with fibrosis. A number of combination serum tests 1997). Interestingly, TGF-β levels also decreased in a group of
have been evaluated for predicting fibrosis stage and outcomes, HCV patients with histologic decreases in necroinflammation
with improving sensitivity and specificity. To date, no single in the absence of changes in fibrosis following IFN treatment
molecule, but rather combinations of different components, (Roulot et al, 1995). Thus TGF-β serum levels not only cor-
demonstrate the best sensitivity and negative predictive values relate with fibrosis scores but may also indicate necroinflam-
to exclude significant fibrosis. mation. PDGF is upregulated following liver injury, and the
The most studied combination serum tests are the AST-to- amount of PDGF correlates with the severity of fibrosis (Yoshida
platelet ratio index (Wai et al, 2003), the Fibrosis-4 index et al, 2014).
(Sterling et al, 2006), the Forns test (Forns et al, 2002), and
the proprietary FibroTest (FT; Biopredictiv, Paris) (Imbert- Proteomics and Glycomics
Bismut et al, 2001). Newer proprietary biomarker scores Proteomics and glycomics are promising new technologies suit-
include the HepaScore (Quest Diagnostics, Madison, NJ) able as noninvasive diagnostic methods. Patterns of proteins or
(Leroy et al, 2014) and the FibroMeter (Cales et al, 2005). All glycoproteins can be assessed by mass spectroscopy of serum
these biochemical tests range from sufficient to excellent in samples (Miller et al, 2014; Paulo et al, 2013; Rodriguez-
ruling out significant fibrosis (F3 to F4) when the proper cutoff Suarez et al, 2012). Profiles of serum protein N-glycans were
value is chosen, but they are less useful in distinguishing mild found to have similar AUROC values to the FT for the diag-
from moderate fibrosis. The sensitivities of the tests vary based nosis of compensated cirrhosis. The combination with the FT
on the etiology of the liver disease. Further difficulties of these was able to increase the sensitivity and specificity of the test
tests are the absence of an ideal gold standard, in view of the (Callewaert et al, 2004).
liver biopsy’s significant sampling variability and interlabora-
tory differences (Gressner et al, 2009). The lack of a true gold Stiffness Assessments
standard means that the utility of serum tests can never be fairly An alternative approach to fibrosis assessment is the measure-
evaluated when compared to biopsy (Mehta et al, 2009). ment of liver stiffness using a number of devices, the first of
Overall, these serum markers’ performances are approxi- which, Fibroscan (FS; Echosens), uses vibration-controlled
mately comparable (Castera et al, 2013). Although it is unlikely transient elastography. Although FS was the first to assess stiff-
that these alone will suffice in assessing short-term disease ness clinically, other technologies are now also available for this
progression in clinical trials of antifibrotic drugs or in manage- purpose, including acoustic radiation force impulse imaging
ment of chronic liver diseases, they do predict long-term out- (Lupsor et al, 2009; Rifai et al, 2011) and shear wave elastog-
comes and therefore may have some utility in the future (Mayo raphy (Ferraioli et al, 2012). As noted above, stiffness can also
et al, 2008; Ngo et al, 2006). be assessed by MR elastography, which has the advantage of
assessing the entire liver (Loomba et al, 2014). However, there
Serum Assays of Extracellular Matrix Molecules is far more experience and published data assessing FS than
Serum assays have been developed that measure circulating the other technologies.
molecules involved either in the deposition of ECM or its deg- For FS, the stiffer the liver tissue, the faster a wave propa-
radation, as well as cytokines involved in fibrogenesis. Among gates. Results are expressed in kilopascals (kPa) (Kettaneh
the better-validated of these tests is the enhanced liver fibrosis et al, 2007; Sandrin et al, 2003). With the FS, the virtual cyl-
(ELF) panel (Parkes et al, 2011). This test incorporates inder of tissue that is assessed is at least 200 times larger than
markers of ECM turnover (TIMP-1, hyaluronic acid, and a biopsy sample and therefore is far more representative of the
amino-terminal peptide of procollagen III [P3NP]). The ELF hepatic parenchyma. Thus the FS may provide a more accurate
test was better at identifying minimal, moderate or severe fibro- and reproducible picture of cirrhosis than liver biopsy. There
sis than a clinical-biochemical panel (age, body mass index, is also potential value of the FS in early stages of disease, where
presence of diabetes or impaired fasting glucose, AST : ALT fibrosis may be unevenly distributed and thus underestimated
ratio, platelets, and albumin), and the combination of both by liver biopsy. Furthermore, FS has very low interobserver
scores showed further improved area under the receiver operat- variability. However, its accuracy is limited in patients with
ing characteristic curve (AUROC) values in the subset of obesity, ascites, or acute hepatitis. FS has been evaluated exten-
patients with an altered clinical-biochemical panel, in reference sively (Alkhouri et al, 2013; Castera et al, 2013) and is licensed
in several European and Asian countries, and in the United clinical use, but dozens are currently in clincal trials. Thus
States and Canada. meaningful success seems imminent.
It is important to remember that stiffness can arise from
edema or inflammation, not only fibrosis, and thus interpreta- Reversibility of Fibrosis: “Point of No Return”
tion of the results in the proper clinical context is essential The concept of fibrosis as an irreversible and constantly pro-
because acute hepatitis can yield stiffness values comparable to gressing state is no longer considered accurate. Liver fibrosis of
cirrhosis. Also the ability to perform the test and the accuracy different etiologies is usually reversible by removing the
of data may be reduced in obese patients, although newer causative agent (Friedman & Bansal, 2006). For example, a
probes may overcome this problem. When the FS has been decrease in the viral load of HBV patients, clearance of HCV
combined with the FT in HCV patients, both tests agreed in with pegylated IFN and ribavirin or direct-acting antivirals
70% to 80% of subjects, with increasing concordance in higher (D’Ambrosio et al, 2012; Marcellin et al, 2013) (but not main-
stages of liver fibrosis. Compared with the liver biopsy, results tenance IFN monotherapy [Di Bisceglie et al, 2008]); cessation
were confirmed in 84% to 94% of cases, with a tendency of of ethanol intake; weight loss or bariatric surgery in patients
FS/FT to underestimate fibrosis (Castera et al, 2005). More with NASH (Lassailly et al, 2014; Tai et al, 2012); and decrease
recent studies are focused on assessing its accuracy in NAFLD in iron or copper, or immunosuppressive therapy in autoim-
and NASH (Alkhouri et al, 2013; Kumar et al, 2013). To assist mune diseases have been shown experimentally and clinically
in this effort, an additional technology has been added to FS, to limit fibrosis progression, and to even regress cirrhosis in
called controlled attenuation parameter, which can estimate some patients. In fact, removing the causative agent is still the
liver fat content (de Ledinghen et al, 2012). most effective antifibrotic therapy.
In patients with more advanced disease, assessment of vas- Although fibrosis, inflammation, and bile duct proliferation
cular changes can be highly predictive of outcomes. Specifi- decrease when the damaging stimulus is withdrawn, regenera-
cally, the measure of the pressure gradient across the liver, or tive nodules may become autonomous and grow progressively.
hepatic venous pressure gradient (HVPG), is highly predictive This raises a key question of whether a point of no return exists,
of clinical deterioration (Ripoll et al, 2007). Although invasive, wherein even complete clearance of the underlying disease will
this technique is increasingly incorporated into clincal trials, no longer yield an improvement in cirrhosis. Increased cross-
but efforts are underway to develop noninvasive imaging linking of the collagen fibrils during time makes the fibrous
approaches that can capture the same information as HVPG septa progressively resistant to proteolysis by metalloprotein-
without the need for hepatic catheterization. ases, and hypoxia stimulates secretion of proangiogenic factors
In aggregate, all these noninvasive approaches can accurately by HSCs, such as VEGF and angiopoietin-1, which induce
distinguish between patients with little or no fibrosis and those proliferation and motility (Nakamura et al, 2007). However,
with advanced disease, but these approaches are less reliable at antagonism of VEGF, in particular, may be deleterious, as it
discriminating intermediate stages of fibrosis. They show a high also plays a role in restoration of liver architecture after cessa-
level of variability due to interlaboratory differences, but on the tion of experimental liver injury (Yang et al, 2014). Overall, as
other hand, the patients at risk for false-positive results are well many as 70% of patients with cirrhosis will have regression of
defined. Although the combination of FT, ActiTest (Biopredic- fibrosis once the primary etiology is mitigated, but in those
tiv), and FS, or the ELF panel, in NAFLD patients seem to be whom disease progresses thereafter, there may be a positive
good noninvasive alternatives to liver biopsy, their value in feedback loop between angiogenesis and fibrogenesis, which
individual patient management during time needs to be estab- persists even when the primary etiology is cleared, leading to
lished. These tests may even be superior to liver biopsy at cor- sustained abnormalities of the intrahepatic vasculature.
rectly staging and grading fibrosis (Poynard et al, 2004).
Despite its limitations, to date no single test can match the Prevention of Hepatocyte Apoptosis in Liver Injury
overall information from liver biopsy histology (inflammation, Apoptosis of hepatocytes during liver injury is a proinflamma-
fibrosis, steatosis, architecture). The role of noninvasive alter- tory event associated with Kupffer cell and HSC activation
natives so far lies in improving the fibrosis staging and grading (Canbay et al, 2004). Thus agents have been developed to
made from liver biopsies and, second, in reducing the number minimize apoptotic death of hepatocytes in chronic liver injury
of liver biopsies by screening patients with abnormal liver tests, by inhibiting the caspases that contribute to the apoptotic
to identify patients with a higher probability of liver fibrosis who cascade. In the surgical setting, reduced hepatocyte damage is
need further evaluation or therapy. There is legitimate hope for pursued by decreasing the time of vascular occlusion (Pringle
prognostic value of noninvasive tests and that disease progres- maneuver) during resection or shortening the ischemia-
sion or response to therapy can be assessed by combinations of reperfusion time in transplantsurgery. Experimental approaches
tests that assess serum markers, liver function, vascular changes, to reduce ischemia-reperfusion injury, (e.g., intermittent clamp-
as well as histology. This, however, remains to be clarified by ing or drugs) merit continued development to preserve hepa-
longitudinal studies. tocyte integrity.
Caspase Inhibitors
THERAPEUTIC STRATEGIES
Apoptosis is a functional antagonist to mitosis. Together they
The elucidation of pathways of hepatic fibrogenesis has pro- regulate homeostatic cell turnover. The two main pathways of
vided a rational framework for developing antifibrotic therapies apoptosis are the extrinsic pathway, which is death ligand–
in chronic liver disease (Fig. 7.4). Methods to attack several death receptor–mediated and is dependent on caspase-8, and
points are under development, and it may ultimately be advan- the intrinsic pathway, which is regulated by BCL2-induced
tageous to combine more than one therapy for maximal effi- mitochondrial dysfunction and downstream activation of
cacy. To date, no antifibrotic therapy has been approved for caspase-9 and its effector caspases: 3, 6, and 7. Pan-caspase
1. Control or cure primary disease 2. Target receptor-ligand interactions
Adiponectin PPARα, δ, γ agonist

NASH Other* CB1R antagonist ET-1 antagonist
ACE-I or ARB Tyrosine kinase antagonists
FXR agonist Immunosuppression Ghrelin FXR agonist
Vitamin E UDCA
PPARγ agonist Remove iron or copper
Lipogenesis inhibitor Alcohol abstinence
Viral suppression SVR Quiescent Activated
HSC HSC
HBV Liver Injury HCV
*e.g., Wilson disease
Autoimmune liver disease
Hereditary hemochromatosis
Alcoholic liver disease
Normal Cirrhotic
liver liver
Liver Injury
Latent TGF-β TGF-β CTGF mAb

↓NF-κB TIMP antagonist
ACE-I LOXL2 mAb
CB1R antagonist
Block
↑Macrophage fibrolytic activity
activation
↑NK-cell activity
↑ Collagen ↑ Proliferation ↑ Matrix Degradation
↓ Matrix Degradation ↑ Apoptosis Prevent Cross-Linking
Stellate cell
3. Inhibit fibrogenesis 4. Promote resolution of fibrosis
FIGURE 7.4. Mechanisms by which antifibrotic therapies may lead to fibrosis regression. 1. Disease-specific therapies that control or cure
the underlying disease are still the most effective antifibrotic approach. 2. Targeting receptor-ligand interactions with either established or experimental
drugs to reduce hepatic stellate cell (HSC) activation will attenuate fibrosis development, with multiple potential strategies under development. 3.
Inhibition of the most potent of the profibrogenic pathways, for example, preventing activation of latent TGF-β, or blocking the activity of CTGF, are
among the more promising antifibrotic strategies. 4. Resolution of fibrosis can be promoted by enhancing the apoptosis of activated hepatic stellate
cells either with drugs or through the activity of either NK cells or fibrolytic macrophages, and by increasing degradation of extracellular matrix, or
by preventing its cross-linking with antagonists to LOXL2. ACE-I, angiotensin converting enzyme-I inhibitor; CB1R; cannabinoid receptor type 1;
CTGF, connective tissue growth factor; ET-1, endothelin 1; FXR, farnesoid X receptor; HBV, HCV, hepatitis B and C virus, respectively; LOXL2, lysyl
oxidase 2; mAb, monoclonal antibody; NASH, nonalcoholic steatohepatitis ; NF-κB, nuclear factor kappa B; NK, natural killer; PPAR, peroxisome
proliferator–activated receptor; SVR, sustained virologic response; TGF-β, transforming growth factor-β; TIMP, tissue inhibitor of metalloproteinase;
UDCA, ursodeoxycholic acid. (From Lee YM, et al: Pathobiology of liver fibrosis—a translational success story, Gut 64:830–841, 2015.)
inhibitors are being used to assess the therapeutic effect of the drug should be antiinflammatory. In the 2 weeks of treat-
apoptosis inhibition. In preclinical studies, the pan-caspase ment, AST and ALT levels were significantly reduced in the
inhibitor Z-VAD-FMK was able to reduce mortality in rats with subset of patients with HCV as their underlying disease, without
acute hepatic failure after major hepatic resection (Yoshida affecting the HCV mRNA levels. No adverse events occurred
et al, 2007). Other caspase inhibitors have reduced ischemia- (Pockros et al, 2007). More recent trials of this agent are
reperfusion injury in rodents, decreased fibrogenesis in mice ongoing in different liver diseases. Caspase inhibitors thus far
after bile duct ligation (Canbay et al, 2004), and also decreased are considered safe, as animals with genetic deletions of death
fibrogenesis in a methionine/choline-deficient (MCD) diet– receptors do not show increased spontaneous tumors. None-
induced NASH mouse model. In the latter, hepatic steatosis theless, the potential emergence of tumors following long-term
and fibrogenesis was improved, however, without improvement administration remains a lingering concern.
of liver injury (Witek et al, 2009).
There are currently clinical trials using a pan-caspase inhibi-
Inhibition of Hepatic Stellate Cells Activation
tor for patients with liver disease of different etiologies. This or Inactivation of Myofibroblasts
agent, Emricasan (formerly called IDN-6556), however, also Oxidant stress in the form of ROS released by injured hepato-
inhibits caspase-1, which has an inflammatory effect, and thus cytes through the action of NADPH represents a major
fibrogenic stimulus. Thus antioxidants including vitamin E Affecting the phosphorylation of CCAAT/enhancer-binding
(Sanyal et al, 2010), silymarin, CYP2E1 inhibitors (Nieto et al, protein (CEBP-B) may lead to caspase activation and enhanced
2002), phosphocholine, cysteamine (Dohil et al, 2011) or HSC apoptosis. CEBP-B is a transcription factor predomi-
S-adenosyl-L-methionine may benefit fibrosis, particularly in nantly expressed in adipose, hepatic, and immune tissue. When
patients with alcohol-induced liver disease and NASH, in which CEBP-B is phosphorylated by ribosomal S-6 kinase (RSK), this
oxidant stress plays an especially important role. In principle, leads to activation of caspase-8, which leads to MFB apoptosis
antioxidants should be efficacious in inflammatory liver disease. (Buck & Chojkier, 2007). Thus activating RSK or caspase-8
However, efforts to establish the activity of antioxidants are directly will lead to MFB apoptosis.
confounded by the uneven quality of commercially available There are also receptor-ligand–mediated pathways of MFB
products, especially as these compounds are typically available apoptosis that are potential therapeutic targets. For example,
over-the-counter and their potency is not monitored. activation of the cannabinoid 1 (CB1) receptor leads to
Given their widespread use in diabetes, PPARγ agonists have increased collagen deposition and protects MFB from apopto-
been tested in clinical trials both in NASH and HCV (Belfort sis, whereas the CB2 receptor is proapoptotic via induction of
et al, 2006), but their activity is modest and is associated with intracellular oxidative stress. Correspondingly, CB1 knockout
unwanted weight gain, especially in patients with NASH mice or CB1 antagonist–treated mice (Teixeira-Clerc et al,
(Sanyal et al, 2010). PPARγ nuclear receptors are expressed 2006), as well as CB2 receptor–stimulated mice, (Munoz-
in HSCs, and PPARγ overexpression can regress MFBs to Luque et al, 2008) display less fibrosis and more MFB apop-
their quiescent state experimentally (Hazra et al, 2004). With tosis than control mice after CCl4 or thioacetamide (TAA)
improved formulation, PPARγ ligands may overcome some of treatment, or bile duct ligation (BDL), and CB2 knockout mice
their unwanted adverse effects in clinical trials and merit further show increased CCl4-induced fibrosis (Julien et al, 2005).
evaluation. Whereas trials of a systemic CB1 receptor antagonist for obesity
WNT signaling has been implicated in pulmonary and renal and NASH were discontinued because of central nervous
fibrosis and has also been reported to promote hepatic fibrosis system (CNS) effects, a new generation of peripheral CB1
by enhancing HSC activation and survival. This suggests that antagonists that do not enter the CNS are under development
WNT antagonism may be a useful target in liver fibrosis and could have a major role as an antifibrotic strategy.
(Boulter et al, 2012; Ren et al, 2013; Zhu et al, 2012), but no NK cells directly interact with the retinoic acid early-1
trials have been conducted, in part because of the complexity (RAE-1) ligand expressed by early activated MFBs and can
of WNT signaling and its divergent activities. lead to apoptosis (Radaeva et al, 2006). This effect is lost,
however, in mature MFBs as these cells lose RAE-1 expression
Induction of Myofibroblast Apoptosis (Radaeva et al, 2007), thereby limiting the potential of thera-
Because the natural resolution of fibrosis leads to apoptosis and peutic NK-cell–mediated MFB death in advanced fibrosis (Gao
clearance of MFBs, approaches that exploit these native path- & Radaeva, 2013).
ways of resolution merit attention. In addition to apoptosis, The adipokines adiponectin and leptin are natural counter-
activated HSCs/MFBs can also revert to an inactivated state, regulators. Leptin is produced by MFBs that contribute to their
driven in part by PPARγ signaling (Kisseleva et al, 2012; activation (Ikejima et al, 2002; Marra, 2002) and has a profi-
Troeger et al, 2012). brogenic effect on MFBs in hepatic injury (Saxena et al, 2002).
One approach to driving MFB apoptosis is TIMP antagonism. In contrast, adiponectin promotes MFB apoptosis (Ding et al,
Because TIMP is antiapoptotic and blocks matrix proteases, 2005) and inhibits liver fibrogenesis in vitro and in vivo
reduced expression or neutralization favors clearance of MFBs (Kamada et al, 2003). Adiponectin may become a useful anti-
through increased apoptosis and enhanced breakdown of scar fibrotic agent, particularly in NASH.
(Iredale et al, 1998). The translation of this approach to humans
will be challenging, however, as human cells appear to have more
Blocking Myofibroblast–Extracellular
enhanced antiapoptotic activity through increased expression of Matrix Interactions
the antiapoptotic protein BCL-2 (Novo et al, 2006). The ECM and MFBs interact in a positive feedback mecha-
NF-κB is a nuclear transcription factor that inhibits apop- nism that could be amenable to therapeutic antagonism. MFBs
tosis of MFBs. Thus any compound that inhibits NF-κB merits interact with ECM via α/β-integrins (Zhou et al, 2004), thereby
evaluation, in particular through the antagonism of JunD decreasing apoptosis and increasing proliferation of MFBs.
(Smart et al, 2006). An example of NF-κB antagonism is bort- This leads to increased collagen type I deposition, which further
ezomib, a proteosomal inhibitor that prolongs the half-life of promotes survival of fibrogenic MFBs (Issa et al, 2003). Thus
inhibitor of kappa B-α (IκB-α), a naturally occurring cytosolic blocking the MFB-ECM interaction could lead to increased
inhibitor of NF-κB (Elsharkawy et al, 2005), preventing it from MFB apoptosis. This has been confirmed by α3β2-integrin
translocating to the nucleus and acting as a transcription factor. disruption with echistatin, neutralizing antibodies, or siRNA
Related compounds, including gliotoxin or sulfalazine, exert (Zhou et al, 2004). Similar studies exploring antagonism of
similar effects by regulating transcription of NF-κB. For integrin α5β6 are ongoing. Related to this has been the pro-
example, treatment of rodents with the NF-κB inhibitor glio- posal to block integrins as a means of attenuating TGF-β acti-
toxin, which with the help of targeting technologies specifically vation (Henderson et al, 2013); such efforts are in preclinical
targets MFBs, selectively induced MFB apoptosis, and acceler- development but have great promise for human studies.
ated regression of fibrosis in experimental liver injury (Douglass
et al, 2008). Further, an NF-κB decoy increased TNF-α– Antagonizing Compounds That Mediate Inflammation
induced MFB apoptosis and decreased carbon tetrachloride As inflammation precedes and stimulates liver fibrosis, the
(CCl4)-induced fibrosis in comparison with the control group use of antiinflammatory drugs has been proposed. A number
(Son et al, 2007). of agents have antiinflammatory activity. For example,
corticosteroids have been used for decades to treat autoimmune chronic hepatic inflammation. The concern regarding use of
hepatitis. Pentoxifylline may exert its antifibrotic activity by TGF-β blocking agents has further increased due to the first
downregulating TGF-β1 and CTGF signaling (Raetsch et al, clinical study using CAT-192, a recombinant human antibody
2002). However, pentoxifylline can upregulate TIMP-1, that neutralizes TGF-β in patients with cutaneous systemic
thereby reducing its antifibrotic effect. It also inhibits NF-κB sclerosis, in which a significant increase in morbidity and mor-
in Kupffer cells, thereby reducing TNF-α production, the tality was shown, with no evidence of a therapeutic effect
impact of which is uncertain. Recent clinical studies of pentox- (Denton et al, 2007; Yingling et al, 2004). Based on growing
fylline in alcoholic liver disease have shown negative results, concerns about systemic TGF-β neutralization, current efforts
and thus there is waning enthusiasm for this agent (Park et al, are directed at local inhibition of TGF-β at the cell surface,
2014). primarily by blocking integrins that participate in cell surface
A more rational antiinflammatory strategy may be the antag- TGF-β activation, as described above.
onism of chemokines because they are increasingly implicated To antagonize PDGF, administration of a PDGF kinase
in human liver disease (Marra and Tacke, 2014). Small-mole- inhibitor with an HSC-selective carrier (mannose-6-phosphate
cule antagonists of chemokine receptors are well tolerated and modified human serum albumin) significantly reduced BDL-
currently in clinical trials based on promising studies in animals induced fibrosis (Gonzalo et al, 2007). Imatinib mesylate
and humans (Ochoa-Callejero et al, 2013). (Gleevec), a clinically used PDGF receptor tyrosine kinase
The renin-angiotensin system may also amplify inflamma- inhibitor, also attenuates proliferation and migration and fibro-
tion and has assumed major importance in the understanding sis in animal models (Gordon & Spiera, 2011; Kim et al, 2012;
of hepatic fibrosis. Angiotensin II is a vasoconstrictive peptide Kuo et al, 2012; Yoshiji et al, 2005), and newer drugs, such as
that is expressed by activated HSC in chronically injured livers nilotinib, show promise as well (Liu et al, 2011). Sorafenib, a
(Bataller et al, 2003b; Paizis et al, 2002). It induces hepatic multikinase inhibitor approved for treatment of liver cancer,
inflammation and stimulates fibrogenic actions of HSCs, also shows antifibrotic activity in animal models (Hong et al,
including cell proliferation, cell migration, secretion of proin- 2013), attesting to the contribution of receptor tyrosine kinases
flammatory cytokines, and collagen synthesis (Bataller et al, such as PGDF-R in fibrosis as well as cancer. However,
2000; Bataller et al, 2003a & 2003c). Inhibitors of this system sorafenib has adverse effects (e.g., rash, diarrhea, hand-foot
have been in clinical use for antihypertensive therapy for a syndrome), which may be acceptable to patients with cancer
substantial time, which makes their use in humans attractive. but will not be acceptable to asymptomatic patients with fibrotic
Preliminary studies in patients with chronic HCV and NASH liver disease. Thus better-tolerated kinase inhibitors are an
suggest a positive effect on fibrosis progression by administering appealing class of antifibrotic compounds.
blocking agents (Colmenero et al, 2009; Oakley et al, 2009; Apart from blocking HSC-stimulating factors, activating
Yokohama et al, 2004). HSC inhibitory factors is yet another possibility. Hepatocyte
Ursodeoxycholic acid (UDCA) has a beneficial effect on growth factor (HGF) inhibits HSC activation, decreases TGF-
fibrosis in primary biliary cirrhosis. Similarly, a nitric oxide– β, and increases HSC apoptosis (Kim et al, 2005). However,
releasing derivative of UDCA reduces inflammation, fibrosis, potential procarcinogenic effects are probable and would limit
and portal pressure in an animal model (Fiorucci et al, 2003). its therapeutic use. Clinical trials with HGF deletion variants
Interestingly, UDCA also activates the pregnane X receptor, and HGF mimetics are underway.
which has antifibrotic properties (Beuers et al, 2009). In rodents, the blockade of the ETA receptor, which leads
More recently, ligands for FXR, another nuclear receptor, to vasocontriction or scar-contraction upon binding of ET-1,
have been developed, which are also antifibrotic in animal and the administration of vasodilators (prostaglandin E2 and
models and in a Phase 2 clinical trial in NASH (Neuschwander- NO donors) have antifibrotic qualities (Feng et al, 2009;
Tetri et al, 2014; Zhang et al, 2009). Larger Phase III trials of Rockey, 2013). At one point, ET receptor blockade was highly
this agent are planned. attractive; however, clinical trials of these drugs for other indi-
cations demonstrated unacceptable liver toxicity (Kenna et al,
Selectively Antagonizing Pathways of Hepatic Stellate 2015), and thus their development was halted until safety con-
Cell Activation cerns can be more thoroughly addressed.
Fibrogenic, proliferative, proangiogenic, vasoconstrictive, and Bone marrow–derived mesenchymal stem cells have an
proinflammatory mediators work synergistically toward hepatic antifibrotic effect—on the one hand, increasing HSC apopto-
fibrogenesis in the setting of chronic liver injury. Thus efforts sis through secretion of HGF, and, on the other hand,
are underway to antagonize the specific mediators driving these by decreasing proliferation of HSCs through release of
pathways. IL-10 and TNF-α upon stimulation by IL-6 from hepatic
Multiple approaches have been directed toward blocking the MFBs (Parekkadan et  al, 2007a; & 2007b). This makes stem
profibrogenic TGF-β signaling pathway. The effects of soluble cell therapy an interesting direction for fibrosis therapies.
TGF-β receptor type II (George et al, 1999), TGF-β blocking However, bone marrow–derived mesenchymal stem cells also
antibodies, TGF-β antisense oligonucleotides, or agents that contribute to scar-forming MFBs in various organs, including
interfere with TGF-β downstream signal transduction have the liver (Russo et  al, 2006). A key problem with autologous
been assessed experimentally (see Fig. 7.4). Systemically block- and stem cell and bone marrow therapies to date has been
ing the TGF pathway has theoretical limitations, however, their incomplete characterization, such that the exact compo-
because apart from stimulating wound healing and fibrosis, sition of cells being transplanted is not standardized (Moore
TGF-β is also a central inhibitor of uncontrolled inflammation et  al, 2014). Therefore a key goal of future studies is to
and essential in inducing epithelial differentiation and in trig- understand and fully control the cellular composition of such
gering apoptosis. This raises safety concerns for the general and therapies to ensure that their effects are predictable and
long-term use of TGF-β inhibition, especially in patients with reproducible.
Although no antifibrotic drug is approved for clinical use in

Enhancing Extracellular Matrix Degradation patients with liver fibrosis, the broad and rational development
A number of rational approaches are under development to of a range of promising compounds means that success is likely
increase ECM degradation rather than solely block its produc- in the near future. The challenge ahead is to find further target-
tion. The most advanced are efforts to block collagen crossing points and to show efficiency of these new drugs in vivo
linking by using an antibody to LOXL2, based on promising and ultimately in clinical trials. Thus long follow-up studies are
animal data (Barry-Hamilton et al, 2010). As mentioned earlier, required. Equally challenging is the need to define clear and
the two main families regulating ECM turnover are the MMPs robust end points for clinical trials to ensure that a therapeutic
that degrade the collagen and noncollagenous ECM substrates, benefit is apparent. Thus finding good noninvasive alternatives
and the TIMPs that inhibit MMP activities and also have an for diagnosis, follow-up and assessing prognosis of liver fibrosis
antiapoptotic effect on MFBs (Murphy et al, 2002). Another are important first steps for developing and evaluating better
approach is to block TIMP-1 with a monoclonal antibody; in treatment options.
an animal model this reverses CCl4-induced fibrosis (Parsons
et al, 2004). References are available at expertconsult.com.
Chapter 7 Liver fibrogenesis: mechanisms and clinical relevance 122.e1
REFERENCES Canbay A, et al: The caspase inhibitor IDN-6556 attenuates hepatic

injury and fibrosis in the bile duct ligated mouse, J Pharmacol Exp
Adams LA, Angulo P: Role of liver biopsy and serum markers of liver Ther 308(3):1191–1196, 2004.
fibrosis in non-alcoholic fatty liver disease, Clin Liver Dis 11(1):25– Canbay A, et al: Apoptosis: the nexus of liver injury and fibrosis, Hepa-
35, viii, 2007. tology 39(2):273–278, 2004.
Adorini L, et al: Farnesoid X receptor targeting to treat nonalcoholic Cao Q, et al: DLPC decreases TGF-beta1-induced collagen mRNA by
steatohepatitis, Drug Discov Today 17(17–18):988–997, 2012. inhibiting p38 MAPK in hepatic stellate cells, Am J Physiol Gastro-
Alkhouri N, et al: Combined paediatric NAFLD fibrosis index and intest Liver Physiol 283(5):G1051–G1061, 2002.
transient elastography to predict clinically significant fibrosis in chil- Castera L, et al: Prospective comparison of transient elastography,
dren with fatty liver disease, Liver Int 33(1):79–85, 2013. Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in
Angulo P, et al: The NAFLD fibrosis score: a noninvasive system chronic hepatitis C, Gastroenterology 128(2):343–350, 2005.
that identifies liver fibrosis in patients with NAFLD, Hepatology Castera L, et al: Noninvasive evaluation of NAFLD, Nat Rev Gastro-
45(4):846–854, 2007. enterol Hepatol 10(11):666–675, 2013.
Bachem MG, et al: The role of thrombocytes in liver fibrogenesis: Charlton M: Evolving aspects of liver transplantation for nonalcoholic
effects of platelet lysate and thrombocyte-derived growth factors on steatohepatitis, Curr Opin Organ Transplant 18(3):251–258, 2013.
the mitogenic activity and glycosaminoglycan synthesis of cultured Chu AS, et al: Lineage tracing demonstrates no evidence of cholangio-
rat liver fat storing cells, J Clin Chem Clin Biochem 27(9):555–565, cyte epithelial-to-mesenchymal transition in murine models of
1989. hepatic fibrosis, Hepatology 53(5):1685–1695, 2011.
Bachem MG, et al: Transforming growth factors (TGF alpha and Colmenero J, et al: Effects of losartan on hepatic expression of non-
TGF beta 1) stimulate chondroitin sulfate and hyaluronate synthesis phagocytic NADPH oxidase and fibrogenic genes in patients with
in cultured rat liver fat storing cells, FEBS Lett 257(1):134–137, chronic hepatitis C, Am J Physiol Gastrointest Liver Physiol 297(4):
1989. G726–G734, 2009.
Banerjee R, et al: Multiparametric magnetic resonance for the non- Crispe IN: Hepatic T cells and liver tolerance, Nat Rev Immunol
invasive diagnosis of liver disease, J Hepatol 60(1):69–77, 2014. 3(1):51–62, 2003.
Barry-Hamilton V, et al: Allosteric inhibition of lysyl oxidase-like-2 Crispe IN: Immune tolerance in liver disease, Hepatology 60(6):2109–
impedes the development of a pathologic microenvironment, Nat 2117, 2014.
Med 16(9):1009–1017, 2010. D’Ambrosio R, et al: A morphometric and immunohistochemical study
Bataller R, et al: Prolonged infusion of angiotensin II into normal rats to assess the benefit of a sustained virological response in hepatitis
induces stellate cell activation and proinflammatory events in liver, C virus patients with cirrhosis, Hepatology 56(2):532–543, 2012.
Am J Physiol Gastrointest Liver Physiol 285(3):G642–G651, 2003a. de Ledinghen V, et al: Non-invasive diagnosis of liver steatosis using
Bataller R, et al: Angiotensin II induces contraction and proliferation controlled attenuation parameter (CAP) and transient elastography,
of human hepatic stellate cells, Gastroenterology 118(6):1149–1156, Liver Int 32(6):911–918, 2012.
2000. De Minicis S, et al: Dysbiosis contributes to fibrogenesis in the course
Bataller R, et al: Activated human hepatic stellate cells express the of chronic liver injury in mice, Hepatology 59(5):1738–1749, 2014.
renin-angiotensin system and synthesize angiotensin II, Gastroenter- Denton CP, et al: Recombinant human anti-transforming growth factor
ology 125(1):117–125, 2003b. beta1 antibody therapy in systemic sclerosis: a multicenter, random-
Bataller R, et al: NADPH oxidase signal transduces angiotensin II in ized, placebo-controlled phase I/II trial of CAT-192, Arthritis Rheum
hepatic stellate cells and is critical in hepatic fibrosis, J Clin Invest 56(1):323–333, 2007.
112(9):1383–1394, 2003c. Di Bisceglie AM, et al: Prolonged therapy of advanced chronic hepatitis
Beaussier M, et al: Prominent contribution of portal mesenchymal cells c with low-dose peginterferon, N Engl J Med 359(23):2429–2441,
to liver fibrosis in ischemic and obstructive cholestatic injuries, Lab 2008.
Invest 87(3):292–303, 2007. Ding X, et al: The roles of leptin and adiponectin: a novel paradigm
Bedossa P, et al: Sampling variability of liver fibrosis in chronic hepa- in adipocytokine regulation of liver fibrosis and stellate cell biology,
titis C, Hepatology 38(6):1449–1457, 2003. Am J Pathol 166(6):1655–1669, 2005.
Belfort R, et al: A placebo-controlled trial of pioglitazone in subjects Dohil R, et al: Enteric-coated cysteamine for the treatment of paediat-
with nonalcoholic steatohepatitis, N Engl J Med 355(22):2297–2307, ric non-alcoholic fatty liver disease, Aliment Pharmacol Ther 33(9):
2006. 1036–1044, 2011.
Beuers U, et al: Medical treatment of primary sclerosing cholangitis: a Dooley S, et al: Expression of Smads during in vitro transdifferentiation
role for novel bile acids and other (post-)transcriptional modulators?, of hepatic stellate cells to myofibroblasts, Biochem Biophys Res
Clin Rev Allergy Immunol 36(1):52–61, 2009. Commun 283(3):554–562, 2001.
Bilzer M, et al: Role of Kupffer cells in host defense and liver disease, Douglass A, et al: Antibody-targeted myofibroblast apoptosis reduces
Liver Int 26(10):1175–1186, 2006. fibrosis during sustained liver injury, J Hepatol 49(1):88–98, 2008.
Bissell DM, et al: Cell-specific expression of transforming growth fac- Dumas ME, et al: Metabolic phenotyping and systems biology
tor-beta in rat liver. Evidence for autocrine regulation of hepatocyte approaches to understanding metabolic syndrome and fatty liver
proliferation, J Clin Invest 96(1):447–455, 1995. disease, Gastroenterology 146(1):46–62, 2014.
Borkham-Kamphorst E, et al: Antisense strategy against PDGF Ehling J, et al: Elastin-based molecular MRI of liver fibrosis, Hepatology
B-chain proves effective in preventing experimental liver fibrogene- 58(4):1517–1518, 2013.
sis, Biochem Biophys Res Commun 321(2):413–423, 2004. Elsharkawy AM, et al: The role and regulation of hepatic stellate cell
Boulter L, et al: Macrophage-derived Wnt opposes Notch signaling to apoptosis in reversal of liver fibrosis, Apoptosis 10(5):927–939, 2005.
specify hepatic progenitor cell fate in chronic liver disease, Nat Med Everson GT, et al: Quantitative liver function tests improve the predic-
18(4):572–579, 2012. tion of clinical outcomes in chronic hepatitis C: results from the
Brunt EM, et al: Nonalcoholic fatty liver disease (NAFLD) activity Hepatitis C Antiviral Long-term Treatment Against Cirrhosis Trial,
score and the histopathologic diagnosis in NAFLD: distinct clinico- Hepatology 55(4):1019–1029, 2012.
pathologic meanings, Hepatology 53(3):810–820, 2011. Fallowfield JA, et al: Scar-associated macrophages are a major source
Buck M, Chojkier M: C/EBPbeta phosphorylation rescues macrophage of hepatic matrix metalloproteinase-13 and facilitate the resolution
dysfunction and apoptosis induced by anthrax lethal toxin, Am J of murine hepatic fibrosis, J Immunol 178(8):5288–5295, 2007.
Physiol Cell Physiol 293(6):C1788–C1796, 2007. Feng HQ, et al: Endothelin antagonism in portal hypertensive mice:
Cales P, et al: A novel panel of blood markers to assess the degree of implications for endothelin receptor-specific signaling in liver disease,
liver fibrosis, Hepatology 42(6):1373–1381, 2005. Am J Physiol Gastrointest Liver Physiol 297(1):G27–G33, 2009.
Callewaert N, et al: Noninvasive diagnosis of liver cirrhosis using DNA Ferraioli G, et al: Accuracy of real-time shear wave elastography for
sequencer-based total serum protein glycomics, Nat Med 10(4):429– assessing liver fibrosis in chronic hepatitis C: a pilot study, Hepatology
434, 2004. 56(6):2125–2133, 2012.
Calvaruso V, et al: Computer-assisted image analysis of liver collagen: Fierer DS, et al: Liver fibrosis during an outbreak of acute hepatitis C
relationship to Ishak scoring and hepatic venous pressure gradient, virus infection in HIV-infected men: a prospective cohort study,
Hepatology 49(4):1236–1244, 2009. J Infect Dis 198(5):683–686, 2008.
Fiorucci S, et al: Nitric oxide and portal hypertension: a nitric Iredale JP, et al: Extracellular matrix degradation in liver fibrosis: bio-
oxide-releasing derivative of ursodeoxycholic actd that selectively chemistry and regulation, Biochim Biophys Acta 1832(7):876–883,
releases nitric oxide in the liver, Dig Liver Dis 35(Suppl):61–69, 2013.
2003. Ishak K, et al: Histological grading and staging of chronic hepatitis,
Forns X, et al: Identification of chronic hepatitis C patients without J Hepatol 22(6):696–699, 1995.
hepatic fibrosis by a simple predictive model, Hepatology 36(4 Pt Issa R, et al: Mutation in collagen-1 that confers resistance to the action
1):986–992, 2002. of collagenase results in failure of recovery from CCl4-induced liver
Fouts DE, et al: Bacterial translocation and changes in the intestinal fibrosis, persistence of activated hepatic stellate cells, and diminished
microbiome in mouse models of liver disease, J Hepatol 56(6):1283– hepatocyte regeneration, FASEB J 17(1):47–49, 2003.
1292, 2012. Iwaisako K, et al: Origin of myofibroblasts in the fibrotic liver in mice,
Friedman SL: Mechanisms of disease: mechanisms of hepatic fibrosis Proc Natl Acad Sci U S A 111(32):E3297–E3305, 2014.
and therapeutic implications, Nat Clin Pract Gastroenterol Hepatol Jiang Z, et al: Hepatic stellate cells promote immunotolerance follow-
1(2):98–105, 2004. ing orthotopic liver transplantation in rats via induction of T cell
Friedman SL: Hepatic stellate cells—protean, multifunctional, and apoptosis and regulation of Th2/Th3-like cell cytokine production,
enigmatic cells of the liver, Physiol Rev 88(1):125–172, 2008a. Exp Ther Med 5(1):165–169, 2013.
Friedman SL: Mechanisms of hepatic fibrogenesis, Gastroenterology Jiao J, et al: Dendritic cell regulation of carbon tetrachloride-induced
134(6):1655–1669, 2008b. murine liver fibrosis regression, Hepatology 55(1):244–255, 2012.
Friedman SL, Bansal MB: Reversal of hepatic fibrosis—fact or fantasy?, Julien B, et al: Antifibrogenic role of the cannabinoid receptor CB2 in
Hepatology 43(2 Suppl 1):S82–S88, 2006. the liver, Gastroenterology 128(3):742–755, 2005.
Friedman SL, et al: Hepatic lipocytes: the principal collagen-producing Kamada Y, et al: Enhanced carbon tetrachloride-induced liver fibrosis
cells of normal rat liver, Proc Natl Acad Sci U S A 82(24):8681–8685, in mice lacking adiponectin, Gastroenterology 125(6):1796–1807,
1985. 2003.
Fuchs BC, et al: Molecular MRI of collagen to diagnose and stage liver Kenna JG, et al: Multiple compound-related adverse properties con-
fibrosis, J Hepatol 59(5):992–998, 2013. tribute to liver injury caused by endothelin receptor antagonists,
Gao B, Radaeva S: Natural killer and natural killer T cells in liver J Pharmacol Exp Ther 352(2):281–290, 2015.
fibrosis, Biochim Biophys Acta 1832(7):1061–1069, 2013. Kesteloot F, et al: ADAM metallopeptidase with thrombospondin type
Gao B, Xu M: Chemokines and alcoholic hepatitis: are chemokines 1 motif 2 inactivation reduces the extent and stability of carbon
good therapeutic targets?, Gut 63(11):1683–1684, 2014. tetrachloride-induced hepatic fibrosis in mice, Hepatology 46(5):
George J, et al: In vivo inhibition of rat stellate cell activation by soluble 1620–1631, 2007.
transforming growth factor beta type II receptor: a potential new Kettaneh A, et al: Features associated with success rate and perfor-
therapy for hepatic fibrosis, Proc Natl Acad Sci U S A 96(22):12719– mance of FibroScan measurements for the diagnosis of cirrhosis
12724, 1999. in HCV patients: a prospective study of 935 patients, J Hepatol
Gonzalo T, et al: Local inhibition of liver fibrosis by specific delivery 46(4):628–634, 2007.
of a PDGF kinase inhibitor to hepatic stellate cells, J Pharmacol Exp Khimji AK, Rockey DC: Endothelin and hepatic wound healing, Phar-
Ther 321:856–865, 2007. macol Res 63(6):512–518, 2011.
Gordon J, Spiera R: Imatinib and the treatment of fibrosis: recent trials Kim Y, et al: Anti-fibrotic activity and enhanced interleukin-6 produc-
and tribulations, Curr Rheumatol Rep 13(1):51–58, 2011. tion by hepatic stellate cells in response to imatinib mesylate, Liver
Gressner AM, et al: Non-invasive biomarkers for monitoring the Int 32(6):1008–1017, 2012.
fibrogenic process in liver: a short survey, World J Gastroenterol Kim WH, et al: Growth inhibition and apoptosis in liver myofibroblasts
15(20):2433–2440, 2009. promoted by hepatocyte growth factor leads to resolution from liver
Guo J, et al: Functional linkage of cirrhosis-predictive single nucleotide cirrhosis, Am J Pathol 166(4):1017–1028, 2005.
polymorphisms of Toll-like receptor 4 to hepatic stellate cell Kisseleva T, et al: Myofibroblasts revert to an inactive phenotype
responses, Hepatology 49(3):960–968, 2009. during regression of liver fibrosis, Proc Natl Acad Sci U S A 109(24):
Hazra S, et al: PPAR gamma and hepatic stellate cells, Comp Hepatol 9448–9453, 2012.
3(Suppl 1):S7, 2004. Kleiner DE, et al: Design and validation of a histological scoring system
Henao-Mejia J, et al: Inflammasome-mediated dysbiosis regulates for nonalcoholic fatty liver disease, Hepatology 41(6):1313–1321,
progression of NAFLD and obesity, Nature 482(7384):179–185, 2005.
2012. Knodell RG, et al: Formulation and application of a numerical scoring
Henderson NC, et al: Targeting of alphav integrin identifies a core system for assessing histological activity in asymptomatic chronic
molecular pathway that regulates fibrosis in several organs, Nat Med active hepatitis, Hepatology 1(5):431–435, 1981.
19(12):1617–1624, 2013. Kubes P, Mehal WZ: Sterile inflammation in the liver, Gastroenterology
Hernandez-Gea V, Friedman SL: Pathogenesis of liver fibrosis, Annu 143(5):1158–1172, 2012.
Rev Pathol 6:425–456, 2011. Kumar R, et al: Liver stiffness measurements in patients with different
Higuchi H, Gores GJ: Mechanisms of liver injury: an overview, Curr stages of nonalcoholic fatty liver disease: diagnostic performance and
Mol Med 3(6):483–490, 2003. clinicopathological correlation, Dig Dis Sci 58(1):265–274, 2013.
Hong F, et al: Antifibrotic activity of sorafenib in experimental hepatic Kuo WL, et al: Imatinib mesylate improves liver regeneration and
fibrosis: refinement of inhibitory targets, dosing, and window of attenuates liver fibrogenesis in CCL4-treated mice, J Gastrointest
efficacy in vivo, Dig Dis Sci 58(1):257–264, 2013. Surg 16(2):361–369, 2012.
Houglum K, et al: TGF-beta and collagen-alpha 1 (I) gene expression Lassailly G, et al: Effects of bariatric surgery on severe liver injury in
are increased in hepatic acinar zone 1 of rats with iron overload, Am morbid obese patients with proven NASH: a prospective study,
J Physiol 267(5 Pt 1):G908–G913, 1994. Hepatology 60(Suppl S1):Abstract 213, 2014.
Huang Y, et al: Image analysis of liver biopsy samples measures fibrosis Lee YM, et al: Pathobiology of liver fibrosis—a translational success
and predicts clinical outcome, J Hepatol 61(1):22–27, 2014. story, Gut 64(5):830–841, 2015.
Huang H, et al: A 7 gene signature identifies the risk of developing Lemoinne S, et al: Origins and functions of liver myofibroblasts,
cirrhosis in patients with chronic hepatitis C, Hepatology 46(2):297– Biochim Biophys Acta 1832(7):948–954, 2013.
306, 2007. Leroy V, et al: Prospective evaluation of FibroTest(R), FibroMeter(R),
Ikejima K, et al: Leptin receptor-mediated signaling regulates hepatic and HepaScore(R) for staging liver fibrosis in chronic hepatitis B:
fibrogenesis and remodeling of extracellular matrix in the rat, Gas- comparison with hepatitis C, J Hepatol 61(1):28–34, 2014.
troenterology 122(5):1399–1410, 2002. Lipson KE, et al: CTGF is a central mediator of tissue remodeling and
Imbert-Bismut F, et al: Biochemical markers of liver fibrosis in patients fibrosis and its inhibition can reverse the process of fibrosis, Fibro-
with hepatitis C virus infection: a prospective study, Lancet genesis Tissue Repair 5(Suppl 1):S24, 2012.
357(9262):1069–1075, 2001. Liu Y, et al: Inhibition of PDGF, TGF-beta, and Abl signaling and
Iredale JP, et al: Mechanisms of spontaneous resolution of rat liver reduction of liver fibrosis by the small molecule Bcr-Abl tyrosine
fibrosis. Hepatic stellate cell apoptosis and reduced hepatic expres- kinase antagonist Nilotinib, J Hepatol 55(3):612–625, 2011.
sion of metalloproteinase inhibitors, J Clin Invest 102(3):538–549, Loomba R, Sanyal AJ: The global NAFLD epidemic, Nat Rev Gastro-
1998. enterol Hepatol 10(11):686–690, 2013.
Loomba R, et al: Magnetic resonance elastography predicts advanced Murphy P, et al: Associations between histologic features of nonalco-
fibrosis in patients with nonalcoholic fatty liver disease: a prospective holic fatty liver disease (NAFLD) and quantitative diffusion-weighted
study, Hepatology 60(6):1920–1928, 2014. MRI measurements in adults, J Magn Reson Imaging 41(6):1629–
Lua I, et al: Mesodermal mesenchymal cells give rise to myofibroblasts, 1638, 2015.
but not epithelial cells, in mouse liver injury, Hepatology 60(1):311– Murphy FR, et al: Inhibition of apoptosis of activated hepatic stellate
322, 2014. cells by tissue inhibitor of metalloproteinase-1 is mediated via effects
Ludwig J, et al: Morphologic features of chronic hepatitis associated on matrix metalloproteinase inhibition: implications for reversibility
with primary sclerosing cholangitis and chronic ulcerative colitis, of liver fibrosis, J Biol Chem 277(13):11069–11076, 2002.
Hepatology 1(6):632–640, 1981. Nagula S, et al: Histological-hemodynamic correlation in cirrhosis—a
Ludwig J, et al: Staging of chronic nonsuppurative destructive cholan- histological classification of the severity of cirrhosis, J Hepatol 44(1):
gitis (syndrome of primary biliary cirrhosis), Virchows Arch A Pathol 111–117, 2006.
Anat Histol 379(2):103–112, 1978. Nakamura T, et al: Significance and therapeutic potential of endothe-
Lupsor M, et al: Performance of a new elastographic method (ARFI lial progenitor cell transplantation in a cirrhotic liver rat model,
technology) compared to unidimensional transient elastography in Gastroenterology 133(1):91–107.e1, 2007.
the noninvasive assessment of chronic hepatitis C. Preliminary Nelson DR, et al: Transforming growth factor-beta 1 in chronic hepa-
results, J Gastrointestin Liver Dis 18(3):303–310, 2009. titis C, J Viral Hepat 4(1):29–35, 1997.
Maher JJ, et al: Acetaldehyde-induced stimulation of collagen synthesis Neuschwander-Tetri BA, et al: Farnesoid X nuclear receptor ligand
and gene expression is dependent on conditions of cell culture: obeticholic acid for non-cirrhotic, non-alcoholic steatohepatitis
studies with rat lipocytes and fibroblasts, Alcohol Clin Exp Res (FLINT); a muticentre, randomised, placebo-controlled trial, Lancet
18(2):403–409, 1994. 2014. in press.
Mann DA: Epigenetics in liver disease, Hepatology 60(4):1418–1425, Ngo Y, et al: A prospective analysis of the prognostic value of biomark-
2014. ers (FibroTest) in patients with chronic hepatitis C, Clin Chem
Mann J, Mann DA: Transcriptional regulation of hepatic stellate cells, 52(10):1887–1896, 2006.
Adv Drug Deliv Rev 61(7–8):497–512, 2009. Nieto N, et al: Stimulation and proliferation of primary rat hepatic
Mann DA, Marra F: Fibrogenic signalling in hepatic stellate cells, stellate cells by cytochrome P450 2E1-derived reactive oxygen
J Hepatol 52(6):949–950, 2010. species, Hepatology 35(1):62–73, 2002.
Manousou P, et al: Digital image analysis of collagen assessment of Nishioka T, et al: Lysyl oxidase: from basic science to future cancer
progression of fibrosis in recurrent HCV after liver transplantation, treatment, Cell Struct Funct 37(1):75–80, 2012.
J Hepatol 58(5):962–968, 2013. Noureddin M, et al: Longitudinal comparison between MRI, MRS
Marcellin P, et al: Regression of cirrhosis during treatment with teno- and histology-determined steatosis in NAFLD patients at two-
fovir disoproxil fumarate for chronic hepatitis B: a 5-year open-label time points in a randomized trial, Hepatology 58(6):1930–1940,
follow-up study, Lancet 381(9865):468–475, 2013. 2013a.
Marra F: Leptin and liver fibrosis: a matter of fat, Gastroenterology Noureddin M, et al: Utility of magnetic resonance imaging versus
122(5):1529–1532, 2002. histology for quantifying changes in liver fat in nonalcoholic fatty
Marra F, Tacke F: Roles for chemokines in liver disease, Gastroenterol- liver disease trials, Hepatology 58(6):1930–1940, 2013b.
ogy 147(3):577–594.e1, 2014. Noureddin M, et al: Clinical and histological determinants of nonalco-
Martin-Vilchez S, et al: The hepatitis B virus X protein induces para- holic steatohepatitis and advanced fibrosis in elderly patients, Hepa-
crine activation of human hepatic stellate cells, Hepatology 47(6): tology 58(5):1644–1654, 2013.
1872–1883, 2008. Novo E, et al: Overexpression of Bcl-2 by activated human hepatic
Mayo MJ, et al: Prediction of clinical outcomes in primary biliary cir- stellate cells: resistance to apoptosis as a mechanism of progressive
rhosis by serum enhanced liver fibrosis assay, Hepatology 48(5):1549– hepatic fibrogenesis in humans, Gut 55(8):1174–1182, 2006.
1557, 2008. Oakley F, et al: Angiotensin II activates I kappaB kinase phosphoryla-
Mazzocca A, et al: Binding of hepatitis C virus envelope protein E2 to tion of RelA at Ser 536 to promote myofibroblast survival and liver
CD81 up-regulates matrix metalloproteinase-2 in human hepatic fibrosis, Gastroenterology 136(7):2334–2344.e1, 2009.
stellate cells, J Biol Chem 280(12):11329–11339, 2005. Ochoa-Callejero L, et al: Maraviroc, a CCR5 antagonist, prevents
Mederacke I, et al: Fate tracing reveals hepatic stellate cells as domi- development of hepatocellular carcinoma in a mouse model, PLoS
nant contributors to liver fibrosis independent of its aetiology, Nat ONE 8(1):e53992, 2013.
Commun 4:2823, 2013. Olaso E, et al: DDR2 receptor promotes MMP-2-mediated prolifera-
Mehta SH, et al: Exceeding the limits of liver histology markers, tion and invasion by hepatic stellate cells, J Clin Invest 108(9):1369–
J Hepatol 50(1):36–41, 2009. 1378, 2001.
METAVIR Cooperative Study Group: Intraobserver and interobserver Paik YH, et al: Toll-like receptor 4 mediates inflammatory signaling by
variations in liver biopsy interpretation in patients with chronic hepa- bacterial lipopolysaccharide in human hepatic stellate cells, Hepatol-
titis C. The French METAVIR Cooperative Study Group, Hepatol- ogy 37(5):1043–1055, 2003.
ogy 20(1 Pt 1):15–20, 1994. Paizis G, et al: Up-regulation of components of the renin-angiotensin
Michelotti GA, et al: NAFLD, NASH and liver cancer, Nat Rev Gas- system in the bile duct-ligated rat liver, Gastroenterology 123(5):1667–
troenterol Hepatol 10(11):656–665, 2013. 1676, 2002.
Michelotti GA, et al: Smoothened is a master regulator of adult liver Parekkadan B, et al: Immunomodulation of activated hepatic stellate
repair, J Clin Invest 123(6):2380–2394, 2013. cells by mesenchymal stem cells, Biochem Biophys Res Commun
Milani S, et al: Differential expression of matrix-metalloproteinase-1 363(2):247–252, 2007a.
and -2 genes in normal and fibrotic human liver, Am J Pathol Parekkadan B, et al: Mesenchymal stem cell-derived molecules reverse
144(3):528–537, 1994. fulminant hepatic failure, PLoS ONE 2(9):e941, 2007b.
Miller MH, et al: The serum proteome of nonalcoholic fatty liver Park SH, et al: Pentoxifylline vs. corticosteroid to treat severe alcoholic
disease: a multimodal approach to discovery of biomarkers of non- hepatitis: a randomised, non-inferiority, open trial, J Hepatol 61(4):
alcoholic steatohepatitis, J Gastroenterol Hepatol 29(10):1839–1847, 792–798, 2014.
2014. Parkes J, et al: Enhanced Liver Fibrosis (ELF) test accurately identifies
Mitchell C, et al: Dual role of CCR2 in the constitution and the resolu- liver fibrosis in patients with chronic hepatitis C, J Viral Hepat
tion of liver fibrosis in mice, Am J Pathol 174(5):1766–1775, 2009. 18(1):23–31, 2011.
Moore JK, et al: Systematic review: the effects of autologous stem cell Parsons CJ, et al: Antifibrotic effects of a tissue inhibitor of metallo-
therapy for patients with liver disease, Aliment Pharmacol Ther proteinase-1 antibody on established liver fibrosis in rats, Hepatology
39(7):673–685, 2014. 40(5):1106–1115, 2004.
Muhanna N, et al: Activation of hepatic stellate cells after phagocytosis Paulo JA, et al: Mass spectrometry-based quantitative proteomic profil-
of lymphocytes: a novel pathway of fibrogenesis, Hepatology 48(3): ing of human pancreatic and hepatic stellate cell lines, Genomics
963–977, 2008. Proteomics Bioinformatics 11(2):105–113, 2013.
Munoz-Luque J, et al: Regression of fibrosis after chronic stimulation Pellicoro A, et al: Liver fibrosis and repair: immune regulation of
of cannabinoid CB2 receptor in cirrhotic rats, J Pharmacol Exp Ther wound healing in a solid organ, Nat Rev Immunol 14(3):181–194,
324(2):475–483, 2008. 2014.
Perepelyuk M, et al: Hepatic stellate cells and portal fibroblasts are Russo FP, et al: The bone marrow functionally contributes to liver
the major cellular sources of collagens and lysyl oxidases in normal fibrosis, Gastroenterology 130(6):1807–1821, 2006.
liver and early after injury, Am J Physiol Gastrointest Liver Physiol Ryan KK, et al: FXR is a molecular target for the effects of vertical
304(6):G605–G614, 2013. sleeve gastrectomy, Nature 509(7499):183–188, 2014.
Pockros PJ, et al: Oral IDN-6556, an antiapoptotic caspase inhibitor, Rygiel KA, et al: Epithelial-mesenchymal transition contributes to
may lower aminotransferase activity in patients with chronic hepatitis portal tract fibrogenesis during human chronic liver disease, Lab
C, Hepatology 46(2):324–329, 2007. Invest 88(2):112–123, 2008.
Powell EE, et al: Host genetic factors influence disease progression in Safadi R, et al: Immune stimulation of hepatic fibrogenesis by CD8
chronic hepatitis C, Hepatology 31(4):828–833, 2000. cells and attenuation by transgenic interleukin-10 from hepatocytes,
Poynard T, et al: Natural history of liver fibrosis progression in patients Gastroenterology 127(3):870–882, 2004.
with chronic hepatitis C. The OBSVIRC, METAVIR, CLINIVIR, Sandrin L, et al: Transient elastography: a new noninvasive method for
and DOSVIRC groups, Lancet 349(9055):825–832, 1997. assessment of hepatic fibrosis, Ultrasound Med Biol 29(12):1705–
Poynard T, et al: Overview of the diagnostic value of biochemical 1713, 2003.
markers of liver fibrosis (FibroTest, HCV FibroSure) and necrosis Sanyal AJ, et al: Endpoints and clinical trial design for nonalcoholic
(ActiTest) in patients with chronic hepatitis C, Comp Hepatol 3(1):8, steatohepatitis, Hepatology 54(1):344–353, 2011.
2004. Sanyal AJ, et al: Pioglitazone, vitamin E, or placebo for nonalcoholic
Pradat P, et al: The cirrhosis risk score predicts liver fibrosis progres- steatohepatitis, N Engl J Med 362(17):1–5, 2010.
sion in patients with initially mild chronic hepatitis C, Hepatology Saxena NK, et al: Leptin in hepatic fibrosis: evidence for increased
51(1):356–357, 2010. collagen production in stellate cells and lean littermates of ob/ob
Pradere JP, et al: Toll-like receptor 4 and hepatic fibrogenesis, Semin mice, Hepatology 35(4):762–771, 2002.
Liver Dis 30(3):232–244, 2010. Schulze-Krebs A, et al: Hepatitis C virus-replicating hepatocytes
Preaux AM, et al: Matrix metalloproteinase-2 activation in human induce fibrogenic activation of hepatic stellate cells, Gastroenterology
hepatic fibrosis regulation by cell-matrix interactions, Hepatology 129(1):246–258, 2005.
30(4):944–950, 1999. Schwabe RF, et al: CD40 activates NF-kappa B and c-Jun N-terminal
Radaeva S, et al: Natural killer cells ameliorate liver fibrosis by killing kinase and enhances chemokine secretion on activated human
activated stellate cells in NKG2D-dependent and tumor necrosis hepatic stellate cells, J Immunol 166(11):6812–6819, 2001.
factor-related apoptosis-inducing ligand-dependent manners, Gas- Seki E, Brenner DA: Toll-like receptors and adaptor molecules in liver
troenterology 130(2):435–452, 2006. disease: update, Hepatology 48(1):322–335, 2008.
Radaeva S, et al: Retinoic acid signaling sensitizes hepatic stellate cells Shi Z, et al: Strain-specific differences in mouse hepatic wound healing
to NK cell killing via upregulation of NK cell activating ligand are mediated by divergent T helper cytokine responses, Proc Natl
RAE1, Am J Physiol Gastrointest Liver Physiol 293(4):G809–G816, Acad Sci U S A 94(20):10663–10668, 1997.
2007. Singh S, et al: Fibrosis progression in nonalcoholic fatty liver vs non-
Raetsch C, et al: Pentoxifylline downregulates profibrogenic cytokines alcoholic steatohepatitis: a systematic review and meta-analysis of
and procollagen I expression in rat secondary biliary fibrosis, Gut paired-biopsy studies, Clin Gastroenterol Hepatol 13(4):643–654.e1–
50(2):241–247, 2002. e9, quiz e39–e40, 2014.
Ramachandran P, Iredale JP: Reversibility of liver fibrosis, Ann Hepatol Smart DE, et al: JunD is a profibrogenic transcription factor regulated
8(4):283–291, 2009. by Jun N-terminal kinase-independent phosphorylation, Hepatology
Ramachandran P, Iredale JP: Macrophages: central regulators of 44(6):1432–1440, 2006.
hepatic fibrogenesis and fibrosis resolution, J Hepatol 56(6):1417– Son G, et al: Selective inactivation of NF-kappaB in the liver using
1419, 2012. NF-kappaB decoy suppresses CCl4-induced liver injury and fibrosis,
Ramachandran P, et al: Differential Ly-6C expression identifies the Am J Physiol Gastrointest Liver Physiol 293(3):G631–G639, 2007.
recruited macrophage phenotype, which orchestrates the regression Sterling RK, et al: Development of a simple noninvasive index to
of murine liver fibrosis, Proc Natl Acad Sci U S A 109(46):E3186– predict significant fibrosis in patients with HIV/HCV coinfection,
E3195, 2012. Hepatology 43(6):1317–1325, 2006.
Ramadori G, Saile B: Portal tract fibrogenesis in the liver, Lab Invest Svegliati-Baroni G, et al: Bile acids induce hepatic stellate cell prolifera-
84(2):153–159, 2004. tion via activation of the epidermal growth factor receptor, Gastro-
Ratziu V, et al: Sampling variability of liver biopsy in nonalcoholic fatty enterology 128(4):1042–1055, 2005.
liver disease, Gastroenterology 128(7):1898–1906, 2005. Tacke F, Yoneyama H: From NAFLD to NASH to fibrosis to HCC:
Regev A, et al: Sampling error and intraobserver variation in liver role of dendritic cell populations in the liver, Hepatology 58(2):494–
biopsy in patients with chronic HCV infection, Am J Gastroenterol 496, 2013.
97(10):2614–2618, 2002. Tacke F, Zimmermann HW: Macrophage heterogeneity in liver injury
Reif S, et al: The role of focal adhesion kinase-phosphatidylinositol and fibrosis, J Hepatol 60(5):1090–1096, 2014.
3-kinase-akt signaling in hepatic stellate cell proliferation and type I Tai CM, et al: Improvement of nonalcoholic fatty liver disease after
collagen expression, J Biol Chem 278(10):8083–8090, 2003. bariatric surgery in morbidly obese Chinese patients, Obes Surg
Ren S, et al: LRP-6 is a coreceptor for multiple fibrogenic signaling 22(7):1016–1021, 2012.
pathways in pericytes and myofibroblasts that are inhibited by DKK- Tang A, et al: Accuracy of MR imaging-estimated proton density
1, Proc Natl Acad Sci U S A 110(4):1440–1445, 2013. fat fraction for classification of dichotomized histologic steatosis
Rifai K, et al: Clinical feasibility of liver elastography by acoustic radia- grades in nonalcoholic fatty liver disease, Radiology 274(2):416–425,
tion force impulse imaging (ARFI), Dig Liver Dis 43(6):491–497, 2014.
2011. Teixeira-Clerc F, et al: CB1 cannabinoid receptor antagonism: a new
Ripoll C, et al: Hepatic venous pressure gradient predicts clinical strategy for the treatment of liver fibrosis, Nat Med 12(6):671–676,
decompensation in patients with compensated cirrhosis, Gastroenter- 2006.
ology 133(2):481–488, 2007. Trepo E, et al: Role of a cirrhosis risk score for the early prediction of
Rockey DC: Translating an understanding of the pathogenesis of fibrosis progression in hepatitis C patients with minimal liver disease,
hepatic fibrosis to novel therapies, Clin Gastroenterol Hepatol 11(3): J Hepatol 55(1):38–44, 2011.
224–231.e5, 2013. Troeger JS, et al: Deactivation of hepatic stellate cells during liver
Rodriguez-Suarez E, et al: Proteomics analysis of human nonalcoholic fibrosis resolution in mice, Gastroenterology 143(4):1073–1083.e22,
fatty liver, Methods Mol Biol 909:241–258, 2012. 2012.
Roh YS, Seki E: Toll-like receptors in alcoholic liver disease, non- Tsukada S, et al: SMAD and p38 MAPK signaling pathways indepen-
alcoholic steatohepatitis and carcinogenesis, J Gastroenterol Hepatol dently regulate alpha1(I) collagen gene expression in unstimulated
28(Suppl 1):38–42, 2013. and transforming growth factor-beta-stimulated hepatic stellate cells,
Roulot D, et al: Quantitative analysis of transforming growth factor J Biol Chem 280(11):10055–10064, 2005.
beta 1 messenger RNA in the liver of patients with chronic hepatitis Vandanmagsar B, et al: The NLRP3 inflammasome instigates obesity-
C: absence of correlation between high levels and severity of disease, induced inflammation and insulin resistance, Nat Med 17(2):179–
Hepatology 21(2):298–304, 1995. 188, 2011.
Vazquez-Torres A, et al: Toll-like receptor 4 dependence of innate and Yingling JM, et al: Development of TGF-beta signalling inhibitors for
adaptive immunity to Salmonella: importance of the Kupffer cell cancer, Nat Rev Drug Discov 3(12):1011–1022, 2004.
network, J Immunol 172(10):6202–6208, 2004. Yokohama S, et al: Therapeutic efficacy of an angiotensin II receptor
Wai CT, et al: A simple noninvasive index can predict both significant antagonist in patients with nonalcoholic steatohepatitis, Hepatology
fibrosis and cirrhosis in patients with chronic hepatitis C, Hepatology 40(5):1222–1225, 2004.
38(2):518–526, 2003. Yoshida S, et al: Extrahepatic platelet-derived growth factor-beta,
Wang FP, et al: High mobility group box-1 promotes the proliferation delivered by platelets, promotes activation of hepatic stellate cells
and migration of hepatic stellate cells via TLR4-dependent signal and biliary fibrosis in mice, Gastroenterology 147(6):1378–1392,
pathways of PI3K/Akt and JNK, PLoS ONE 8(5):e64373, 2013. 2014.
Wang Y, et al: Hepatic stellate cells, liver innate immunity, and hepa- Yoshida N, et al: Improvement of the survival rate after rat massive
titis C virus, J Gastroenterol Hepatol 28(Suppl 1):112–115, 2013. hepatectomy due to the reduction of apoptosis by caspase inhibitor,
Watanabe A, et al: Inflammasome-mediated regulation of hepatic stel- J Gastroenterol Hepatol 22(11):2015–2021, 2007.
late cells, Am J Physiol Gastrointest Liver Physiol 296(6):G1248– Yoshiji H, et al: Imatinib mesylate (STI-571) attenuates liver fibrosis
G1257, 2009. development in rats, Am J Physiol Gastrointest Liver Physiol 288(5):
Wells RG: The portal fibroblast: not just a poor man’s stellate cell, G907–G913, 2005.
Gastroenterology 147(1):41–47, 2014. Yu MC, et al: Inhibition of T-cell responses by hepatic stellate cells via
Winau F, et al: Ito cells are liver-resident antigen-presenting cells for B7-H1-mediated T-cell apoptosis in mice, Hepatology 40(6):1312–
activating T cell responses, Immunity 26(1):117–129, 2007. 1321, 2004.
Witek RP, et al: Pan-caspase inhibitor VX-166 reduces fibrosis in an Zhang Z, et al: High mobility group box 1 activates Toll like receptor
animal model of nonalcoholic steatohepatitis, Hepatology 50(5):1421– 4 signaling in hepatic stellate cells, Life Sci 91(5–6):207–212,
1430, 2009. 2012.
Wobser H, et al: Lipid accumulation in hepatocytes induces fibrogenic Zhang S, et al: Farnesoid X receptor agonist WAY-362450 attenuates
activation of hepatic stellate cells, Cell Res 19(8):996–1005, 2009. liver inflammation and fibrosis in murine model of non-alcoholic
Wree A, et al: From NAFLD to NASH to cirrhosis-new insights into steatohepatitis, J Hepatol 51(2):380–388, 2009.
disease mechanisms, Nat Rev Gastroenterol Hepatol 10(11):627–636, Zhou X, et al: Expression of matrix metalloproteinase-2 and -14 per-
2013. sists during early resolution of experimental liver fibrosis and might
Wree A, et al: NLRP3 inflammasome activation results in hepatocyte contribute to fibrolysis, Liver Int 24(5):492–501, 2004.
pyroptosis, liver inflammation and fibrosis, Hepatology 59(3):898– Zhou X, et al: Engagement of alphavbeta3 integrin regulates prolifera-
910, 2014. tion and apoptosis of hepatic stellate cells, J Biol Chem 279(23):23996–
Wynn TA: Cellular and molecular mechanisms of fibrosis, J Pathol 24006, 2004.
214(2):199–210, 2008. Zhu NL, et al: Hepatic stellate cell-derived delta-like homolog 1
Xie G, et al: Cross-talk between Notch and Hedgehog regulates hepatic (DLK1) protein in liver regeneration, J Biol Chem 287(13):10355–
stellate cell fate in mice, Hepatology 58(5):1801–1813, 2013. 10367, 2012.
Yang L, et al: Vascular endothelial growth factor promotes fibrosis
resolution and repair in mice, Gastroenterology 146(5):1339–1350.e1,
2014.
CHAPTER 8
Bile secretion and pathophysiology of biliary
tract obstruction
Henry Anthony Pitt and Attila Nakeeb
OVERVIEW whereas splanchnic stimulation causes vasoconstriction with

decreased hepatic blood flow and thus results in diminished
Bile secretion is one of the major functions of the liver, bile secretion. Gastrointestinal hormones—secretin, cholecys-
which serves two major purposes: (1) the excretion of tokinin, gastrin, and glucagon—all increase bile flow, primarily
hepatic metabolites—including bilirubin, cholesterol, drugs, by increasing water and electrolyte secretion. This action prob-
and toxins—and (2) the facilitation of intestinal absorption of ably occurs at a site distal to the hepatocyte. Finally, the most
lipids and fat-soluble vitamins. More recently, through their important factor in regulating the volume of bile flow is the rate
interaction with the gut microbiome, bile acids also have been of bile salt synthesis by hepatocytes. This rate is regulated
found to have important signaling functions. Through receptor by the return of bile salts to the liver by the enterohepatic
activation, bile acids regulate lipid, glucose, and energy metab- circulation.
olism. Alterations in bile secretion also may contribute to cho-
lelithiasis (see Chapter 32) and its potential complications, such Bile Composition
as cholecystitis (see Chapter 33) and choledocholithiasis (see The components of hepatic and gallbladder bile are essentially
Chapters 36 and 37). On the other hand, obstruction of bile the same, but the concentration varies considerably because of
flow results in alterations of coagulation, the immune system, the ability of the gallbladder to absorb water (Table 8.1). The
and all organ functions. This chapter will discuss the physiology gallbladder absorbs water both actively via sodium-hydrogen
of bile secretion, the pathophysiology of bile obstruction, and (Na+/H+) pumps and passively through aquaporin channels.
the management of obstructive jaundice. Both chloride (Cl−) and bicarbonate (HCO3−) are absorbed by
the gallbladder epithelium via the cystic fibrosis transmembrane
BILE SECRETION regulator (CFTR; Swartz-Basile et al, 2007). The secretion of
hydrogen ions and the absorption of bicarbonate by the gall-
Bile Formation bladder alter the acid-base balance from basic in hepatic bile
The two primary roles of bile in normal physiology are the excre- to acidic in gallbladder bile.
tion of organic compounds, such as bilirubin and cholesterol, The gallbladder mucosa also absorbs calcium (Ca+2) and
and the intestinal absorption of lipids. Bile secretion results from magnesium (Mg+2). However, calcium absorption is not as effi-
the active transport of solutes into the canaliculus, followed by cient as the absorption of sodium and water, which leads to a
the passive flow of water. Water constitutes approximately 85% significantly greater relative increase in the concentration of
of the volume of bile. The major organic solutes in bile are bili- calcium in the gallbladder. Similarly, the concentration of bili-
rubin, bile salts, phospholipids, and cholesterol. Bilirubin, the rubin, which is not actively absorbed by the gallbladder, may
breakdown product of spent red blood cells, is conjugated with be as high as 10-fold. Thus precipitation of calcium bilirubinate
glucuronic acid by the hepatic enzyme glucuronyl transferase crystals, the major component of pigment gallstones, is much
and is excreted actively into the adjacent canaliculus. Normally, more likely to occur within the gallbladder. In addition, the
a large enzyme reserve exists to handle excess bilirubin produc- biliary lipids, bile salts, phospholipids, and cholesterol all
tion, which might exist in hemolytic states. become more concentrated in the gallbladder. While gallblad-
Bile salts are steroid molecules synthesized by hepatocytes. der bile becomes concentrated, several changes occur in the
The primary bile salts in humans, cholic and chenodeoxycholic capacity of bile to solubilize cholesterol. The solubility in
acid, account for approximately 80% of those produced. The the micellar fraction is increased, but the stability of the
primary bile salts, which are then conjugated with either taurine phospholipid-cholesterol vesicles is greatly decreased. Because
or glycine, can undergo bacterial alteration in the intestine to cholesterol crystal precipitation occurs preferentially by vesicu-
form the secondary bile salts, deoxycholate and lithocholate. lar, rather than micellar, mechanisms, the net effect of concen-
The purpose of bile salts is to solubilize lipids and facilitate their trating bile is an increased tendency to form cholesterol crystals
absorption. Phospholipids are synthesized in the liver in con- (Klein et al, 1996).
junction with bile salt synthesis, and lecithin is the primary
phospholipid in human bile, constituting more than 95% of its Bile Salt Secretion
total. The final major solute of bile is cholesterol, which is also Bile is secreted from the hepatocyte into canaliculi, which drain
produced primarily by the liver with a small contribution from their contents into small bile ducts. Secretion of bile salts is the
dietary sources. major osmotic force for the generation of bile flow. Bile acids
The normal volume of bile secreted daily by the liver is 750 are formed at a rate of 500 to 600 mg per day. The bulk of the
to 1000 mL. Bile flow depends on neurogenic, humoral, and bile salt pool is maintained in the gallbladder, followed by the
chemical control. Vagal stimulation increases bile secretion, liver, the small intestine, and the extrahepatic bile ducts. Bile
123
124 PART 1 LIVER, BILIARY AND PANCREATIC ANATOMY AND PHYSIOLOGY
acids are synthesized from cholesterol via two main pathways: mediated by sodium-dependent and sodium-independent
a classic pathway leads to the formation of cholic acid, and an mechanisms. The sodium-dependent pathway accounts for
alternative pathway results in the synthesis of chenodeoxycholic more than 80% of taurocholate uptake but less than 50% of
acid. The classic pathway is the predominant mode of bile acid cholate uptake (Meier & Stieger, 2002). In recent years, a
synthesis in humans. As a result, 60% to 70% of the bile acid number of transport proteins have been identified that play a
pool consists of cholic acid and its metabolite deoxycholic acid, key role in this process (Fig. 8.1). The bile salt transporter
with chenodeoxycholic acid occurring less commonly in human is termed the sodium-taurocholate cotransporting polypeptide
bile (Holm et al, 2013; Kullak-Ublick et al, 2004). (NTCP), and is exclusively expressed in the liver and located
In plasma, bile acids circulate bound to either albumin or in the basolateral membrane of the hepatocyte. Sodium-
lipoproteins. In the space of Disse within the liver, bile salt independent hepatic uptake of bile acids is mediated primarily
uptake into the hepatocytes is very efficient. This process is by a family of transporters termed the organic anion transporting
polypeptides (OATPs). In contrast to NTCP, these transporters
have a broader substrate affinity and transport a variety of
TABLE 8.1 Composition of Hepatic and Gallbladder Bile organic anions, including the bile salts. OATP-C is the major
Characteristics* Hepatic Bile Gallbladder Bile sodium-independent bile salt uptake system, but OATP-A
also takes up bile acids, and OATP-8 mediates taurocholate
Sodium 160 270
uptake.
Potassium 5 10
Intracellular bile acid transport occurs within a matter of
Chloride 90 15 seconds. Two mechanisms may be responsible for bile acid
Bicarbonate 45 10 transcellular movement: One involves transfer of bile acids
Calcium 4 25 from the basolateral membrane to the canalicular membrane
Magnesium 2 4 via bile acid–binding proteins (Crawford, 1996); the other
Bilirubin 1.5 15 moves cellular bile salts through vesicular transport. In con-
Proteins 150 200 trast, the transport of bile salts across the canalicular membrane
Bile acids 50 150 of hepatocytes represents the rate-limiting step in the overall
Phospholipids 8 40 secretion of bile salts from the blood into bile.
Cholesterol 4 18 Bile salt concentrations are 1000-fold greater within the
canaliculi than in the hepatocytes. This gradient necessitates an
Total solids — 125
active transport mechanism, which is an adenosine triphos-
pH 7.8 7.2
phate (ATP)-dependent process. The ATP-binding cassette
Significant ranges may be seen.
transporter ABCB 11 (formerly known as the bile salt export
*All determinations are milliequivalents per liter, except for pH. pump [BSEP]) plays a key role in this process (Henkel et al,
Bile Salt Transport Biliary Lipids/Bilirubin/Drug Transport
Hepatocyte Hepatocyte
Bile Canaliculus
Bile Salt ??Biliary Lipid Uptake
Uptake BSEP MDR3
NTCP BS and Intracellular Transport??
HBAB BL BL
BS BS
Na+ BS
MDR1 OATP-A
D/OA
??Bile Salt Intracellular Transport?? Bile D/OA
BRCP
D/OA
BS B
OATPs MRP2 MRP2 OATP-C
BS B, D/OA
BS B
Vesicles
ER/Golgi ER
Na+ BS, B, D/OA

MRP3
Na+/K+-ATPase
Basolateral (Sinusoidal) Domain

K+
FIGURE 8.1. Bile formation in human liver. ATP, Adenosine triphosphate; B, bilirubin; BL, biliary lipids; BRCP, breast cancer–related protein; BS,
bile salts; BSEP, bile salt export pump; D/OA, drugs/organic anions; ER, endoplasmic reticulum; HBAB, hepatic bile acid–binding protein; MDR1 and
MDR3, multidrug-resistance proteins 1 and 3, respectively; MRP3, MDR-related protein-3; NTCP, Na+-taurocholate cotransporting polypeptide; OATPs
(A, C, 8), organic anion transporting polypeptides.
Chapter 8 Bile secretion and pathophysiology of biliary tract obstruction 125
2013). The ABC transporters mediate the transport of metabo- HEPATIC BILE
lites, peptides, fatty acids, cholesterol, and lipids in the liver, Micelles
intestines, pancreas, lungs, kidneys, brain, and in macrophages. Unilamellar vesicle
Although ABCB 11 is the major transporter for monovalent (phospholipid-rich)
bile salts into the canaliculus, MDR-related protein-2 (MRP2),
a member of the multidrug-resistant protein family, also trans-
ports sulfated and glucuronidated bile salts into the canaliculus.
MRP2 also mediates the export of multiple other organic
anions, including conjugated bilirubin, leukotrienes, glutathi-
one disulfide, chemotherapeutic agents, uricosurics, antibiotics, – H2O
GALLBLADDER
toxins, and heavy metals (Gerk & Vore, 2002).
BILE
Recent studies suggest that bile acids are signaling molecules
that regulate lipid, glucose, and energy metabolism (Li &
Chang, 2015). This function of bile acids is mediated primarily
by the nuclear receptor farnesoid X receptor (FXR) and the Unilamellar vesicle
G-protein–coupled receptor TGR5. Bile acids in the small and (cholesterol-enriched)
large intestine regulate the gut microbiome, incretin secretion,
and fibroblast growth factors 15 and 19 (FGF15/FGF19) pro- Fusion
duction. These FGFs, in turn, modulate lipid, glucose, and
energy metabolism and may play a role in the rapid improve-
ment in glycemic control after gastric bypass surgery. In addi- Large, cholesterol-rich
tion, FXR and TGR5 receptors exist in other tissues, such as multilamellar vesicles
the heart and the kidneys, and therefore may help to explain
the dysfunction that occurs in these organs with biliary obstruc-
tion (Swann et al, 2011).
Biliary Lipid Secretion

Cholesterol monohydrate
Compared with bile salts, the biliary lipids, phospholipids and crystal nucleation
cholesterol, play a secondary role in the formation of bile.
Phospholipids and cholesterol are formed primarily from low- FIGURE 8.2. Concentration of bile leads to net transfer of phospho-
density lipoproteins circulating in plasma and from de novo lipids and cholesterol from vesicles to micelles. Phospholipids are trans-
synthesis by hepatocytes. Less is known about the secretion of ferred more efficiently than cholesterol, leading to cholesterol enrichment
biliary lipids compared with bile salt secretion; however, biliary of the remaining (remodeled) vesicles. Aggregation of these cholesterol-
lipid secretion is crucial for cholesterol disposal, intestinal rich vesicles forms multilamellar liquid crystals of cholesterol monohy-
absorption of dietary lipids, and cytoprotection against bile drate. (From Vessey DA, 1990: Metabolism of drugs and toxins by the
human liver. In Zakin D, Boyer TD [eds]: Hepatology: a textbook of Liver
acid–induced hepatocyte and cholangiocyte injury (Arrese &
Disease, 2nd ed. Philadelphia, WB Saunders, p 1492.)
Accatino, 2002).
Phospholipid secretion involves the delivery of phospholip-
ids to the inner leaflet of the canalicular plasma membrane
(Elferink & Groen, 2000). In humans, the MDR3 transporter The concept of mixed micelles as the only cholesterol carrier
translocates phospholipids from the inner to the outer leaflet of has been challenged by the demonstration that much of the
the canalicular membrane. Progressive familial intrahepatic biliary cholesterol exists in a vesicular form. Structurally, these
cholestatis type 3 develops in humans with an MDR3 defi- vesicles are made up of lipid bilayers of cholesterol and phos-
ciency with an MDR3 deficiency (Kullak-Ublick et al, 2004). pholipids. In their simplest and smallest form, the vesicles are
These patients have no phosphatidylcholine in bile and there- unilamellar, but an aggregation may take place, leading to mul-
fore do not form mixed micelles with bile salts. As a result, tilamellar vesicles. Present theory suggests that in states of
toxic bile salts injure the biliary epithelium, resulting in neona- excess cholesterol production, these large vesicles also may
tal cholestasis, cholestasis of pregnancy, and cirrhosis in adults. exceed their capability to transport cholesterol, and crystal pre-
Less is known about the role of transporter proteins in cho- cipitation may occur (Fig. 8.2).
lesterol secretion, but the ABC transporters ABCG5 and
ABCG8 have been demonstrated to be involved in the elimina- Bilirubin Secretion
tion of plant steroids (Lee et al, 2001). Cholesterol is highly Heme is released at the time of degradation of senescent eryth-
nonpolar and insoluble in water; thus it is insoluble in bile. The rocytes by the reticuloendothelial system. Heme is the source
key to maintaining cholesterol in solution is the formation of of approximately 80% to 85% of the bilirubin that is produced
micelles, a bile salt–phospholipid–cholesterol complex. Bile daily. The remaining 15% to 20% is derived largely from the
salts are amphipathic compounds that contain both a hydro- breakdown of hepatic hemoproteins. Both enzymatic and non-
philic and hydrophobic portion. In aqueous solutions, bile salts enzymatic pathways for the formation of bilirubin have been
are oriented with the hydrophilic portion outward. Phospholip- proposed. Although both may be important physiologically, the
ids are incorporated into the micellar structure, allowing cho- microsomal enzyme hemeoxygenase—found in high concentra-
lesterol to be added to the hydrophobic central portion of the tion throughout the liver, spleen, and bone marrow—plays a
micelle. In this way, cholesterol can be maintained in solution major role in the initial conversion of heme to biliverdin, which
in an aqueous medium. is then reduced to bilirubin by the cytosolic enzyme biliverdin
reductase before being released into the circulation. In this

“unconjugated” form, bilirubin has a very low solubility and is Cholesterol
bound avidly to plasma proteins, primarily albumin, before
uptake and further processing by the liver. The liver is the sole Newly synthesized
organ capable of removing the albumin-bilirubin complex from LIVER bile acid
Blood
the circulation and esterifying the potentially toxic bilirubin to (≈ 0.6 g/24 hr)
(cholesterol)
water-soluble, nontoxic, monoconjugated and deconjugated
derivatives. SMALL
BOWEL
In the sinusoidal membrane of the hepatocyte, bilirubin is
taken up by OATP-C, a membrane transporter belonging to 2-4 g bile acid
the OATP family (Cui et al, 2001). OATP-C is involved with pool cycling
the uptake of both conjugated and unconjugated bilirubin, but PORTAL 6-10x/day
VEIN
unconjugated bilirubin also can cross hepatic sinusoidal mem-
branes by a diffusion process. In the hepatocyte, bilirubin binds COLON
to a driver of gluthathione-S-transferase and is catalyzed by
bilirubin uridine-5′-diphosphate glycosyltransferase to form
bilirubin glucuronides. Mutations in the gene encoding biliru-
bin UDP-glycosyltransferase are associated with the unconju-
Fecal bile acids
gated hyperbilirubin syndromes, Crigler-Najjar and Gilbert (≈ 0.6 g/24 hr)
syndromes (Iganagi et al, 1990).
FIGURE 8.3. Enterohepatic circulation of bile salts. Cholesterol is
Bilirubin glucuronides are excreted into the bile canaliculus
taken up from plasma by the liver. Bile acids are synthesized at a rate
primarily via MRP2, which also plays a role in the transport of
of 0.6 g/24 hr and are excreted through the biliary system into the small
glucuroniductal bile salts and a wide spectrum of organic bowel. Most of the bile salts are reabsorbed in the terminal ileum and
anions, including the antibiotic ceftriaxone. MRP3, which is are returned to the liver to be extracted and reextracted. (Modified from
expressed in the basolateral membrane of hepatocytes and chol- Dietschy JM: The biology of bile acids, Arch Intern Med 130:482–474,
angiocytes, also participates in the transport of bilirubin mono- 1972.)
glucuronide. In addition, MRP3 may prevent intracellular
accumulation of conjugated bilirubin, bile salts, and other
organic anions in cholestatic situations. Enterohepatic Circulation
Bile salts are synthesized and conjugated in the liver; secreted
Bile Flow into bile; stored temporarily in the gallbladder; passed from the
The bile ducts, gallbladder, and sphincter of Oddi act in concert gallbladder into the duodenum; absorbed throughout the small
to modify, store, and regulate the flow of bile. Bile flow is primar- intestine, especially in the ileum; and returned to the liver via
ily driven by bile salt secretion. During its passage through the the portal vein. This cycling of bile acids between the liver and
bile ductules, canalicular bile is modified by the absorption and the intestine is referred to as the enterohepatic circulation (Fig.
secretion of electrolytes and water. Bicarbonate secretion by the 8.3). The total amount of bile acids in the enterohepatic circu-
bile ducts plays an important role in bile salt–independent bile lation is defined as the circulating bile pool. In this highly efficient
flow. The gastrointestinal hormone secretin increases bile flow system, nearly 95% of bile salts are reabsorbed. Thus, of the
primarily by increasing the active secretion of chloride-rich fluid total bile salt pool of 2 to 4 g, which recycles through the
by the bile ducts. Bile duct secretion also is stimulated by other enterohepatic cycle 6 to 10 times daily, only about 600 mg is
hormones, such as cholecystokinin and gastrin. The bile duct actually excreted into the colon. Bacterial action in the colon
epithelium is capable of water and electrolyte absorption, which on the two primary bile salts, cholate and chenodeoxycholate,
may be of primary importance in the storage of bile during results in the formation of the secondary bile salts, deoxycholate
fasting in patients who have previously undergone cholecystec- and lithocholate. In fact, the bile acid signature of an individual
tomy. The main functions of the gallbladder are to concentrate is very dependent on gut microbial modification (Swann et al,
and store hepatic bile during the fasting state and deliver bile 2011). Bacterial enzymes modify primary bile acids through
into the duodenum in response to a meal. The usual capacity of deconjugation, dehydrogenation, dehydroxylation, and sul-
the human gallbladder is about 40 to 50 mL. Only a small frac- fation reactions. In turn, bile acids restrict bacterial prolifera-
tion of the bile produced each day would be stored, were it not tion and overgrowth. However, the physiology of bile salts,
for the gallbladder’s remarkable absorptive capacity. biliary lipids, bilirubin, bile flow, and the enterohepatic circula-
The enterohepatic circulation provides an important nega- tion is dramatically altered when the bile ducts become
tive feedback system on bile salt synthesis. Should the recircula- obstructed.
tion be interrupted by resection of the terminal ileum or by
primary ileal disease, abnormally large losses of bile salts occur. BILIARY OBSTRUCTION
This situation increases bile salt production to maintain a
normal bile salt pool. Similarly, if bile salts are lost through an The evaluation and management of the patient with biliary
external biliary fistula, increased bile salt synthesis is necessary. obstruction is a common problem facing the general surgeon.
However, except for those unusual circumstances in which Over the past 40 years, significant advances have been made in
excessive losses occur, bile salt synthesis matches losses, main- our understanding of the pathophysiology, diagnosis, and man-
taining a constant bile salt pool size. During fasting, approxi- agement of the jaundiced patient. Similarly, advances have
mately 90% of the bile acid pool is sequestered in the been made in perioperative and operative management that
gallbladder. have resulted in improved survival of the jaundiced patient.
TABLE 8.2 Classification of Jaundice

Defect in Bilirubin Metabolism Predominant Hyperbilirubinemia Examples
Increased production Unconjugated Congenital hemoglobinopathies, hemolysis, multiple

transfusions, sepsis, burns
Impaired hepatocyte uptake Unconjugated Gilbert’s disease, drug induced
Reduced conjugation Unconjugated Neonatal jaundice, Crigler-Najjar syndrome
Altered transport and excretion Conjugated Hepatitis, cirrhosis, Dubin-Johnson syndrome, Rotor’s syndrome
Biliary obstruction Conjugated Choledocholithiasis, benign strictures, chronic pancreatitis,
sclerosing cholangitis, periampullary cancer, biliary
malignancies
Obstructive jaundice affects multiple organ systems, including

BOX 8.1 Lesions Commonly Associated With Biliary Tract
hepatic, renal, cardiovascular, hematologic, and immune
systems. This section will review the causes, pathophysiology, Obstruction
and management of biliary obstruction. Type I: Complete Obstruction
Tumors of the head of the pancreas
Causes of Jaundice Common bile duct ligation
Jaundice may result from (1) increased production of bilirubin, Cholangiocarcinoma
Gallbladder cancer
(2) impaired hepatocyte uptake of bilirubin, (3) reduced con-
Parenchymal liver tumors (primary or secondary)
jugation of bilirubin, (4) altered transport or excretion of
bilirubin into the bile canaliculus, or (5) obstruction of the Type II: Intermittent Obstruction
intrahepatic or extrahepatic biliary tree (Table 8.2). Overpro- Choledocholithiasis
duction, impaired uptake, and reduced conjugation of bilirubin Periampullary tumors
all lead to a predominantly unconjugated hyperbilirubinemia. Duodenal diverticula
Altered transport and excretion and biliary ductal obstruction Choledochal cyst
result in hyperbilirubinemia that is primarily conjugated. Some Polycystic liver disease
patients have multiple defects in normal metabolism. For Biliary parasites
example, secondary hepatocellular dysfunction may develop in Hemobilia
a patient with biliary obstruction from a tumor. Therefore these
Type III: Chronic Incomplete Obstruction
classification systems may be simplifications of more complex
Strictures of the common bile duct
disease processes. Congenital biliary atresia
Diseases that cause bile duct obstruction may be further Traumatic (iatrogenic)
divided into conditions that cause (1) complete obstruction, Sclerosing cholangitis
(2) intermittent obstruction, (3) chronic incomplete obstruc- Post radiotherapy
tion, or (4) segmental duct obstruction (Box 8.1). Patients with Stenosis of biliary-enteric anastamosis
complete biliary obstruction will have clinical jaundice, and Chronic pancreatitis
those with intermittent obstruction may experience symptoms Cystic fibrosis
(pain, pruritus, fevers) and biochemical changes without neces- Sphincter of Oddi stenosis
sarily experiencing clinical jaundice. Hepatic fibrosis can even-
Type IV: Segmental Obstruction
tually develop in patients with chronic incomplete obstruction
Traumatic
(see Chapter 7) and biliary cirrhosis (see Chapter 76). Intrahepatic stones
Sclerosing cholangitis
Pathophysiology Cholangiocarcinoma
Biliary obstruction produces local effects on the bile ducts,
which lead to derangements of hepatic function and, ultimately,
to widespread systemic effects. Jaundiced patients are at
increased risk for hepatic dysfunction, renal failure, cardiovas- distorted and swollen. In patients with long-standing obstruc-
cular impairments, nutritional deficiencies, bleeding problems, tion, intrahepatic bile ductule proliferation occurs with an
infections, and wound complications, and their perioperative increase in the length and tortuosity of the canaliculi.
mortality and morbidity are increased. The biliary system normally has a low pressure (5 to 10 cm
H2O); however, in the setting of complete or partial biliary
Hepatobiliary obstruction, biliary pressure can approach 30 cm H2O. While
Hepatocytes are arranged in plates along which blood flows biliary pressure increases, the tight junctions between hepato-
from portal to central veins. Within these plates, the small cytes and bile duct cells are disrupted, resulting in an increase
apical domains of adjacent hepatocytes form a tubular lumen, in bile duct and canalicular permeability. Bile contents can
the canaliculus, which is the site of primary bile formation. freely reflux into liver sinusoids, causing a marked polymorpho-
From the canalicular network, bile flows to the small ductules nuclear leukocyte infiltration into the portal triads. This inflam-
and subsequently to the larger ducts. With biliary obstruction, matory response is followed by increased fibrinogenesis with
the bile canaliculi become dilated, and the microvilli are deposition of reticulin fibers, which undergo conversion to type
I collagen (see Chapters 5 to 7). The extrahepatic bile ducts Cardiovascular

exhibit mucosal atrophy and squamous metaplasia, followed by In addition to hepatic dysfunction, obstructive jaundice may
inflammatory infiltration and fibrosis in the subepithelial layers cause severe hemodynamic and cardiac disturbances. Experi-
of the bile duct (Karsten et al, 1991). mental animals with obstructive jaundice tend to be hypoten-
In addition to the structural effects of biliary obstruction on sive and exhibit an exaggerated hypotensive response to
the bile ducts, elevated biliary pressure can alter bile production hemorrhage. Studies in experimental animals have demon-
by hepatocytes. In the setting of biliary obstruction and elevated strated that bile duct ligation (1) decreases cardiac contractility;
biliary pressure, bile becomes less lithogenic because of a rela- (2) reduces left ventricular pressures; (3) impairs response to
tive decrease in cholesterol and phospholipid secretion com- β-agonist drugs, such as isoproterenol and norepinephrine; and
pared with bile acid secretion. With the relief of biliary (4) decreases peripheral vascular resistance (Ma et al, 1999).
obstruction and the normalization of biliary pressures, the In a study of nine patients with obstructive or cholestatic jaun-
recovery of cholesterol and phospholipid secretion is more dice, Lumlertgul and colleagues (1991) showed a significantly
rapid than bile acid secretion; therefore bile is more lithogenic blunted response in left ventricular ejection fraction compared
in this setting. This phenomenon may lead to premature occlu- with normal volunteers following the infusion of the positive
sion of decompressive bilary stents placed for the management inotrope dobutamine. Padillo and others (2001) also have
of obstructive jaundice. shown in 13 patients a negative correlation between serum bili-
Several authors have reported impairment of both macro- rubin and left ventricular systolic work. Successful internal
vascular and microvascular perfusion of the liver in obstructive biliary drainage in these patients was associated with a signifi-
jaundice. Intravital fluorescence microscopy has shown a sig- cant increase in cardiac output, compliance, and contractility.
nificant increase in the number of nonperfused sinusoids after The combination of depressed cardiac function and decreased
3 days of extrahepatic obstruction. Moreover, in perfused sinu- total peripheral resistance most likely makes the jaundiced
soids, a 35% decrease in the mean diameter and a 25% decrease patient more susceptible to the development of postoperative
in flow velocity were noted (Koeppel et al, 1997). This altera- shock than nonjaundiced patients.
tion in hepatic perfusion may help explain the increased risk of
hepatocellular dysfunction when performing liver resections in Renal
patients with obstructive jaundice (see Chapter 103). Extrahe- The association between jaundice and postoperative renal
patic biliary obstruction and jaundice also can alter important failure has been known for many years. The reported incidence
secretory, metabolic, and synthetic functions of the liver. When of postoperative acute kidney injury has been reported to be as
biliary pressure rises higher than 20 cm H2O, hepatic bile secre- high as 10% but varies depending on the nature of the proce-
tion is diminished, and hepatocytes cannot excrete efficiently dure. Moreover, the mortality rate in patients with jaundice in
against the high pressure. As a result, excretory products of the whom renal failure developed has been reported to be as high
hepatocytes reflux directly into the vascular system, leading to as 70% (Fogarty et al, 1995). Important factors that may play
systemic toxicity. a role in the development of renal failure in obstructive jaundice
Patients with jaundice have a decreased capacity to excrete include (1) depressed cardiac function, (2) hypovolemia, (3)
drugs, such as antibiotics, that are normally secreted into bile bile salt–mediated effects on renal FXR and TGR5 receptors,
(Blenkharn et al, 1985). The increased concentration of bile and (4) endotoxemia.
acids associated with obstructive jaundice results in inhibition The decreased cardiac function associated with obstructive
of the hepatic cytochrome P450 enzymes and therefore a jaundice leads to a decrease in renal perfusion. Decreased
decrease in the rate of oxidative metabolism in the liver. In cardiac output also may result in stretching of the atrium and
addition, bile acids in abnormally high concentrations can increased production of atrial natriuretic peptide (ANP), a
induce apoptosis (programmed cell death) in hepatocytes (Patel hormone known to cause natriuresis, to counter the action of
et al, 1994). The synthetic function of the hepatocyte also is water- and sodium-retaining hormones, to inhibit the thirst
decreased with obstructive jaundice, as evidenced by decreased mechanism, and to produce peripheral vasodilation. Plasma
plasma levels of albumin, clotting factors, and secretory immu- levels of ANP have been shown to be increased in both experi-
noglobulins (IgA). mental animals and in patients with extrahepatic biliary obstruc-
Kupffer cells are tissue macrophages that are the predomi- tion (Padillo et al, 2001).
nant cell type of the hepatic reticuloendothelial system (see In addition to the direct effects of jaundice on the heart and
Chapter 10). Normally, infectious agents, damaged blood cells, peripheral vasculature discussed earlier, the increased serum
cellular debris, fibrin degradation products, and endotoxin levels of bile acids associated with obstructive jaundice have a
absorbed or formed in the portal circulation are effectively fil- direct diuretic and natriuretic effect on the kidney that results
tered by Kupffer cells and removed from the systemic circula- in significant extracellular volume depletion and hypovolemia.
tion. Kupffer cells also play an interactive role with hepatocytes, In dogs, the infusion of bile into the renal artery results in
modulating synthesis of hepatic proteins. Obstructive jaundice increased urine flow, natriuresis, and kaliuresis. This diuretic
has been shown to have profound effects on Kupffer cells, effect may be mediated by increased prostaglandin E2 produc-
including decreased endocytosis, phagocytosis, clearance of tion by the kidney (Green & Better, 1994) and/or by their effect
bacteria and endotoxin, and expression of the major histocom- on FXR and TGR5 receptors (Swann et al, 2011).
patibility complex class II antigen, with a consequent dimin- Another significant factor in the development of renal failure
ished ability to process antigen (Nehez & Anderson, 2002; is endotoxemia (Fig. 8.4). Approximately 50% of patients with
Puntis & Jiang, 1996). Biliary obstruction also has been shown obstructive jaundice have endotoxin in their peripheral blood
to increase levels of proinflammatory cytokines, including (Hunt et al, 1982). This phenomenon may be the result of
tumor necrosis factor-α and interleukin-6 (see Chapters 10 decreased hepatic clearance of endotoxin by Kupffer cells and
and 11). a lack of bile salts in the gut lumen that normally prevent
Obstructive jaundice
↓ Kupffer cell ↑ Systemic bile salts ↑ Systemic bilirubin

clearance ↓ Gut bile salts
↓ Clearance of Endotoxemia
cardiotoxins?
Cardiovascular
system
Renal system Coagulation system
ANP and ANP and Prostaglandins Peritubular fibrin

hypodipsia cytokines and cytokines deposition
↓ LV ↓ Plasma ↓ Peripheral
function volume resistance
↓ Blood Altered intrarenal Direct ↑ Renal vascular resistance

pressure hemodynamics parenchyma ↓ Renal permeability
toxicity
Renal impairment or
acute renal failure
FIGURE 8.4. Obstructive jaundice leads to renal impairment or acute kidney injury. ANP, Atrial natriuretic peptide; LV, left ventricular.
absorption of endotoxins and inhibit anaerobic bacterial growth. degradation product levels before surgery are at increased risk
Endotoxin also causes renal vasoconstriction and redistribution for hemorrhagic complications. In addition to problems with
of renal blood flow away from the cortex, and disturbances in endotoxemia, patients with coexisting cirrhosis often have addi-
coagulation that include the activation of complement, macro- tional problems related to thrombocytopenia from hypersplen-
phages, leukocytes, and platelets (Hunt et al, 1982). As a ism and fibrinolysis. Portal hypertension in patients with
result, glomerular and peritubular fibrin is deposited. This cirrhosis also exacerbates these coagulation disorders.
factor, in combination with reduced renal cortical blood flow,
results in the tubular and cortical necrosis observed in jaun- Immune System
diced patients with renal failure. Surgery in patients with jaundice is associated with a higher
rate of postoperative septic complications compared with those
Coagulation without jaundice, due in large measure to defects in cellular
Disturbances of blood coagulation are commonly present in immunity that make them more prone to infection (see Chap-
jaundiced patients. The most frequently observed abnormality ters 10 and 12). Cainzos and colleagues (1992) have demon-
is prolongation of the prothrombin time. This problem results strated an association between jaundice and altered delayed-type
from impaired vitamin K absorption from the gut, secondary hypersensitivity. Only 16% of 118 patients with jaundice were
to a lack of intestinal bile. This coagulopathy is usually revers- immunocompetent, compared with 76% of 59 healthy con-
ible with parenteral administration of vitamin K. Decreased bile trols, when tested with a panel of seven skin antigens. Several
levels in the small intestine may result in diminished absorption authors have shown impaired T-cell proliferation (Thompson
of other fat-soluble vitamins and fats, which results in weight et al, 1990), decreased neutrophil chemotaxis (Andy et al,
loss and loss of calcium. This latter factor, as well as the earlier- 1992), defective bacterial phagocytosis (Scott-Conner et al,
mentioned increase in circulating endotoxin, may further con- 1993), and suppression of natural killer–cell activity (Lane
tribute to clotting abnormalities. et al, 1996) following bile duct ligation in animals. As men-
In experimental animals, endotoxin affects metabolism of tioned earlier, the ability of the reticuloendothelial system, spe-
factors XI and XII and causes platelet and direct endothelial cifically Kupffer cells, to clear bacteria and endotoxin from the
damage (Hunt et al, 1982). Moreover, endotoxin release in circulation also is reduced in obstructive jaundice. Studies in
patients with jaundice results in a low-grade, disseminated humans also have demonstrated decreased T-lymphocyte pro-
intravascular coagulation with increased fibrin degradation liferation (Fan et al, 1994), decreased expression of adhesion
products. Hunt and colleagues have shown that patients molecules (Plusa e al, 1996), and altered monocyte functions
with jaundice with circulating endotoxin or increased fibrin (Lago et al, 2006).
The absence of bile from the intestinal tract also plays a with extrahepatic bile duct obstruction and 123 nonjaundiced
role in the infectious complications seen in patients with control patients undergoing cholecystectomy. The patients with
obstructive jaundice. Bacterial translocation from the gut is jaundice had only 11% of the skin propylhydroxylase activity
increased in the setting of bile duct obstruction (Deitch et al, of the controls. In the subgroup of patients with jaundice sec-
1990). Obstruction causes a disruption of the enterohepatic ondary to malignancy, propylhydroxylase activity was less than
circulation and results in loss of the emulsifying antiendotoxin 7% of controls. With relief of obstruction, the activity increased
effect of bile acids; therefore a larger pool of endotoxin is to 22% of controls. Interestingly, in patients with jaundice
available within the intestine for absorption into the portal secondary to benign obstruction, the activity increased to 100%
circulation. The combination of reduced or absent bile in of controls.
the intestine and impaired cellular immunity and reticuloen-
dothelial cell function appears to be a major factor contribut- Other Factors
ing to more frequent infective complications in the jaundiced Other problems that face patients with jaundice are anorexia,
patient. weight loss, and malnutrition. Appetite is adversely influenced
Acute cholangitis is a bacterial infection of the biliary ductal by the lack of bile salts in the intestinal tract. In addition,
system, and it varies in severity from mild and self-limited to patients with pancreatic or periampullary malignant lesions
severe and life threatening (see Chapter 43). The clinical triad may have partial duodenal obstruction or abnormal gastric
associated with cholangitis—fever, jaundice, and pain—was emptying, in some cases secondary to tumor infiltration of the
first described in 1877 by Charcot. Cholangitis results from a celiac nerve plexus. Patients with pancreatic or ampullary
combination of two factors: significant bacterial concentrations tumors also may have pancreatic endocrine and exocrine insuf-
in the bile and biliary obstruction. Although bile from the gall- ficiency. This latter problem may further compound other
bladder and bile ducts is usually sterile, in the presence of nutritional defects that, in turn, may exacerbate the immune
common bile duct stones or other obstructing pathology, the deficits of the patient with jaundice.
incidence of positive cultures increases; likewise, instrumenta- Several recent observations suggest that the many physio-
tion of the biliary tree also greatly increases rates of bile colo- logic derangements induced by obstructive jaundice take a long
nization. The most common organisms recovered from the bile time to reverse. For example, Koyama and colleagues (1981)
in patients with cholangitis include Escherichia coli, Klebsiella have shown that hepatic mitochondrial function does not return
pneumonia, the enterococci, and Bacteroides fragilis (Thompson to normal even 7 weeks after relief of obstruction. This same
et al, 1990a). However, even in the presence of high biliary prolonged effect of obstructive jaundice has been noted with
bacterial concentrations, clinical cholangitis and bacteremia lymphocyte, polymorphonuclear, and Kupffer cell function.
will not develop, unless obstruction causes elevated intraductal Therefore even patients who have had temporary relief of
pressures (Lipsett & Pitt, 1993). biliary obstruction via percutaneous or endoscopic stents are
Normal biliary pressures range from 7 to 14 cm H2O. In the likely to remain at risk for significant complications following
presence of bacteribilia and normal biliary pressures, hepatic surgery because derangements in hepatic function are likely still
venous blood and perihepatic lymph are sterile. However, with present at the time of operation. Moreover, a recent analysis by
partial or complete biliary obstruction, intrabiliary pressures Strasberg and colleagues (2014) suggests that preoperative
rise to 20 to 30 cm H2O, and organisms rapidly appear in both jaundice may adversely affect long-term survival in patients
the blood and lymph. The fever and chills associated with with resected pancreatic cancer.
cholangitis are the result of systemic bacteremia caused by
cholangiovenous and cholangiolymphatic reflux. Management
The most common causes of biliary obstruction are choledo- Historically, the only option for the relief of obstructive
cholithiasis (see Chapter 36), benign strictures (see Chapter jaundice was operative intervention; however, with the develop-
42), biliary-enteric anastomotic strictures (see Chapters 31), ment of therapeutic techniques such as percutaneous (see
and periampullary or proximal biliary cancers (see Chapters 49 Chapters 30 and 52) and endoscopic stenting (see Chapter 29),
to 51 and 59). Before 1980, choledocholithiasis was the cause balloon dilation, and endoscopic sphincterotomy, many non-
of approximately 80% of the reported cases of cholangitis. In operative options for the relief of obstructive jaundice are now
recent years, however, malignant strictures have become a fre- available. The surgeon must determine the safest and most
quent cause, especially after the placement of biliary stents. effective therapy for each individual patient and must ade-
Endoscopic cholangiography, percutaneous transhepatic chol- quately prepare each patient for surgery or nonoperative thera-
angiography, and stent placement via either the endoscopic or peutic intervention.
percutaneous route are all known to cause bacteremia, and Patients with obstructive jaundice and those with hepatocel-
these procedures are frequently performed in patients with a lular disease severe enough to cause jaundice are prone to
presumptive diagnosis of malignant obstruction. many secondary problems. Patients with jaundice are at
increased risk for the development of kidney injury, gastroin-
Wound Healing testinal bleeding, infections, and wound complications (see
Delayed wound healing and a high incidence of wound dehis- earlier section on Pathophysiology). Cardiac, pulmonary, and
cence and incisional hernia have been observed in patients with renal function must be considered in every patient undergoing
jaundice undergoing surgery. Patients with obstructive jaundice major abdominal surgery. In addition, special attention must
have decreased activity of the enzyme propylhydroxylase in be focused on the patient with jaundice’s nutritional status,
their skin. Propylhydroxylase is necessary for the incorporation coagulability, immune function, and presence or absence of
of the amino acid proline into collagen, and its activity has been biliary sepsis. Complications related to portal hypertension,
used as a measure of collagen synthesis. Grande and colleagues such as ascites, varices, and encephalopathy, may also develop
(1990) measured skin propylhydroxylase activity in 95 patients in patients with chronic liver disease and cirrhosis, and these
abnormalities may require specific treatment (see Chapters 76 which should include biliary drainage in patients with cholan-
and 79 to 82). gitis in addition to systemic antibiotics.
In cirrhotic patients, clotting abnormalities are often multi-
Cardiopulmonary factorial and include thrombocytopenia secondary to hyper-
In assessing cardiopulmonary status, the patient’s age, history splenism, prolongation of PT and partial thromboplastin time
of recent myocardial infarction, and the presence of congestive (PTT), and fibrinolysis. Vitamin K should be administered if
heart failure, significant valvular heart disease, or a disturbance the PT is prolonged. If no effect is seen and/or if the PTT is
of normal cardiac rhythm all have been correlated with increased also prolonged, fresh frozen plasma should be given. Throm-
operative risk (Goldman et al, 1977). In addition, patients with bocytopenia usually can be managed by intraoperative platelet
severe pulmonary disease may not be candidates for extensive infusions. If the patient has a shortened clot lysis time and
abdominal surgery. hypofibrinogenemia, ε-aminocaproic acid may be indicated.
Renal Pruritus
Patients with jaundice especially those with cirrhosis and chol- Pruritus is often a distressing problem in the jaundiced patient.
angitis, are at increased risk for renal insufficiency. The main- The exact cause of pruritus remains obscure, but increased
tenance of adequate blood volume and the correction of circulating levels of bile salts, histamines, and central nervous
dehydration are extremely important if renal complications are system opiate receptors have been implicated. In some patients,
to be avoided. Fluid management can be quite complex in relief from itching can be obtained by bile salt–binding agents,
patients with jaundice, because both excess and insufficient such as cholestyramine. Various sedatives and antihistamines
fluids may be problematic. Therefore selected patients may also can provide relief of itching in jaundiced patients. However,
benefit from invasive hemodynamic monitoring with central relief of biliary obstruction remains the most effective method
venous catheters and, less commonly, pulmonary artery cath- for managing pruritus, and improvement can occur rapidly
eters, to assist in assessing intravascular volume. after stent placement, although occasionally can take a week
Certain oral bile salts have been shown to be efficacious in or so.
preventing the development of postoperative kidney injury. In
a study by Cahill (1983) 54% of 24 patients with jaundice not Cholangitis
given oral bile salts before surgery were found to have systemic Biliary sepsis also has been identified as a major risk in jaun-
endotoxemia, which was associated with renal impairment in diced patients (see Chapter 43). Cholangitis occurs when
two thirds of the cases. In comparison, none of the 8 patients partial or complete obstruction of the bile duct exists, resulting
with jaundice given 500 mg of sodium deoxycholate every 8 in increased intraluminal pressure and infected bile proximal to
hours for 48 hours before surgery had portal or systemic endo- the obstruction. Patients with cholangitis may present with
toxemia. Moreover, none of these eight patients had evidence right upper quadrant abdominal pain, fever, and/or jaundice
of renal impairment. (Charcot’s triad). Patients with “toxic” cholangitis—Charcot’s
triad plus shock and mental confusion (Reynold’s pentad)—
Nutrition have significant mortality with appropriate antibiotic therapy
Malnutrition is a significant risk factor for surgery in the setting alone and therefore require emergent biliary decompression.
of obstructive jaundice (see Chapter 26). Halliday and col- Gigot and colleagues (1989) identified seven prognostic
leagues (1988) noted that patients who died in the postopera- factors that are indications for urgent biliary decompression:
tive period following surgery for obstructive jaundice had a (1) acute kidney injury, (2) liver abscess, (3) cirrhosis, (4) high
significant reduction in body weight, midarm circumference, malignant stricture, (5) percutaneous transhepatic cholangiog-
total body potassium, and reactivity to skin test antigens prior raphy, (6) female gender, and (7) advanced age. However,
to operation. In a study from Italy (Foschi et al, 1986), enteral emergent surgical treatment is associated with significant mor-
hyperalimentation was found to significantly decrease operative bidity and mortality; therefore both percutaneous and endo-
morbidity and mortality in a group of patients treated with 20 scopic biliary drainage have been proposed as effective therapy
days of preoperative percutaneous biliary drainage. Although for the 5% to 10% of patients with cholangitis who are unre-
most patients with benign biliary problems are adequately sponsive to conservative therapy. Lai and colleagues (1992)
nourished, various degrees of malnutrition are frequently have shown, in a series of 82 patients with severe acute chol-
present in patients with malignant obstruction. Therefore angitis, that endoscopic drainage is associated with a lower
patients with malignant obstructive jaundice should be evalu- morbidity (34% vs. 66%) and mortality (10% vs. 32%) than
ated for evidence of malnutrition and have nutritional support operative drainage. A more complete discussion of the bacteri-
instituted if necessary. ology and antibiotic management of patients with cholangitis
can be found in Chapters 12 and 43.
Coagulation
Patients with obstructive jaundice, cholangitis, or cirrhosis all Preoperative Drainage
are prone to excessive intraoperative bleeding. The most The preoperative relief of jaundice and the reversal of its sys-
common clotting defect in patients with obstructive jaundice is temic effects by either endoscopic or transhepatic biliary
prolongation of the prothrombin time (PT), which is usually decompression has been proposed as a method to decrease the
reversible by the administration of parenteral vitamin K. risk of surgery in jaundiced patients. However, several prospec-
Patients with severe jaundice and/or cholangitis also may tive randomized studies have shown that the routine use of
develop disseminated intravascular coagulation (DIC), which preoperative biliary drainage (PBD) does not reduce operative
may require infusion of platelets and fresh frozen plasma. morbidity or mortality in patients with obstructive jaundice
Reversal of DIC also requires control of the underlying sepsis, (Hatfield et al, 1982; Lai et al, 1994; McPherson et al, 1984;
TABLE 8.3 Prospective Randomized Trials of Preoperative Biliary Drainage (PBD)

First Author, Year Type of Drainage Mortality, PBD Mortality, No PBD Morbidity, PBD Morbidity, No PBD
Hattfield, 1982 Percutaneous 4/29 4/28 7/29 4/28

McPherson, 1984 Percutaneous 11/34 6/31 17/34 13/31
Pitt, 1985 Percutaneous 3/37 2/38 30/37 20/38
Lai, 1994 Endoscopic 6/43 6/44 28/43 18/44
Wig, 1999 Percutaneous 1/20 4/20 12/20 10/20
van der Gaag, 2010 Endoscopic 15/102 12/94 75/102 37/94
Total (%) 40/265 (15.1) 34/255 (13.3) 169/265 (63.8)* 104/255 (40.0)
*P < .001 vs. no PBD.
Pitt et al, 1985; van der Gaag et al, 2010, Wig et al, 1999) patients with periampullary malignancy. Patients with proximal
(Table 8.3). In addition, meta-analyses also concluded that biliary obstruction undergoing major hepatic resection repre-
preoperative biliary drainage increased (P < .001), rather than sent an entirely different subgroup, but a multicenter European
decreased, overall complications from surgery and the drainage study suggests that those requiring right hepatectomy benefit
procedure and provided no benefit in terms of reduced mortal- from preoperative drainage of the future liver remnant (see
ity or decreased hospital stay (Fang et al, 2013, Sewnath et al, Chapters 51 and 103B and C). Preoperatively placed transhe-
2002). In fact, several retrospective studies have documented patic catheters also may be of value in the operating room
a higher incidence of infectious complications (wound infec- during difficult biliary dissections in patients with proximal
tion, pancreatic fistula) and even mortality in patients undergo- biliary tumors or strictures, and they may aid in the placement
ing pancreatic or biliary tract resection after preoperative biliary of long-term transhepatic stents. Finally, preoperative drainage
decompression (Sewnath et al, 2002; Sohn et al, 2000). is required in patients receiving neoadjuvant therapy. In these
A criticism of some of the prospective studies is that the patients, metal stents have fewer problems with cholangitis at
duration of preoperative drainage (10 to 18 days) may have no additional cost (Walter et al, 2015).
been inadequate to reverse the multiple metabolic and immu-
nologic abnormalities associated with severe obstructive jaun- SUMMARY
dice. Both animal and human studies demonstrate that the
recovery of various metabolic and immune functions requires Over the past few decades, tremendous strides have been made
at least 6 weeks to pass after the relief of biliary obstruction in our understanding of bile secretion and our ability to care
(Clements et al, 1993; Kennedy et al, 1994; Thompson et al, for the jaundiced patient. Clinicians now have a better under-
1990b). Similarly, animal studies strongly suggest that return standing of normal bile salt, biliary lipid, and bilirubin physiol-
of bile to the intestinal tract has significant advantages over ogy and can classify the diseases that cause jaundice as defects
external biliary drainage (Roughneen et al, 1986). in normal metabolism. Similarly, investigators have elucidated
Although the data suggest that routine PBD may be of the multiple pathophysiologic effects of jaundice that explain
limited benefit, PBD may have some value in selected patients why jaundiced patients are at risk for increased perioperative
with advanced malnutrition, biliary sepsis, and hilar malignan- morbidity and mortality.
cies that require liver resection (Farges et al, 2013). Regarding
the latter, most of the published data regarding PBD focus on References are available at expertconsult.com.
Chapter 8 Bile secretion and pathophysiology of biliary tract obstruction 132.e1
REFERENCES Kullak-Ublick GA, et al: Enterohepatic bile salt transporters in normal

physiology and liver disease, Gastroenterology 126:322–342, 2004.
Andy OJ, et al: Peritoneal neutrophil chemotaxis is impaired in biliary Lago MT, et al: Phagocytosis and production of H2O2 by human
obstruction, Am Surg 58:28–31, 1992. peripheral blood mononuclear cells from patients with obstructive
Arrese M, Accatino L: From blood to bile: recent advances in hepato- jaundice, Pancreatology 6:273–278, 2006.
biliary transport, Ann Hepatol 1:64–71, 2002. Lai ECS, et al: Endoscopic biliary drainage for severe acute cholangitis,
Blenkharn JI, et al: Decreased biliary excretion of piperacillin after N Engl J Med 326:1582–1586, 1992.
percutaneous relief of extrahepatic obstructive jaundice, Antimicrob Lai ECS, et al: Preoperative endoscopic drainage for malignant
Agents Chemother 28:778–780, 1985. obstructive jaundice, Br J Surg 81:1195–1198, 1994.
Cahill CJ: Prevention of postoperative renal failure in patients with Lane DR, et al: Suppression of natural killer cell activity in biliary
obstructive jaundice: the role of bile salts, Br J Surg 70:590–595, obstruction, Ann Surg 62:259–262, 1996.
1983. Lee MH, et al: Identification of a gene, ABCG5, important in the regu-
Cainzos M, et al: Hyperbilirubinemia, jaundice and anergy, Hepatogas- lation of dietary cholesterol absorption, Nat Genet 27:79–83, 2001.
troenterology 39:330–332, 1992. Li T, Chiang J: Bile acids as metabolic regulators, Curr Opin Gastroen-
Clements WD, et al: Effects of extrahepatic obstructive jaundice on terol 31:159–165, 2015.
Kuppfer cell clearance capacity, Arch Surg 128:200–204, 1993. Lipsett PA, Pitt HA: Biliary infection: prophylaxis and treatment. In
Crawford JM: Role of vesicle-mediated transport pathways in hepato- Toouli J, editor: Surgery of the biliary tract, Edinburgh, 1993,
cellular bile secretion, Semin Liver Dis 16:169–189, 1996. Churchill Livingstone, pp 59–70.
Cui Y, et al: Hepatic uptake of bilirubin and its conjugates by the Lumlertgul D, et al: The jaundiced heart: evidence of a blunted
human organic anion transport SLC21A6, J Biol Chem 276:9626– response to positive inotropic stimulation, Ren Fail 13:15–22, 1991.
9630, 2001. Ma Z, et al: Differential effects of jaundice and cirrhosis on
Deitch EA, et al: Obstructive jaundice promotes bacterial translocation β-adrenoceptor signaling in three rat models of cirrhotic cardiomy-
from the gut, Am J Surg 159:79–84, 1990. opathy, J Hepatol 30:485–491, 1999.
Elferink O, Groen AK: Mechanisms of biliary lipid secretion and their McPherson GAD, et al: Preoperative percutaneous transhepatic biliary
role in lipid homeostasis, Semin Liver Dis 20:293–305, 2000. drainage: the results of a controlled trial, Br J Surg 71:371–375,
Fan ST, et al: T lymphocytces with malignant biliary obstruction, 1984.
J Gastroenterol Hepatol 9:391–395, 1994. Meier PJ, Stieger B: Bile salt transporters, Annu Rev Physiol 64:635–
Fang Y, et al: Meta-analysis of randomized clinical trials on safety and 661, 2002.
efficiency of biliary drainage before surgery for obstructive jaundice, Nehez L, Anderson R: Compromise of immune function in obstructive
Br J Surg 100:1589–1596, 2013. jaundice, Eur J Surg 168:315–328, 2002.
Farges O, et al: Multicenter European study of preoperative Padillo J, et al: Improved cardiac function in patients with obstructive
biliary drainage for hilar cholangiocarcinoma, Br J Surg 100:274– jaundice after internal biliary drainage: hemodynamic and hormonal
283, 2013. assessment, Ann Surg 234:652–656, 2001.
Fogarty BJ, et al: Renal dysfunction in obstructive jaundice, Br J Surg Patel T, et al: Increase in intracellular magnesium promotes
82:877–884, 1995. glycodeoxycholate-induced apoptosis in rat hepatocytes, J Clin Invest
Foschi D, et al: Hyperalimentation of jaundiced patients on percutane- 94:2183–2192, 1994.
ous transhepatic biliary drainage, Br J Surg 73:716–719, 1986. Pitt HA, et al: Does preoperative percutaneous biliary drainage reduce
Gerk PM, Vore M: Regulation of expression of the multidrug resistance- operative risk or increase hospital cost?, Ann Surg 201:545–552,
associated protein 2 (MRP2) and its role in drug disposition, 1985.
J Pharmacol Exp Ther 302:407–415, 2002. Plusa S, et al: Obstructive jaundice causes reduced expression of poly-
Gigot JF, et al: Acute cholangitis: multivariate analysis of risk factors, morphonuclear leukocyte adhesion response to bacterial wall prod-
Ann Surg 209:435–438, 1989. ucts in vitro, Gut 38:784–787, 1996.
Goldman L, et al: Multifactorial index of cardiac risk in noncardiac Puntis MCA, Jiang WG: Plasma cytokine levels and monocyte activa-
surgical procedures, N Engl J Med 297:845–850, 1977. tion in patients with obstructive jaundice, J Gastroenterol Hepatol
Grande L, et al: Obstructive jaundice and wound healing, Br J Surg 11:7–13, 1996.
77:440–442, 1990. Roughneen PT, et al: Impaired specific cell-mediated immunity in
Green J, Better OS: Circulatory disturbance and renal dysfunction in experimental biliary obstruction and its reversibility by internal
liver disease and in obstructive jaundice, Isr J Med Sci 30:48–65, biliary drainage, J Surg Res 41:113–125, 1986.
1994. Scott-Conner CE, et al: Impaired bacterial killing in early obstructive
Halliday AW, et al: Nutritional risk factors in major hepatobiliary jaundice, Am J Surg 166:308–310, 1993.
surgery, JPEN J Parenter Enteral Nutr 12:43–48, 1988. Sewnath ME, et al: Meta-analysis on the efficacy of preoperative biliary
Hatfield AR, et al: Preoperative external biliary drainage in obstructive drainage for tumors causing obstructive jaundice, Ann Surg 236:17–
jaundice, Lancet 2:896–899, 1982. 27, 2002.
Henkel AS, et al: Hepatic overexpression of Abcb 11 in mice promotes Sohn TA, et al: Do preoperative biliary stents increase
the conversion of bile acids within the entrohepatic circulation, Am postpancreaticoduo-denectomy complications?, J Gastrointest Surg
J Physiol Gastrointest Liver Physiol 304:G221–G226, 2013. 4:258–268, 2000.
Holm R, et al: Bile salts and their importance for drug absorption, Int Strasberg SM, et al: Jaundice: an important, poorly recognized risk
J Pharm 453:44–55, 2013. factor for diminished survival in patients with adenocarcinoma of the
Hunt DR, et al: Endotoxemia, disturbance of coagulation, and obstruc- head of the pancreas, HPB (Oxford) 16:150–156, 2014.
tive jaundice, Am J Surg 144:325–329, 1982. Swann JR, et al: Systemic gut microbial modulation of bile acid metab-
Iganagi T, et al: Biochemical and molecular aspects of genetic disorders olism in host tissue compartments, Proc Natl Acad Sci USA
of bilirubin metabolism, Biochim Biophys Acta 1407:173–184, 1990. 108:4523–4530, 2011.
Karsten TM, et al: Morphological changes of extrahepatic bile ducts Swartz-Basile DA, et al: Leptin regulates gallbladder genes related to
during obstruction and subsequent decompression by endoprosthe- absorption and secretion, Am J Physiol 293:84–90, 2007.
sis, Surgery 111:562–568, 1991. Thompson JE, et al: Broad spectrum penicillin as adequate therapy for
Kennedy JA, et al: Modulation of immune function and weight loss by acute cholangitis, Surg Gynecol Obstet 171:275–282, 1990a.
L-arginine in obstructive jaundice in the rat, Br J Surg 81:1199– Thompson RL, et al: Development and reversibility of T lymphocyte
1201, 1994. dysfunction in experimental obstructive jaundice, Br J Surg 77:1229–
Klein A, et al: Liver, biliary tract, and pancreas. In O’Leary J, editor: 1232, 1990b.
Physiologic basis of surgery, Baltimore, 1996, Wilkins & Williams, Van der Gaag NA, et al: Preoperative biliary drainage for cancer of the
pp 441–478. head of the pancreas, N Engl J Med 362:129–137, 2010.
Koeppel TA, et al: Extrahepatic biliary obstruction impairs microvas- Walter D, et al: Cost efficacy of metal stents for palliation of extrahe-
cular perfusion and increases leukocyte adhesion in rat liver, Hepatol- patic bile duct obstruction in a randomized controlled trial, Gastro-
ogy 26:1085–1091, 1997. enterology 149:130–138, 2015.
Koyama K, et al: Experimental and clinical studies on the effect of Wig JD, et al: Usefulness of percutaneous tranhepatic biliary drainage
biliary drainage in obstructive jaundice, Am J Surg 142:293–299, in patients with surgical jaundice: a prospective randomized study,
1981. J Assoc Physicians India 47:271–274, 1999.
CHAPTER 9A
Advances in the molecular characterization
of liver tumors
Matthias S. Matter and Jesper B. Andersen
OVERVIEW
transforming growth factor-β (TGF-β), promote cholangiocyte
Primary liver cancers are among the most rapidly evolving growth along with accrual of molecular aberrations. Impaired
malignant tumors worldwide. An underlying chronic inflamma- deoxyribonuclease (DNA) damage repair (DDR) and tumor
tory liver disease, which precedes liver cancer development for suppressor genes, as well as activation of protooncogenes, cause
several decades and creates a protooncogenic microenviron- malignant transformation together with a deregulated plethora
ment, frequently impairs progress in therapeutic approaches. of key signaling pathways, for instance, cyclooxygenase-2
Depending on the cellular target of malignant transformation, (COX2), phosphatidylinositol-3-kinase (PI3K)/AKT/mTOR,
a large spectrum of molecular and morphologic patterns is (PI3K: Phosphoinositide 3-kinase; AKT: v-akt murine thymoma
observed. As such, it is crucial to advance our existing under- viral oncogene homolog 1/Protein kinase B; mTOR: Mechanis-
standing of the molecular pathogenesis, particularly the genomic tic Target of Rapamycin) epidermal growth factor receptor
heterogeneity of these malignancies, to improve current clinical (EGFR), ERB-B2 receptor tyrosine-kinase receptor (ERBB2),
strategies and patient outcome. This has been achieved for RAS/RAF/mitogen-activated protein kinase (MAPK), hepato-
other cancers, such as breast carcinoma, facilitated by the delin- cyte growth factor, Janus activating kinase/signal transducer and
eation of patient subsets and individualization for precision activator of transcription (JAK/STAT), IL-6, and vascular
therapies. In hepatobiliary cancers, many questions persist as endothelial growth factor (VEGF) (Andersen, 2015; Andersen
to the evolutionary process and cellular origin of the initial & Thorgeirsson, 2014; Marquardt & Andersen, 2014). Interac-
transforming incident, the context of tumor plasticity, and the tion between the epithelial and stromal compartments contrib-
causative features driving the disease. Next-generation sequenc- utes to uncontrolled proliferation, survival, angiogenesis,
ing (NGS) and other novel molecular as well as cytologic/ invasion, and metastasis.
histologic techniques have begun to underline the persistent Many studies have evaluated the role of single genes in CCA
alterations, which may trigger the inherited drug resistance (a (particularly using immunohistochemical techniques), includ-
hallmark of liver, biliary tract, and pancreatic tumors), metas- ing their potential impact in prognosis and as diagnostic markers
tasis, and disease recurrence. However, a complex issue remains (Andersen, 2015; Andersen & Thorgeirsson, 2014; Sia et al,
for interpreting the heterogeneous pool of co-occurring aberra- 2013b). Recent integrative genomics approaches (microarray-
tions in the tumor genotype, in which chromosomal proximity based studies) have broadly assessed molecular factors in CCA
to a causal aberration likely influences, for example, the thera- (Andersen et al, 2012; Oishi et al, 2012; Sia et al, 2013a; Wang
peutic response. In this chapter, we will present a comprehen- et al, 2013). A 238-gene signature was shown to have clinical
sive overview of the molecular achievements that currently are relevance correlating with a high-risk group of patients and
advancing the characterization of hepatobiliary cancers. We will predicted overall and recurrence-free survival (Andersen et al,
focus on biliary tract cancer, that is, cholangiocarcinoma 2012). This prognostic gene signature included 36 independent
(CCA), a broadly treatment-refractory tumor type whose survival genes, which mostly are cholangio specific and repre-
molecular pathogenesis is very poorly understood (see also sent the β-catenin/myelocytomatosis, tumor necrosis factor,
Chapters 9B, 9C, and 9D). We will discuss how genomic and vascular endothelial growth factor receptor (VEGFR)/
achievements can be applied clinically to advance diagnosis and ERBB pathways.
therapy, as well as ultimately improve patient outcome.
Chromosomal Aberrations
MOLECULAR ALTERATIONS IN Several comparative genomic hybridization (array CGH)
CHOLANGIOCARCINOMA studies have been used to examine the chromosomal imbalance
of hepatobiliary tumors. A meta-analysis of five studies ana-
The pathogenesis of CCA is a multistep process (Andersen, lyzed 98 intrahepatic CCA (iCCA) cases and identified common
2015) that often originates in a state of chronic biliary inflamma- chromosomal gains at 1q, 5p, 7p, 8q, 17q, and 20q, as well as
tion and, consequently, biliary damage (cholestasis) triggered by losses at 1p, 4q, 8p, 9p, 17p, and 18q (Andersen & Thorgeirs-
external stimuli. Obstruction of the bile acid results in aberrant son, 2012). McKay and colleagues (2011) analyzed 32 patients
cytokine signaling and promotes cholangiocyte proliferation (7 iCCAs, 13 perihilar cholangiocarcinomas [pCCA] and 12
though the activation of growth factors. Release of proinflam- distal cholangiocarcinomas [dCCA], and they found common
matory mediators, for instance, interleukin-6 (IL-6) and gains among the CCA subtypes at 16q, 17p, 17q, 19p, and
133
19q. Interestingly, these chromosomal regions encode two (19%), and PBRM1 (polybromo 1) (17%) (Jiao et al, 2013).
important genes: ERBB2/HER2 (chr17q12) and MAP2K2/ Interestingly, the authors determined that 47% of iCCA cases
MEK2 (chr19p13) in the development of hepatobiliary tumors have somatic alterations in at least one chromatin-modifying
and represent favorite drug targets in many clinical trials. In gene, which is not mutually exclusive, and further suggests a
other diseases, single nucleotide polymorphisms (SNPs) have continuum of epigenetic changes. As such, patients with a
been implicated as diagnostic and prognostic markers. Large- mutated IDH gene were found to have a significantly reduced
scale genome-wide association studies in CCA are limited; 3 year survival rate of 33% compared with 81% for patients
however, several such projects have been completed both with wild-type IDH status. These patients, including patients
in hepatocellular carcinoma (HCC) (Li et al, 2012b) and with mutations in chromatin-remodeling factors, may, as a
pancreatic cancer, for instance, the PanScan consortium result of their tumor genotype, experience an altered sensitivity
(Amundadottir et al, 2009; Petersen et al, 2010). Using SNP to drugs, for instance, demethylating agents and histone deacet-
arrays, Sia and colleagues (2013a) analyzed 149 iCCA cases ylase inhibitors.
and identified several focal copy number alterations, including In a recent study that focused on patients with Chinese eth-
gains at 1q and 7p as well as losses at 3p, 4q, 6q, 8p, 9p, 9q, nicity, the authors sequenced the complete exome of seven
13q, 14q, 17p, and 21q. In all, these data show common iCCAs and their surrounding nontumoral tissues (Gao et al,
genetic trends, such as gains on the chromosomal arms 1q and 2014). Interestingly, these patients were all selected as treat-
7p, as well as a tendency for a structural overlap with HCC, ment naïve and, besides, diagnosed without any background
with a gain at 1q, 8q, 11q, and 17q and a loss at 4q, 8p, 13q, liver or biliary diseases. The study showed a range of 7 to 192
and 17p (Andersen & Thorgeirsson, 2012; Marquardt & mutations per tumor, which is comparable to other current
Andersen, 2012). studies. However, the authors found a prevalence of transver-
sions (A : C or G : T variants) compared with transitions (A : G
Detection of Somatic Mutations Following the Era of or C : T variants) at a ratio of 3.7 : 1, including a predominant
Deep Sequencing change of G/C nucleotides. Interestingly, transversions are also
In CCA, deep-sequencing approaches were first used to most frequently described in HCC (Guichard et al, 2012;
describe the genetic variants of liver fluke-related tumors (Ong Huang et al, 2011), suggesting a common origin of at least a
et al, 2012). The authors performed whole-exome sequencing subset of primary liver tumors. A key finding in this study, which
of eight fluke-related tumors and the matched normal tissues. included a prevalence screening of 124 tumors, was a high
They validated their results using Sanger sequencing for a total incidence of activating somatic mutations in the gene for protein
of 206 somatic mutations, affecting 187 genes. The average tyrosine phosphatase nonreceptor type 3 (PTPN3) (41%) (Gao
number of nonsynonymous (protein-altering) variants was 26 et al, 2014), associated with increased expression of the PTPN3
per tumor, ranging from 19 to 34. The predominant somatic protein and disease recurrence. Recurrent alterations were
substitution was a C : G > T : A transition (COSMIC database: found in nine members (PTPRB, PTPRQ, PTPRS, PTPRZ1,
333/545, 61% of cases). Frequent somatic mutations were SBF1, SBF2, MTMR3, and EYA1) of the PTP family, which
found in genes such as TP53 (44.4%), KRAS (16.7%), and increased the overall mutational frequency to 51.6%, and
SMAD4 (16.7%). Additionally, alterations in 10 previously making the PTP family the most affected pathway in iCCA.
unrecognized genes included inactivating mutations in MLL3, Recently, Morris and colleagues (2011) showed that a deletion
ROBO2, RNF43, and PEG3, as well as activating mutations in in protein tyrosine phosphatase receptor type S (PTPRS) led to
the GNAS oncogene. These genes are involved in histone mod- the activation of EGFR and PI3K, two pathways often deregu-
ification, genomic instability, and G protein–coupled signaling. lated in iCCA, implying that the PTP family may represent an
Analysis of 15 non–fluke-associated CCAs (10 iCCAs and 5 important therapeutic target (Table 9A.1).
extrahepatic CCAs) showed, when compared with fluke-related Direct analysis of targeted regions, hot-spot mutations, or
CCA, a distinct genetic pattern that correlated with this panels of cancer-related genes are likely more relevant for
regional-specific risk factor (Chan-On et al, 2013). The study implementation to the clinic. A recent target-specific exome
also included a prevalence screen of 108 CCAs associated with sequencing approach (hybridization-capture libraries) was used
fluke infestation and 101 cases of non–fluke-related etiology. to evaluate 182 cancer-related genes from 28 formalin-fixed
The mean somatic mutational rate of non–fluke-infested CCA iCCA cases (Ross et al, 2014). This study concluded that
was 16 per tumor (range, 1 to 62), a significantly lower muta- actionable genomic aberrations were found in two thirds of the
tional burden compared with fluke-infested tumors, with an examined tumors, suggesting that there is, in at least 67% of
average of 26 somatic variants. In part, this may explain the patients, a potential for genomics-driven approaches to either
elevated risk of disease caused by the parasitic infection. Impor- individualize or influence clinical decision making. Moreover,
tantly, Chan-On and colleagues (2013) identified a significant a NGS-based analysis of 75 CCA cases performed in hot-spot
degree of variation in chromatin-modifying genes, such as regions, representing 236 genes, classified the identified altera-
BAP1 (BRCA1-associated protein 1) (10%) and ARID1A (AT- tions into a few select cellular pathways, including MAPK
rich interaction domain 1A) (10%), as well as confirmed the (35% to 55%), mechanistic target of rapamycin (mTOR) (25%
presence of key somatic mutations in the genes IDH1 (isocitrate to 40%), DNA repair (16% to 45%), chromatin remodeling
dehydrogenase 1) and IDH2. Besides these genetic variants, (15% to 33%), and fibroblast growth factor (5% to 13%) sig-
recurrent mutations were identified in genes AGPAT6, ATP10A, naling (Churi et al, 2014). Interestingly, patients enrolled in
BRPF3, CCT8L2, GPR112, HMCN1, and LRRIQ1. Analysis this study were referred to best-fit clinical trials based on the
of an additional 32 iCCAs, with a mean variant rate of 39 genetic aberrations, including seven cases with wild-type KRAS
somatic mutations per genome (range, 13 to 300), identified (sustained stable disease on erlotinib), one patient with a KRAS
inactivating mutations in multiple chromatin-remodeling mutation (sustained stable disease on pazopanib and tra-
factors, with a high prevalence in BAP1 (25%), ARID1A metinib), two patients with B-Raf proto-oncogene, serine/
Chapter 9A Advances in the molecular characterization of liver tumors 135
TABLE 9A.1 Key Somatic Variants in Liver, Biliary, and Pancreatic Tumors
Target Genes CCA HCC PDAC||
Telomere Maintainence
TERT promoter nr 20%-60% nr
WNT/β-Catenin
CTNNB1, AXIN1/2, APC 0% 2%-35% 9%
Cell Cycle and DNA Repair
TP53 7%-36% 18%-35% 41%-85%
CKN2A/B, CDKs, RB1, ATM 4%-7% 4%-12% 3%-25%
IRF2 0% 0%-5% nr
Apoptosis
TNFRSF10A/B, TRADD, CASP, XIAP, 0%-21% 8%-20% nr
MCL1*
Chromatin-Remodeling Factors
ARID1A 10%-36% 10%-16% 0%-9%
ARID2 nr 2%-18% nr
BAP1 11%-20% nr 4%-25%
MLLs 0%-17% 1%-8% 8%-9%
PBRM1 0%-13% 0%-3.5% 0%-2%
PI3K/mTOR/RAS Pathway
ERBB1-3 0%-25% 0%-10% nr
KRAS 0%-40% 0%-1.5% 90%-100%
NRAS 7%-17% nr nr
HRAS nr 0%-1% nr
RPS6KA3 nr 0%-9% nr
PIK3CA, PIK3C2G 4%-17% 0%-2% 0%-2%
PTEN 7%-11% 0%-2% 0%-9%
TSC1 0%-4% 0%-1% 0%-9%
Epigenetics
IDH1, IDH2, IDH3A 0%-36% 0%-1% nr
Protein Tyrosine Phosphatase
PTPN family 41%-52% nr 0%-2%
FGF Pathway
FGF19 nr 4%-15% nr
FGFR2 0%-14% 0%-2% nr
FGFR2 fusion gene products† 50% nr nr
JAK/STAT Pathway‡
JAK1 nr 0%-9% 0%-17%
IL6R nr 0%-26% nr
IL6ST nr 2%-3% nr
Oxidative Stress Response
NRF2 nr 0%-8% nr
KEAP1 nr 0%-6% nr
TGF-β/SMAD Pathway
TGF-β, TGFBR1/2 0%-4% 0%-1% 0%-6%
§
SMAD4 0%-17% 0%-1% 20%-27%
*Gene amplification found in iCCA.
†
FGFR2 gene fusions with BICC1, KIAA1598, AHCYL1, MGEA5 in iCCA etc.
‡
No somatic variants were reported in iCCA; however, the pathway is commonly activated in more than 50% of iCCA cases.
§
SMAD4 mutations reported in iCCA associated with liver fluke infestation.
||
Overlap of somatic mutations observed in pancreatic cancer with hepatobiliary tumors. Key mutations in PDACs are KRAS, TP53, and SMAD4).
CCA, Cholangiocarcinoma; HCC, hepatocellular carcinoma; iCCA, intrahepatic cholangiocarcinoma; nr, somatic mutation not reported; PDACs, pancreatic ductal adenocarcinomas.
threonine kinase (BRAF) mutations (reached a partial response 2011). A genome-wide analysis of 28 CCA, which utilized the
on a BRAF inhibitor) and one patient received a c-MET inhibi- Illumina 27K methylation array (Illumina, San Diego), identi-
tor and experienced a significant metabolic response according fied a total of 1610 differentially expressed CpG sites compared
to the fluorodeoxyglucose/positron-emission tomography scan. with adjacent normal tissue (Sriraksa et al, 2013). These CpG
In contrast, two patients with ERBB2 mutations who were loci involved 603 hypermethylated genes, including the homeo-
treated with either trastuzumab or lapatinib experienced no box (HOX) genes and targets of the polycomb repressive
drug efficacy. The limited drug effect may suggest the need for complex 2 (PRC2), EED, and SUZ12, as well as histone meth-
combinatorial therapy to achieve a successful response, inhibit- ylation. In a comprehensive genome-wide analysis of differen-
ing signaling pathways downstream of the primary target, such tially methylated regions (DMRs), Goeppert and colleagues
as the MAPK, PI3K/AKT/mTOR, or prostaglandin E2 (PGE2)/ (2014) evaluated 10 iCCAs and 8 extrahepatic CCAs. Interest-
COX2 pathways. ingly, the authors found a significant enrichment of genes
involved in WNT, TGFb, PI3K, MAPK, and NOTCH signal-
Translocations ing, key pathways commonly altered in iCCA.
Recent NGS data have shown a high prevalence of FGFR2 gene Frequent mutations have been shown in both IDH1 and
fusions reported in CCA, including FGFR2-BICC1 (2/4) (Wu IDH2 (Borger et al, 2012; Wang et al, 2013). Mutations in
et al, 2013), FGFR2-KIAA1598 (1/28) (Ross et al, 2014), IDH are associated with hypermethylation enriched in CpG
FGFR2-TACC3 (1/28) (Ross et al, 2014), FGFR2-AHCYL1 promoter shores, which suggests a global deregulation within
(7/66) (Arai et al, 2014), and FGFR2-MGEA5 (1/6) (Borad the transcriptional program (Wang et al, 2013). In support,
et al, 2014; Graham et al, 2014). Interestingly, Borad and IDH1 was recently emphasized as an epigenetic rheostat that,
colleagues (2014) identified FGFR2 gene fusions in 50% when mutated, was proposed to reshape the genomic landscape
(3/6) of iCCA cases. This is currently the only study to include with a global consequence on the transcriptional machinery,
both transcriptome (RNAseq) and whole-genome sequencing triggering an altered state in the cellular process of differentia-
of tumors from the same patient. Arai and colleagues (2014) tion (Turcan et al, 2012). Importantly, mutations in IDH were
concluded from a cohort of more than 100 CCAs that the recently shown to cause the deregulation of hepatocyte nuclear
FGFR2 rearrangements occurred in approximately 14% (9/66) factor 4a (HNF4a), blocking hepatocytic differentiation and
intrahepatic cases. Interestingly, FGFR2 leads to the activation thus promoting bile duct cancer (Saha et al, 2014).
of MAPK, a pathway activated in the majority of CCA cases. The contribution of chronic liver inflammation alongside an
These data underscore the need to evaluate the efficacy of aberrant epigenetic landscape provides survival signals to the
FGFR inhibitors, such as BGJ398 and PD173074, in CCA. tumor. The cyclin-dependent kinase inhibitor p16 is frequently
Importantly, it was shown that FGFR2 translocations can be silenced by epigenetic modifications and was proposed as a
detected by fluorescence in situ hybridization (FISH) (Graham prognostic marker in patients with primary sclerosing cholan-
et al, 2014), making it a possible diagnostic tool. gitis (Ahrendt et al, 1999), suggesting that this change is an
early event in the malignant process. Also, inflammatory signals,
CONSEQUENCE OF EPIGENETIC ALTERATIONS for instance, v-akt murine thymoma viral oncogene homolog 1/
Protein kinase B (AKT), can trigger the release of cytokines
An aberrant epigenome is a common hallmark of human through the JAK/STAT pathway and promote tumorigenesis
cancers, including the molecular pathogenesis of hepatobiliary (Isomoto et al, 2005; Kobayashi et al, 2005). In fact, the sup-
tumors. Epigenetic mechanisms involved in gene regulation pressor of cytokine signaling 3 (SOCS3), which is silenced by
include histone modification, DNA methylation, and noncod- promoter hypermethylation, is a common target in CCA
ing RNAs. There is limited information related to the impact (Isomoto et al, 2007). Reduced expression of SOCS3 results
of altered histone modifications in the transformation of cholin activation of IL-6/STAT3 signaling and a subsequent auto-
angiocytes as well as the response of CCA to epigenetic-based crine feedback loop, as well as paracrine signaling within the
therapies. However, in the past few years, NGS studies have tumor microenvironment.
highlighted the likely importance of chromatin-remodeling
factors as well as IDH in hepatobiliary cancer development. Alterations in the Noncoding RNA Landscape
MicroRNAs (miRs) function as critical rheostats of the genome,
Aberrant DNA Cytosine-Guanine Dinucleotide regulating key properties such as cell survival, autophagy, stem-
(CpG) Methylation ness, and response to therapy.A number of studies have inves-
Promoter hypermethylation has been shown in numerous tigated the biologic significance and aberrant expression of
important cancer-associated genes in CCA (Isomoto, 2009; miRs in CCA cell lines (Braconi et al, 2010; Meng et al, 2006;
Sandhu et al, 2008). Epigenetic silencing of tumor suppressor Meng et al, 2008; Mott et al, 2007), demonstrating a link
genes, such as p16INK4a/CDKN2A (17% to 83%), SOCS3 between tumor growth, response to therapy, and expression of
(62%), RASSF1A (31% to 69%), p15 (54%), CDH1 (17% to inflammatory cytokines. Kawahigashi and colleagues (2009)
49%), hMLH1 (19% to 45%), APC (27% to 47%), p14ARF provided the first comprehensive profile of differentially
(19% to 30%), and GSTP1 (15% to 31%) are among some of expressed miRs in iCCA-derived cells (HuCCT1 and MEC)
the most frequent events. compared with normal bile duct epithelial cells (HIBEpiC). A
Studies of DNA methylation have been restricted to select unique 27-miR signature was identified, which included down-
candidate gene approaches. A recent study analyzed 26 selected regulation of 8 miRs (i.e., miR22, miR125a, miR127, miR199a,
genes in 102 liver fluke-related iCCAs and found a unique miR199a*, miR214, miR376a, and miR424) specific to the
promoter hypermethylation of DcR1/Dicer1 and the gene normal bile duct epithelium. (If a microRNA from each arm
encoding: opioid binding protein/cell adhesion molecule-like [-3p or 5p] of the same pre-miRNA hairpin exists in uneven
(OPCML), compared with adjacent tissue (Sriraksa et al, amounts, the least abundant miR is denoted with an asterisk.)
Chapter 9A Advances in the molecular characterization of liver tumors 137
In two recent studies, miRs were profiled in iCCA (Chen et al, on chromosomes, and as such, further assist in the cytologic
2009; Selaru et al, 2009). Selaru and colleagues (2009) showed diagnosis by increasing the overall sensitivity (Barr Fritcher
that the expression level of miR21 could distinguish iCCA from et al, 2014). FISH utilizes fluorescently labeled probes that
normal bile ducts. Also, increasing expression of miR21 cor- bind to a complementary specimen DNA and can be visualized
related with a decreased expression of programmed cell death using a fluorescence microscope. FISH probes can be designed
4 and tissue inhibitor of metalloproteinase 3, suggesting that to specifically demonstrate chromosomal aberrations, such as
miR21 may act as an oncomir. This miR was also found to loss or gain of chromosomes, which indicates potential malig-
regulate phosphatase and tensin homolog-dependent activation nant transformation. In addition, FISH probes can be locus
of PI3K signaling, which affects the chemosensitivity of CCA specific and applied to detect structural variation, such as chro-
(Meng et al, 2006). Another study found a cluster of 38 miRs mosomal amplifications (e.g., HER2 or MET) or translocations
differentially expressed in 27 iCCAs, compared with 10 normal (e.g., anaplastic lymphoma kinase–echinoderm microtubule-
cholangiocytes (Chen et al, 2009). Interestingly, hierarchical associated protein-like 4 or FGFR2). In principle, FISH probes
clustering grouped the clinical samples into two clusters, dis- can be randomly assorted and are only limited by the number
tinguishing the patients according to the level of the CA19-9 of fluorescent signals. Studies evaluating the efficacy of FISH
antigen. This marker is a serum-secreted mucin-type glycopro- analysis in cytopathologic diagnosis have mainly been focused
tein commonly used as a predictor for CCA and shown to be on the UroVysion FISH method (Abbott Molecular, Des
associated with poor prognosis after surgical resection (Huang Plaines, IL). This commercial assay was initially developed for
et al, 2004). In addition, the status of vascular invasion was diagnosis of bladder cancer and contains probes directed
significantly correlated with the patient stratification. Recently, against chromosome 3, 7, and 17, as well as a locus-specific
miR200c was found in iCCA shown to prevent epithelial-to- probe for 9p21 (the INK4 locus, where the genes encoding
mesenchymal transition (EMT) (Oishi et al, 2012). Integrative p16/14 and p15 reside). Several publications have demon-
analysis of miR and gene expression profiles linked miR200c strated that the specificity of FISH in the diagnosis of CCA is
to the expression of neural cell adhesion molecule 1 (NCAM1), the same as that of routine cytology, whereas its sensitivity is
a marker of the hepatic stemlike phenotype. Another microRNA significantly higher (Barr Fritcher et al, 2014; Vasilieva et al,
(miR204) was found in Chinese iCCAs and shown to play a 2012). Because the ancillary UroVysion FISH technique was
critical role in controlling EMT by regulating the expression of not developed directly for CCA, it may be worthwhile to design
Slug, as well as E-cadherin and vimentin (Qiu et al, 2013). an assay that takes into account genetic aberrations more per-
Similarly, inhibition of miR214 was shown to promote EMT tinent to CCA, which likely would increase the sensitivity of
in iCCA cell lines (Huh28 and HuCCT1) by directly targeting the cytopathologic diagnosis. Moreover, FISH analysis can be
Twist (Li et al, 2012a). Further studies of “miR-to-gene” regu- combined, for instance, with measurement of the CA19-9
lation are warranted to determine the contribution of miRs in serum level or analysis of mutations in KRAS, to increase
EMT and iCCA development, their predictive and prognostic accuracy in the diagnosis of CCA (Barr Fritcher et al, 2013;
significance, as well as potential in therapeutics. Kipp et al, 2010). Other techniques involve analysis of cells
obtained from ERCP to measure the DNA content by flow
FUTURE DIRECTIONS FOR MOLECULAR cytometry. In addition, digital imaging analysis has been
ANALYSIS OF CCA described and gives a qualitative account of the cellular con-
stituents using spectroscopic data (Rumalla et al, 2001; Ryan
Molecular analysis of CCA is currently not performed routinely and Baldauf, 1994). However, both of these techniques are
in the diagnosis and clinical management of CCA patients experimental but warrant further evaluation to support their
(Bridgewater et al, 2014). However, we may predict that the value in assisting cytopathologic diagnosis.
growing knowledge of the underlying molecular pathogenesis Similar to supporting the diagnosis of CCA, no molecular
of CCA in the near future will have broad implications in clini- analysis is currently performed with the aim to direct manage-
cal practice and therapeutics. ment of CCA patients, whereas genomics in lung or colon
The diagnosis of CCA may be supported by ancillary molec- cancer routinely is performed to reveal predictive markers in
ular analysis, in particular on cytologic specimens. CCA tends the guidance of targeted therapy. However, it is likely just a
to grow longitudinally along the bile duct system, and therefore question of time before molecular aberrations in drugable path-
it is often difficult to obtain a representative biopsy. As a result, ways, such as KRAS, MAPK/MEK, ERBBs, IL-6/STAT3,
cytologic brushings obtained during endoscopic retrograde IDH1/2, FGFR2, and MET, will be routinely tested in the
cholangiopancreatography (ERCP) is the method of choice in clinic; however, it remains unclear if proteomic profiles as mea-
the evaluation of biliary strictures under suspicion of malig- sured by matrix-assisted laser desorption/ionization imaging
nancy. These specimens are rated by a cytopathologist as (1) mass spectrometry will play a role in patient management in
benign or negative for malignancy, (2) equivocal, or (3) positive the near future (Le Faouder et al, 2014).
for malignancy. However, cytopathologic evaluation of biliary
cells to diagnose a CCA has poor clinical sensitivity, which is IMPLEMENTATION OF NEXT-GENERATION
typically below 50% (Barr Fritcher et al, 2014). In contrast, the SEQUENCING TO THE CLINIC
specificity is nearly 100%. The low sensitivity of cytology using
ERCP brushings can be explained by sampling errors and cyto- The inception of NGS technologies and the accompanying
logic interpretation, which can be difficult, particularly in the opportunities have not only generated great opportunities but
background of chronic inflammation and if only a few tumor also considerable challenges for translational applications in
cells are present in the specimen. cancer research. Besides the associated costs, the amount of
FISH is a cytogenetic technique that can be used to detect generated data and depths of analyses currently exceed the
and localize the presence or absence of specific DNA sequences mainstream ability to interpret and subsequently transform the
data into useful clinical information (Stadler et al, 2014). culprit cannot necessarily be assumed in hepatobiliary cancers.
Further, application of these technologies for translational This is a concern for future clinical trial designs. Thus, very
applications requires both training and interaction in multidis- large cohorts are required to obtain a sufficient amount of
ciplinary teams consisting of not only health care providers but material for identification and treatment of patients who share
also scientific specialists, such as molecular biologists and bio- a similar molecular profile.
informaticians, as well as patients themselves. The difficulty of During the past decade, a detailed map has been generated
establishing ethically and clinically valid standards for the of the structural variation in the human cancer genome. This
appropriate use of the NGS technology and integration in our map delineates how tumors in general may develop as the
diagnostic pipeline remain a major question and burden consequence of intragenic mutations in roughly 140 genes
(McCarthy et al, 2013). that belong to 12 distinct signaling pathways with common
In primary liver cancer, several problems unique to the cellular processes (cell fate, cell survival, and genome mainte-
disease have to be overcome for successful implementation of nance). Alterations in this framework are major genetic
genomics into clinical routine. First, the application of transla- drivers that promote cellular transformation (Vogelstein et al,
tional genomics requires availability of well-preserved speci- 2013). In the clinic, these genes may represent a tailored
mens in addition to routine pathology, which highlights the NGS panel for routine mutational profiling, which may be
need for mandatory biopsies and, consequently, adaptation of offered as a sensitive diagnostic test and guidance for thera-
current guidelines. Further, molecular and cellular (e.g., peutic decisions.
stromal infiltration) tumor heterogeneity as well as resistance In the past few years, technical progress has greatly advanced
to current treatment modalities might require sequential analy- our understanding of hepatocarcinogenesis by increasing our
sis of the respective genomic landscape to successfully guide knowledge of the molecular complexity and intratumoral het-
individual treatment decisions. The underlying liver disease, as erogeneity. To achieve a paradigm shift toward precision medi-
well as the cellular diversity, might require additional sampling cine in the management of hepatobiliary cancers, multicenter
of the noncancerous microenvironment, which further compli- consortia and international collaborations are crucial for future
cates analysis and, at the end, decisions. success. Given the failure of recent phase 3 trials (Worns and
The spectrum of molecular alterations is highly dependent Galle, 2014), gastroenterology and hepatology will need to
on the ethnicity, etiology, and regional background of the liver reach the “NGS-age.”
disease. As a consequence, significant molecular heterogeneity
is increasingly recognized, corroborating that a clear oncogene References are available at expertconsult.com.
Chapter 9A Advances in the molecular characterization of liver tumors 138.e1
REFERENCES Kawahigashi Y, et al: MicroRNA profiling of human intrahepatic

cholangiocarcinoma cell lines reveals biliary epithelial cell-specific
Ahrendt SA, et al: Chromosome 9p21 loss and p16 inactivation microRNAs, J Nippon Med Sch 76:188–197, 2009.
in primary sclerosing cholangitis-associated cholangiocarcinoma, Kipp BR, et al: Comparison of KRAS mutation analysis and FISH for
J Surg Res 84:88–93, 1999. detecting pancreatobiliary tract cancer in cytology specimens col-
Amundadottir L, et al: Genome-wide association study identifies vari- lected during endoscopic retrograde cholangiopancreatography,
ants in the ABO locus associated with susceptibility to pancreatic J Mol Diagn 12:780–786, 2010.
cancer, Nat Genet 41:986–990, 2009. Kobayashi S, et al: Interleukin-6 contributes to Mcl-1 up-regulation
Andersen JB: Molecular pathogenesis of intrahepatic cholangiocarci- and TRAIL resistance via an Akt-signaling pathway in cholangiocar-
noma, J Hepatobiliary Pancreat Sci 22:101–113, 2015. cinoma cells, Gastroenterology 128:2054–2065, 2005.
Andersen JB, et al: Genomic and genetic characterization of cholangio- Le Faouder J, et al: Tumoral heterogeneity of hepatic cholangiocarci-
carcinoma identifies therapeutic targets for tyrosine kinase inhibitors, nomas revealed by MALDI imaging mass spectrometry, Proteomics
Gastroenterology 142:1021–1031, e1015, 2012. 14:965–972, 2014.
Andersen JB, Thorgeirsson SS: Genetic profiling of intrahepatic chol- Li B, et al: Down-regulation of miR-214 contributes to intrahepatic
angiocarcinoma, Curr Opin Gastroenterol 28:266–272, 2012. cholangiocarcinoma metastasis by targeting Twist, FEBS J 279:
Andersen JB, Thorgeirsson SS: A perspective on molecular therapy in 2393–2398, 2012a.
cholangiocarcinoma: present status and future directions, Hepat Li S, et al: GWAS identifies novel susceptibility loci on 6p21.32 and
Oncol 1:143–157, 2014. 21q21.3 for hepatocellular carcinoma in chronic hepatitis B virus
Arai Y, et al: Fibroblast growth factor receptor 2 tyrosine kinase fusions carriers, PLoS Genet 8:e1002791, 2012b.
define a unique molecular subtype of cholangiocarcinoma, Hepatol- Marquardt JU, Andersen JB: Next-generation sequencing: application
ogy 59:1427–1434, 2014. in liver cancer-past, present and future?, Biology (Basel) 1:383–394,
Barr Fritcher EG, et al: FISHing for pancreatobiliary tract malignancy 2012.
in endoscopic brushings enhances the sensitivity of routine cytology, Marquardt JU, Andersen JB: Liver cancer oncogenomics: Opportuni-
Cytopathology 25:288–301, 2014. ties and dilemma for clinical applications, Hepat Oncol 2:79–93,
Barr Fritcher EG, et al: Primary sclerosing cholangitis with equivocal 2014.
cytology: fluorescence in situ hybridization and serum CA 19-9 McCarthy JJ, et al: Genomic medicine: a decade of successes, chal-
predict risk of malignancy, Cancer Cytopathol 121:708–717, 2013. lenges, and opportunities, Sci Transl Med 5:189sr184, 2013.
Borad MJ, et al: Integrated genomic characterization reveals novel, McKay SC, et al: Array comparative genomic hybridization identifies
therapeutically relevant drug targets in FGFR and EGFR pathways novel potential therapeutic targets in cholangiocarcinoma, HPB
in sporadic intrahepatic cholangiocarcinoma, PLoS Genet 10: (Oxford) 13:309–319, 2011.
e1004135, 2014. Meng F, et al: Involvement of human micro-RNA in growth and
Borger DR, et al: Frequent mutation of isocitrate dehydrogenase response to chemotherapy in human cholangiocarcinoma cell lines,
(IDH)1 and IDH2 in cholangiocarcinoma identified through broad- Gastroenterology 130:2113–2129, 2006.
based tumor genotyping, Oncologist 17:72–79, 2012. Meng F, et al: Epigenetic regulation of microRNA-370 by interleukin-6
Braconi C, et al: MicroRNA-dependent regulation of DNA in malignant human cholangiocytes, Oncogene 27:378–386, 2008.
methyltransferase-1 and tumor suppressor gene expression by Morris LG, et al: Genomic dissection of the epidermal growth factor
interleukin-6 in human malignant cholangiocytes, Hepatology 51: receptor (EGFR)/PI3K pathway reveals frequent deletion of the
881–890, 2010. EGFR phosphatase PTPRS in head and neck cancers, Proc Natl
Bridgewater J, et al: Guidelines for the diagnosis and management of Acad Sci U S A 108:19024–19029, 2011.
intrahepatic cholangiocarcinoma, J Hepatol 60:1268–1289, 2014. Mott JL, et al: mir-29 regulates Mcl-1 protein expression and apopto-
Chan-On W, et al: Exome sequencing identifies distinct mutational sis, Oncogene 26:6133–6140, 2007.
patterns in liver fluke-related and non-infection-related bile duct Oishi N, et al: Transcriptomic profiling reveals hepatic stem-like gene
cancers, Nat Genet 45:1474–1478, 2013. signatures and interplay of miR-200c and epithelial-mesenchymal
Chen L, et al: The role of microRNA expression pattern in human transition in intrahepatic cholangiocarcinoma, Hepatology 56:1792–
intrahepatic cholangiocarcinoma, J Hepatol 50:358–369, 2009. 1803, 2012.
Churi CR, et al: Mutation profiling in cholangiocarcinoma: prognostic Ong CK, et al: Exome sequencing of liver fluke-associated cholangio-
and therapeutic implications, PLoS ONE 9:e115383, 2014. carcinoma, Nat Genet 44:690–693, 2012.
Gao Q, et al: Activating mutations in PTPN3 promote cholangiocar- Petersen GM, et al: A genome-wide association study identifies pan-
cinoma cell proliferation and migration and are associated with creatic cancer susceptibility loci on chromosomes 13q22.1, 1q32.1
tumor recurrence in patients, Gastroenterology 146:1397–1407, and 5p15.33, Nat Genet 42:224–228, 2010.
2014. Qiu YH, et al: miR-204 inhibits epithelial to mesenchymal transition
Goeppert B, et al: Global alterations of DNA methylation in cholan- by targeting slug in intrahepatic cholangiocarcinoma cells, Cell
giocarcinoma target the Wnt signaling pathway, Hepatology 59:544– Physiol Biochem 32:1331–1341, 2013.
554, 2014. Ross JS, et al: New routes to targeted therapy of intrahepatic cholan-
Graham RP, et al: Fibroblast growth factor receptor 2 translocations in giocarcinomas revealed by next-generation sequencing, Oncologist
intrahepatic cholangiocarcinoma, Hum Pathol 45:1630–1638, 2014. 19:235–242, 2014.
Guichard C, et al: Integrated analysis of somatic mutations and focal Rumalla A, et al: Improved diagnostic yield of endoscopic biliary brush
copy-number changes identifies key genes and pathways in hepato- cytology by digital image analysis, Mayo Clin Proc 76:29–33, 2001.
cellular carcinoma, Nat Genet 44:694–698, 2012. Ryan ME, Baldauf MC: Comparison of flow cytometry for DNA
Huang JL, et al: Outcomes after resection of cholangiocellular carci- content and brush cytology for detection of malignancy in pancre-
noma, Am J Surg 187:612–617, 2004. aticobiliary strictures, Gastrointest Endosc 40:133–139, 1994.
Huang Q, et al: RNA-Seq analyses generate comprehensive transcrip- Saha SK, et al: Mutant IDH inhibits HNF-4alpha to block hepatocyte
tomic landscape and reveal complex transcript patterns in hepatocel- differentiation and promote biliary cancer, Nature 513:110–114,
lular carcinoma, PLoS ONE 6:e26168, 2011. 2014.
Isomoto H: Epigenetic alterations associated with cholangiocarcinoma Sandhu DS, et al: Epigenetic DNA hypermethylation in cholangiocar-
(review), Oncol Rep 22:227–232, 2009. cinoma: potential roles in pathogenesis, diagnosis and identification
Isomoto H, et al: Interleukin 6 upregulates myeloid cell leukemia-1 of treatment targets, Liver Int 28:12–27, 2008.
expression through a STAT3 pathway in cholangiocarcinoma cells, Selaru FM, et al: MicroRNA-21 is overexpressed in human cholangio-
Hepatology 42:1329–1338, 2005. carcinoma and regulates programmed cell death 4 and tissue inhibi-
Isomoto H, et al: Sustained IL-6/STAT-3 signaling in cholangiocarci- tor of metalloproteinase 3, Hepatology 49:1595–1601, 2009.
noma cells due to SOCS-3 epigenetic silencing, Gastroenterology Sia D, et al: Integrative molecular analysis of intrahepatic cholangio-
132:384–396, 2007. carcinoma reveals 2 classes that have different outcomes, Gastroen-
Jiao Y, et al: Exome sequencing identifies frequent inactivating muta- terology 144:829–840, 2013a.
tions in BAP1, ARID1A and PBRM1 in intrahepatic cholangiocar- Sia D, et al: Intrahepatic cholangiocarcinoma: pathogenesis and ratio-
cinomas, Nat Genet 45:1470–1473, 2013. nale for molecular therapies, Oncogene 32:4861–4870, 2013b.
Sriraksa R, et al: Aberrant DNA methylation at genes associated with Vogelstein B, et al: Cancer genome landscapes, Science 339:1546–
a stem cell-like phenotype in cholangiocarcinoma tumors, Cancer 1558, 2013.
Prev Res (Phila) 6:1348–1355, 2013. Wang P, et al: Mutations in isocitrate dehydrogenase 1 and 2 occur
Sriraksa R, et al: CpG-island methylation study of liver fluke-related frequently in intrahepatic cholangiocarcinomas and share hyper-
cholangiocarcinoma, Br J Cancer 104:1313–1318, 2011. methylation targets with glioblastomas, Oncogene 32:3091–3100,
Stadler ZK, et al: Cancer genomics and inherited risk, J Clin Oncol 2013.
32:687–698, 2014. Worns MA, Galle PR: HCC therapies-lessons learned, Nat Rev Gas-
Turcan S, et al: IDH1 mutation is sufficient to establish the glioma troenterol Hepatol 11:447–452, 2014.
hypermethylator phenotype, Nature 483:479–483, 2012. Wu YM, et al: Identification of targetable FGFR gene fusions in diverse
Vasilieva LE, et al: Modern diagnostic approaches to cholangiocarci- cancers, Cancer Discov 3:636–647, 2013.
noma, Hepatobiliary Pancreat Dis Int 11:349–359, 2012.
CHAPTER 9B
Molecular pathology of pancreatic cancer and
premalignant tumors
Jaclyn F. Hechtman and Christine Iacobuzio-Donahue
PANCREATIC CANCER OVERVIEW (PanIN-3) that consist of full-thickness dysplasia/carcinoma in

situ (see Fig. 9B.4C). Although low-grade PanIN lesions are
Pancreatic ductal adenocarcinoma (PDA), commonly referred common incidental findings, high-grade PanIN lesions are more
to as “pancreatic cancer” is the fourth most common cause of common in the pancreata of patients with PDA.
cancer in both men and women (American Cancer Society, Key evidence supporting that PanIN lesions are precursors
2014). PDA remains a devastating diagnosis, with an overall to PDA is that similar hallmark molecular defects are found in
survival rate of no greater than 6% (Siegel et al, 2013). The PanIN lesions adjacent to invasive cancers (Biankin et al, 2003;
high mortality of PDA is reflected in statistics from the Ameri- Luttges et al, 2001a; Maitra et al, 2002; van Heek et al, 2002;
can Cancer Society. In 2014 approximately 46,420 Americans Wilentz et al, 2000a). Kirsten rat sarcoma oncogene (KRAS)
were diagnosed with PDA, and approximately 39,590 will die mutations are frequently found in early PanIN lesions, includ-
of it (American Cancer Society, 2014). ing PanIN-1 (Shi et al, 2009), whereas genes involved in
The past two decades have seen an exponential increase in deoxyribonuclease (DNA) repair mechanisms or transforming
our understanding of the molecular basis and etiology behind growth factor-β (TGF-β) signaling, such as for tumor protein
PDA (Hruban et al, 2001; Jones et al, 2008, 2009; Wu et al, 53 (TP53) and SMAD4, respectively, are altered in the latter
2011). Still, the clinical management of this disease, including stages of this progression model (Murphy et al, 2013). It has
primary prevention, early detection, and better targeted treat- also been shown that a higher frequency of PanIN lesions may
ment options, has not changed significantly during the past be found in pancreata of patients with an inherited risk of PDA,
decade. Currently, the only cure for this disease is surgical again supporting the hypothesis that PanINs are true precursors
resection. Unfortunately, only approximately 20% of the patient to PDA (Shi et al, 2009).
population is seen with resectable disease (Allen et al, 2013). Once a PDA has formed, additional genetic changes con-
This chapter aims (1) to show that even though every tumor tinue to occur with time, thereby creating subclones (intratu-
has a number of common molecular events, it is the differences moral heterogeneity) that seed metastases (e.g., peritoneal or
among tumors that have clinical implications and (2) to set the distant) (Yachida et al, 2010). It has been estimated that it
stage for the future, which will include a discussion of success- takes an average of 6.8 years for a parental pancreatic ductal
ful early detection and treatment strategies for this deadly adenocarcinoma (PDAC) clone to give rise to a given meta-
disease. static lesion (Yachida et al, 2013).
Progression Model of Pancreatic Ductal Intraductal Papillary Mucinous Neoplasm
Adenocarcinoma Intraductal papillary mucinous neoplasms (IPMNs) are a well-
In the first era of pancreatic cancer research (Figs. 9B.1 and accepted clinical and pathologic entity (Fig. 9B.4D) (Hruban
9B.2), the fields of molecular biology and pathology combined et al, 2004). IPMNs typically produce radiographically identifi-
to establish a paradigm that PDA culminates from a multistep able pancreatic ductal dilatation, which may predominantly
progression model (Hruban et al, 2001). The second era of involve the main pancreatic ducts (main-duct type IPMN), the
pancreatic research (see Figs. 9B.1 and 9B.2) led to the identi- secondary ducts (branch-duct type IPMN), or both types of
fication of targetable recurrent alterations in PDA and possible ducts (mixed type). The distinction between the branch-duct
targets for therapy (Fig. 9B.3). This slow, sequential process may type and main-duct type is important, because the former are
be the reason PDA is primarily a disease of people in their sixth more likely to involve the head and uncinate process of the
and seventh decades of life. Definable pathologic markers on this pancreas and are associated with lower-grade dysplasia and
stepwise progression, which follows a similar model first devel- fewer invasive carcinomas (Correa-Gallego et al, 2013).
oped in colon carcinogenesis, are lesions referred to as pancreatic Approximately 30% to 40% of resected IPMNs harbor an
intraepithelial neoplasia (PanIN) (Hruban et al, 2001; Figs. invasive adenocarcinoma, and adenocarcinoma is most strongly
9B.4A-C). These lesions are believed to be precursor lesions to associated with main-duct IPMNs. Approximately half of inva-
pancreatic cancer. PanIN lesions are thought to develop years sive carcinomas arising within IPMNs are so-called colloid
before the emergence of PDA and are pathologically graded as (mucinous) carcinomas, and most of the remainder are tubular
low-grade PanIN lesions without cytologic dysplasia (PanIN-1) adenocarcinomas, the latter being histologically indistinguish-
(see Fig. 9B.4A) to intermediate lesions (PanIN-2) with cyto- able from invasive ductal adenocarcinomas that arise in the
logic abnormalities such as pseudostratification, crowding, or setting of PanINs (Adsay et al, 2002). Colloid carcinomas asso-
nuclear enlargement (see Fig. 9B.4B) to high-grade lesions ciated with IPMNs have a relatively good prognosis compared
139
with other pancreatic carcinomas of the ductal type, and have adenocarcinomas (Adsay et al, 2001) and (2) IPMNs have a
a 5 year survival of 60% (Maire et al, 2002). propensity to be multifocal lesions; therefore patients who
PanINs and IPMNs show some overlapping features. For undergo partial pancreatectomy and are left with a remnant
example, both are inherently intraductal lesions composed pre- pancreas need to be followed for life, even when the lesion
dominantly of columnar, mucin-producing cells that may grow originally resected was a noninvasive IPMN (Sohn et al,
in a flat configuration or may produce papillae; these lesions 2004).20
show a range of cytologic and architectural atypia and can give Genetic analyses of IPMNs have disclosed abnormalities in
rise to invasive adenocarcinomas of the pancreas (Fig. 9B.4E). many of the same genes altered in conventional ductal adeno-
An important feature that distinguishes the two lesions is carcinoma, including mutations in the KRAS2, TP53, and
that PanINs are microscopic lesions and IPMNs are macro- CDKN2A genes, although the frequency and stage of neoplastic
scopic. Nevertheless, recognition of an IPMN and its distinc- progression at which these alterations occur in IPMNs differ
tion from a PanIN lesion is important for two reasons: (1) from PanINs (Hruban et al, 2004; Sessa et al, 1994). For
IPMN-associated colloid carcinomas have a significantly better example, contrary to PanINs, IPMNs harbor KRAS mutations
prognosis than either PanIN- or IPMN-associated tubular in only half of analyzed cases (Amato et al, 2014). Moreover,
abnormalities in SMAD4, which are present in 30% of PanIN-3
and 55% of PDA, are rare in IPMNs (Iacobuzio-Donahue et al,
2000). IPMNs may also contain genetic alterations of genes
1st era 2nd era
that are specific to this form of neoplasia. Activating mutations
in GNAS have been found in more than 70% of IPMNs, and
1988 1994 1996 2000 2005 2010
a subset show genetic inactivation of RNF43 (Amato et al,
2014). Interestingly, correlations between phenotypic differen-
tiation of IPMNs (described later) and mutations have been
identified: GNAS mutations are more common in gastric and
KRAS DPC4 FANC gene
p53 intestinal-type IMPNs than in pancreaticobiliary-type IPMNs,
mutations inactivation mutations
whereas KRAS mutations are more common in gastric and
p16 inactivation Sporadic and familial Pancreatic
pancreatic genome carcinoma pancreaticobiliary type IPMNs (Amato et al, 2014). Molecular
BRAC2 inactivation surveyed sequenced testing of pancreatic cyst fluid for GNAS and KRAS mutations
FIGURE 9B.1. Genetic landmark discoveries in pancreatic cancer may help support a diagnosis of IPMN and distinguish it from
research. The first era is highlighted by the discovery of tumor suppres- other cystic lesions, including neuroendocrine tumors with
sor genes such as DPC4 and loss of p16. The second era is highlighted cystic degeneration, benign pseudocysts, and solid and cystic
by genomic profiling of discovery and validating sets of isolated and pseudopapillary neoplasms. However, a negative result does not
purified tumors from pancreatic cancer patients. rule out mucinous cystic neoplasm (Singhi et al, 2014).
first era second era
Primary tumor Generated

cell line
Xenografted 24 discovery tumors
Generated Sequencing Oligonucleotide Gene

cell line arrays expression
analysis
39 genes in Deletions and

DNA analysis: mRNA analysis: another 90 tumors amplifications
Sequencing LOH SAGE arrays
Gene cloning
Discovery of core
signaling pathways
disrupted in PDA
Validation studies:
Functional work
Immunohistochemistry on
clinical specimens
FIGURE 9B.2. Simplified flowchart of genetic discoveries performed in the two eras of genetic research. In the first era (left), resected
pancreatic cancers were xenografted, and genomic DNA was isolated for genetic analysis of a region or putative tumor suppressor or oncogene. In
the second era (right), combining advances in equipment and techniques and high-throughput sequencing allowed a genomic survey of the pancreatic
cancer genome. LOH, Loss of heterozygosity; PDA, pancreatic ductal carcinoma; SAGE, serial analysis of gene expression.
Chapter 9B Molecular pathology of pancreatic cancer and premalignant tumors 141
Surgery Surgery
Standard Treatment (current) Molecular Testing
Gemcitabine
BRCA2//FANC High cyto HuR Hedgehog cellular

signaling pathway
Targeted Treatment (future)

DNA damaging agents:
Mitomycin C
Cisplatin
PARP-inhibitors Gemcitabine IPI-9
Variable Outcomes Improved Outcomes

FIGURE 9B.3. Targeted treatment strategy of PDA. Left, The current strategy. Right, The future. IPI-9 is the abbreviation for IPI-926, an agent
that inhibits the Hedgehog cellular signaling pathway (the stromal component of PDA). cyto, cytoplasmic; HuR, human antigen R; PDA, pancreatic
ductal adenocarcinoma; PARP, poly (ADP-ribose) polymerase. (Courtesy Jennifer Brumbaugh, Thomas Jefferson University, Philadelphia.)
Another distinction between PanINs and IPMNs relates to telomeric abnormalities, and epigenetic silencing by methyla-
the expression of the caudal differentiation factor CDX2, a tion of defined promoter DNA sequences. An individual pan-
marker of intestinal differentiation. Most IPMNs express creatic tumor contains on average 63 genetic alterations,
CDX2, in particular IPMNs associated with an invasive colloid primarily point mutations. Only a small subset of these muta-
carcinoma, whereas this is uncommon both in PanINs and in tions is required for tumorigenesis (Jones et al, 2008).
the subset of IPMNs that give rise to invasive cancers resem- The field of analyzing the genetics of pancreatic cancer can
bling ductal adenocarcinomas (Adsay et al, 2004). CDX2 be broken down chronologically (see Fig. 9B.1). First, land-
expression in IPMNs is generally associated with expression of mark studies starting in the late 1980s and spanning nearly two
MUC2, an intestinal epithelial apomucin, whereas the absence decades are highlighted by the discovery of KRAS activation,
of CDX2 expression usually is associated with expression of SMAD4 and BRCA2 mutations, and CDKN2A silencing (see
MUC1, a biliary apomucin, and concomitant lack of expression Figs. 9B.1 and 9B.2, left). Some of these discoveries spurred
of MUC2. These findings have suggested that there may be new lines of investigation in the field of pancreatic cancer as
two divergent pathways of carcinogenesis within the pancreatic well as specific classification of PDA subtypes (Iacobuzio-
ducts (Adsay et al, 2002). The first is a so-called intestinal Donahue et al, 2009). More recently, with the help of advanced
pathway that gives rise to CDX2- and MUC2-expressing DNA sequencing technology, investigators have been able to
IPMNs that progress to the better-prognosis colloid carcino- sequence the entire genomes of various pancreatic cancer sub-
mas. The second is a pancreatobiliary pathway that gives rise types (see Figs. 9B.1 and 9B.2, right).
to CDX2-negative, MUC2-negative, MUC1-expressing PanINs
and a subset of IPMNs, both of which can progress to the Copy-Number Aberrations
poorer-prognosis conventional ductal adenocarcinomas. Although now considered primitive, valuable cytogenetic analy-
sis performed more than a decade ago found that chromosomal
aberrations occur in virtually every pancreatic cancer. Cytoge-
GENETICS OF PANCREATIC DUCTAL netic analyses of pancreatic cancers have shown multiple,
ADENOCARCINOMA nonrandom numeric and structural changes (Griffin et al,
1995; Sirivatanauksorn et al, 2001). The most common
Genomic (DNA) Alterations in Pancreatic Cancer numeric abnormalities include losses of chromosomes 6, 12,
The multitude of genetic abnormalities in pancreatic cancer 13, and 18 and gains of chromosomes 7 and 20. Structural
have many characteristics similar to other solid tumors; thus abnormalities (intrachromosomal break points) frequently
they include point mutations in critical genes, chromosomal involve 1p and 1q, 3p, 4q, 6q, 7q, 17p, 11p, 11q, 15q, 16q,
(copy number) aberrations, mitochondrial DNA mutations, and 19q (Rigaud et al, 2000). The technical limitations of
A B
C D
E
FIGURE 9B.4. Pathologic features of precursor lesions of pancreatic ductal adenocarcioma (PDA). A, Pancreatic intraepithelial neoplasia
(PanIN). PanIN-1, flat lesion composed of columnar cells with basally located benign-appearing nuclei and supranuclear mucin. B, PanIN2 showing
mild cytologic and architectural atypia, including crowding, nuclear enlargement, and tufting. C, PanIN-3 showing complex papillary architecture with
budding of epithelial cells into lumen and severe cytologic atypia of lining cells. D, Intraductal papillary mucinous neoplasm (IPMN) with low-grade
dysplasia showing well-formed papillae lined by mucin-containing cells. D, IPMN with high-grade dysplasia with complex branching papillae and cells
showing marked cytologic atypia. E, Invasive, moderately differentiated PDA forming angular abortive glands and invading perineural spaces.
conventional cytogenetics have presented challenges for identi-

fying genes that are affected by chromosomal breaks. Allelotyp- Specific Gene Mutations
ing identifies areas of gross chromosomal loss by using Allelotyping provides insight into areas harboring tumor sup-
polymorphic microsatellite markers to determine regions of pressor genes, but it cannot qualify areas of enhanced genomic
genomic loss compared with matched healthy tissues, also expression, as happens with oncogene activation. Simplistically,
known as loss of heterozygosity (LOH) analysis. Allelotyping wild-type tumor suppressor genes can “put the brakes” on the
operates on the basic principle of the two-hit hypothesis, which speeding vehicle (the cell), but if mutated, these “brakes”
postulates that tumor-suppressor genes require biallelic inacti- become defective, and the vehicle cannot stop. Using a similar
vation. This most commonly happens by intragenic mutation automobile analogy, protooncogenes, in a mutated form known
in one allele, followed by loss of genetic material in the other as oncogenes, become the “accelerators,” and these “go signals”
allele. Identifying regions of single or biallelic loss or mutation are often critical in transforming normal cells to a malignant
holds the potential to understand the role of neighboring novel phenotype.
and well-known tumor-suppressor genes. A landmark allelo- Much like other solid tumors, genes altered in pancreatic
type analysis of pancreatic cancers was performed using cancer include three functional classes: oncogenes, tumor suppres-
approximately 80 pancreatic cancer xenografts and 386 micro- sor genes, and caretaker genes. A family of caretaker genes recog-
satellite markers (Iacobuzio-Donahue et al, 2004). This work nized as being disrupted in pancreatic cancer are genes of
discovered allelic losses in chromosome regions in proximity to the Fanconi anemia complementation group, which are involved
tumor suppressor genes CDKN2A, TP53, and SMAD4. Allelo- in homologous recombination-based DNA damage repair
type analysis of PanIN lesions also has been performed using (D’Andrea et al, 2003). Patients with Fanconi anemia present
microdissected samples, and as expected, LOH is seen in many with a wide variety of clinical issues, including aplastic anemia
of the same chromosomal regions as invasive cancer, including and a high risk of developing cancer. BRCA2 is a member of
9p, 17p, and 18q (Luttges et al, 2001b; Yamano et al, 2000). this DNA repair pathway and is mutated in a subset of familial
Although the changes are conserved in most synchronous prepancreatic cancers (Murphy et al, 2002). This has led to the
cursor lesions (i.e., the same allele is lost in PanIN and associ- search for mutations in other Fanconi anemia genes in pancre-
ated cancers), there is occasional clonal divergence between atic cancer. Somatic mutations of two genes in the core complex,
high-grade PanIN lesions harboring genetically distinct changes FANCC and FANCG, were discovered but are rare in sporadic
from the synchronous invasive cancer (Yamano et al, 2000). pancreatic cancers (van der Heijden et al, 2003). Through
These findings may have important clinical implications in other modern techniques, Jones and colleagues (2009) discov-
regard to tumor heterogeneity and clonal cancer cell drug ered FANCN (PALB2) as another mutated gene found in
resistance. familial pancreatic cancers.
Comparative genome hybridization (CGH) identifies Mutations in this core complex and in this DNA repair
genomic amplifications and deletions and differentially labels mechanism have major therapeutic implications (Fig. 9B.5,
normal and tumor genomic sequences with different dyes. The
relative ratio of the two dyes indicates regions of cancer- 500
associated losses or gains, with a ratio of 1 : 1 consistent with
no change in copy number compared with healthy DNA. Con-
ventional CGH is performed on metaphase spreads and suffers
from both low resolution and the inability to precisely map the
various regions of amplifications and deletions (Mahlamaki
et al, 1997). The resolution of array CGH is significantly better 300
than the conventional technique, ranging from 500 to 30 kb,
%
permitting the precise mapping of deletion and amplicon

boundaries and genes targeted therein. Array technology also
provides the ability to more efficiently use probes to study
amplification of a larger number of genes. Array CGH analysis
of pancreatic cancers has identified numerous recurrent copy-
100
number aberrations, including amplifications of myelocytoma-
tosis oncogene (c-MYC) (8q), epidermal growth factor receptor
gene (EGFR) (7p), KRAS (12p), AKT2 (19q), and NCOA3
(20q) and deletions of SMAD4 (18q), CDKN2A (9p), FHIT
10 20
(3p), and MAP2K4 (17p) (Aguirre et al, 2004; Calhoun et al,
2006; Holzmann et al, 2004). Days
Utilizing high-density single-nucleotide polymorphism FIGURE 9B.5. Preclinical model shows an example of a successful
arrays, Calhoun and colleagues (2006) surveyed all the targeted treatment strategy against a Fanconi-deficient tumor. A single-
commercially available pancreatic cancer cell lines. In brief, dose treatment with mitomycin C (5 mg/kg) of pancreatic cancer cell
this study provided high-resolution and detailed break-point lines xenografted into nude mice. Note the hypersensitivity and tumor
regression in the FANCC-deficient PL11 cells (squares) compared with
mapping of these cell lines and found two subclasses of cancer
the retrovirally corrected FANC-proficient PL11 cells (triangles). Solid
cell lines, original chromosomal instability (CIN) and holey CIN lines indicate treated mice; gray lines indicate no-treatment controls.
genotypes (Calhoun et al, 2006). Perhaps global classification Similar sensitivity was seen in the BRCA2-deficient CAPAN1 xenografted
of tumor cells with high-density arrays will become part of cells. (From van der Heijden MS, et al: In vivo therapeutic responses
a prognostic or predictive molecular signature panel in the contingent on Fanconi anemia/BRCA2 status of the tumor, Clin Cancer
future. Res 11(20):7508–7515, 2005.)
TABLE 9B.1 Genetic Syndromes with Inherited Predisposition to Pancreatic Cancer

Genetic Testing Risk of Pancreatic Available Targeted Therapy/Clinical
Syndrome/Disease Gene(s) Considerations Adenocarcinoma Correlates
Hereditary breast/ BRCA1 BRCA2 NCCN test criteria require BRCA1: 2.26-fold* Cross-linking chemotherapeutics
ovarian cancer personal or family BRCA2: 3- to 9-fold† (mitomycin C, cisplatin,
syndrome (HBOCS) history of breast and chorambucil, melphalan) PARP
ovarian cancer inhibitors
Peutz-Jeghers STK11 (LKB1) Diagnosis of index case 132-fold; lifetime risk Reports of a PJS- associated cancer
syndrome (PJS) is generally based on ~36%§ with loss of the wild-type STK11
clinical findings/ allele, together with a germline
working definition‡ mutation in the other allele
Some sporadic PDAs exhibit somatic
mutations of STK11||
Hereditary pancreatitis PRSS1, Testing guidelines are 50- to 67-fold; Tumor susceptibility is presumably due
SPINK1, based on symptoms lifetime risk 44%a,b to mitogenic stimulation and clonal
CFTR, CTRC with or without family outgrowth of PDA cells as part of
history of pancreatitis¶ the normal healing responses that
occur subsequent to repeated
rounds of tissue destructionb
FAMM melanoma CDKN2A Documented patients/ 13- to 39-fold† Somatic p16 alterations were identified
syndrome families with multiple in 80% of PDAsc
melanomas
HNPCC-Lynch MLH1, MSH2, Bethesda Guidelines† MSI-H pancreatic cancer may have a
syndrome MSH6, PMS2 (tumor MSI/IHC) and better prognosis after resection,
Amsterdam Clinical possibly because of intensive
Criteria II (germline immunoreaction to the tumord
studies)
Familial adenomatous APC APC is considered in Relative risk 4.46e Some theorize that pancreaticobiliary
polyposis (FAP) individuals with ≥20 Lifetime risk ~2%f secretions affect the development of
colon adenomas adenomas and cancer in this areag
Cystic fibrosis (CF) CFTR Genotyping identifies Relative risk 5.3i Modifier genes or environmental
patients with class IV factors may also be important in
and V mutations, stratifying risk (e.g., mucin genes
which are likely to are found in both CF and PDA)j
represent those with a
functioning pancreash
*From Thompson D, et al: Cancer incidence in BRCA1 mutation carriers, J Natl Cancer Inst 94(18):1358–1365, 2002.
†
From Brand RE, et al: Advances in counseling and surveillance of patients at risk for pancreatic cancer, Gut 56(10):1460–1469, 2007.
‡
From Giardiello FM, et al: Increased risk of cancer in the Peutz-Jeghers syndrome, N Engl J Med 316(24):1511–1514, 1987.
§
From Giardiello FM, et al: Very high risk of cancer in familial Peutz-Jeghers syndrome, Gastroenterology 119(6):1447–1453, 2000.
||
From Su GH, et al: Germline and somatic mutations of the STK11/LKB1 Peutz-Jehgers gene in pancreatic and biliary cancers, Am J Pathol 154(6):1835–1840, 1999.
¶
From Ellis et al: Genetic testing for hereditary pancreatitis: guidelines for indications, counseling, consent, and privacy issues, Pancreatology 1(5):405–415, 2001.
a
From Lowenfels AB, et al: Hereditary pancreatitis and the risk of pancreatic cancer: International Hereditary Pancreatitis Study Group, J Natl Cancer Inst 89(6):442–446, 1997.
b
From Howes N, et al: Clinical and genetic characteristics of hereditary pancreatitis in Europe, Clin Gastroenterol Hepatol 2(3):252–261, 2004.
c
From Rozenblum E, et al: Tumor-suppressive pathways in pancreatic carcinoma, Cancer Res 57(9):1731–1734, 1997.
d
From Nakata B, et al: Prognostic value of microstaellite instability in resectable pancreatic cancer, Clin Cancer Res 8(8):2536–2540, 2002.
e
From Giardiello FM, et al: Increased risk of thyroid and pancreatic carcinoma in familial adenomatous polyposis, Gut 34(10):1394–1396, 1993.
f
From Burt RW: Colon cancer screening, Gastroenterology 119(3):837–853, 2000.
g
From Wallace MH, et al: Upper gastrointestinal disease in patients with familial adenomatous polyposis, Br J Sur 85(6):742–750, 1998.
h
From Krysa J, et al: Pancreas and cystic fibrosis: the implications of increased survival in cystic fibrosis, Pancreatology 7(5-6):447–450, 2007.
i
From Maisonneuve P, et al: Risk of pancreatic cancer in patients with cystic fibrosis, Gut 56(9):1327–1378, 2007.
j
From Singh AP, et al: MUC4 expression is regulated by cystic fibrosis transmembrane conductance regulator in pancreatic adenocarcinoma cells via transcriptional and post-transcriptional mechanisms,
Oncogene 26(1):30–41, 2007.
FAMM, familial multiple mole; HNPCC, hereditary nonpolyposis colorectal cancer; IHC, immunohistochemical; MIS-H, microsatellite instability, high frequency; MSI, microsatellite instability; NCCN, National
Comprehensive Cancer Network; PARP, poly (ADP-ribose) polymerase; PDA, pancreatic ductal adenocarcinoma.
Modified from Showalter SI, et al: Identifying pancreatic cancer patients for targeted treatment: The challenges and limitations of the current selection process and vision for the future, Expert Opin Drug
Deliv 7(3):1–12, 2010.
Table 9B.1; see Familial Pancreatic Cancer section) (van der defects in other FANC genes yet to be thoroughly investigated
Heijden et al, 2004). Unlike most in vivo experiments, xeno- (i.e., FANCA) are the direct cause of some familial and spo-
grafted mice with isogenic cell lines (FANCC deficient and radic PDAs.
proficient) experienced regression of tumor after a single dose
of the available mitomycin C intrastrand cross-linking drug Oncogenes
(van der Heijden et al, 2005). Although other mutations in the Perhaps the best evidence that KRAS activation is an early and
Fanconi complementation group have not yet been described, important event in tumorigenesis that comes from decades of
other than FANN, FANCC, and FANCG, and the frequency research on pancreatic cancer. The KRAS oncogene on chromo-
of these mutations in PDA appear to be low, it is likely that some 12p is the most commonly altered oncogene, with as many
as 90% of pancreatic cancers containing mutations on codons gene in approximately 30% of pancreatic cancers, which offers
12, 13, and 61 (Caldas et al, 1994b). Activating mutations potential therapeutic benefit because targeted therapies have
impair the intrinsic guanosine triphosphate (GTP)ase activity of been developed that specifically inhibit the growth of MTAP-
the KRAS gene product, resulting in a protein that is constitu- deficient cells (Chen et al, 1996).
tively active in intracellular signal transduction. KRAS mutation As many as 80% of pancreatic cancers contain an inactiva-
happens early in the pathway to oncogenesis, with approximately tion of the TP53 gene on chromosome 17p. Such inactivation
30% of PanIN-1 lesions harboring KRAS mutations (Hruban most often occurs via intragenic mutation combined with loss
et al, 2000; Moskaluk et al, 1997). The first mouse model of of the second allele, although homozygous deletions also occur
pancreatic cancer was generated by constitutive overexpression in some PDAs. The TP53 (p53) protein leads to cell-cycle
of mutant KRAS2 in murine pancreatic ductal epithelium, arrest and activates apoptosis in the presence of DNA damage.
underscoring its importance in pancreatic oncogenesis (Hingo- It is believed that loss of TP53 function allows cells to survive
rani et al, 2003). This model was further developed into a power- and divide, despite the presence of damaged DNA, leading to
ful and useful preclinical model for PDA progression (Hingorani accumulation of additional genetic abnormalities and, eventu-
et al, 2005). Several good sources on mouse modeling and pan- ally, to neoplasia. Nuclear overexpression of the TP53 protein
creatic cancer have been published (Frese et al, 2007; Karreth does not correlate well with mutation status. By immunohisto-
et al, 2009; Olive et al, 2006; Tuveson et al, 2005). chemistry, TP53 accumulation is seen only in the advanced
Rarely, pancreatic cancers with wild-type KRAS genes PanIN-3 lesions, consistent with TP53 being a “late” genetic
harbor point mutations of BRAF, another gene in the event in pancreatic cancer progression (Maitra et al, 2003).
RAS/RAF/mitogen-activated protein kinase (MAPK) signaling PDA cells with a mutant TP53 have been shown to have a
pathway, thereby explaining why mutations of these genes greater propensity for metastases (Morton et al, 2010). Of
occur in mutually exclusive patterns in pancreatic cancer note, in an experimental model, the BRCA2 gene could not be
(Calhoun et al, 2003). This highlights the importance of iden- artificially disrupted in cancer cells with intact wild-type TP53
tifying different molecular targets that lead to similar pathways status (Gallmeier et al, 2007; Hucl et al, 2008).
in pancreatic cancer development and of finding a drug that Inactivation of the SMAD4 gene on chromosome 18q21
can target one pathway, not one gene. Studies have shown that occurs in 55% of pancreatic cancers by homozygous deletions
targeting KRAS may have potential in modulating angiogenesis (30%) or by intragenic mutations and loss of the second allele
in tumorigenesis. Matsuo and colleagues (2009) showed that (25%). Loss of SMAD4 function interferes with intracellular
oncogenic overexpression of KRAS increases production of signaling cascades downstream of the TGF family of cell surface
angiogenesis, promoting CXC chemokines and vascular endo- receptors, leading to decreased growth inhibition and uncon-
thelial growth factor (VEGF) from human pancreatic duct epi- trolled proliferation. Although SMAD4 alterations are most
thelial cells. This upregulation acts through the MAPK pathway common in PDA, they are also frequently seen in other carci-
and c-JUN signaling (Matsuo et al, 2009). KRAS mutation has nomas, occurring in approximately 15% of colorectal carci-
been shown to be associated with increased VEGFA expression noma and 10% of gastric carcinoma (Cerami et al, 2012).
and poorer prognosis in pancreatic carcinoma (Ikeda et al, Similar to TP53, loss of SMAD4 function is a late genetic event
2001). Yet to date, targeting KRAS activation in PDA patients in pancreatic carcinoma progression, with loss of SMAD4 seen
has shown no success. Perhaps KRAS activation is a critical only in a few PanIN-3 lesions (Maitra et al, 2003). Examination
early event in pancreatic tumorigenesis, but once cells become of resected tumor specimens found that SMAD4 inactivation
malignant, there is no need for constitutive KRAS activation, portends a poorer prognosis and greater potential to metasta-
or for oncogenic addiction, for that matter. size (Blackford et al, 2009; Iacobuzio-Donahue et al, 2009).
Other oncogenes implicated in pancreatic cancers include Moreover, in a study of PDA patients who underwent autopsy,
MYC and EGFR, which can be mutated (GNAS) or amplified loss of SMAD4 was highly correlated with extensive metastatic
(MYC) in various subsets of cancers. Overexpression of MYC burden. Identification of downstream targets might allow res-
transcripts occurs in approximately 50% to 60% of pancreatic toration of SMAD4-dependent signaling in pancreatic cancer,
cancers and has been shown to cooperate with KRAS (Aguirre yielding an improved prognosis (Cao et al, 2008).
et al, 2004; Han et al, 2002; Stellas et al, 2014). Several tumor-suppressor genes are inactivated in smaller
numbers (5% to 10%) of pancreatic cancers, including STK11
Tumor Suppressor Genes (chromosome 19p) (Sahin et al, 2003), TGFBR1 (chromo-
CDKN2A, on chromosome 9p, is the most commonly inacti- some 9q), TGFBR2 (chromosome 3p), RB1 (chromosome
vated gene in pancreatic cancers, occurring in 90% of patients 13q) (Jaeger et al, 1997), and MAP2K4 (chromosome 17p) (Su
(Caldas et al, 1994a; Schutte et al, 1997). CDKN2A belongs et al, 1998). MAP2K4 function has been explored in a number
to the cyclin-dependent kinase inhibitor family and inhibits of models, yet the main reason for its loss in pancreatic cancer
cell-cycle progression through the G1-S checkpoint mediated is still unknown (Cunningham et al, 2006).
by cyclin-dependent kinases such as CDK4 and CDK6. Homo- Separately, DNA-level abnormalities in the switch/sucrose
zygous deletions (40%), intragenic mutation with loss of the nonfermentable (SWI/SNF) complex gene member AT-rich
second allele (40%), and epigenetic silencing by promoter interaction domain 1A (ARID1A) have been reported in approx-
methylation (10% to 15%) all contribute to gene inactivation. imately 14% of PDAs in provisional TCGA data, mostly gene
Loss of CDKN2A function occurs throughout the process of deletions and truncating mutations, whereas another 6% of PDA
oncogenesis, with lesions appearing in different PanINs: 30% have decreased messenger RNA (mRNA) ARID1A levels
of PanIN-1A and PanIN-1B, 55% of PanIN-2, and 71% of without corresponding DNA abnormality, suggesting an epigen-
PanIN-3 lesions show loss of nuclear p16 protein expression etic aberration (Biankin et al, 2012; Cerami et al, 2012).
(Wilentz et al, 1998). The CDKN2A homozygous deletions ARID1A and other SWI/SNF members remodel chromatin,
encompass the methylthioadenosine phosphorylase (MTAP) thus controlling the transcription and expression of various
genes. Abnormalities in ARID1A have been associated with mutations and allelic loss, it has become apparent since the
upregulation of the phosphoinositide-3-kinase (PI3K) pathway 1990s that epigenetic mechanisms of silencing are probably as
as well as sensitivity to PI3K and AKT inhibition (Samartzis important in terms of frequency and prevalence in many cancers
et al, 2014). (Baylin et al, 2000). Epigenetic silencing occurs predominantly
Mixed-lineage leukemia 3 (MLL3 or KMT2C) is a gene through hypermethylation of so-called CpG islands in the pro-
involved in histone methylation and transcriptional coactiva- moter region of tumor suppressor genes, leading to transcrip-
tion. It functions as a tumor suppressor and is recurrently tional abrogation. In cancers, preferential hypermethylation of
mutated in approximately 18% of pancreatic carcinoma, with the promoter occurs in the neoplastic cells with consequent
the most common mutation type being truncating mutations downregulation of gene expression, but this does not occur in
(frameshift and nonsense mutations) (Biankin et al, 2012; the corresponding normal counterpart. Epigenetic silencing is
Cerami et al, 2012). seen frequently in pancreatic cancers and tends to involve genes
that function in tumor suppression or in critical homeostatic
Other Caretaker Genes pathways (e.g., CDKN2A, E-cadherin, retinoic acid β, osteo-
In addition to classic oncogenes or tumor suppressor genes, nectin, SOCS1) or in both (Sato et al, 2003a; Ueki et al, 2000).
caretaker genes (beyond Fanconi anemia–related genes) have Aberrant methylation of genes is also found in precursor
been shown to play a role in oncogenesis. In theory, caretaker lesions of pancreatic cancers and tends to occur in intermedi-
genes do not influence cell growth and proliferation directly but ate- or late-stage lesions (PanIN-2 and PanIN-3) (Fukushima
rather prevent the accumulation of DNA damage and cumula- et al, 2002). Although there has been extensive work on the role
tive mutations within key exonic sequences that make up the of promoter hypermethylation in the pathogenesis of cancers,
human genome. Loss of function of the DNA damage repair more recent data suggest that promoter hypomethylation in
genes (MLH1, hMSH2) occurs in a small subset of pancreatic candidate genes also may be important in cancer development
cancers in the familial setting, but has been reported to occur and progression. Genes showing preferential hypomethylation
in approximately 17% of sporadic, nonfamilial cases (Borelli in pancreatic cancers—SERPINB5, S100P, MSLN, PSCA, and
et al, 2014; Ghimenti et al, 1999; Nakata et al, 2002). Histo- CLDN4—are usually overexpressed in the cancers compared
logically, these microsatellite instability (MSI) cancers com- with healthy pancreas, suggesting that epigenetic mechanisms
prise poorly differentiated cancers with a syncytial growth can affect gene expression in either direction (Sato et al, 2003b).
pattern, expanding tumor margins, extensive necrosis, and Aberrant epigenetic silencing by promoter methylation also
intratumoral lymphocytic infiltrates. This uncommon variant has been reported in IPMNs, including methylation of the
has been termed medullary cancer to distinguish it from the more SOCS1 and CDKN2A genes (Sato et al, 2002). Global analyses
common PDAC (Wilentz et al, 2000b). Although findings in of gene expression in IPMNs have revealed the overexpression
colon cancer have attempted to correlate MSI status with of LCN2, LGALS3, CTSE, CLDN4, and three members of the
response to 5-fluorouracil, other studies have questioned these trefoil factor family, TFF1, TFF2, and TFF3 (Sato et al, 2004a;
claims (Brody et al, 2009a). Terris et al, 2002). These global analyses also have shown that
CLDN4, CXCR4, S100A4, and MSLN all are expressed at
Telomere Length Abnormalities significantly higher levels in invasive IPMNs than in noninva-
Telomeres are hexameric repeats of the sequence TTAGGG at sive IPMNs, suggesting that these proteins may contribute to
the ends of chromosome arms that confer stability to chromo- the process of invasion (Sato et al, 2004b).
somes during cell division and prevent the ends from becoming
“promiscuous” (Gisselsson, 2003). In other words, intact telo-
mere structure guards against chromosomal fusion and thus
Core Signaling Pathways Disrupted in
may prevent chromosomal instability (Greenberg et al, 2005; Pancreatic Cancer
Raynaud et al, 2008). In fact, telomeric dysfunction has been The modern era of molecular biology has progressed to allow
hypothesized to be one of the more important gateways of high-throughput surveying of the pancreatic cancer genome
chromosomal instability, a signature of most solid cancers char- (see Figs. 9B.1 and 9B.2, right), bringing some clarity to our
acterized by aneuploidy and extensive chromosomal rearrange- understanding of the interactions of molecular pathways in
ments. The development of direct visualization of in situ tumorigenesis. This high-throughput analysis revealed that
telomere length was a breakthrough for understanding telomere pancreatic cancers contain an average of 63 genes that are
length abnormalities and cancer development (Meeker et al, genetically altered (Jones et al, 2008). In this work, Jones and
2004). A study by van Heek and colleagues (2002) showed colleagues used a combination of modern molecular techniques
that telomere length abnormalities are one of the earliest to report that 67% to 100% of all pancreatic cancer genomes
demonstrable genetic aberrations in pancreatic cancer, with surveyed had a genetic abnormality in 12 core signaling path-
greater than 90% of the lowest-grade PanIN lesions showing ways and processes. These pathways, confirmed by later studies
marked shortening of telomeres, compared with normal duct as well, include apoptosis, DNA damage control, regulation
epithelium (van Heek et al, 2002). It has been hypothesized of G1-to-S phase transition, Hedgehog signaling, hemophilic
that intact telomeres may serve as “caretakers” in the pancreatic cell adhesion, integrin signaling, c-JUN N-terminal kinase sig-
ducts and that the loss of telomeres in PanIN lesions sets the naling, KRAS signaling, regulation of invasion, small GTPase-
stage for progressive accumulation of chromosomal abnormali- dependent signaling (other than KRAS), TGFB signaling, and
ties, eventually culminating in neoplasia. WNT/NOTCH signaling (Biankin et al, 2012; Jones et al,
2008).
Alternative Genetic Silencing: Epigenetic Abnormalities Recently, genetic aberrations in the axon guidance pathway
Although the classic two-hit hypothesis postulated that tumor genes have been identified in genes in the SLIT/ROBO pathway
suppressor gene silencing occurs by a combination of intragenic (mutations and deletions in SLIT, ROBO1, and ROBO2),
ephrins (EPHA5 and EPHA7), and class 3 semaphorins (ampli- Identification of mutations in the FANC-BRCA pathway in
fications and mutations in SEMA3A and SEMA3E) (Biankin familial cancers has the potential to shed light on the treatment
et al, 2012). These genes have been implicated in cell growth, of sporadic cancers as well. It has been suggested that perhaps
metastasis, and invasion (Mehlen et al, 2011), overlapping with as many as 25% of sporadic breast and ovarian cancers manifest
some of the pathways listed in the previous paragraph. a BRCA-like phenotype. (Turner et al, 2004). The data have
Unlike certain subtypes of leukemia that are driven by single been derived from BRCA1, FANCC, FANCG, and FANCF
“targetable” oncogenes, we have learned that pancreatic cancers methylation studies (Turner et al, 2004). Further studies are
result from genetic alterations of large numbers of genes that warranted, but future banking of sporadic pancreatic cancers
function through a distinct number of pathways. The study by to study all of the tumor characteristics—posttranscriptional
Jones and colleagues (2008) suggests that it may be beneficial modification, polymorphisms, CGH analysis—along with thor-
to target the physiologic effects of disrupted pathways rather ough analysis of family history may reveal a “BRCAness”
than individual target genes. In fact, targeting multiple path- among certain sporadic pancreatic tumors, aiding in the per-
ways or multiple points in the pathway may be in line with the sonalization of therapy (Martin et al, 2010). Focused DNA-
early preclinical and clinical success stories of the concept of repair microarray analysis may also shed light on other cancer
“synthetic lethality” (Ashworth, 2008). susceptibility genes.
Of note, a recent report demonstrated that PALB2, formerly
FAMILIAL PANCREATIC CANCER known as FANCN, is a gene inherited in a mutant form that
produces a stop codon in a small percentage of familial PDAs.
Approximately 10% of pancreatic cancers show familial aggrega- This gene was discovered to be mutated in 3 of 96 familial
tion (Petersen et al, 2003). The presence of two first-degree pancreatic cancers, each producing a different stop codon
family members with pancreatic cancer confers a 6- to 18-fold (Jones et al, 2009). Truncating mutations in PALB2 were not
increased risk of the disease in other first-degree relatives. This found in any of the 1084 patients of similar ethnicity used as a
risk is increased 32- to 57-fold in families with three or more control cohort in a similar study, thus ruling out a polymorphic
first-degree relatives with pancreatic cancer (Klein et al, 2004; sequence variant. This information suggests that next to
Tersmette et al, 2001). Only a minority of these familial cancers BRCA2, PALB2 is the second most commonly mutated gene
is caused by a recognized cancer syndrome associated with in hereditary pancreatic cancer (Jones et al, 2009). Thus most
germline mutations in known genes. Included in Table 9B.1 are familial pancreatic cancers have no known genetic bases at this
possible “targeted therapies” for the genes and disorders germane time, although many believe an autosomal dominant inheri-
to pancreatic cancer.Thus this table underscores the significance tance of a rare mutant allele is the most likely cause of these
for understanding the inherited lesion that may have contributed cancers. Of apparently “sporadic” (nonfamilial) pancreatic
to the pancreatic cancers found in these families. cancer patients, 7% harbor germline mutations in the BRCA2
For example, as mentioned earlier in this chapter, Fanconi gene, and this low-penetrance pattern is peculiar to cancers
anemia is characterized as a rare, autosomally recessive cancer arising in the Ashkenazi Jewish population (Goggins et al,
syndrome that results initially from a mutation in one of the 1996). Perhaps no single gene is responsible for the other
multiple FANC/BRCA complementation groups in the FANC/ familial forms of PDA, carcinogenesis of which may be the best
BRCA pathway (Mathew, 2006). One gene in this pathway, example that core pathways collaborate with the environment
BRCA2, is associated with a greatly increased risk of cancer (Yeo et al, 2009); thus no single gene from one pathway will
when deleted via a biallelic mutation. The BRCA genes play a prove to be disrupted in the complex process of tumorigenesis
critical role in DNA repair via RAD51 repair pathways (Gud- in this familial form of pancreatic cancer.
mundsdottir et al, 2006). Loss of functional BRCA1 and Several other known genetic syndromes associated with
BRCA2 conveys chromosomal instability to cells by impairing PDA exist, including familial atypical mole and multiple mela-
the critical function of DNA double-stranded break repair noma syndrome (FAMM) and Peutz-Jeghers syndrome (PJS).
(Ashworth, 2008). It has been well established that DNA dam- FAMM results from the microdeletion of CDKN2 on chromo-
aging agents such as mitomycin C or cisplatin effectively kill some 9p21.3, particularly p16INK4a (Gruis et al, 1995; Ranade
cells with loss of BRCA2 or related genes (see Fig. 9B.5) et al, 1995). As a result, CDK4/6 function is uninhibited.
(Hussain et al, 2004). FAMM kindred have also been reported to harbor CDK4 muta-
Currently, poly (adenosine diphosphate-ribose) polymerase tions that prevent CDKN2 binding, as opposed to CDKN2
(PARP) inhibitors have similar promising results as intrastrand microdeletions, in some cases (Zuo et al, 1996). Patients with
cross-linking agents in early-phase trials in other cancer types FAMM have an approximately 80% lifetime risk of melanoma
(ovarian, breast) (Farmer et al, 2005). Certainly, pancreatic and a 20% lifetime risk of PDA (Rustgi et al, 2014). PJS is an
cancer is a logical tumor system in which to test novel PARP autosomal dominant syndrome caused by mutations of the
inhibitors in combination with other DNA-damaging agents. tumor suppressor STK11 (LKB1). This syndrome is best
Targeting the FANC-BRCA pathway with PARP inhibitors known for polyps throughout the small and large intestine with
similarly has been shown to lead to synthetic lethality, creating an arborizing pattern of musculature, and these patients develop
a convenient and fortuitous therapeutic window (Ashworth, various types of carcinoma, including PDA. PDA develops in
2008). The makeup of this therapeutic window relies on the approximately one fourth of patients with PJS by age 75 years
fact that healthy cells will have an intact DNA repair mecha- (Korsse et al, 2013).
nism and thus will be capable of managing and repairing Another condition that may progress to PDA is hereditary
the damage put forth by a DNA-damaging agent. In contrast, pancreatitis. These patients carry germline mutations in the
the tumor will be unable to repair such damage due to loss PRSS1 gene, which encodes cationic trypsinogen. Multiple
of a key aspect of this repair mechanism (i.e., BRCA2; see mutations have been described; the original was R117H, which
Fig. 9B.5). resulted in the elimination of a hydrolysis site in trypsin and an
inability to inactivate trypsin (Whitcomb et al, 1996). Patients (Lopez de Silanes et al, 2005). It is has been shown that during
with hereditary pancreatitis are at a 35-fold relative risk for times of certain cellular stressors—brought on by agents such
PDA by age 75 years. Diet modification, including lowering as ultraviolet C (UVC) irradiation, heat shock, hypoxia, tamox-
triglyceride intake, as well as abstaining from smoking or drink- ifen (Hostetter et al, 2008), and actinomycin D—HuR can
ing, are advised to decrease the risk of progression to chronic bind to certain apoptotic or survival mRNA transcripts by
pancreatitis. binding to AU-rich elements in the 3′ untranslated region
(UTR) of these mRNAs. In regard to tumorigenesis, HuR has
TRANSCRIPTOMIC (RNA) ABNORMALITIES IN been shown to bind to and stabilize proteins such as p21, p53,
PANCREATIC CANCER and cyclin A (Gorospe, 2003). For instance, Gorospe and col-
leagues (2003) showed that HuR can enhance translation of
Several studies have analyzed pancreatic cancers and compared proteins such as p53 under stress from UVC irradiation. Thus
their gene expression profile with healthy pancreas tissue to HuR’s role in cellular stress and damage gives it a likely pivotal
identify differentially overexpressed and underexpressed genes role in both the tumorigenesis process and in the acute cellular
(Argani et al, 2001a; Crnogorac-Jurcevic et al, 2001; Geng response to chemotherapy in pancreatic cancer cells.
et al, 1998; Iacobuzio-Donahue et al, 2002, 2003a, 2003b; In vitro and in vivo studies have shown that PDAs with HuR
Ryu et al, 2001, 2002). A comprehensive analysis of pancreatic overexpression were dramatically sensitive to gemcitabine, the
cancer using high-density oligonucleotide microarrays identi- standard chemotherapeutic treatment for pancreatic cancer,
fied 217 genes as overexpressed threefold or greater in cancers when compared with a control group (Constantino et al, 2009;
versus healthy tissue (Iacobuzio-Donahue et al, 2003a). Six Richards et al, 2010). Patients who had low cytoplasmic HuR
genes—keratin 19, retinoic acid–induced 3, secretory leukocyte levels had a sevenfold increase in mortality compared with
protease inhibitor, stratifin, tetraspan 1, and transglutaminase patients who had elevated cytoplasmic HuR levels (Fig. 9B.6)
2—were found to be overexpressed in pancreatic cancer by (Constantino et al, 2009).
three platforms: oligonucleotide, cDNA microarrays, and serial
analysis of gene expression (SAGE). The future role of one or MicroRNAs
all of these six genes in early detection or therapy remains to MicroRNAs (miRNAs) are defined as short, noncoding regions
be elucidated. of RNA sequences (22 nucleotides) that can potently regulate
The identification of differentially expressed genes not only gene expression patterns; miRNAs have been shown to regulate
serves the further understanding of the basic biology of pancre- a number of disease- and developmental-related genes, and
atic cancers, it also provides a fertile ground to identify markers they are tissue specific in expression (Rosenfeld et al, 2008).
for early diagnosis, imaging, and novel therapeutic strategies. These miRNA-specific attributes make them putative, power-
Mesothelin (MSLN) was identified by SAGE as a gene over- ful, and unique candidate biomarkers (Mardin et al, 2009).
expressed in pancreatic cancers, and it was confirmed by immu- Discovering the presence of miRNAs in pancreatic cancer may
nohistochemistry to be restricted to the neoplastic epithelium be extremely valuable, although understanding the significance
(Argani et al, 2001b). This identification led to the develop- of these miRNAs may be more difficult and tedious, as miRNAs
ment of a pancreatic cancer vaccine targeted to the mesothelin have been shown to distinguish between various disease states
antigen and the development of antimesothelin antibody- and tissues, including pancreatitis, PDA, IPMN, and healthy
conjugated immunotoxins (Thomas et al, 2004). Phase 1 clini-
cal trials showed that investigated antimesothelin drugs are well
1.0 Cytoplasmic HuR
tolerated, and patients with advanced cancers often achieve
stable disease on antimesothelin therapy (Hassan et al, 2007; High
Hassan et al, 2010). In 2013, the U.S. Food and Drug Admin- + Low
0.8
istration (FDA) granted the antimesothelin drug CRS-207
approval for use in combination therapy with GVAX, a drug + ++ + ++ + +
Proportion alive
that stimulates granulocyte-macrophage colony-stimulating

0.6
factor (Le et al, 2012).
Posttranscriptional Regulation
0.4
In recent years, strong evidence has shown that posttranscrip-
tional regulation of genes can directly affect both the tumori-
genesis process (Lopez de Silanes et al, 2003, 2005) and cancer 0.2 +
cell susceptibility to chemotherapy (Brody et al, 2009a; Con- ++
stantino et al, 2009; Gorospe, 2003). Posttranscriptional gene
regulation can have the same effect on gene expression as a 0.0
genetic mutation or methylation of a promoter. One potent 0 200 400 600 800 1000 1200 1400
mechanism of posttranscriptional regulation involves RNA- Days postsurgery
binding proteins. One such RNA-binding protein that has been
FIGURE 9B.6. Tumor human antigen R (HuR) cytoplasmic status can
shown to be important in a number of tumor systems is Hu
stratify pancreatic cancer patients treated with standard-of-care chemo-
antigen R (HuR), a ubiquitously expressed member of the HU therapy (gemcitabine) into two groups: responders (high cytoplasmic
family that mediates cellular response to stress and DNA HuR) versus nonresponders (low cytoplasmic HuR). (From Costantino
damage by posttranscriptional regulation (Hinman et al, 2008). CL, et al: The role of HuR in gemcitabine efficacy in pancreatic cancer:
Elevated HuR cytoplasmic expression is detected in tumors HuR up-regulates the expression of the gemcitabine metabolizing
with poor pathologic features and poor predicted outcomes enzyme deoxycytidine kinase, Cancer Res 69(11):4567–4572, 2009.)
specimens (Mardin et al, 2009). Further, it has been shown rarely (approximately 3%) harbor TP53 mutations. Instead, the
that a miRNA molecular signature can stratify long- and short- sporadic forms of this tumor frequently harbor death domain–
term survivors (Bloomston et al, 2007). associated protein/gene (DAXX)/alpha-thalassemia X-linked
mental retardation protein/gene (ATRX) mutations (43%) or
multiple endocrine neoplasia type 1 (MEN1) mutations (44%)
MOLECULAR GENETICS OF OTHER (Jiao et al, 2011).
PANCREATIC NEOPLASMS In addition to sporadic forms, pancreatic neuroendocrine
tumors are also seen as a component of various inherited tumor
Acinar Cell Carcinoma syndromes associated with germline mutations, including
Unlike PDA, acinar cell carcinomas (ACC) activating muta- MEN1, due to mutations that cause loss of function of the
tions in KRAS are uncommon (Chmielecki et al, 2014). MEN1 and VHL genes in MEN and von Hippel Lindau syn-
Whole-exome sequencing of 17 ACCs revealed hot-spot–acti- dromes, respectively (Charlesworth et al, 2012; Thakker,
vating mutations, including GNAS p.R201C in two tumors and 2014). PanNETs occurring in association with these syndromes
BRAF p.V600E in one tumor, as well as mutations in tumor are thought to follow a less aggressive course more often.
suppressors, including SMAD4, TP53, retinoblastoma 1 (RB1),
phosphatase and tensin homolog (PTEN), and ARID1A (Jiao FINAL THOUGHTS AND PERSPECTIVES
et al, 2014). In addition to point mutations and indels, SND1-
BRAF fusions were identified in 6 of 44 (14%) either pure or We are currently at an interesting and critical time in studying
mixed differentiation acinar cell carcinomas. Transfectants the molecular aspects of pancreatic cancer. The research com-
expressing this fusion have shown increased MAPK pathway munity has incredible resources at its disposal, ranging from
activity as well as sensitivity to the MAP/extracellular signal- patient databases to complex sequencing equipment. This
regulated protein kinase (ERK) kinase (MEK) inhibitor tra- coming of age of pancreatic cancer research must include sur-
metinib (Chmielecki et al, 2014). Epigenetic changes have also geons, pathologists, molecular biologists, and medical oncolo-
been identified, including microsatellite instability in 10% to gists collaborating toward ultimately better and more personalized
20% of ACCs (Jiao et al, 2014; Liu et al, 2014). patient care. Current research will also need to provide better
early detection markers so that physicians can have more oppor-
Pancreatic Neuroendocrine Tumors tunities to prevent cancer from forming, instead of attempting
Pancreatic neuroendocrine tumors (PanNETs) also have differ- to cure it before it is too late.
ent molecular profiles than either PDAs or ACCs. PanNETs
lack KRAS, SMAD4, and CDKN2A mutations, and only very References are available at expertconsult.com.
Chapter 9B Molecular pathology of pancreatic cancer and premalignant tumors 149.e1
REFERENCES Charlesworth M, et al: Pancreatic lesions in von Hippel-Lindau

disease? A systematic review and meta-synthesis of the literature,
Adsay NV, et al: Colloid (mucinous noncystic) carcinoma of the pan- J Gastrointest Surg 16(7):1422–1428, 2012.
creas, Am J Surg Pathol 25(1):26–42, 2001. Chen ZH, et al: Gene deletion chemoselectivity: codeletion of the
Adsay NV, et al: The dichotomy in the preinvasive neoplasia to invasive genes for p16(INK4), methylthioadenosine phosphorylase, and the
carcinoma sequence in the pancreas: differential expression of alpha- and betainterferons in human pancreatic cell carcinoma lines
MUC1 and MUC2 supports the existence of two separate pathways and its implications for chemotherapy, Cancer Res 56(5):1083–1090,
of carcinogenesis, Mod Pathol 15(10):1087–1095, 2002. 1996.
Adsay NV, et al: Pathologically and biologically distinct types of epi- Chmielecki J, et al: Comprehensive genomic profiling of pancreatic
thelium in intraductal papillary mucinous neoplasms: delineation of acinar cell carcinomas identifies recurrent RAF fusions and frequent
an “intestinal” pathway of carcinogenesis in the pancreas, Am J Surg inactivation of DNA repair genes, Cancer Discov 4(12):1398–1405,
Pathol 28(7):839–848, 2004. 2014.
Aguirre AJ, et al: High-resolution characterization of the pancreatic Correa-Gallego C, et al: Predicting dysplasia and invasive carcinoma
adenocarcinoma genome, Proc Natl Acad Sci U S A 101(24):9067– in intraductal papillary mucinous neoplasms of the pancreas: devel-
9072, 2004. opment of a preoperative nomogram, Ann Surg Oncol 20(13):4348–
Allen VB, et al: Diagnostic accuracy of laparoscopy following com- 4355, 2013.
puted tomography (CT) scanning for assessing the resectability with Constantino CL, et al: The role of HuR in gemcitabine efficacy in
curative intent in pancreatic and periampullary cancer, Cochrane pancreatic cancer: HuR up-regulates the expression of the gem-
Database Syst Rev (11):CD009323, 2013. citabine metabolizing enzyme deoxycytidine kinase, Cancer Res
Amato E, et al: Targeted next-generation sequencing of cancer genes 69(11):4567–4572, 2009.
dissects the molecular profiles of intraductal papillary neoplasms of Crnogorac-Jurcevic T, et al: Gene expression profiles of pancreatic
the pancreas, J Pathol 233(3):217–227, 2014. cancer and stromal desmoplasia, Oncogene 20(50):7437–7446,
American Cancer Society: Cancer Facts and Figures 2014. Atlanta: 2001.
American Cancer Society, 2014. Available at: http://www.cancer.org/ Cunningham SC, et al: Targeted deletion of MKK4 in cancer cells: a
acs/groups/content/@research/documents/webcontent/acspc- detrimental phenotype manifests as decreased experimental metas-
042151.pdf. Accessed May 21, 2014. tasis and suggests a counterweight to the evolution of tumor-
Argani P, et al: Discovery of new markers of cancer through serial suppressor loss, Cancer Res 66(11):5560–5564, 2006.
analysis of gene expression: prostate stem cell antigen is overex- D’Andrea AD, et al: The Fanconi anaemia/BRCA pathway, Nat Rev
pressed in pancreatic adenocarcinoma, Cancer Res 61(11):4320– Cancer 3(1):23–34, 2003.
4324, 2001a. Farmer H, et al: Targeting the DNA repair defect in BRCA mutant
Argani P, et al: Mesothelin is overexpressed in the vast majority of cells as a therapeutic strategy, Nature 434(7035):917–921, 2005.
ductal adenocarcinomas of the pancreas: identification of a new Frese KK, et al: Maximizing mouse cancer models, Nat Rev Cancer
pancreatic cancer marker by serial analysis of gene expression 7(9):645–658, 2007.
(SAGE), Clin Cancer Res 7(12):3862–3868, 2001b. Fukushima N, et al: Aberrant methylation of preproenkephalin and
Ashworth A: A synthetic lethal therapeutic approach: poly(ADP) ribose p16 genes in pancreatic intraepithelial neoplasia and pancreatic
polymerase inhibitors for the treatment of cancers deficient in DNA ductal adenocarcinoma, Am J Pathol 160(5):1573–1581, 2002.
double-strand break repair, J Clin Oncol 26(22):3785–3790, 2008. Gallmeier E, et al: Gene-specific selection against experimental fanconi
Baylin SB, et al: DNA hypermethylation in tumorigenesis: epigenetics anemia gene inactivation in human cancer, Cancer Biol Ther
joins genetics, Trends Genet 16(4):168–174, 2000. 6(5):654–660, 2007.
Biankin AV, et al: Molecular pathogenesis of precursor lesions of pan- Geng M, et al: Isolation of differentially expressed genes by combining
creatic ductal adenocarcinoma, Pathology 35(1):14–24, 2003. representational difference analysis (RDA) and cDNA library arrays,
Biankin AV, et al: Pancreatic cancer genomes reveal aberrations in axon Biotechniques 25(3):434–438, 1998.
guidance pathway genes, Nature 491(7424):399–405, 2012. Ghimenti C, et al: Microsatellite instability and mismatch repair gene
Blackford A, et al: SMAD4 gene mutations are associated with poor inactivation in sporadic pancreatic and colon tumours, Br J Cancer
prognosis in pancreatic cancer, Clin Cancer Res 15(14):4674–4679, 80(1-2):11–16, 1999.
2009. Gisselsson D: Chromosome instability in cancer: how, when, and why?
Bloomston M, et al: MicroRNA expression patterns to differentiate Adv Cancer Res 87:1–29, 2003.
pancreatic adenocarcinoma from normal pancreas and chronic pan- Goggins M, et al: Germline BRCA2 gene mutations in patients with
creatitis, JAMA 297(17):1901–1908, 2007. apparently sporadic pancreatic carcinomas, Cancer Res 56(23):5360–
Borelli I, et al: A founder MLH1 mutation in Lynch syndrome families 5364, 1996.
from Piedmont, Italy, is associated with an increased risk of pancre- Gorospe M: HuR in the mammalian genotoxic response: post-
atic tumours and diverse immunohistochemical patterns, Fam transcriptional multitasking, Cell Cycle 2(5):412–414, 2003.
Cancer 13(3):401–413, 2014. Greenberg RA, et al: Telomere structural dynamics in genome integrity
Brody JR, et al: Limits to thymidylate synthase and TP53 genes as control and carcinogenesis, Adv Exp Med Biol 570:311–341,
predictive determinants for fluoropyrimidine sensitivity and further 2005.
evidence for RNA-based toxicity as a major influence, Cancer Res Griffin CA, et al: Consistent chromosome abnormalities in adenocar-
69(3):984–991, 2009a. cinoma of the pancreas, Cancer Res 55(11):2394–2399, 1995.
Cao D, et al: Differential expression of multiple genes in association Gruis NA, et al: Homozygotes for CDKN2 (p16) germline mutation
with MADH4/DPC4/SMAD4 inactivation in pancreatic cancer, Int in Dutch familial melanoma kindreds, Nat Genet 10(3):351–353,
J Clin Exp Pathol 1(6):510–517, 2008. 1995.
Caldas C, et al: Frequent somatic mutations and homozygous deletions Gudmundsdottir K, et al: The roles of BRCA1 and BRCA2 and associ-
of the p16 (MTS1) gene in pancreatic adenocarcinoma, Nat Genet ated proteins in the maintenance of genomic stability, Oncogene
8(1):27–32, 1994a. 25(43):5864–5874, 2006.
Caldas C, et al: Detection of K-ras mutations in the stool of patients Han H, et al: Identification of differentially expressed genes in pancre-
with pancreatic adenocarcinoma and pancreatic ductal hyperplasia, atic cancer cells using cDNA microarray, Cancer Res 62(10):2890–
Cancer Res 54(13):3568–3573, 1994b. 2896, 2002.
Calhoun ES, et al: BRAF and FBXW7 (CDC4, FBW7, AGO, SEL10) Hassan R, et al: Phase I study of SS1P, a recombinant anti-mesothelin
mutations in distinct subsets of pancreatic cancer: potential thera- immunotoxin given as a bolus I.V. infusion to patients with
peutic targets, Am J Pathol 163(4):1255–1260, 2003. mesothelin-expressing mesothelioma, ovarian, and pancreatic
Calhoun ES, et al: Identifying allelic loss and homozygous deletions in cancers, Clin Cancer Res 13(17):5144–5149, 2007.
pancreatic cancer without matched normals using high-density Hassan R, et al: Phase I clinical trial of the chimeric anti-mesothelin
single-nucleotide polymorphism arrays, Cancer Res 66(16):7920– monoclonal antibody MORAb-009 in patients with mesothelin-
7928, 2006. expressing cancers, Clin Cancer Res 16(24):6132–6138, 2010.
Cerami E, et al: The cBio cancer genomics portal: an open platform Hingorani SR, et al: Preinvasive and invasive ductal pancreatic
for exploring multidimensional cancer genomics data, Cancer Discov cancer and its early detection in the mouse, Cancer Cell 4(6):437–
2(5):401–404, 2012. 450, 2003.
Hingorani SR, et al: Trp53R172H and KrasG12D cooperate to Lopez de Silanes I, et al: Role of the RNA-binding protein HuR in
promote chromosomal instability and widely metastatic pancreatic colon carcinogenesis, Oncogene 22(46):7146–7154, 2003.
ductal adenocarcinoma in mice, Cancer Cell 7(5):469–483, 2005. Lopez de Silanes I, et al: HuR: post-transcriptional paths to malig-
Hinman MN, et al: Diverse molecular functions of Hu proteins, Cell nancy, RNA Biol 2(1):11–13, 2005.
Mol Life Sci 65(20):3168–3181, 2008. Luttges J, et al: The immunohistochemical mucin expression pattern
Holzmann K, et al: Genomic DNA-chip hybridization reveals a higher distinguishes different types of intraductal papillary mucinous neo-
incidence of genomic amplifications in pancreatic cancer than con- plasms of the pancreas and determines their relationship to muci-
ventional comparative genomic hybridization and leads to the iden- nous noncystic carcinoma and ductal adenocarcinoma, Am J Surg
tification of novel candidate genes, Cancer Res 64(13):4428–4433, Pathol 25(7):942–948, 2001a.
2004. Luttges J, et al: Allelic loss is often the first hit in the biallelic inactiva-
Hostetter C, et al: Cytoplasmic accumulation of the RNA binding tion of the p53 and DPC4 genes during pancreatic carcinogenesis,
protein HuR is central to tamoxifen resistance in estrogen receptor- Am J Pathol 158(5):1677–1683, 2001b.
positive breast cancer cells, Cancer Biol Ther 7(9):1496–1506, 2008. Mahlamaki EH, et al: Comparative genomic hybridization reveals fre-
Hruban RH, et al: Progression model for pancreatic cancer, Clin Cancer quent gains of 20q, 8q, 11q, 12p, and 17q, and losses of 18q, 9p,
Res 6(8):2969–2972, 2000. and 15q in pancreatic cancer, Genes Chromosomes Cancer 20(4):383–
Hruban RH, et al: Molecular pathology of pancreatic cancer, Cancer J 391, 1997.
7(4):251–258, 2001. Maire F, et al: Prognosis of malignant intraductal papillary mucinous
Hruban RH, et al: An illustrated consensus on the classification of tumours of the pancreas after surgical resection. Comparison with
pancreatic intraepithelial neoplasia and intraductal papillary muci- pancreatic ductal adenocarcinoma, Gut 51(5):717–722, 2002.
nous neoplasms, Am J Surg Pathol 28(8):977–987, 2004. Maitra A, et al: Cyclooxygenase 2 expression in pancreatic adenocar-
Hucl T, et al: A syngeneic variance library for functional annotation of cinoma and pancreatic intraepithelial neoplasia: an immunohisto-
human variation: application to BRCA2, Cancer Res 68(13):5023– chemical analysis with automated cellular imaging, Am J Clin Pathol
5030, 2008. 118(2):194–201, 2002.
Hussain S, et al: Direct interaction of FANCD2 with BRCA2 in DNA Maitra A, et al: Multicomponent analysis of the pancreatic adenocar-
damage response pathways, Hum Mol Genet 13(12):1241–1248, cinoma progression model using a pancreatic intraepithelial neopla-
2004. sia tissue microarray, Mod Pathol 16(9):902–912, 2003.
Iacobuzio-Donahue CA, et al: Dpc-4 protein is expressed in virtually Mardin WA, et al: MicroRNAs: novel diagnostic and therapeutic tools
all human intraductal papillary mucinous neoplasms of the pancreas: for pancreatic ductal adenocarcinoma?, Ann Surg Oncol 16(11):3183–
comparison with conventional ductal adenocarcinomas, Am J Pathol 3189, 2009.
157(3):755–761, 2000. Martin SA, et al: Genomic instability and the selection of treatments
Iacobuzio-Donahue CA, et al: Exploring the host desmoplastic for cancer, J Pathol 220(2):281–289, 2010.
response to pancreatic carcinoma: gene expression of stromal and Mathew CG: Fanconi anaemia genes and susceptibility to cancer,
neoplastic cells at the site of primary invasion, Am J Pathol 160(1):91– Oncogene 25(43):5875–5884, 2006.
99, 2002. Matsuo Y, et al: K-Ras promotes angiogenesis mediated by immortal-
Iacobuzio-Donahue CA, et al: Highly expressed genes in pancreatic ized human pancreatic epithelial cells through mitogen-activated
ductal adenocarcinomas: a comprehensive characterization and protein kinase signaling pathways, Mol Cancer Res 7(6):799–808,
comparison of the transcription profiles obtained from three major 2009.
technologies, Cancer Res 63(24):8614–8622, 2003a. Meeker AK, et al: Telomere length abnormalities occur early in the
Iacobuzio-Donahue CA, et al: Exploration of global gene expression initiation of epithelial carcinogenesis, Clin Cancer Res 10(10):3317–
patterns in pancreatic adenocarcinoma using cDNA microarrays, 3326, 2004.
Am J Pathol 162(4):1151–1162, 2003b. Mehlen P, et al: Novel roles for slits and netrins: axon guidance cues
Iacobuzio-Donahue CA, et al: Large-scale allelotype of pancreaticobili- as anticancer targets?, Nat Rev Cancer 11(3):188–197, 2011.
ary carcinoma provides quantitative estimates of genome-wide allelic Morton JP, et al: Mutant p53 drives metastasis and overcomes growth
loss, Cancer Res 64(3):871–875, 2004. arrest/senescence in pancreatic cancer, Proc Natl Acad Sci U S A
Iacobuzio-Donahue CA, et al: DPC4 gene status of the primary carci- 107(1):246–251, 2010.
noma correlates with patterns of failure in patients with pancreatic Moskaluk CA, et al: p16 and K-ras gene mutations in the intraductal
cancer, J Clin Oncol 27(11):1806–1813, 2009. precursors of human pancreatic adenocarcinoma, Cancer Res
Ikeda N, et al: The association of K-ras gene mutation and vascular 57(11):2140–2143, 1997.
endothelial growth factor gene expression in pancreatic carcinoma, Murphy KM, et al: Evaluation of candidate genes MAP2K4, MADH4,
Cancer 92(3):488–499, 2001. ACVR1B, and BRCA2 in familial pancreatic cancer: deleterious
Jaeger D, et al: The role of synaptic and voltage-gated currents in BRCA2 mutations in 17%, Cancer Res 62(13):3789–3793, 2002.
the control of Purkinje cell spiking: a modeling study, J Neurosci Murphy SJ, et al: Genetic alterations associated with progression from
17(1):91–106, 1997. pancreatic intraepithelial neoplasia to invasive pancreatic tumor,
Jiao Y, et al: DAXX/ATRX, MEN1, and mTOR pathway genes are Gastroenterology 145(5):1098–1109, 2013.
frequently altered in pancreatic neuroendocrine tumors, Science Nakata B, et al: Prognostic value of microsatellite instability in
331(6021):1199–1203, 2011. resectable pancreatic cancer, Clin Cancer Res 8(8):2536–2540,
Jiao Y, et al: Whole-exome sequencing of pancreatic neoplasms with 2002.
acinar differentiation, J Pathol 232(4):428–435, 2014. Olive KP, et al: The use of targeted mouse models for preclinical testing
Jones S, et al: Core signaling pathways in human pancreatic cancers of novel cancer therapeutics, Clin Cancer Res 12(18):5277–5287,
revealed by global genomic analyses, Science 33:1801–1806, 2008. 2006.
Jones S, et al: Exomic sequencing identifies PALB2 as a pancreatic Petersen GM, et al: Familial pancreatic cancer: where are we in 2003?,
cancer susceptibility gene, Science 324(5924):217, 2009. J Natl Cancer Inst 95(3):180–181, 2003.
Karreth FA, et al: Modelling oncogenic Ras/Raf signalling in the Ranade K, et al: Mutations associated with familial melanoma impair
mouse, Curr Opin Genet Dev 19(1):4–11, 2009. p16INK4 function, Nat Genet 10(1):114–116, 1995.
Klein AP, et al: Prospective risk of pancreatic cancer in familial pan- Raynaud CM, et al: Telomere length, telomeric proteins and genomic
creatic cancer kindreds, Cancer Res 64(7):2634–2638, 2004. instability during the multistep carcinogenic process, Crit Rev Oncol
Korsse SE, et al: Pancreatic cancer risk in Peutz-Jeghers syndrome Hematol 66(2):99–117, 2008.
patients: a large cohort study and implications for surveillance, J Med Richards NG, et al: HuR status is a powerful marker for prognosis and
Genet 50(1):59–64, 2013. response to gemcitabine-based chemotherapy for resected pancreatic
Le DT, et al: A live-attenuated Listeria vaccine (ANZ-100) and a live- ductal adenocarcinoma patients, Ann Surg 252(3):499–505, 2010.
attenuated Listeria vaccine expressing mesothelin (CRS-207) for Rigaud G, et al: Allelotype of pancreatic acinar cell carcinoma, Int J
advanced cancers: phase I studies of safety and immune induction, Cancer 88(5):772–777, 2000.
Clin Cancer Res 18(3):858–868, 2012. Rosenfeld N, et al: MicroRNAs accurately identify cancer tissue origin,
Liu W, et al: DNA mismatch repair abnormalities in acinar cell carci- Nat Biotechnol 26(4):462–469, 2008.
noma of the pancreas: frequency and clinical significance, Pancreas Rustgi AK: Familial pancreatic cancer: genetic advances, Genes Dev
43(8):1264–1270, 2014. 28(1):1–7, 2014.
Chapter 9B Molecular pathology of pancreatic cancer and premalignant tumors 149.e3
Ryu B, et al: Invasion-specific genes in malignancy: serial analysis of Tersmette AC, et al: Increased risk of incident pancreatic cancer
gene expression comparisons of primary and passaged cancers, among first-degree relatives of patients with familial pancreatic
Cancer Res 61(5):1833–1838, 2001. cancer, Clin Cancer Res 7(3):738–744, 2001.
Ryu B, et al: Relationships and differentially expressed genes among Thakker RV: Multiple endocrine neoplasia type 1 (MEN1) and type 4
pancreatic cancers examined by large-scale serial analysis of gene (MEN4), Mol Cell Endocrinol 2014 386(1-2):2–15, 2014.
expression, Cancer Res 62(3):819–826, 2002. Thomas AM, et al: Mesothelin-specific CD8(+) T cell responses
Sahin F, et al: Loss of Stk11/Lkb1 expression in pancreatic and biliary provide evidence of in vivo cross-priming by antigen-presenting cells
neoplasms, Mod Pathol 16(7):686–691, 2003. in vaccinated pancreatic cancer patients, J Exp Med 200(3):297–306,
Samartzis EP, et al: Loss of ARID1A expression sensitizes cancer cells 2004.
to PI3K- and AKT-inhibition, Oncotarget 5(14):5295–5303, 2014. Turner N, et al: Hallmarks of “BRCAness” in sporadic cancers, Nat
Sato N, et al: Aberrant methylation of CpG islands in intraductal Rev Cancer 4(10):814–819, 2004.
papillary mucinous neoplasms of the pancreas, Gastroenterology Tuveson DA, et al: Ductal pancreatic cancer in humans and mice, Cold
123(1):365–372, 2002. Spring Harb Symp Quant Biol 70:65–72, 2005.
Sato N, et al: Discovery of novel targets for aberrant methylation in Ueki T, et al: Hypermethylation of multiple genes in pancreatic adeno-
pancreatic carcinoma using high-throughput microarrays, Cancer Res carcinoma, Cancer Res 60(7):1835–1839, 2000.
63(13):3735–3742, 2003a. van der Heijden MS, et al: Fanconi anemia gene mutations in young-
Sato N, et al: Frequent hypomethylation of multiple genes overex- onset pancreatic cancer, Cancer Res 63(10):2585–2588, 2003.
pressed in pancreatic ductal adenocarcinoma, Cancer Res 63(14): van der Heijden MS, et al: Functional defects in the fanconi anemia
4158–4166, 2003b. pathway in pancreatic cancer cells, Am J Pathol 165(2):651–657,
Sato N, et al: Gene expression profiling identifies genes associated with 2004.
invasive intraductal papillary mucinous neoplasms of the pancreas, van der Heijden MS, et al: In vivo therapeutic responses contingent
Am J Pathol 164(3):903–914, 2004a. on Fanconi anemia/BRCA2 status of the tumor, Clin Cancer Res
Sato N, et al: Gene expression profiling identifies genes associated with 11(20):7508–7515, 2005.
invasive intraductal papillary mucinous neoplasms of the pancreas, van Heek NT, et al: Telomere shortening is nearly universal in pancre-
Am J Pathol 164(4):903–914, 2004b. atic intraepithelial neoplasia, Am J Pathol 161(5):1541–1547,
Schutte M, et al: Abrogation of the Rb/p16 tumor-suppressive pathway 2002.
in virtually all pancreatic carcinomas, Cancer Res 57(15):3126–3130, Wilentz RE, et al: Inactivation of the p16 (INK4A) tumor-suppressor
1997. gene in pancreatic duct lesions: loss of intranuclear expression,
Sessa F, et al: Intraductal papillary-mucinous tumours represent a dis- Cancer Res 58(20):4740–4744, 1998.
tinct group of pancreatic neoplasms: an investigation of tumour cell Whitcomb DC, et al: Hereditary pancreatitis is caused by a mutation
differentiation and K-ras, p53 and c-erbB-2 abnormalities in 26 in the cationic trypsinogen gene, Nat Genet 14(2):141–145, 1996.
patients, Virchows Arch 425(4):357–367, 1994. Wilentz RE, et al: Loss of expression of Dpc4 in pancreatic intraepi-
Shi C, et al: KRAS2 mutations in human pancreatic acinar-ductal thelial neoplasia: evidence that DPC4 inactivation occurs late in
metaplastic lesions are limited to those with PanIN: implications for neoplastic progression, Cancer Res 60(7):2002–2006, 2000a.
the human pancreatic cancer cell of origin, Mol Cancer Res 7(2):230– Wilentz RE, et al: Genetic, immunohistochemical, and clinical features
236, 2009. of medullary carcinoma of the pancreas: a newly described and
Siegel R, et al: Cancer statistics, 2013, CA Cancer J Clin 63(1):11–30, characterized entity, Am J Pathol 156(5):1641–1651, 2000b.
2013. Wu J, et al: Recurrent GNAS mutations define an unexpected pathway
Singhi AD, et al: Preoperative GNAS and KRAS testing in the diag- for pancreatic cyst development, Sci Transl Med 3(92):92ra66,
nosis of pancreatic mucinous cysts, Clin Cancer Res 20(16):4381– 2011.
4389, 2014. Yachida S: Evolution and dynamics of pancreatic cancer progression,
Sirivatanauksorn V, et al: Non-random chromosomal rearrangements Oncogene 32(45):5253–5260, 2016.
in pancreatic cancer cell lines identified by spectral karyotyping, Int Yachida S, et al: Distant metastasis occurs late during the genetic evolu-
J Cancer 91(3):350–358, 2001. tion of pancreatic cancer, Nature 467(7319):1114–1117, 2010.
Stellas D, et al: Therapeutic effects of an anti-myc drug on mouse Yamano M, et al: Genetic progression and divergence in pancreatic
pancreatic cancer, J Natl Cancer Inst 106(12):2014. doi: 10.1093/ carcinoma, Am J Pathol 156(6):2123–2133, 2000.
jnci/dju320. Yeo TP, et al: Assessment of “gene-environment” interaction in cases
Sohn TA, et al: Intraductal papillary mucinous neoplasms of the pan- of familial and sporadic pancreatic cancer, J Gastrointest Surg
creas: an updated experience, Ann Surg 239(6):788–797, 2004. 13(8):1487–1494, 2009.
Su GH, et al: Alterations in pancreatic, biliary, and breast carcinomas Zuo L, et al: Germline mutations in the p16INK4a binding domain of
support MKK4 as a genetically targeted tumor suppressor gene, CDK4 in familial melanoma, Nat Genet 12(1):97–99, 1996.
Cancer Res 58(11):2339–2342, 1998.
Terris B, et al: Characterization of gene expression profiles in intra-
ductal papillary-mucinous tumors of the pancreas, Am J Pathol
160(5):1745–1754, 2002.
CHAPTER 9C
Molecular pathogenesis of biliary tract cancer
Jason K. Sicklick and Paul T. Fanta
BILIARY TRACT CANCERS Study Group of Japan, 2000). Intrahepatic tumors have been
subcategorized by their growth characteristics into three groups,
Cholangiocarcinoma (CC), or biliary tract adenocarcinoma, is namely, mass-forming, periductal-infiltrating, or intraductal-
the sixth leading gastrointestinal tumor in the Western world. growing types (Liver Cancer Study Group of Japan, 2000).
First described by Durand-Fardel in 1840 (Olnes & Erlich, Until recently, International Classification of Disease (ICD)
2004), this malignancy arises within cholangiocytes of the codes combined IHCC with hepatocellular carcinoma under
biliary tree. As a result, tumors may occur within the liver (i.e., the code for primary liver tumor (Khan et al, 2002a, 2002b).
intrahepatic cholangiocarcinoma [IHCC]), within the extrahe- But these are clearly different entities, and the second and third
patic biliary tree (i.e., extrahepatic cholangiocarcinoma editions of the International Classification of Diseases for
[EHCC]), or within the gallbladder (i.e., gallbladder cancer Oncology (ICD-O-2/3) have attempted to correct for this issue.
[GBCA]). Together, these biliary tract cancers constitute a rare However, in ICD-O-2, Klatskin tumors were assigned a unique
set of malignancies with poor prognoses. Patients diagnosed histology code, but this was cross-referenced to the topography
with these tumors often are seen late in their clinical course, code for intrahepatic rather than extrahepatic tumors. Under
and chemotherapeutic regimens lack significant response rates. the third ICD-O-3 edition, Klatskin tumors are cross-
As a result, in patients with advanced disease treatment goals referenced to either location (Khan et al, 2012). In addition to
are often palliative in nature (Ishak et al, 1994) (see Chapters the aforementioned coding issues, many tumors previously
47, 49, 50, and 51). Due to the rarity of these entities, as referred to as liver adenocarcinoma of unknown primary site were
well as their common cell of origin, treatment for all three likely unrecognized IHCCs. Together, these changes in ICD-
tumor types has been identical. However, with fairly recent classification have influenced observed changes in the incidence
developments in next-generation sequencing (NGS) and other rates of IHCC and EHC.
molecular techniques, differentiation among these tumor types
can be refined into distinct diseases with unique genetic signa- EPIDEMIOLOGY
tures. Examination of these molecular signatures may be impor-
tant in clinical trial design using drugs targeting specific Although rare, CC has a distinctly higher incidence in certain
pathways. demographic groups and geographic regions (see Chapters 50,
51, and 59). The peak incidence of CC is the seventh decade
CLASSIFICATION of life, with a slightly higher male predilection (Olnes & Erlich,
2004). The annual reported incidences of IHCC and EHCC
The vast majority of biliary tracts tumors are adenocarcinomas in the United States are each approximately 0.6 to 1.0 cases
(Nakeeb et al, 1996), and they most often arise at or near the per 100,000 people (Welzel et al, 2006), with an estimated
hepatic duct confluence. The latter fall under the general cate- 5800 new cases each year. Interestingly, U.S. Surveillance, Epi-
gory of EHCCs (see Chapter 51), which are now further subclas- demiology, and End Results (SEER) data suggest that age-
sified in the current American Joint Committee on Cancer standardized incidence rates (ASIR) for IHCC gradually rose
seventh edition (Edge, 2010) and National Comprehensive from 0.59 per 100,000 people in 1990 to 0.91 in 2001 (Khan
Cancer Network (NCCN) guidelines as perihilar cholangiocarci- et al, 2012). This was followed by a sharp fall before plateauing
noma, or Klatskin tumors, after the Yale University pathologist at 0.60 per 100,000 people in 2007. The ASIR for EHCC
who was one of the first to report a series of patients with this remained relatively stable at approximately 0.80 per 100,000
condition, and distal cholangiocarcinoma (Benson et al, 2014). people until 2001. It then began increasing to 0.97 per 100,000
Bismuth (1992) further categorized perihilar cholangiocarcino- people in 2007. These trends coincide with a switch from
mas by their precise location with reference to the biliary bifurca- ICD-O-2 to ICD-O-3 in 2001. Therefore, although the inci-
tion and the hepatic lobar ducts. This classification is most useful dence of EHCC has remained constant, the incidence of IHCC
for descriptive purposes and for operative planning. In contrast has increased in the United States.
to the perihilar tumors, the distal cholangiocarcinomas account Outside of the United States, CC incidence rates vary mark-
for a relatively small fraction of all bile duct tumors, and midbile edly, presumably reflecting differences in infectious causes, envi-
duct tumors are even rarer (see Chapter 59). Finally, less com- ronmental risk factors (i.e., sedentary lifestyles, alcohol,
monly seen than the aforementioned is disease that has diffuse smoking, and diet) exposure to toxic chemicals, and genomics.
involvement of the entire biliary tree. The highest disease incidence rates are in northeast Thailand
Cholangiocarcinoma may also arise from the intrahepatic (96/100,000 men), where it occurs approximately 100 times
bile ducts, giving rise to the subgroup known as intrahepatic or more often than in the West. For uncertain reasons, epidemio-
peripheral cholangiocarcinoma (see Chapter 50) (Liver Cancer logic studies have shown that the mortality of intrahepatic
150
Chapter 9C Molecular pathogenesis of biliary tract cancer 151
biliary tract neoplasms is rising globally, but those of extrahe- IHCC, the authors concluded that these conditions might
patic tumors are static or falling. Examination of the World explain the divergent incidence trends of IHCC and EHCC.
Health Organization mortality data for the United States, These data also support an earlier report by Palmer and col-
United Kingdom, France, Italy, Japan, and Australia from 1979 leagues who performed a literature review and meta-analysis of
to 1998 (Ishak et al, 1994) demonstrates that mortality rates for case-control studies on IHCC, demonstrating that major risk
IHCC were increased in both sexes in all countries except factors for this cancer include cirrhosis, chronic hepatitis B and
among Japanese women. More recently, a U.S. study examined C, alcohol use, diabetes, and obesity (Palmer et al, 2012).
trends in IHCC incidence using data from the SEER program, Based upon the aforementioned predisposing factors, a
which represents greater than 10% of the total United States common theme of chronic biliary epithelial inflammation
population. Data from 1976 to 2000 were analyzed by age, sex, appears to be a predisposing factor for the development of
and ethnicity. The incidence of IHCC increased by 165%, from biliary tract cancers. Primary sclerosing cholangitis (PSC; see
0.32 per 100,000 (1975 to 1979) to 0.85 per 100,000 Chapter 41) is the most common condition predisposing to
(1995 to 1999). This increase was reflected in all groups but was CC, with rates of 8% to 40% reported in patients with PSC
highest in black men (139%), followed by white men (124%) (Khan et al, 2002a). Compared with sporadic cases, PSC
and white women (111%). The increased incidence may be patients tend to be seen with CC earlier in life (e.g., most com-
attributable, in part, to increased detection (Olnes & Erlich, monly in the 30- to 50-year-old age groups) (Bergquist et al,
2004). However, increased detection of a tumor is usually asso- 1998; Pitt et al, 1995). Experimental and epidemiologic data
ciated with an increase in the proportion of patients with early- also suggest a pathogenic association between CC and liver
stage disease or smaller tumors. But the rise in intrahepatic fluke infestation, especially Opisthorcis viverrini and, less defini-
cancers was not associated with a significant change in the pro- tively, Clonorchis sinensis (see Chapter 45) (Watanapa, 1996).
portion of early-stage cancers, histologically confirmed tumors, Congenital abnormalities of the biliary tree, congenital hepatic
or smaller lesions (Broome et al, 1996). Furthermore, the inci- fibrosis, and choledochal cysts (cystic dilatations of the bile
dence did not seem to be plateauing, as would be expected ducts), also carry a 15% risk of malignant change after the
if the increase were due to an improvement in diagnostic modal- second decade of life, with an average age of onset at 34 years
ities, such as magnetic resonance imaging and computed (see Chapter 46) (Scott et al, 1980). Furthermore, with
tomography (Khan et al, 2002a; Shaib & El-Serag, 2004; Taylor- untreated choledochal cysts the risk increases to 28% (Lipsett
Robinson et al, 2001). Conversely, the incidence of EHCC has et al, 1994; Ohtsuka et al, 2001). Biliary stasis, reflux of pan-
shown no such change and may be declining. According to creatic juice, activation of bile acids (van Mil et al, 2001), and
SEER data, U.S. age-standardized mortality rates for extrahe- deconjugation of carcinogens are also speculated as mechanistic
patic tumors fell from 0.6 per 100,000 in 1979 to 0.3 per drivers of carcinogenesis related to the theme of chronic inflam-
100,000 in 1998; and age-standardized incidence rates decreased mation. Finally, biliary adenomas and papillomatosis have been
from 1.08 per 100,000 to 0.82 per 100,000 in the same period. associated with the development of CC.
Despite improved imaging modalities, most patients are seen
with unresectable disease and typically die within 12 months of Biology of Clinical Risk Factors
diagnosis. In addition to a lack of highly efficacious systemic Several underlying mechanisms play a role in the induction of
therapies, sepsis from cholangitis, frequently related to inter- chronic biliary inflammation and cholestasis.
ventions performed for biliary obstruction and progressive liver
failure, contribute to the high mortality (Shaib & El-Serag, Bile Content and Deconjugation of Xenobiotics
2004). Although biliary tract tumors remain relatively rare, Polymorphisms in bile salt transporter proteins (i.e., BSEP,
there has been increased interest in studying the biology of ATP8B1, and ABCB4) can lead to unstable bile content and
these diseases in recent years (Khan et al, 2002a; Patel, 2001, deconjugation of environmental toxins (i.e., xenobiotics) previ-
2002; Taylor-Robinson et al, 2001). ously conjugated in the liver (Jacquemin, 2001; Meier et al,
2004; Thompson & Strautnieks, 2001). In the background of
CHRONIC BILIARY INFLAMMATION AND congenital bile duct abnormalities, this process increases the
CHOLESTASIS risk of CC (Kubo et al, 1995). Individuals who are heterozy-
gous for bile salt transporter polymorphisms are thought to
Clinical Risk Factors have an increased predisposition to CC as adults, following
Reported risk factors for biliary tract cancers comprise a diverse exposure to cofactors that result in chronic inflammation in the
group of conditions that include infectious causes, congenital biliary tree (Kubo et al, 1995).
conditions, inflammatory diseases, drugs, environmental expo-
sures, and toxins (Patel, 2014). In a study of the SEER data- DNA Mutagens
base, Welzel and colleagues found that in addition to established Promutagenic deoxyribonuclease (DNA) adducts have been
risk factors (e.g., choledochal cysts and cholangitis), biliary identified in CC tissue, indicating exposure to DNA-damaging
cirrhosis, cholelithiasis, hepatolithiasis, alcoholic liver disease, agents (Khan et al, 2003). Thorotrast, a radiologic contrast
nonspecific cirrhosis, diabetes, thyrotoxicosis and chronic pan- agent banned in the 1960s, has been strongly associated with
creatitis, obesity, chronic nonalcoholic fatty liver disease the development of CC many years after exposure and may
(NAFLD), hepatitis B virus infection, hepatitis C virus (HCV) increase the risk of cholangiocarcinogenesis 300-fold (Hardell
infection, human immunodeficiency virus infection, and et al, 1984; Sahani et al, 2003; Shaib & El-Serag, 2004). Expo-
smoking were associated with the development of these tumors sures to byproducts from the rubber and chemical industries,
(Welzel et al, 2006). Given that three of these factors (i.e., HCV including dioxins and nitrosamines (Sorensen et al, 1998), as
infection, chronic NAFLD, and obesity) have increased in inci- well as those from alcohol and smoking (Chalasani et al, 2000),
dence in the United States and were only associated with have been implicated.
Inherited Syndrome advanced molecular analyses, including NGS, rapid progress is

Lynch syndrome, an autosomal dominant predisposition for being made in our understanding of the genomic basis of these
DNA mismatch repair, is associated with a high incidence of malignancies. Tumor profiling of biliary tract carcinomas has
colorectal, endometrial, stomach, ovarian, pancreatic, ureter also revealed distinct molecular alterations that differentiate
and renal pelvis, bile duct, and brain tumors (Shigeyasu et al, IHCC, EHCC, and GBCA, perhaps identifying distinct thera-
2014). The associated lifetime risk for bile duct cancer in peutic options in each tumor subtype. It is beyond the scope of
patients with Lynch syndrome is approximately 2%. Although this chapter to document every known molecular alteration
hereditary, and not environmental, DNA repair is a recurrent reported or associated with CC, but rather focuses upon recur-
theme in the development of biliary tract carcinomas. rent themes in altered signaling pathways that together result
in the pathogenic phenotype (Fig. 9C.1).
MOLECULAR PATHOGENESIS Biology of Biliary Epithelial Injury and Repair
The molecular pathogenesis of biliary tract tumors is now Similar to the development of other tumors, biliary tract carci-
becoming an area of vigorous investigation. In the era of nogenesis is thought to be a multistep process dependent on
Chronic biliary Reactive cellular

Malignant
inflammation + Cellular injury repair + clonal Tumor growth + metastasis
transformation
cholestasis proliferation
Genomic Alterations Dysregulated Signaling Pathways

ARID1A IDH1/2 Worse Prognosis Better Prognosis
BAP1 KRAS COX-2 Chemokines
BRAF MET EGFR Interleukins
CDKN2A/B loss PBRM1 ERBB2 JAK/STAT
ERBB2 PIK3CA HGF/MET
Risk Factors FGFR1-3 fusions TP53 Hippo
Alcohol Liver flukes IDH1/2
Cholangitis NAFLD Epigenetic Alterations IGF
Choledochocal cyst Obesity APC p14ARF PI3K/AKT/mTOR
Cirrhosis PSC CDH1 p15 PLK1/2
CDKN2A p16INK4a RAS/RAS/MEK/ERK
Diabetes Smoking
GSTP1 RASSF1A
HBV/HCV Thyrotoxicosis
hMLH1 SOCS-3
Lifestyle Toxins
MicroRNA (miR) Alterations
miR21 miR199a
Inflammatory miR22 miR199a*
miR125a miR376a
signaling Biliary tract adenocarcinoma
miR127 miR424
Mitogenic Factors
Bcl-2 ERBB2 Proliferative epithelium
Biliary epithelium COX-2 IL-6/STAT3 Transforming
CXCR4 Notch-1 events
Epithelial-Stromal Interactions Epithelial-Mesenchymal

IL-6/STAT TGF-β Transition (EMT)
miR200c miR214
miR204
Tumor-Stromal Interactions
Hedgehog (Hh) PDGF
Notch PLK
Stromal cells Mesenchymal cancer cell

(Hepatic stellate cells +
myofibroblasts)
FIGURE 9C.1. Vogelgram of biliary carcinogensis. The progression from benign biliary epithelium to biliary tract adenocarcinoma occurs through
a series of stages, including chronic biliary inflammation and cholestasis caused by several risk factors, followed by cellular injury, reactive cellular
repair, clonal proliferation, malignant transformation, tumor growth, and metastasis. Each one of these steps is regulated by many factors, including
epithelial-stromal interactions, mitogens, genomic alterations, epigenetic alterations, microRNAs, dysregulated signaling pathways, epithelial-to-
mesenchymal transitions, and tumor-stromal interactions. APC, adenomatosis polyposis coli; ARID1A, AT-rich interaction domain 1A; BAP1, BRCA1-
associated protein 1; Bcl-2, B-cell chronic lymphocytic leukemia/lymphoma; BRAF, B-Raf protooncogene; CDH1, cadherin 1, CDKN, cyclin-dependent
kinase inhibitor; COX-2, cyclooxygenase-2; EGFR, epidermal growth factor receptor; ERBB2, ERB-B2 receptor tyrosine kinase 2; FGFR1-3, fibroblast
growth factor receptor 1-3; GSTP1, glutathione-S-transferase pi 1; HBV/HCV, hepatitis B and C virus, respectively; HGF, hepatocyte growth factor;
hMLH1, human mutL homolog 1; IDH1/2, isocitrate dehydrogenase 1/2; IGF, insulin-like growth factor; IL-6, interleukin-6; JAK, Janus activating
kinase; KRAS, Kirsten rat sarcoma; mTOR, mechanistic target of rapamycin; NAFLD, nonalcoholic liver disease; PBRM1, polybromo 1; PDGF,
platelet-derived growth factor; PI3K, phosphatidylinositol-3-kinase; PIK3CA, phosphatidylinositol-4,5-bisphosphate-3-kinase catalytic subunit alpha;
PLK, polo-like kinase; PSC, primary sclerosing cholangitis; SOC-3, general sugar transporter; STAT, signal-transducer and activator of transcription;
TGF-β, transforming growth factor-β; TP53, tumor protein 53.
the interaction between environmental factors and host genetic glycosylase (NEIL)1 genes encode DNA glycosylases involved
factors. Most of the putative environmental risk factors for CC in repair of oxidative base damage, and mutations in these genes
result in chronic biliary inflammation leading to repair mecha- are associated with CC.
nisms and, ultimately, carcinogenesis.
Conceptually, exposure to an inflammatory stimulus would Activation-Induced Cytidine Deaminase
not have the same effect on each cell due to changes in perfu- Other work has also suggested that chronic inflammation can
sion (e.g., centrilobular versus periportal), as well as differen- play a critical role leading up to cholangiocarcinogenesis.
tial levels of cytochrome P450 expression, exposure to bile salt Komori and colleagues (2008) found that the proinflammatory
concentrations, and exposure to inflammatory components cytokine-induced aberrant production of activation-induced
(e.g., cytokines and immune surveillance cellular components, cytidine deaminase (AID), a member of the DNA/RNA-editing
such as Kupffer cells and hepatic stellate cells) (see Chapters 7 enzyme family, might link bile duct inflammation to an
and 10). Based on this, the concept of heterogeneity can be enhanced genetic susceptibility to mutagenesis that leads to
inferred where distinctive clonal populations may arise based CC. Ectopic AID production is induced in response to tumor
upon differential response to stimuli. In this section, we review necrosis factor-α (TNF-α) stimulation via a nuclear factor
the underlying host factors associated with bile tract cancers. kappa B (NF-κB)–dependent pathway. Aberrant expression of
AID in biliary cells results in the generation of somatic muta-
Genetic Polymorphism at Cytochrome P450 tions in tumor-related genes, including tumor protein 53
Genetic polymorphisms exist in the cytochrome (CYP) P450 (TP53), c–myelocytomatosis viral oncogene (c-MYC), and the
enzyme complex, a large family of constitutive and inducible promoter region of the cyclin-dependent kinase inhibitor A
enzymes that play a central role in the oxidative metabolism of (CDKN2A) sequences. In human tissue specimens, reverse
both environmental toxins and endogenous compounds. These transcription polymerase chain reaction analyses revealed that
polymorphisms play a critical role in how endogenous and AID was significantly increased in 28 of 30 CC tissues (93%),
exogenous toxins are biotransformed by the liver. Similar to whereas only trace amounts of AID were detected in the normal
many other cancers, which rely upon a sequence of chronic liver. Immunohistochemistry showed that all of the CC tissue
injury and repair, the development of CC may be partially regu- samples examined showed overproduction of endogenous AID
lated by the host ability to respond toxic insults. protein in cancer cells. Moreover, immunostaining for AID was
Several CYPs are involved in metabolism of oxysterols, detectable in 16 of 20 bile epithelia in PSC.
which are cholesterol oxidation products whose expression may
be dysregulated in inflammation-related diseases, including Human CYP1A2 and Arylamine N-Acteyltransferases
cancer. A recent study focused on CYP39A1, which can metab- (NAT1 and NAT2)
olize 24-hydroxycholesterol and plays an important role in the The CYP1A2, NAT1, and NAT2 genes have been shown to
inflammatory response and oxidative stress (Khenjanta et al, be potential modifiers of an individual’s susceptibility to
2014). Immunohistochemistry showed that 70% of CC patients certain types of cancers. Prawan and colleagues (2005) evalu-
had low CYP39A1 expression, which correlated with metasta- ated the relationship between CYP1A2, NAT1, and NAT2
sis. In the setting of Opisthorchis-associated CC, the expression polymorphisms in Thai patients with CC. A total of 216 CC
and function of CYP2A6 and CYP2E1 were altered in the livers patients and 233 control subjects were genotyped using PCR.
of 13 patients (Yongvanit et al, 2012). This report provides Two CYP1A2 alleles (CYP1A2*1A wild type and *1F), six
evidence that enhanced CYP2A6 activity and diminished NAT1 alleles (NAT1*4 wild type, *3, *10, *11, *14A, and
CYP2E1 activity may be involved in the progression of CC. *14B), and seven NAT2 alleles (NAT2*4 wild type, *5, *6A,
Finally, recent molecular profiling of EHCC specimens dem- *6B, *7A, *7B, and *13), were analyzed. The CYP1A2*1A/*1A
onstrated significant enrichment of CYP-metabolic pathways, genotype conferred a decreased risk of the cancer (adjusted
including transcription factors such as glutathione-S-transfer- odds ratio [OR], 0.28; 95% confidence interval [CI], 0.08 to
ase α1 (GSTA1) and GSTA3, which may cause abnormal gene 0.94), compared with CYP1A2*1F/1*F. Frequency distribu-
expression and tumorigenesis through CYP450-metabolic tions of rapid NAT2*13 and two slow alleles, *6B and *7A,
pathways (Qi et al, 2014). Therefore differential CYP activity were associated with lower CC risk. This study suggests that
may be involved in the initiation and/or progression of disease the NAT2 polymorphism might be a modifier of individual
via modulation of chronic inflammation, viral hepatitis, para- risk of CC.
sitic infestation, and recurrent cholangitis (Khan et al, 2005).
Trefoil Factor Family
MRP2/ABCC2 Trefoil factor family 1 (TFF1) is critical for mucosal protection
Hoblinger and colleagues (2009) reported that multidrug and tumor suppression in the stomach. To examine its role in
resistance–associated protein 2 (MRP2/ABCC2) is one of the cholangiocarcinogenesis, specimens with varying degrees of
adenosine triphosphate-binding cassette (ABC) transporters dysplasia were examined. These included IHCC as a result of
expressed on the apical membrane of hepatocytes and cholan- hepatolithiasis, biliary epithelial dysplasia with hepatolithiasis,
giocytes. ABCC2 plays an important role in the biliary clear- hepatolithiasis without dysplasia or carcinoma, IHCC without
ance of endogenous and exogenous toxic compounds. The hepatolithiasis, and control normal livers (Sasaki et al, 2003).
ABCC2 variant c.3972C>T in exon 28 has been shown to be TFF1 expression in the biliary epithelium was increased in
associated with the risk of carcinogenesis. hepatolithiasis compared with control livers (P < .01). In biliary
epithelial dysplasia and noninvasive IHCC with hepatolithiasis,
MUTYH and NEIL1 TFF1 was extensively expressed, and MUC5AC gastric mucin
Forsbring and colleagues (2009) recently found that the human was usually colocalized with TFF1. However, TFF1 expression
mutY DNA glycosylase (h-MUTYH) and Nei-like DNA was significantly decreased in invasive IHCC, despite preserved
expression of MUC5AC. A total of four missense mutations Gallbladder Intrahepatic

were detected: Three were found in two noninvasive IHCC with
hepatolithiasis (29%) and one in invasive IHCC (11%). Loss
of heterozygosity of the TFF1 gene was not detectable. The
decreased expression of TFF1 in invasive IHCC may be
explained by the methylation of the TFF1 promoter region.
Upregulation of TFF1 coupled with MUC5AC in biliary epi- 291
thelium in hepatolithiasis, biliary epithelial dysplasia, and non- 790 1633
invasive IHCC may reflect gastric metaplasia and early
neoplastic lesion. Under such conditions, decreased TFF1
expression may lead to increased cell proliferation and then to 165
the invasive character of IHCC. 35 265
Biliary Epithelial Proliferation

Ultimately, the underlying chronic biliary injury, which may be
modified by toxin exposure, along with modulation by endog-
enous host factors or inherited gene polymorphisms, may lead
80
to an exaggerated repair response that results in cholangiocyte
proliferation.
Mitogenic Factors
Identification of these mitogenic stimuli with resultant increases
in DNA, RNA, protein synthesis, as well as increases in immune Extrahepatic
modulation, have been implicated in the carcinogenesis of FIGURE 9C.2. Common gene expression alterations in biliary tract
biliary tract tumors. The liver fluke Opisthorchis viverrini can cancers. A Venn diagram is used to depict the relationship of transcrip-
generate mitogenic substances, such as glutathione-S-transfer- tional changes among biliary cancer subtypes. There were 165 common
ase (GST), which act as a secretory product that may play an genes with significantly altered expression in intrahepatic cholangiocar-
important role in promoting the genesis of CC. GST has a cinoma, extrahepatic cholangiocarcinoma, and gallbladder cancer.
proliferative function on NIH-3T3 murine fibroblasts and
MMNK1 nontumorigenic human bile duct epithelial cells in a
dose-dependent manner with subsequent in vitro activation of Malignant Transformation
both phospho-AKT (v-akt murine thymoma viral oncogene Gene Expression Analysis
homolog 1) (pAKT) and phospho-ERK (extracellular signal- In 2009, Miller and colleagues (2009) investigated the molecular
regulated kinase) (pERK) (Daorueang et al, 2012). Other alterations in carcinomas of the biliary tree, including IHCC,
mitogens can activate these pathways. Recombinant human EHCC, and GBCA, using frozen specimens from patients who
TFF2 can stimulate proliferation and trigger phosphorylation underwent surgical resection. Unsupervised hierarchical cluster-
of epidermal growth factor receptor (EGFR) with downstream ing analysis revealed that cancers from these different sites did
ERK activation, displaying potential mitogenic impact in CC not cluster separately, implying that there was no difference in
via EGFR/mitogen-activated protein kinase (MAPK) activation the global gene expression patterns between the biliary cancer
(Kosriwong et al, 2011). Moreover, TFF family members have subgroups. Although the individual cancer subtypes did not
been assessed as potential biomarkers in CC. Gene copy cluster separately, unique patterns of differential gene expression
number, messenger ribonucleic acide (mRNA) levels, and were observed when compared with normal biliary epithelium.
protein expression were evaluated in bile duct epithelium biop- The relationship between gene transcriptional changes among
sies collected from individuals with CC, precancerous bile duct the three biliary cancer subtypes is depicted in a Venn diagram
dysplasia, and from disease-free control participants. The (Fig. 9C.2). Overall, 165 probe sets were commonly differen-
TFF1, TFF2, and TFF3 mRNA levels were significantly tially expressed in all three cancer subtypes. Selected commonly
increased in CC tissue. differentially expressed genes are listed in Table 9C.1.
In healthy tissues, cellular senescence results in irreversible
growth arrest. However, this is prevented in malignant cells by Copy-Number Alterations
maintenance of chromosomal length via telomerase activity. The same study (Miller et al, 2009) also investigated alterations
This is observed in CC cells but not in normal cholangiocytes. in gene copy number using array-based comparative genome
Data would suggest that interleukin-6 (IL-6) is partially respon- hybridization (aCGH) analysis. Considerable heterogeneity
sible for this because it acts as an autocrine promoter of CC was observed in the extent of chromosomal instability between
growth. IL-6 stimulation leads to enhanced telomerase, patients, even within the same subtype (i.e., IHCC, EHCC,
decreased cellular senescence, and thereby increased CC and GBCA). For example, some patients had alterations in
growth (Yamagiwa et al, 2006). Moreover, in conjunction with nearly every chromosomal arm, and other patients with the
hepatocyte growth factor (HGF), IL-6 increases CC cell growth same diagnosis had few structural chromosomal changes.
in vitro, as well as induces a rapid release of prostaglandin Despite the heterogeneity, segments of chromosomes 3p, 6q,
synthesis, followed by downstream signal transduction via 8p, 9p, and 14q were commonly lost across all three biliary
MAPKs, protein kinase C, and calmodulin (Wu et al, 2002). cancers subtypes. Commonly amplified chromosomal regions
Taken together, these, and probably other mitogens, drive the across the three subtypes include segments of 1q, 3q, 5p, 7p,
proliferation of CC. 7q, 8q, and 20q. Based upon the expression analysis data and
TABLE 9C.1 Top 10 Genes with Significantly Altered Expression in Gallbladder Cancer (GB), Intrahepatic Cholangiocarcinoma (Intra), and
Extrahepatic Cholangiocarcinoma (Extra) Along with the Fold Change (Δ) in Expression in Each Cancer Subtype
Chromosomal
Δ GB Δ Intra Δ Extra Gene Gene Title Location Functional Pathway
58 11 11 RRM2 Ribonucleotide reductase M2 2p25-p24 Nucleotide metabolism

41 6 10 PTTG1 Pituitary tumor–transforming 1 5q35.1 Cell cycle
26 9 7 TYMS Thymidylate synthetase 18p11.32 Nucleotide metabolism
19 4 5 CDK1 Cell division cycle 2, G1 to S, and G2 to M 10q21.1 Cell cycle
18 4 7 CCNB2 Cyclin B2 15q22.2 Cell cycle
−13 −27 −14 CDKN1C Cyclin-dependent kinase inhibitor 1C (p57, Kip2) 11p15.5 G1-to-S cell cycle
−61 −27 −26 ALDH1A2 Aldehyde dehydrogenase 1 family, member A2 15q21.2 Metabolism/biosynthesis
−67 −20 −7 CNN1 Calponin 1, basic, smooth muscle 19p13.2-p13.1 Muscle contraction
−85 −51 −18 CES1 Carboxylesterase 1 16q22.2 Irinotecan pathway
−102 −30 −24 DES Desmin 2q35 Muscle contraction
aCGH data, this work suggested that these three biliary tract 130 site-specific gene mutations. Mutations were identified
tumors were fairly similar in their molecular signature. within several of these genes, including KRAS (35%), TP53
However, a more recent study demonstrated that even within (22%), phosphatidylinositol-4,5-bisphosphate-3-kinase cata-
IHCC, there are distinct patterns. They analyzed 149 formalin- lytic subunit alpha (PIK3CA, 10%), B-Raf protooncogene
fixed IHCC cases by single nucleotide polymorphism array, and (BRAF, 7%), adenomatosis polyposis coli (APC, 6%), neuro-
identified several similar and different copy-number alterations blastoma RAS viral oncogene (NRAS,3%), isocitrate dehydro-
(CNAs), including gains at 1q and 7p as well as losses at 3p, genase 1 (IDH1,2%), v-AKT murine thymoma viral oncogene
4q, 6q, 8p, 9p, 9q, 13q, 14q, 17p, and 21q (Sia D. et al, 2013). homolog 1 (AKT1, 1%), β1-catenin (CTNNB1, 1%), and phos-
Although gains in chromosomal arms 1q and 7p seem frequent, phatase and tensin (PTEN, 1%). Although IDH1 mutations
there are clear discrepancies between studies where DNA copy- were rare in other common gastrointestinal malignancies, they
number gains range from 0% to 73% (Endo et al, 2002; Sirica, were identified in three tumors (25%) of an initial series of 12
2008; Sirica et al, 2002). biliary tract carcinomas. To better define both IDH1 and IDH2
mutations, an additional 75 bile duct cancers, making 87 total
Pre–Next-Generation Sequencing tumors (IHCC [N = 40], EHCC [N = 22], and GBCA [N =
With the revolutionary changes that are occurring in genomics, 25)] were examined (Table 9C.2).
the cost, sequencing time, and analysis time of NGS has sig- Combining these cohorts of biliary tract cancers, mutations
nificantly decreased during the last several years. Whole-exome in IDH1 and IDH2 were found in 10.3% of cases. On subset
sequencing and targeted sequencing of several hundred cancer- analysis (Table 9C.3), only IHCC (9/40, 23%) had IDH1
specific genes has provided significant insight and a deeper (20%) or IDH2 (3%) mutations, whereas none were identified
understanding of oncogenes and tumor-suppressor genes in EHCC or GBCA. Therefore IDH1 mutations were defined
involved in biliary tract carcinogenesis. In the era preceding as a molecular feature of IHCC, as well as suggested as a
NGS, several studies showed abnormal expression of the potential therapeutic target specific to IHCC. In contrast,
Kirsten rat sarcoma (KRAS) oncogene in 21% to 100% of KRAS (23%) and TP53 (14%) predominated in EHCC,
cases, as well as alteration in the TP53 tumor-suppressor gene whereas PIK3CA (12%) mutations were the most common in
in up to 37% of archival CC specimens (Isa et al, 2002). These GBCA. For the first time, this study defined a specific genetic
genetic alterations were associated with a more aggressive phe- abnormality in this lethal cancer that represented a potentially
notype in biliary tract tumors (Isa et al, 2002). KRAS and TP53 new target for therapy. Because the original publications were
mutations were also been identified in bile and pancreatic juice unable to molecularly distinguish IHCC, EHCC, and GBCA,
of affected patients (Isa et al, 2002; Itoi et al, 1999), but neither this study represented a paradigm shift in which we now rec-
KRAS nor TP53 mutational analysis were shown to be superior ognize that a subset of genes are frequently and specifically
to conventional cytopathology in the diagnosis of pancreatico- altered in each subtype of biliary tract cancer.
biliary tumors. However, combined pathologic analysis and Since this publication in 2012, several additional studies
mutation analysis increased diagnostic sensitivity (Aishima have been reported using whole-exome sequencing analysis. In
et al, 2002; Isa et al, 2002; Itoi et al, 1999). a U.S. study of 32 IHCCs, the investigators identified frequent
inactivating mutations in multiple chromatin-remodeling genes,
Post–Next-Generation Sequencing including BRCA1-associated protein 1 (BAP1, 17.5%), AT-rich
Compared with older studies, which relied upon sequencing interaction domain 1A (ARID1A, 15.0%), and polybromo 1
one gene a time, more recent studies have capitalized upon (PBRM1, 17.5%) (see Table 9C.2) (Jiao et al, 2013). Similar
NGS to narrowly or broadly characterize tumors. In 2012, to the earlier report, they also identified frequent mutations at
Borger and colleagues studied 287 tumors from gastrointestinal previously reported hot spots in the IDH1 (10%) and IDH2
cancer patients, including biliary tract, colorectal, gastroesoph- (5%) genes encoding metabolic enzymes. However, in contrast
ageal, hepatic, pancreatic, and small intestine carcinomas to the earlier study, TP53 (20%) was the most frequently altered
(Borger et al, 2012). They evaluated 15 known cancer genes for gene in the nine GBCA specimens studied.
TABLE 9C.2 Most Common Somatic Mutations in Biliary Tract Tumors

Author (Year) Borger (2012) Jiao (2013) Chan-On (2013) Chan-On (2013)
(N) (N = 40 IHCC + 22 EHCC + 25 GBCA) (N = 32 IHCC + 8 GBCA) (N = 10 IHCC + 5 EHCC) (N = 8 Opisthorchis viverrini–related IHCC)
ARID1A Unknown 15.0% 26.7% 12.5%

BAP1 Unknown 17.5% 26.7% 0.0%
BRAF 1.1% 0.0% 0.0% 0.0%
FGFR2 Unknown 10.0% 0.0% 0.0%
IDH1 9.2% 10.0% 20.0% 0.0%
IDH2 1.1% 5.0% 0.0% 0.0%
KRAS 9.2% 2.5% 20.0% 37.5%
PBRM1 Unknown 17.5% 0.0% 0.0%
PIK3CA 3.4% 7.5% 0.0% 0.0%
TP53 6.9% 20.0% 20.0% 50.0%
ARID1A, AT-rich interaction domain 1A; BAP1, BRCA1-associated protein 1; BRAF, B-Raf protooncogene; EHCC, extrahepatic cholangiocarcinoma; FGFR2, fibroblast growth factor receptor 2; GBCA,
gallbladder carcinoma; IDH1/2, isocitrate dehydrogenase 1/2; IHCC, intrahepatic cholangiocarcinoma; KRAS, Kirsten rat sarcoma; PBRM1, polybromo 1; PIK3CA, phosphatidylinositol-4,5-bisphosphate-3-
kinase catalytic subunit alpha; TP53, tumor protein 53.
TABLE 9C.3 Most Common Somatic Mutations by Biliary Tract TABLE 9C.4 Most Common Somatic Mutations and
Tumor Subtype Immunohistochemical Changes by Biliary Tract Tumor Subtype
IHCC (N = 40) EHCC (N = 22) GBCA (N = 25) IHCC EHCC GBCA
Gene Sequencing (N = 120) (N = 25) (N = 64)
AKT1 3% 0% 0%
APC 0% 0% 4% IDH1 14% 0% 1.5%
BRAF 3% 0% 0% KRAS 17% 28% 13%
CTNNB1 0% 0% 4% TP53 8% 44% 41%
IDH1 20% 0% 0% IHCC EHCC GBCA
IDH2 3% 0% 0% Immunohistochemistry (IHC) (N = 434) (N = 126) (N = 244)
KRAS 5% 23% 4%
ERBB2 overexpression 1.5% 18% 15%
NRAS 5% 0% 4%
PBRM1 loss 21% 15% 53%
PIK3CA 0% 0% 12%
EHCC, Extrahepatic cholangiocarcinoma; ERBB2, ERB-B2 receptor tyrosine kinase 2; GBCA,
PTEN 3% 0% 0%
gallbladder carcinoma; IDH1, isocitrate dehydrogenase 1; IHCC, intrahepatic cholangiocarcinoma;
TP53 5% 14% 4% KRAS, Kirsten rat sarcoma; PBRM1, polybromo 1.
AKT1, v-AKT murine thymoma viral oncogene homolog 1; APC, APC, adenomatosis polyposis coli; From Holcombe RF, et al: Tumor profiling of biliary tract carcinomas to reveal distinct molecular
BRAF, B-Raf protooncogene; CTNNB1, β1-catenin; IDH1/2, isocitrate dehydrogenase 1/2; KRAS, alterations and potential therapeutic targets, J Clin Oncol 33(Suppl 3):abstr 285, 2015.
Kirsten rat sarcoma; NRAS, neuroblastoma RAS viral (v-ras) oncogene homolog; PIK3CA,
phosphatidylinositol-4,5-bisphosphate-3-kinase catalytic subunit alpha; PTEN, phosphatase and
tensin; TP53, tumor protein 53. With the evolution of personalized medicine, many more
From Borger DR, et al: Frequent mutation of isocitrate dehydrogenase (IDH)1 and IDH2 in patients are having molecular analyses performed on their
cholangiocarcinoma identified through broad-based tumor genotyping, Oncologist 17:72–79, 2012. tumors. Two large studies analyzing biliary tract carcinomas
that utilized different molecular platforms were recently pre-
sented (Holcombe et al, 2015; Ross et al, 2015). Together,
In a separate study from Singapore, Chan-On and col- these studies provide further insight into the distinct molecular
leagues (2013) profiled 209 CCs from Asia and Europe, includ- alterations identified in each subtype (Tables 9C.4 and 9C.5).
ing 108 cases caused by infection with the liver fluke O. viverrini, In the first study, 815 cases (IHCC (N = 434), EHCC [N =
as well as 101 cases caused by non–Opisthorchis-related etiolo- 126], GBCA [N = 244], and not otherwise specified [N = 11])
gies. Whole-exome sequencing (IHCC [N = 10] and EHCC [N were evaluated using a commercial multiplatform profiling
= 5]) identified recurrent somatic mutations in BAP1 (26.7%) service (Caris Life Sciences, Phoenix, AZ) (Holcombe et al,
and ARID1A (26.7%), as well as frequent mutations in IDH1 2015). Testing included sequencing (Sanger sequencing and
(20%), KRAS (20%), and TP53 (20%). Compared with intra- NGS) and protein expression analysis (immunohistochemistry,
hepatic non–Opisthorchis-related tumors, the fluke-related IHC). In this analysis, 24 of 47 genes tested had mutations,
IHCC had higher rates of TP53 (50%) and KRAS (37.5%) with the highest rates in TP53 (28%), KRAS (18%), IDH1
mutations. On the other hand, they had lower ARID1A (12.5%) (9%), and SMAD4 (6%). Breast cancer 1/2 (BRCA1/2) muta-
mutations and no mutations in BAP1 or IDH1. This study cor- tions were seen in 3 of 41 (7.3%) and 5 of 40 (12.5%) cases,
roborated the finding of frequent somatic alterations in specific respectively. On analysis of individual tumor types (see Table
genes identified in the U.S. study, as well as demonstrated that 9C.4), IDH1 (14%) mutations remained a molecular feature of
different causative etiologies are associated with distinct sets of IHCC, whereas KRAS (28%) and TP53 (44%) predominated
somatic mutations (e.g., BAP1, IDH1, KRAS, TP53) within the in EHCC. Additionally, TP53 (41%) mutations were the
same tumor type. most common in GBCA. Mutually exclusive protein loss of
BAP1, IDH1, KRAS, PBRM1, and TP53 mutations. EHCC

TABLE 9C.5 Most Common Somatic Genomic Alterations by
have frequent ARID1A, KRAS, and TP53 mutations. GBCAs
Biliary Tract Tumor Subtype have frequent PIK3CA and TP53 mutations.
IHCC EHCC GBCA Taken together, there is a diverse somatic landscape of
(N = 412) (N = 57) (N = 85) genomic alterations in biliary tract cancers that can serve to
Total genomic alterations/patient 2.9 4.4 4
distinguish them, as well as possibly provide clinically rational
targets for therapies.
ARID1A 17% 12% 13%
BRAF 5% 3% 1% Epigenetic Alterations
CDKN2A/B loss 18% 17% 19% Many human cancers have aberrant epigenetic alterations. Epi-
ERBB2 amplification 3% 11% 16% genetic mechanisms involved in gene regulation include DNA
FGFR1-3 fusions and 11% 0% 3% methylation, histone modification, and noncoding (nc) RNAs.
amplifications Studies defining the role of these epigenetic alterations in the
IDH1/2 20% 0% 0% genesis of biliary tract tumors are beginning to emerge. Thus
KRAS 22% 42% 11% there are limited data on the role of altered histone modifica-
MET amplification 4% 0% 0% tions in the cholangiocyte transformation. Aberrant promoter
PI3KCA 5% 7% 14% hypermethylation of several genes, including the tumor suppres-
ARID1A, AT-rich interaction domain 1A; BRAF, B-Raf protooncogene; CDKN2A/B, cyclin-
sor genes, such as p16INK4a, Ras association domain family
dependent kinase inhibitor A/B; EHCC, extrahepatic cholangiocarcinoma; ERBB2, ERB-B2 receptor member 1 (RASSF1A), and APC has been reported in biliary
tyrosine kinase 2; FGFR1-3, fibroblast growth factor receptor 1-3; GBCA, gallbladder carcinoma; tract tumors. Moreover, point mutations in CpG islands of the
IDH1/2, isocitrate dehydrogenase 1/2; IHCC, intrahepatic cholangiocarcinoma; KRAS, Kirsten rat p16INK4a promoter may function as a methylation equivalent
sarcoma; MET, also called c-MET or hepatocyte growth factor receptor (HGFR); PIK3CA,
phenomenon resulting in gene inactivation in PSC (Taniai et al,
phosphatidylinositol-4,5-bisphosphate-3-kinase catalytic subunit alpha.
2002). Furthermore, promoter hypermethylation of SOCS-3
From Ross JS, et al: Comprehensive genomic profiling of biliary tract cancers to reveal
tumor-specific differences and genomic alterations, J Clin Oncol 33(Suppl 3):abstr 231, 2015.
(general sugar transporter), which is implicated in IL-6/ signal-
transducer and activator of transcription 3 (STAT3) activation,
has been noted in 27% of CCs. Other relevant aberrantly
chromatin modifiers BAP1 and PBRM1 were seen in 17% and methylated genes include runt-related transcription factor 3
27%, respectively. EHCC (18%) and GBCA (15%) also had (RUNX3), which is altered in 42% of IHCC, and p14ARF, which
significantly higher HER2 overexpression than IHCC. Finally, prevents TP53 degradation and hence cell-cycle arrest, has been
GBCA showed a high frequency of PBRM1 underexpression. reported as altered in 18% of tumors (Patel, 2014). The list
Therefore multiplatform cancer profiling revealed additional continues. In fact, since 2010, more than 40 studies have evalu-
distinct biomarker characteristics of biliary tract carcinomas ated nearly as many genes with influence on CC formation.
that offer insights into disease biology and potential therapeutic Unfortunately, the studies evaluating diagnostic and prognostic
interventions. markers are often inconclusive due to small cohort sizes, het-
In the second study, 554 cases (IHCC [N = 412], EHCC erogeneity of patients, and their underlying risk factors (Ander-
[N = 57], and GBCA [N = 85]) were evaluated using a sen et al, 2012; Andersen et al, 2013; Andersen et al, 2014; Sia
commercial comprehensive genomic profiling (GCP) service et al, 2013).
(Foundation Medicine, Cambridge, MA) (Ross et al, 2015). In line with emerging genomic platforms, a few recent inte-
CGP was performed on hybridization-captured, adaptor grative genomics studies have simultaneously analyzed chromo-
ligation–based libraries to a mean coverage depth of greater some alterations, genomic changes, epigenetic changes, as well
than 600X for 3,230 exons of 182 cancer-related genes plus 37 as transcriptomics (Andersen et al, 2012; Oishi et al, 2012; Sia
introns from 14 genes frequently rearranged in cancer. The D. et al, 2013; Wang et al, 2013). In one study, Andersen and
CGP assay included base substitutions, insertions/deletions colleagues (2012) identified 238 genes that correlated with
(INDELs), CNAs, and fusions/rearrangements. All three biliary high-risk features. This gene signature was further honed to 36
tract carcinoma subtypes shared genomic alterations in cell- genes, which represented independent survival genes that are
cycle regulation (e.g., CDKN2A/B loss, 17% to 19%) and chro- mostly cholangiocyte specific. These included CTNNB1, MYC,
matin remodeling (ARID1A, 12% to 17%). But IHCC more TNF, vascular endothelial growth factor receptor (VEGFR),
commonly had fibroblast growth factor receptor (FGFR)1-3 and ERBB.
fusions/amplifications (11%), IDH1/2 mutations (20%), BRAF
mutations, (5%), and MET (i.e., c-MET or HGFR [hepatocyte MicroRNA Alterations
growth factor receptor]) amplifications (4%), with a low KRAS In recent years, emerging evidence suggests that the expression
mutational frequency (22%). EHCC and GBCA had higher of nc RNAs, such as microRNAs (miRs), may be important in
ERBB2 (i.e., HER2 [human epidermal growth factor receptor carcinogenesis because they can modulate the expression of
2 or HER2/neu]) amplifications rates (11% and 16%, respec- many genes that regulate critical properties, such as cell sur-
tively), consistent with the prior report of HER2 overexpression, vival, autophagy, stemness, and response to therapy. As a result,
as well as an earlier report of increased expression of MET and miRs have been linked to tumor heterogeneity, as well as sig-
ERBB2 protooncogenes (Aishima et al, 2002). Finally, KRAS nificant determinants of genomics-based patient stratification
(42%) remained predominant in EHCC, whereas PIK3CA (Coulouarn et al, 2009; Oishi et al, 2012). In vitro, several
(14%) mutations were the most common in GBCA consistent groups have studied the biologic significance and aberrant
with the study by Borger and colleagues (2012). expression of miRs (Braconi C et al, 2010; Meng et al, 2006;
Overall, we can summarize the NGS data demonstrating the Meng et al, 2008; Mott et al, 2007 ), linking them to tumor
following common findings: IHCC have frequent ARID1A, growth, response to therapy, and inflammatory cytokines
expression (Stutes et al, 2007). Several miRs, including onco- as well as promote angiogenesis by regulating the expression of
genic miRNAs (e.g., miR-214 and miR-21) that regulate secreted pro-angiogenic proteins (Marti et al, 2015). Moreover,
PTEN-dependent activation of phosphatidylinositol-3-kinase nuclear YAP expression was shown to represent a biomarker of
(PI3K) signaling (He et al, 2013; Kawahigashi et al, 2009), as response to FGFR-directed therapy (Rizvi et al, 2016). Overall,
well as miRs associated with aberrant signaling pathways (e.g., these developmental signaling pathways appear to be important
HGF/MET and IL-6), and stem cell traits (e.g., miR-200c), in disease progression, whereas cross talk between these path-
have been reported. Others have described the regulation of ways needs further investigation.
epithelial-to-mesenchymal transition (EMT)–related genes,
including neural cell adhesion molecule 1, Slug/E-cadherin/ Epidermal Growth Factor Receptor
vimentin, and Twist, by miR-200c, miR-204, and miR214, Although somatic mutations in EGFR family members are rare,
respectively (Li et al, 2012; Oishi et al, 2012; Qiu et al, 2013). overexpression of the receptors occurs in 10% to 32% of CCs.
Comparison of CC lines with cholangiocytes also revealed Furthermore, aberrant phosphorylation of EGFR activates
downregulation of eight miRs (e.g., miR22, miR125a, miR127, MAPK, and p38 signaling can increase cyclooxygenase-2
miR199a, miR199a*, miR214, miR376a, and miR424). (COX-2). In turn, this can inhibit apoptosis while enhancing
Although, the exact role of each miR as either onco-miRs or tumor growth. However, although in vitro EGFR inhibition
prognostic markers remains to be determined, they appear to with erlotinib has been shown to CC cell proliferation, in vivo
be serving important roles in the genesis of biliary cancers. dual blockage of EGFR and ERBB1/ERBB2 with lapatinib is
necessary.
Tumor Growth and Metastasis
Dysregulated Signaling Pathways Hepatocyte Growth Factor (HGF)/MET
The progression and metastasis of biliary tract tumors appear The HGF/MET pathway is rarely mutated in biliary tract
to be driven by a variety of cellular signaling pathways involved tumors, but amplification of MET, the HGF receptor, has been
in responses to embryonic/stem cell signaling pathways (e.g., reported in IHCC. In turn, HGF/MET can activate many
Wnt, Hh, Notch, Hippo), growth factors (e.g., EGF, HGF/ pathways, including MAPK, PI3K, and STAT, as well as has
MET, VEGF), intracellular signal transduction (e.g., KRAS/ to stimulate migration and invasion in CC cells.
MAPK), cytokine signaling (e.g., IL-6/STAT), and cell-cycle
progression (e.g., polo-like kinases [PLKs]) (Sia et al, 2013). Vascular Endothelial Growth Factor and Angiogenic Signaling
VEGF is a signal protein produced by cells that stimulates
Embryonic Signaling angiogenesis. Alterations occur in almost half of IHCCs and
It is increasingly recognized that embryonic signaling pathways correlate with a poor prognosis. Although the application of
are important in the genesis of numerous types of malignancies. targeted therapies such as sorafenib, which targets wild-type
The Hedgehog (Hh), Notch, Wnt/β-catenin, and Hippo signal- BRAF and vascular endothelial growth factor receptor
ing pathways are important regulators of proliferation, survival, (VEGFR), has been studied, the preclinical data have been
self-renewal, and development in embryos and cancers in disappointing in CC models.
adults. The Hh pathway was initially reported be overexpressed
in CC (Berman et al, 2003). Hh has been also shown to regu- IL-6/JAK/STAT Cytokine Signaling
late tumor-stromal interactions in the liver that stimulate the IL-6 is an inflammatory cytokine that is overexpressed in CC
proliferation, migration, and invasion of CC cells (Kim et al, and regulates growth via either autocrine or paracrine mecha-
2014). Hh also directly regulates the viability of CC (El Khatib nisms. The overexpression of IL-6 observed in CC may result
M et al, 2013). Similarly, the Notch pathway has been shown from epigenetic silencing of SOCS-3 (Isomoto et al, 2007).
to regulate cell proliferation, migration, invasion, as well as Binding of IL-6 to its receptor (gp130) results in heterodimer-
EMT while working in concert with TP53 to regulate cell ization with the Janus kinases (JAK1, JAK2, or TYK2). In turn,
viability (El Khatib et al, 2013). Aberrant expression of Notch this drives, activation of STAT3 (i.e., the JAK/STAT pathway)
receptors 1 and 3 play a role during cancer progression, and and/or the MAPK pathway. In turn, IL-6 can regulate the
the Notch pathway protein DLL4 correlates with poor survival methylation of growth factor receptors, telomerase activity, and
in EHCC and GBCA (Yoon et al, 2011). The Wnt/β-catenin miRNA expression (Meng et al, 2008; Wehbe et al, 2006). As
pathway may also be of importance, although genomic muta- a result, inhibiting IL-6 signaling has been proposed as a novel
tions in genes, including APC, are rare. Thus the Wnt/β-catenin therapeutic target (Braconi et al, 2010).
pathway may not be as critical as other pathways in biliary tract
tumors. Finally, the Hippo signaling pathway is a tumor sup- Polo-Like Kinases
pressor pathway in which deletion of the macrophage- The PLKs are a family of serine/threonine kinases involved in
stimulating 1/2(MST1/2) gene was recently shown to cause key regulatory processes, including cell-cycle progression (G2/M
hepatocellular carcinoma (HCC) (Zhou et al, 2009). Inactiva- transition) and cytokinesis. Targeting PLK-1 has been shown to
tion of MST1/2 resulted in activation of the yes-associated increase the efficacy of 5-fluorouracil (Thrum et al, 2011),
protein 1 (YAP1) oncogene and resistance to FAS receptor whereas PLK-2 is a mediator of Hh signaling in CC (Fingas
(FasR)-induced apoptosis. FasR is also known as apoptosis et al, 2013). Currently, PLK inhibitors are in development.
antigen 1 (APO-1), cluster of differentiation 95 (CD95),
or tumor necrosis factor receptor superfamily member 6 Epithelial-to-Mesenchymal Transition
(TNFRSF6). Furthermore, overexpression of YAP1-mediated Epithelial-to-mesenchymal transitions (EMT), mesenchymal-
activation of Notch and Wnt signaling induced the expression to-epithelial transitions (MET), and, epithelial-mesenchymal
of downstream Notch targets. More recently, YAP was shown interactions (EMI) are often lumped together under the term
to regulate proliferation and anti-apoptotic mechanisms in CC, EMT (Sicklick, 2013). However, in the former phenomenon,
epithelial cells lose their polarity and cell-cell adhesion, while the proliferation and activation of CC cells and HSC expression
gaining migratory and invasive properties to become mesenchy- of the angiotensin II type 1 (AT1) receptor in autocrine and
mal. This is thought to be involved in the initiation of metasta- paracrine fashions, respectively (Okamoto et al, 2010). In vitro,
sis. As discussed above, miR200c was determined to prevent CC cell migration and survival are also modulated by the cross-
EMT (Oishi et al, 2012), whereas miR204 regulated genes talk between stromal cell–derived factor-1 (SDF-1) produced
associated with EMT (Zha et al, 2014), and miR214 inhibition by HSC and CC cells bearing the SDF-1 receptor, CXCR-4
was found to promote EMT (Li et al, 2012). Further studies (C-X-C chemokine receptor type 4) (Gentilini et al, 2012).
of miR gene regulation are warranted to determine the contri- Interestingly, Ang II also enhances EMT through the interac-
bution of miRs in biliary carcinogenesis, prognosis, and tion between activated HSC and the SDF-1/CXCR4 axis
therapeutics. (Okamoto et al, 2012). In addition, IL-1β produced by HSC
induces the expression of CXCL5 (C-X-C motif chemokine 5)
Tumor-Stromal Interactions in CC cells. In tumors, CXCL5 expression correlates with a
Hepatic stellate cells (HSCs) are stromal cells in the benign worse overall survival after curative hepatic resection (Okabe
hepatic parenchyma that possess both neural and myofibro et al, 2012). Taken together, the regulation of paracrine tumor-
blastic features (see Chapter 7). HSCs are the major cell stromal pathways may be a useful target in controlling IHCC
type involved in hepatic fibrosis and cirrhosis. In addition, progression.
portal myofibroblasts can contribute to hepatic fibrosis. It is
increasingly realized that tumor-stromal interactions play posi- SUMMARY
tive and negative roles in hepatic carcinogenesis and fibrosis
(Fingas et al, 2011; Kim et al, 2014; Magistri et al, 2014). In the last decade, we have gained significant insight into the
Because CC is a very desmoplastic tumor (Kajiyama et al, environmental risk factors, genomic alterations, tumor hetero-
1999), and the microenvironment is a key component of the geneity, and epithelial-mesenchymal interaction/transitions
tumor development and progression, these interactions are associated with the development of biliary tract adenocarcino-
most relevant to IHCC. As a result, the hepatic stromal com- mas. Presented is a comprehensive review of the molecular
partment is increasingly recognized in the pathogenesis of pathogenesis of CC, which relies upon the underlying themes
IHCC (Andersen et al, 2012; Chuaysri et al, 2009; Massani of chronic inflammation to the biliary epithelium, host-mediated
et al, 2013). response, and subsequent development of the malignant phe-
Several pathways appear to regulate this process. For notype. Many new candidates for targeted therapy based on
instance, the Hh pathway regulates HSC (Sicklick et al, 2005). molecular analyses have emerged, including the MET, EGFR,
In turn, HSCs stimulate the proliferation, migration, and inva- ERBB2, FGFR, JAK/STAT, RAS/RAF/MAPK, PI3K/AKT/
sion of CC cells, as well as promote angiogenesis through Hh mTOR, Wnt/Hh/Notch/Hippo, and IDH pathways. Data has
pathway activation. This renders CC cells more susceptible also emerged on the role of epigenetics and miRs, providing
to necrosis by Hh inhibition (Kim et al, 2014). Moreover, the potential for further studies in these areas. Identifying and
myofibroblast-derived platelet-derived growth factor-BB pro- cataloging somatic alterations and associating these alterations
tects CC cells from TRAIL (TNF-α–related apoptosis-inducing with clinical outcomes may assist in the development novel
ligand) cytotoxicity by a Hh-dependent process (Fingas et al, therapeutic interventions, enhancing early diagnosis, identify-
2011). Finally, expression of the Hh target gene, osteopontin, ing at-risk individuals, and ultimately improving survival in
is an independent predictor of survival in IHCC patients these devastating malignancies.
(Sulpice et al, 2013).
In addition, other mechanisms are emerging in this cross- References are available at expertconsult.com.
talk. Angiotensin II (Ang II) produced in IHCC tissue regulates
Chapter 9C Molecular pathogenesis of biliary tract cancer 159.e1
Gentilini A, et al: Role of the stromal-derived factor-1 (SDF-1)-

REFERENCES CXCR4 axis in the interaction between hepatic stellate cells and
cholangiocarcinoma, J Hepatol 57:813–820, 2012.
Aishima SI, et al: c-erbB-2 and c-Met expression relates to cholangio- Hardell L, et al: Aetiological aspects on primary liver cancer with
carcinogenesis and progression of intrahepatic cholangiocarcinoma, special regard to alcohol, organic solvents and acute intermittent
Histopathology 40:269–278, 2002. porphyria: an epidemiological investigation, Br J Cancer 50:389–397,
Andersen JB, Thorgeirsson SS: Genetic profiling of intrahepatic chol- 1984.
angiocarcinoma, Curr Opin Gastroenterol 28:266–272, 2012. He Q, et al: Ars2 is overexpressed in human cholangiocarcinomas and
Andersen JB, Thorgeirsson SS: Genomic decoding of intrahepatic chol- its depletion increases PTEN and PDCD4 by decreasing
angiocarcinoma reveals therapeutic opportunities, Gastroenterology microRNA-21, Mol Carcinog 52:286–296, 2013.
144:687–690, 2013. Hoblinger A, et al: Association of the c.3972C>T variant of the mul-
Andersen JB, Thorgeirsson SS: A perspective on molecular therapy in tidrug resistance-associated protein 2 gene (MRP2/ABCC2) with
cholangiocarcinoma: present status and future directions, Hepat susceptibility to bile duct cancer, Digestion 80:36–39, 2009.
Oncol 1:143–157, 2014. Holcombe RF, et al: Tumor profiling of biliary tract carcinomas to
Andersen JB, et al: Genomic and genetic characterization of cholangio- reveal distinct molecular alterations and potential therapeutic targets,
carcinoma identifies therapeutic targets for tyrosine kinase inhibitors, J Clin Oncol 33(Suppl 3):abstr 285, 2015.
Gastroenterology 142:1021–1031, e1015, 2012. Isa T, et al: Analysis of microsatellite instability, K-ras gene mutation
Benson AB 3rd, et al: Hepatobiliary cancers, version 2.2014, J Natl and p53 protein overexpression in intrahepatic cholangiocarcinoma,
Compr Canc Netw 12:1152–1182, 2014. Hepatogastroenterology 49:604–608, 2002.
Benson AB, et al: NCCN Clinical Practice Guidelines in Oncology (NCCN Ishak KK, et al: Histological typing of tumours of the liver:WHO interna-
Guidelines®) Hepatobiliary Cancers. © 2015 National Comprehensive tional histological classification of tumours, Berlin, 1994, Springer-
Cancer Network, Inc. Available at NCCN.org, 2016. Verlag.
Bergquist A, et al: Risk factors and clinical presentation of hepatobiliary Isomoto H, et al: Sustained IL-6/STAT-3 signaling in cholangiocarci-
carcinoma in patients with primary sclerosing cholangitis: a case- noma cells due to SOCS-3 epigenetic silencing, Gastroenterology
control study, Hepatology 27:311–316, 1998. 132:384–396, 2007.
Berman DM, et al: Widespread requirement for Hedgehog ligand Itoi T, et al: K-ras codon 12 and p53 mutations in biopsy specimens
stimulation in growth of digestive tract tumours, Nature 425:846– and bile from biliary tract cancers, Pathol Int 49:30–37, 1999.
851, 2003. Jacquemin E: Role of multidrug resistance 3 deficiency in pediatric and
Bismuth H, et al: Management strategies in resection for hilar cholan- adult liver disease: one gene for three diseases, Semin Liver Dis
giocarcinoma, Ann Surg 215:31–38, 1992. 21:551–562, 2001.
Borger DR, et al: Frequent mutation of isocitrate dehydrogenase Jiao Y, et al: Exome sequencing identifies frequent inactivating muta-
(IDH)1 and IDH2 in cholangiocarcinoma identified through broad- tions in BAP1, ARID1A and PBRM1 in intrahepatic cholangiocar-
based tumor genotyping, Oncologist 17:72–79, 2012. cinomas, Nat Genet 45:1470–1473, 2013.
Braconi, et al: Targeting the IL-6 dependent phenotype can identify Kajiyama K, et al: The significance of stromal desmoplasia in intrahe-
novel therapies for cholangiocarcinoma, PLoS ONE 5:e15195, patic cholangiocarcinoma: a special reference of “scirrhous-type”
2010. and “nonscirrhous-type” growth, Am J Surg Pathol 23:892–902,
Braconi C, et al: MicroRNA-dependent regulation of DNA 1999.
methyltransferase-1 and tumor suppressor gene expression by Kawahigashi Y, et al: MicroRNA profiling of human intrahepatic chol-
interleukin-6 in human malignant cholangiocytes, Hepatology 51: angiocarcinoma cell lines reveals biliary epithelial cell-specific
881–890, 2010. microRNAs, J Nippon Med Sch 76:188–197, 2009.
Broome U, et al: Natural history and prognostic factors in 305 Swedish Khan SA, et al: Changing international trends in mortality rates
patients with primary sclerosing cholangitis, Gut 38:610–615, 1996. for liver, biliary and pancreatic tumours, J Hepatol 37:806–813,
Chalasani N, et al: Cholangiocarcinoma in patients with primary scle- 2002a.
rosing cholangitis: a multicenter case-control study, Hepatology Khan SA, et al: Guidelines for the diagnosis and management of chol-
31:7–11, 2000. angiocarcinoma, Gut 51(Suppl 6):VI1–VI9, 2002b.
Chan-On W, et al: Exome sequencing identifies distinct mutational Khan SA, et al: DNA adducts, detected by 32P postlabelling, in human
patterns in liver fluke-related and non-infection-related bile duct cholangiocarcinoma, Gut 52:586–591, 2003.
cancers, Nat Genet 45:1474–1478, 2013. Khan SA, et al: p53 Mutations in human cholangiocarcinoma: a
Chuaysri C, et al: Alpha-smooth muscle actin-positive fibroblasts review, Liver Int 25(704–716):2005.
promote biliary cell proliferation and correlate with poor survival in Khan SA, et al: Rising trends in cholangiocarcinoma: is the ICD clas-
cholangiocarcinoma, Oncol Rep 21:957–969, 2009. sification system misleading us?, J Hepatol 56:848–854, 2012.
Coulouarn C, et al: Loss of miR-122 expression in liver cancer corre- Khenjanta C, et al: Association of CYP39A1, RUNX2 and oxidized
lates with suppression of the hepatic phenotype and gain of meta- alpha-1 antitrypsin expression in relation to cholangiocarcinoma pro-
static properties, Oncogene 28:3526–3536, 2009. gression, Asian Pac J Cancer Prev 15:10187–10192, 2014.
Daorueang D, et al: Secreted Opisthorchis viverrini glutathione Kim Y, et al: Hedgehog signaling between cancer cells and hepatic
S-transferase regulates cell proliferation through AKT and ERK stellate cells in promoting cholangiocarcinoma, Ann Surg Oncol
pathways in cholangiocarcinoma, Parasitol Int 61:155–161, 2012. 21:2684–2698, 2014.
Edge S: AJCC cancer staging manual, New York, 2010, Springer. Komori J, et al: Activation-induced cytidine deaminase links bile duct
El Khatib, et al: Activation of Notch signaling is required for cholan- inflammation to human cholangiocarcinoma, Hepatology 47:888–
giocarcinoma progression and is enhanced by inactivation of p53 in 896, 2008.
vivo, PLoS ONE 8:e77433, 2013. Kosriwong K, et al: Trefoil factors: tumor progression markers and
El Khatib M, et al: Inhibition of hedgehog signaling attenuates carci- mitogens via EGFR/MAPK activation in cholangiocarcinoma, World
nogenesis in vitro and increases necrosis of cholangiocellular carci- J Gastroenterol 17:1631–1641, 2011.
noma, Hepatology 57:1035–1045, 2013. Kubo S, et al: Hepatolithiasis associated with cholangiocarcinoma,
Endo K, et al: ERBB-2 overexpression and cyclooxygenase-2 World J Surg 19:637–641, 1995.
up-regulation in human cholangiocarcinoma and risk conditions, Li B, et al: Down-regulation of miR-214 contributes to intrahepatic
Hepatology 36:439–450, 2002. cholangiocarcinoma metastasis by targeting Twist, FEBS J 279:
Fingas CD, et al: Myofibroblast-derived PDGF-BB promotes Hedge- 2393–2398, 2012.
hog survival signaling in cholangiocarcinoma cells, Hepatology Lipsett PA, et al: Choledochal cyst disease: a changing pattern of pre-
54:2076–2088, 2011. sentation, Ann Surg 220:644–652, 1994.
Fingas CD, et al: Polo-like kinase 2 is a mediator of hedgehog Liver Cancer Study Group of Japan: The general rules for the clinical
survival signaling in cholangiocarcinoma, Hepatology 58:1362–1374, and pathological study of primary liver cancer, ed 4, Tokyo, 2000,
2013. Kanehra.
Forsbring M, et al: Catalytically impaired hMYH and NEIL1 mutant Magistri P, et al: The glypican 3 hepatocellular carcinoma marker regu-
proteins identified in patients with primary sclerosing cholangitis and lates human hepatic stellate cells via Hedgehog signaling, J Surg Res
cholangiocarcinoma, Carcinogenesis 30:1147–1154, 2009. 187:377–385, 2014.
Massani M, et al: Isolation and characterization of biliary epithelial and Shigeyasu K, et al: Early detection of metachronous bile duct cancer
stromal cells from resected human cholangiocarcinoma: a novel in in Lynch syndrome: report of a case, Surg Today 44:1975–1981,
vitro model to study tumor-stroma interactions, Oncol Rep 30:1143– 2014.
1148, 2013. Sia, et al: Intrahepatic cholangiocarcinoma: pathogenesis and rationale
Meier Y, et al: Hepatocellular malignancy in ABCB11/BStP disease for molecular therapies, Oncogene 32:4861–4870, 2013.
(progressive familial intrahepatic cholestasis, type 2): four patients, Sia D, et al: Integrative molecular analysis of intrahepatic cholangio-
Hepatology 40(Supp1 1):471A, 2004. carcinoma reveals 2 classes that have different outcomes, Gastroen-
Meng F, et al: Involvement of human micro-RNA in growth and terology 144:829–840, 2013.
response to chemotherapy in human cholangiocarcinoma cell lines, Sicklick JK: Correcting the misnomers of epithelial-mesenchymal rela-
Gastroenterology 130:2113–2129, 2006. tions, J Surg Res 182:36–39, 2013.
Meng F, et al: Epigenetic regulation of microRNA-370 by interleukin-6 Sicklick JK, et al: Role for hedgehog signaling in hepatic stellate cell
in malignant human cholangiocytes, Oncogene 27:378–386, 2008. activation and viability, Lab Invest 85:1368–1380, 2005.
Miller G, et al: Genome wide analysis and clinical correlation of chro- Sirica AE: Role of ErbB family receptor tyrosine kinases in intrahepatic
mosomal and transcriptional mutations in cancers of the biliary tract, cholangiocarcinoma, World J Gastroenterol 14:7033–7058, 2008.
J Exp Clin Cancer Res 28:62, 2009. Sirica AE, et al: Cyclooxygenase-2 and ERBB-2 in cholangiocarcinoma:
Mott JL, et al: mir-29 regulates Mcl-1 protein expression and apopto- potential therapeutic targets, Semin Liver Dis 22:303–313, 2002.
sis, Oncogene 26:6133–6140, 2007. Sorensen HT, et al: Risk of liver and other types of cancer in patients
Nakeeb A, et al: Cholangiocarcinoma: a spectrum of intrahepatic, peri- with cirrhosis: a nationwide cohort study in Denmark, Hepatology
hilar, and distal tumors, Ann Surg 224:463–473, 1996. 28:921–925, 1998.
Ohtsuka T, et al: Carcinoma arising in choledochocele, Endoscopy Stutes M, et al: Genetic and epigenetic changes associated with chol-
33:614–619, 2001. angiocarcinoma: from DNA methylation to microRNAs, World J
Oishi N, et al: Transcriptomic profiling reveals hepatic stem-like gene Gastroenterol 13:6465–6469, 2007.
signatures and interplay of miR-200c and epithelial-mesenchymal Sulpice L, et al: Molecular profiling of stroma identifies osteopontin as
transition in intrahepatic cholangiocarcinoma, Hepatology 56:1792– an independent predictor of poor prognosis in intrahepatic cholan-
1803, 2012. giocarcinoma, Hepatology 58:1992–2000, 2013.
Okabe H, et al: Identification of CXCL5/ENA-78 as a factor involved Taniai M, et al: p16INK4a promoter mutations are frequent in primary
in the interaction between cholangiocarcinoma cells and cancer- sclerosing cholangitis (PSC) and PSC-associated cholangiocarci-
associated fibroblasts, Int J Cancer 131:2234–2241, 2012. noma, Gastroenterology 123:1090–1098, 2002.
Okamoto K, et al: Angiotensin II induces tumor progression and fibro- Taylor-Robinson SD, et al: Increase in mortality rates from intrahe-
sis in intrahepatic cholangiocarcinoma through an interaction with patic cholangiocarcinoma in England and Wales 1968-1998, Gut
hepatic stellate cells, Int J Oncol 37:1251–1259, 2010. 48:816–820, 2001.
Okamoto K, et al: Angiotensin II enhances epithelial-to-mesenchymal Thompson R, Strautnieks S: BSEP: function and role in progressive
transition through the interaction between activated hepatic stellate familial intrahepatic cholestasis, Semin Liver Dis 21:545–550, 2001.
cells and the stromal cell-derived factor-1/CXCR4 axis in intrahe- Thrum S, et al: Polo-like kinase 1 inhibition as a new therapeutic
patic cholangiocarcinoma, Int J Oncol 41:573–582, 2012. modality in therapy of cholangiocarcinoma, Anticancer Res 31:3289–
Olnes MJ, Erlich R: A review and update on cholangiocarcinoma, 3299, 2011.
Oncology 66:167–179, 2004. van Mil SW, et al: FIC1 disease: a spectrum of intrahepatic cholestatic
Palmer WC, Patel T: Are common factors involved in the pathogenesis disorders, Semin Liver Dis 21:535–544, 2001.
of primary liver cancers? A meta-analysis of risk factors for intrahe- Wang P, et al: Mutations in isocitrate dehydrogenase 1 and 2 occur
patic cholangiocarcinoma, J Hepatol 57:69–76, 2012. frequently in intrahepatic cholangiocarcinomas and share hyper-
Patel T: Increasing incidence and mortality of primary intrahepatic methylation targets with glioblastomas, Oncogene 32:3091–3100,
cholangiocarcinoma in the United States, Hepatology 33:1353–1357, 2013.
2001. Wehbe H, et al: Interleukin-6 contributes to growth in cholangiocarci-
Patel T: Worldwide trends in mortality from biliary tract malignancies, noma cells by aberrant promoter methylation and gene expression,
BMC Cancer 2:10, 2002. Cancer Res 66:10517–10524, 2006.
Patel T: New insights into the molecular pathogenesis of intrahepatic Welzel TM, et al: Impact of classification of hilar cholangiocarcinomas
cholangiocarcinoma, J Gastroenterol 49:165–172, 2014. (Klatskin tumors) on the incidence of intra- and extrahepatic chol-
Pitt HA, et al: Malignancies of the biliary tree, Curr Probl Surg 32:1–90, angiocarcinoma in the United States, J Natl Cancer Inst 98:873–875,
1995. 2006.
Prawan A, et al: Association between genetic polymorphisms of Welzel TM, et al: Risk factors for intrahepatic and extrahepatic chol-
CYP1A2, arylamine N-acetyltransferase 1 and 2 and susceptibility angiocarcinoma in the United States: a population-based case-
to cholangiocarcinoma, Eur J Cancer Prev 14:245–250, 2005. control study, Clin Gastroenterol Hepatol 5:1221–1228, 2007.
Qi DC, et al: Analysis of differentially expressed genes in malignant Wu T, et al: Involvement of 85-kd cytosolic phospholipase A(2) and
biliary strictures, Genet Mol Res 13:2674–2682, 2014. cyclooxygenase-2 in the proliferation of human cholangiocarcinoma
Qiu YH, et al: miR-204 inhibits epithelial to mesenchymal transition cells, Hepatology 36:363–373, 2002.
by targeting slug in intrahepatic cholangiocarcinoma cells, Cell Yamagiwa Y, et al: Interleukin-6 decreases senescence and increases
Physiol Biochem 32:1331–1341, 2013. telomerase activity in malignant human cholangiocytes, Life Sci
Ross JS, et al: Comprehensive genomic profiling of biliary tract cancers 78:2494–2502, 2006.
to reveal tumor-specific differences and genomic alterations, J Clin Yongvanit P, et al: Hepatic cytochrome P450 2A6 and 2E1 status in
Oncol 33(Suppl 3):abstr 231, 2015. peri-tumor tissues of patients with Opisthorchis viverrini-associated
Sahani D, et al: Thorotrast-induced cholangiocarcinoma: case report, cholangiocarcinoma, Parasitol Int 61:162–166, 2012.
Abdom Imaging 28:72–74, 2003. Yoon HA, et al: Clinicopathological significance of altered Notch sig-
Sasaki M, et al: Aberrant expression of trefoil factor family 1 in biliary naling in extrahepatic cholangiocarcinoma and gallbladder carci-
epithelium in hepatolithiasis and cholangiocarcinoma, Lab Invest noma, World J Gastroenterol 17:4023–4030, 2011.
83:1403–1413, 2003. Zha R, et al: Genome-wide screening identified that miR-134 acts as
Scott J, et al: Bile duct carcinoma: a late complication of congenital a metastasis suppressor by targeting integrin beta1 in hepatocellular
hepatic fibrosis: case report and review of literature, Am J Gastroen- carcinoma, PLoS ONE 9:e87665, 2014.
terol 73:113–119, 1980. Zhou D, et al: Mst1 and Mst2 maintain hepatocyte quiescence and
Shaib Y, El-Serag HB: The epidemiology of cholangiocarcinoma, suppress hepatocellular carcinoma development through inactivation
Semin Liver Dis 24:115–125, 2004. of the Yap1 oncogene, Cancer Cell 16:425–438, 2009.
CHAPTER 9D
Molecular biology of liver carcinogenesis
and hepatitis
Masafumi Shimoda and Jack R. Wands
OVERVIEW OF MOLECULAR ETIOLOGY are most prevalent include West and Central Africa and East
and Southeast Asia, with China alone accounting for more than
Recent advances in molecular genetics have emphasized the 50% of the world cases (Ferlay et al, 2015). It is noteworthy
multistep process of tumorigenesis. It is evident that cancer is that HBV is highly endemic in these areas, with the exception
a genetic disease involving aberrant chromosome rearrange- of Japan, where HCV is more prevalent. Oceania, North and
ments, genetic mutations, and epigenetic silencing of tumor South America, and Northern and Eastern Europe are low-rate
suppressor genes (Farazi & DePinho, 2006). Independent of areas.
the etiology, hepatocellular carcinoma (HCC; see Chapter 91) Trends in HCC incidence are likely to be different in regions
generally develops where sustained hepatocyte turnover occurs of high and low persistence of HBV and HCV infection
in the setting of injury-inflammation-regeneration, which leads (El-Serag & Rudolph, 2007). Comparative studies performed
to the accumulation of chromosomal aberrations. In this between 1977 and 1982 and between 1993 and 1997 show that
context, monoclonal populations of hepatocytes become pre- the incidence of HCC in Hong Kong, Shanghai, Singapore, and
neoplastic and, following additional genomic alterations, change Japan has begun to decrease (Parkin et al, 2002). The fall in
into dysplastic cells and eventually HCC (Thorgeirsson & incidence is apparently due to vaccination against HBV, which
Grisham, 2002). Accumulated genetic alterations in preneo- has been accomplished in greater than 80% of newborns
plastic lesions and HCC result in the activation, as well as (Chang et al, 2009), because chronic HBV infection in those
inactivation, of many growth factor signal transduction path- countries is usually acquired through mother-to-newborn or
ways involved in hepatic transformation. It is believed that sibling-to-sibling transmission at a young age. In contrast, the
increased hepatocyte turnover associated with chronic liver incidence of HCC has rapidly increased in some countries, such
injury may be a major feature of hepatic oncogenesis. However, as Australia, the United States, and the United Kingdom, as a
another central question is whether hepatitis viruses, the leading result of chronic HCV infection. For example, HCC is the
cause of HCC worldwide, directly contribute to the develop- fastest growing cause of cancer-related deaths in the United
ment of this disease. Accumulating evidence suggests that States. The annual incidence of liver cancer increased from 2.6
chronic hepatitis B (HBV) and hepatitis C virus (HCV) infec- per 100,000 population for the years 1978 to 1980 to 8 per
tion play a direct role in the molecular pathogenesis of HCC 100,000 in 2010, of which at least 3 of 4 cases are accounted
through specific viral–cellular protein interactions (Branda & for by HCC (El-Serag & Kanwal, 2014). Reasons for this
Wands, 2006; see Chapter 70). increased incidence are not entirely clear but may reflect a
greater prevalence and role of persistent HCV infection
EPIDEMIOLOGY (McGlynn et al, 2001).
Age-specific rates of HCC peak at 75 years and older in most
Primary liver cancer is the fifth most common cancer in men regions of the world (El-Serag & Rudolph, 2007). With the
and the ninth most common in women worldwide. It is esti- exception of Africa, the peak incidence in women occurs 5 years
mated that 782,000 new patients with the disease were diag- later than that found in men. In the United States, recent trends
nosed in 2012 (Ferlay et al, 2015). The 5-year survival rate is have revealed a peak incidence shifting toward a relatively
less than 15% in developed countries, and the United States younger age group (El-Serag & Kanwal, 2014).
has a survival rate of 16.6%, making liver cancer the second Significant gender and ethnic variation in incidence, as well
most fatal tumor after pancreatic cancer (SEER Cancer Statis- as mortality from HCC, has also been found; male rates are
tics Factsheets 2007-2011). Presumably because of its poor nearly triple that of females (Ferlay et al, 2015). The most likely
prognosis, liver cancer is the second leading cause of cancer explanation for gender variation is that men have more risk
death in men and the sixth among women in the world. It is factors, such as exposure to hepatitis virus infection, excessive
estimated that about 745,000 individuals worldwide died from alcohol intake, smoking, and increased iron stores in the liver
this disease in 2012 (Ferlay et al, 2015). (El-Serag & Rudolph, 2007). In addition, androgen receptor
Primary liver cancers consist of numbers of histologically accelerates the progression of HCC through interaction with
distinct types of tumors that arise from hepatocytes, biliary the HBV genome (Ma et al, 2008; Wu et al, 2010). The inci-
epithelial cells, and fibroblasts. The most common is HCC, dence of HCC also varies with race and ethnicity in the same
which accounts for 70% to 85% of all hepatic tumors (Perz area. In the United States, the incidence and subsequent mor-
et al, 2006). Approximately 80% of HCC worldwide is caused tality rates from HCC are two times greater in Asians than
by chronic infection with HBV or HCV or both. The HCC African Americans, which are two times greater than those
burden is unevenly distributed worldwide; areas where tumors found in whites (El-Serag & Rudolph, 2007). These variations
160
Chapter 9D Molecular biology of liver carcinogenesis and hepatitis 161
are explained in part by the accumulation of major risk factors contaminate corn, rice, and peanuts in China and sub-Saharan
in each ethnic group. Africa. High rates of dietary exposure to AFB1 increase the risk
of HCC 4-fold. When people with chronic HBV infection are
RISK FACTORS exposed to AFB1, the relative risk for HCC dramatically
increases to about 60-fold (Kew, 2003). This synergistic effect
Unlike most malignancies, HCC has well-established extrinsic between AFB1 exposure and chronic HBV infection is an
risk factors that account for at least 80% of tumors, namely important observation because in some regions of the world,
chronic infection with HBV or HCV (see Chapter 70). Key AFB1 exposure and chronic HBV infection rates are high.
epidemiologic aspects of HBV- and HCV-induced HCC are Excessive ethanol consumption (>50 to 70 g/day) is another
summarized in Table 9D.1. Chronic HBV infection is the well-defined risk factor for HCC. In the United States, ethanol-
leading cause of HCC, and it has been estimated that there are induced HCC accounted for more than 20% of HCC patients
350 to 400 million HBV carriers, which account for 5% of the between 1996 and 1999 (Davila et al, 2004). Donato and col-
global population. About 59% of HCC patients in developing leagues (2002) showed a linear increase of a relative risk for
countries and 23% of HCC patients in developed countries are HCC by 5-fold when 60 to 140 g/day alcohol is taken. The
chronically infected with HBV (American Cancer Society, 5-year cumulative HCC incidence in alcoholic cirrhosis without
2011). Reported relative risks of HCC among HBV carriers HBV and HCV infection is 8% (Fattovich et al, 2004); however,
range between 5-fold and 100-fold compared with the general it is unlikely that ethanol itself has a direct carcinogenic effect.
uninfected population. This wide range of the estimated values Rather, excessive ethanol ingestion indirectly affects hepatocar-
is caused by the dependence of the relative risk on multiple cinogenesis through the promotion of cirrhosis. Indeed, greater
factors, including HBV load, presence of cirrhosis, and expo- than 80% of HCC tumors found in alcoholics develop in the
sure to aflatoxin B1 (AFB1) (El-Serag, 2012). The 5-year background of a cirrhotic liver. A synergistic effect between
cumulative incidence rates of HCC from HBV-related cirrhosis heavy alcohol consumption and hepatitis virus infection is
are 17% in highly endemic areas and 10% in Europe and the observed in several studies. The relative risk of HCC attribut-
United States (Fattovich et al, 2008). In addition, approxi- able to heavy alcohol consumption alone was only 2.4-fold,
mately 70% to 90% of HBV-related HCC develops in patients whereas in combination with chronic HCV infection, it
with cirrhosis. increased to 50-fold (Hassan et al, 2002). Others have reported
Chronic HCV infection is the second leading cause of HCC. that the concomitant HCV infection in alcoholics increases the
The estimated number of HCV carriers worldwide is 180 risk for HCC 2-fold, whereas HBV infection moderately
million, which accounts for 2% of the global population. increases this risk 1.2- to 1.5-fold (Donato et al, 2002; Jee et al,
Approximately 33% of HCC tumors in developing countries 2004).
and 20% of HCC in developed countries are attributable to Growing evidence now suggests that metabolic dysfunction,
persistent HCV infection (American Cancer Society, 2011). including obesity, diabetes, and nonalcoholic fatty liver disease
According to cross-sectional and case-control studies, HCC (NAFLD), are important risk factors for HCC, especially in
risk is increased 15- to 20-fold in HCV-infected persons com- developed countries (see Chapter 71). Several large cohort
pared with the HCV-negative population (El-Serag, 2012). The studies revealed that obesity is a definitive risk factor for HCC
5-year cumulative incidence of HCC with HCV-related cir- with the 1.5- to 4-fold increased risk. Men are more susceptible
rhosis in developed countries is 17%, with an exception for to obesity-associated HCC than women. Diabetes mellitus also
Japan, where the 5-year cumulative incidence is 30% (Fattovich has established as a moderately strong risk factor for HCC, with
et al, 2004). The high incidence of HCV-related HCC in Japan a two- to four-times higher risk (El-Serag & Rudolph, 2007;
may be due to the prevalence of HCV genotype 1b. AFB1 Starley et al, 2010). More recently, nonalcoholic steatohepatitis
is produced by Aspergillus flavus and related fungi that (NASH), the more aggressive form of NAFLD, is considered
to be a cause of a large proportion of cryptogenic cirrhosis,
which is liable for these patients to develop HCC. However,
TABLE 9D.1 Comparison of Epidemiologic Features between the overall incidence of HCC in patients with NAFLD is much
HBV- and HCV-Induced HCC lower than in patients with well-established etiologies (Starley
HBV HCV et al, 2010).
Other risk factors for HCC include hemochromatosis and
Virus carriers (% of 350-400 million 180 million (2%) hepatic porphyria. It should be noted that daily coffee intake
global population) (5%) reduces the incidence of HCC between 25% and 75% in the
Highly prevalent Asia, sub-Saharan Africa, South and general population in a dose-dependent manner (Inoue et al,
areas Africa, Melanesia, East Asia, South
Micronesia America
2005). This effect may be attributable to the inhibition of trans-
forming growth factor-β (TGF-β) signaling by methylxanthine
Relative risk of HCC 5- to 100-fold* 15- to 20-fold
caffeine, which reduces liver fibrosis (Gressner, 2009).
5-year cumulative 10% (Europe and 17% (Europe and
incidence rates United States) United States)
of HCC from 17% (East Asia) 30% (Japan) GENETIC AND EPIGENETIC ALTERATIONS
cirrhosis
HBV, Hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus.
Chronic inflammation accompanied by sustained cycles of
injury and regeneration of hepatocytes over 20 to 40 years
*Depends on multiple factors including HBV load, presence of cirrhosis, and exposure to
aflatoxin B1. promotes the development of liver fibrosis, cirrhosis, and even-
From El-Serag HB, Kanwal F: Epidemiology of hepatocellular carcinoma in the United States: Where
tually HCC (Fig. 9D.1A) (see Chapter 7). Pathologically, HCC
are we? Where do we go? Hepatology 60:1767–1775, 2014; and El-Serag HB: Epidemiology of occurs early within cirrhotic nodules, which can form in areas
viral hepatitis and hepatocellular carcinoma, Gastroenterology 142:1264–1273, 2012. of adenomatous hyperplasia or dysplasia. These cells eventually
6 months 10-40 years 20-40 years
Cirrhotic nodule
Hyperplasia
Dysplasia
Malignant
Normal liver Chronic hepatitis Cirrhosis transformation Hepatocellular carcinoma
Telomere shortening
Chromosome instability
β-Catenin mutation
A Loss of p53
Injury
Inflammation Hypermethylations
Regeneration
Oxidative stress Telomere dysfunction
Chromosome instability/genetic alterations
Chromosome segregation defects Impairment of DNA damage-repair systems

-p53/MDM2/gankyrin pathway
-RB1/p16INK4 pathway
Enhanced proliferation and survival
B Hepatocellular carcinoma
FIGURE 9D.1. A, Diagram showing progression of liver disease and genetic events accompanied by chronic inflammation. B, Common mechanisms
underlying hepatocarcinogenesis. A number of factors contribute to chromosome instability and other genetic alterations, which lead to the formation
of hepatocellular carcinoma.
become more atypical, and the malignant transformation from investigations performed in TERC knockout mice (Farazi
process becomes complete (Theise et al, 2002). Hepatocellular et al, 2003). This study indicates that telomerase reactivation
carcinoma is like other malignancies and represents a DNA occurred and appeared to be necessary for late-stage tumor
disease with accumulation of many alterations in oncogenes progression; thus the reactivated telomerase enzyme maintains
and tumor suppressor genes. The accumulation of genetic aber- the shortened telomere length in these tumor cells and prevents
rations that induces cellular transformation may take 20 to 40 them from undergoing apoptosis.
years, suggesting that liver carcinogenesis involves a multistep DNA damage-response pathways are safeguards that regu-
process. late cell-cycle checkpoints and prevent DNA-damaged cells
Cirrhosis may represent end-stage liver disease as a result of from further proliferation. Several studies report that the func-
persistent HBV or HCV infection. Because 80% to 90% of tions of key regulatory molecules, including p53, mouse double-
HCC tumors originate from cirrhotic liver, it is evident that minute 2 (MDM2), retinoblasoma 1 (RB1), p16INK4a (/also
continuous rounds of cellular injury, followed by regeneration known as CDKN2A [cyclin-dependent kinase inhibitor 2A],
in the milieu of chronic inflammation, fundamentally contrib- and gankyrin, were impaired in human HCC. The p53 protein,
utes to the oncogenic processes. The sustained cycles of injury encoded by the transforming protein 53 (TP53) gene, is a
and repair increase the chance of genomic alteration. Further- master molecule that maintains genome integrity by inducing
more, the host inflammatory response to viral infection, includ- cell-cycle arrest, followed by activation of DNA repair systems.
ing activation of stellate cells, causes the release of When excessive DNA damage occurs, p53 initiates apoptosis.
proinflammatory cytokines, which accelerate hepatic carcino- TP53 is one of the most mutated genes in HCC, and the muta-
genesis by augmenting oxidative stress and DNA damage tions often result in loss of function. The frequency ranges
(Bataller & Brenner, 2005; Giannelli et al, 2005; Ogata et al, between 11% and 35%, depending on regions (Tornesello et al,
2006). Continuous rounds of this process in the presence of 2013). The regional difference in the frequency is attributable,
inflammation not only increase the chance of genomic altera- in part, if not completely, to a specific mutation caused by
tions but also produce chromosome instability. For example, AFB1 exposure. TP53 missense mutations in codon 249
hyperploidy has been observed in 43% of dysplastic peritu- (R249S) were found in more than 50% of AFB1-related HCC
moral regions and in about 50% of HCC tumors (Laurent-Puig (Gouas et al, 2009), and it was the main cause of AFB1-induced
& Zucman-Rossi, 2006). Molecular mechanisms underlying liver cancer (Bressac et al, 1991). Therefore TP53 mutations in
such genomic instability include telomerase dysfunction, defec- HCC can be frequently found in Africa and Asia where people
tive segregation of chromosomes, and an impaired DNA are exposed to a high level of AFB1. Other mutations of TP53
damage response (Fig. 9D.1B). are found in 20% to 40% of HCC without molecular evidence
Recent advances in genome-wide analysis, including whole- of AFB1 exposure (El-Serag & Rudolph, 2007). The aberration
genome sequencing and DNA array technologies, have proof the p53 pathway can be caused by molecules that inappro-
vided more detailed information on the alterations of oncogenic priately regulate p53 functions. MDM2 is an E3 ubiquitin
or tumor suppressive genes. These sophisticated studies drew ligase targeting tumor suppressor proteins, including p53 and
an important conclusion that HCC does not have a clear addic- RB1. Strikingly, gankyrin (encoded by PSMD10), which pro-
tion to a specific gene, and that genomic alterations in HCC motes such protein degradation by MDM2, was overexpressed
are quite extensive. Accordingly, development of molecularly in 100% (n = 34) of human HCC (Higashitsuji et al, 2000,
targeted therapy for HCC will be challenging. Nevertheless, 2005). More recently, genome-wide copy number variation
several frequently mutated genes were found in HCC genomes, analysis identified interferon regulatory factor 2 (IRF2) as a
including those for telomerase reverse transcriptase (TERT), novel tumor suppressor gene in HCC that activates the p53
p53, and β-catenin. pathway, and showed that IRF2 loss-of-function mutations
Telomere shortening is a key feature of chronic liver disease were frequently and exclusively found in HBV-related HCC
that allows sustained proliferation of hepatocytes (Urabe et al, (Guichard et al, 2012). The CDKN2A gene encodes for two
1996). In human HCC, telomere shortening has been shown splice-variant products, including p16INK4a and p14ARF,
to have a positive correlation with increased chromosome which positively regulate the p53 and RB1 signaling pathways.
instability—chromosomal gains, losses, and translocations—by The expression of the CDKN2A gene was suppressed in 30%
promoting chromosomal fusions (Plentz et al, 2004). A study to 70% of human HCC as a result of methylation of the pro-
of telomerase-deficient mice reveals that telomere dysfunction moter region (Jin et al, 2000; Liew et al, 1999; Matsuda et al,
initiates tumor formation (Farazi et al, 2003). It is noteworthy 1999; Weihrauch et al, 2001). Loss of heterogeneity (LOH) of
that 90% of human HCCs show robust activation of telomerase chromosome 9p, where CDKN2A is located, was found in 15%
(Lee et al, 2004; Nagao et al, 1999; Shimojima et al, 2004). In to 20% of tumors (Boige et al, 1997; Laurent-Puig et al, 2001;
some HBV-induced HCC tumors, the viral genome was found Nagai et al, 1997). Interestingly, CDKN2A deletion rarely
to be integrated into the TERT locus, which results in increased occurs in HCC when TP53 is also mutated (Tannapfel et al,
expression of telomerase (Ferber et al, 2003; Fujimoto et al, 2001). Taken together, impairment of the p53 and RB1 path-
2012; Murakami et al, 2005). More recently, direct sequencing ways is a common genetic feature of HCC.
of the TERT promoter region revealed that 59% of HCCs had The canonical WNT (wingless type) pathway plays a central
recurrent somatic mutations, which may result in the activation role in the development of many cancer types and is a key regu-
of TERT (Nault et al, 2013). Other findings related to telom- lator of the signal is β-catenin, which is encoded by the CTNNB1
erase biology indicate amplification of telomerase RNA com- gene. CTNNB1 is the most frequently mutated gene in HCC,
ponent gene (TERC) mRNA and allelic loss of chromosome ranging from 20% to 40% (Cieply et al, 2009; Tornesello et al,
10p, where a putative telomerase inhibitor resides (Nishimoto 2013). The activating mutations frequently occur in exon 3,
et al, 2001; Takeo et al, 2001). Such telomerase reactivation which result in the nuclear accumulation of β-catenin, thereby
and telomere shortening in HCC cells may be explained in part inducing WNT-responsive gene expression. Indeed, nuclear
accumulation of β-catenin in HCC cells and in clinical HCC because of chromosomal gains and losses and by point muta-
samples is associated with a more aggressive phenotype and tions in the genes. However, recent studies have revealed that
with poor prognosis of HCC patients (Inagawa et al, 2002). epigenetic mechanisms—such as DNA methylation and short,
Another frequently mutated gene in the canonical WNT noncoding RNA (21 to 23 nucleotides) species, or microRNA
pathway is AXIN1, a gene encoding a cytoplasmic protein that (miRNA)—also contribute to aberrant expression of oncogenes
negatively regulates WNT signaling. AXIN1 mutations are and tumor suppressor genes. In human HCC, aberrant DNA
found in 6.8% to 15.2% of HCC (Guichard et al, 2012; Satoh methylation patterns have been detected (Kanai et al, 1999;
et al, 2000). Loss-of-function mutations of AXIN1 lead to a Thorgeirsson & Grisham, 2002; Yu et al, 2003). More impor-
decrease in degradation of β-catenin and to nuclear localization tant, hypermethylation has been observed at the earliest stages
of this protein. of HCC development, and the extent of hypermethylation
Epigenetic regulation of the genome is a fundamental deter- tends to increase with tumor progression (Lee et al, 2003).
minant of global gene expression. Epigenetic regulators have Specific gene targets for hypermethylation include CDKN2A,
come to be recognized as tumor suppressors because next- PTGS2, CDH1, PYCARD, GADD45B, and DLC1. Among
generation sequencing of cancer genomes has defined frequent these genetic elements, it has been shown that CDKN2A,
mutations in epigenetic regulators, including chromatin remod- GADD45B, and PTGS2 expression were directly affected by
eling proteins and histone-modification proteins. In HCC, the methylation, using human HCC cell lines (Higgs et al, 2010;
AT-rich interactive domain (ARID) family, including ARID1A, Liew et al, 1999; Matsuda et al, 1999; Murata et al, 2004).
ARID1B, and ARID2, was found to be mutated in 6% to 17% In addition, miRNA contributes to messenger (mRNA)
of HCC genomes (Fujimoto et al, 2012; Guichard et al, 2012; instability by hybridizing with its complementary target
Li et al, 2011). Through the interaction with the switch/sucrose sequence, followed by mRNA degradation, so that a protein
nonfermentable (SWI/SNF) chromatin-remodeling complex, cannot be generated. Several studies reveal aberrant expression
ARID family proteins bind transcription factors and recruit the of some miRNAs in human HCCs compared with the adjacent,
remodeling activity to a specific gene. Mechanisms underlying nontumorous counterparts. For example, miR-21 was expressed
tumor development by loss-of-function mutations of ARID in human HCC; it targets the phosphatase and tensin (PTEN)
family genes are not known in detail, but it is highly likely that tumor suppressor gene (Meng et al, 2007). MicroRNA-122,
ARID mutations are associated with tumor proliferation, dedif- which targets the cyclin G1 cell-cycle regulator (Gramantieri
ferentiation, and inhibition of apoptosis (Wu & Roberts, 2013). et al, 2007), is abundant in normal hepatocytes and is essential
MLL and MLL3 are histone methyltransferases that positively for homeostasis of hepatocytes. Its implication in hepatocar-
regulate gene transcription. MLL gene translocations are fre- cinogenesis was revealed by using knockout mice and clinical
quently found in infant leukemia, but the significance of loss- samples (Kojima et al, 2011; Tsai et al, 2012). More recently,
of-function mutations of MLL family in HCC remains to be evidence was presented that links aberrant expression of miRNA
determined (Nakagawa & Shibata, 2013). Key genetic altera- to the multistep process of hepatocarcinogenesis. The expres-
tions in HCC are summarized in Table 9D.2. sion of miR-26a was diminished in murine and human tumors,
It is evident that the expression and function of oncogenes resulting in enhanced activity of cyclin D2 and E2 to promote
and tumor suppressor genes are affected by copy number cell proliferation. Moreover, when exogenous miR-26a was
overexpressed in mice prone to form multiple HCCs, substan-
tial protection from disease progression was observed, indicat-
TABLE 9D.2 Genetic Alterations Frequently Found in HCC ing a possible therapeutic approach for this disease (Kota et al,
Gene Symbol Genetic Alteration Frequency (%) 2009). These findings indicate that epigenetic and posttran-
scriptional regulation of gene expression plays an important
Telomere Maintenance role in hepatic oncogenesis.
TERT Promoter activation 20-60
WNT/β-Catenin
SIGNAL TRANSDUCTION PATHWAYS
CTNNB1 Gain-of-function mutation 20-40
AXIN1 Loss-of-function mutation 3-16 Genetic alterations of oncogenes and tumor suppressor genes
Cell Cycle impinge on a wide variety of signal transduction pathways
TP53 Loss-of-function mutation 11-35 involved in proliferation and tumor cell viability. Although the
spectrum of affected signal transduction pathways in HCC cells
CDKN2A Loss of heterozygosity 15-20
Promoter inactivation 30-70 is more heterogeneous compared with that of affected signals
Proliferation in other tumor types, key pathways are commonly dysregulated
in human HCC, such as the WNT/β-catenin, erythroblastosis
IRF2 Loss-of-function mutation 5
(ERB)-B receptor tyrosine kinase (ERBB)/extracellular signal-
IGF2R Allelic loss 0-13
regulated protein kinase (ERK)/phosphatidylinositol-3-kinase
Epigenetic Modifier
(PI3K), and insulin-like growth factor (IGF)/insulin receptor
ARID1A Loss-of-function mutation 10-17 substrate (IRS)/ERK/PI3K cascades (Fig. 9D.2).
ARID1B Loss-of-function mutation 7 The WNT/β-catenin pathway regulates cell proliferation,
ARID2 Loss-of-function mutation 6-16 motility, and differentiation. WNT proteins are ligands that
MLL3 Loss-of-function mutation 4 bind to Frizzled (FZD) cell-surface receptors to stabilize
HCC, Hepatocellular carcinoma. β-catenin in the cytoplasm, followed by translocation to the
From Ding J, Wang H: Multiple interactive factors in hepatocarcinogenesis, Cancer Lett
nucleus, where it upregulates WNT-responsive genes (Thomp-
346:17–23, 2014; and Marquardt JU, Thorgeirsson SS: SnapShot: hepatocellular carcinoma, son & Monga, 2007). In the absence of WNT signaling, the
Cancer Cell 25:550, 2014. amount of cytosolic β-catenin is low as a result of proteolytic
WNT/β-catenin pathway ERBB pathway IGF/IRS pathway
Without WNT ligands With WNT ligands
IGF
LRP-5/6
LRP-5/6
TGF-α
WNT
Receptor
ERBB
FZD FZD
Cytoplasm
DVL DVL DVL
IGF-1R
AXIN IRS
AXIN
GRB2
APC P
GSK-3β β-Catenin SOS
P RAS-
β-Catenin PI3K AKT
GTP
β-Catenin IKK
β-Catenin
P RAF
β-Catenin
β-Catenin P GSK-3β
β-Catenin MEK mTOR
Proteasomal degradation ERK
Nucleus
Groucho β-Catenin ERK
TCF/LEF TCF/LEF Transcr.

factor
Target gene expression Target gene expression

CCND1, MYC, PTGS2, JUN FOS, JUN, MYC
FIGURE 9D.2. Schematic diagram showing the major components of three signal-transduction pathways involved in hepatic oncogenesis. In the
WNT/β-catenin cascade, accumulation of β-catenin is regulated by WNT ligands. The ERBB and IGF/IRS pathways both use downstream activation
of several kinases, including the RAS/RAF/MEK/ERK and PI3K/AKT cascades. AKT, v-Akt murine thymoma viral oncogene homolog; APC, antigen-
presenting cell; AXIN, ***; ERK, extracellular signal-regulated protein kinase; DVL, Disheveled (receptor); ERBB, erythroblastosis (ERB)-B receptor
tyrosine kinase; FZD, Frizzled (receptor); GRB2, growth factor receptor–bound protein 2; GSK, glycogen synthetase kinase; GTP, guanosine triphos-
phate; IGF-1R, insulin-like growth factor-1 receptor; IKK, inhibitor of κB kinase; I/R, ischemia/reperfusion; IRS, insulin receptor substrate; LEF, leukocyte
enhancer factor; MEK, mitogen-activated protein kinase; mTOR, mammalian target of rapamycin; SOS, son of sevenless homolog; TCF, T-cell factor;
TGF, transforming growth factor; transcr., transcription; WNT, wingless type.
degradation produced by the action of glycogen synthetase immunohistochemical staining (Fujito et al, 2004; Inagawa
kinase-3β (GSK-3β)/ adenomatosis polyposis coli (APC)/ et al, 2002; Ishizaki et al, 2004; Mao et al, 2001; Wong et al,
AXIN kinase destruction complex. However, when WNT 2001). These findings suggest that nuclear accumulation of
ligands bind to the Frizzled/ low density lipoprotein receptor- β-catenin is an excellent biomarker for activation of WNT sig-
related protein-5/6 (LRP-5/6)/Dishevelled (DVL) receptor naling in HCC. More important, β-catenin activation occurs at
complex, phosphorylation of β-catenin by GSK-3β is inhibited an earlier stage of the oncogenic process in dysplastic cells,
to allow its accumulation in the cytoplasm. The β-catenin suggesting that the WNT/β-catenin cascade is directly involved
molecules are then transported into the nucleus and bind to in tumor formation (Calvisi et al, 2001, 2004; de la Coste et al,
T-cell factor (TCF)/leukocyte enhancer factor (LEF) tran- 1998; Merle et al, 2004; Thorgeirsson & Grisham, 2002).
scription factors; this complex acts as transcriptional regula- The other cell-surface molecules that play a major role in
tors. Finally, the TCF/LEF/β-catenin complex promotes the activation of WNT/β-catenin signaling are called FZD
activation of target genes, including cyclin D1 (CCND1), receptors FZD. There are 10 FZD receptors (FZD1 to FZD10)
myelocytomatosis viral oncogene (MYC), prostaglandin- in humans. Indeed, studies reveal that 23% to 59% of HBV-
endoperoxide synthase 2 (PTGS2), and JUN, which leads to related HCCs overexpress the FZD7. Through the interaction
proliferation of HCC cells. with a FZD-ligand WNT3, overexpressed FZD7 leads to the
As discussed above, gain-of-function mutations of CTNNB1 activation of this pathway in human tumors and HCC cell lines
or loss-of-function mutations of AXIN1 result in the nuclear (Kim et al, 2008; Pez et al, 2013). The functional consequences
accumulation of β-catenin. The frequency of β-catenin nuclear of FZD7 overexpression were enhanced cell motility and inva-
accumulation varies between 17% and 75%, as determined by sion. In murine HCC models, overexpression of FZD7 occurred
in dysplastic nodules and HCC tissue but not in normal liver IGF-binding proteins (IGFBPs). IGF-2R is a decoy receptor
(Merle et al, 2004, 2005). These findings illustrate that FZD7 to which IGF-2 exclusively binds, but no activation of down-
overexpression was a common event during hepatic oncogen- stream molecules occurs. IGFBP-3, a predominant form of
esis and that it promotes tumor cell motility and invasion. IGFBPs, is a neutralizing peptide that binds to circulating
Receptor tyrosine kinases (RTKs) and subsequent signal IGF-1 and IGF-2 (Pollak, 2008).
transduction, including the mitogen-activated protein kinase Many lines of evidence have established that dysregulation
(MAPK) pathway and the phosphatidylinositol-3-kinase of IGF pathway is involved in the malignant transformation,
(PI3K)/AKT pathway, play a central role in tumor cell prolif- cancer development, and even resistance to anticancer agents.
eration and survival. The ERBB family (especially ERBB1), The same is true for HCC. Comprehensive analysis of 104
MET proto-oncogene, IGF-1 receptor (IGF-1R) are the RTKs HCC cases revealed that IGF pathway activation, namely the
frequently activated in HCC cells. The ERBB family consists presence of phosphorylated IGF-1R, was found in 21% of the
of four RTKs, including ERBB1/EGFR/HER1, ERBB2/HER2/ cases (Tovar et al, 2010). Overexpression of IGF-2, downregu-
NEU, ERBB3/HER3, and ERBB4/HER4 (Yarden & Sli- lation of IGFBP-3, or allelic losses of IGF2R, which all lead to
wkowski, 2001). When ligands bind to the ERBB1 receptor, the activation of the pathway, was found in 25% of the cases.
autophosphorylation occurs that leads to an association with Another report showed that the IGF-2 ligand was overex-
the growth factor receptor–bound protein 2 (GRB2) adaptor pressed in 16% to 40% of tumors, as well as in dysplastic tissue,
molecule and PI3K. When phosphorylated ERBB1 binds to suggesting that IGF-2 may act by autocrine and/or paracrine
GRB2, the complex activates the RAS (rat sarcoma) oncopro- mechanisms (Breuhahn et al, 2006). IRS adaptor molecules are
tein, resulting in enhancement of RAF serine/threonine kinase also important in HCC development. IRS-1 is overexpressed
activity. Activated RAF kinase triggers MEK/ERK kinases, as in the majority of human HCCs (Cantarini et al, 2006; Nishi-
well as ERK, which translocates to the nucleus and upregulates yama & Wands, 1992); overexpression was associated with
the transcription of oncogenes such as FBJ murine osteosar- increased tumor size and progression via activation of the
coma viral oncogene homolog (FOS), JUN, and MYC. When MEK/ERK cascade, which promotes cell proliferation (Tanaka
phosphorylated ERBB1 binds to PI3K, ERBB1 activates PI3K et al, 1997). Constitutive MEK/ERK pathway activation may
and the downstream AKT kinase. AKT phosphorylates a also occur via downregulation of a RAF kinase inhibitor protein
number of important molecules, including mechanistic target (RKIP); this event promotes HCC cell proliferation and migra-
of rapamycin (mTOR), GSK-3β, and inhibitor of kappa B tion (Lee et al, 2006). Indeed, downregulation of RKIP was
kinase (IKK) to promote proliferation and viability of tumor found in 90% of human HCC and suggests that it plays a role
cells. as a tumor suppressor protein in this disease. IRS-2 has also
It has been established that ERBB1 and ERBB3 were over- been found to be overexpressed in the majority of human
expressed in 68% and 84% of HCC, respectively, which cor- tumors and in HCC cell lines (Boissan et al, 2005). The finding
relates with a more clinically aggressive phenotype (Ito et al, of IGF signaling in HCC has led to the development of several
2001a). Ligands for ERBB include epidermal growth factor monoclonal antibodies that bind and neutralize the IGF-1
(EGF), TGF-α, heparin-binding EGF (HB-EGF), amphiregu- receptor and of small molecules that antagonize IGF-1R as a
lin, β-cellurin, and epiregulin. In this regard, TGF-α and potential therapeutic approach; such agents are undergoing
HB-EGF proteins may play a role in the pathogenesis of HCC. preclinical and early-phase clinical trials for solid tumors and
TGF-α was frequently upregulated at an earlier stage of tumor HCC (Pollak, 2008).
formation (Schaff et al, 1994; Yeh et al, 1987; Zhang et al, The angiogenic pathway is an emerging and promising
2004). In murine models, HCC develops in TGF-α–overex- molecular target in HCC because HCC is usually a highly
pressing transgenic mice, whereas TGF-α–knockout mice were vascular tumor. Angiogenesis is a complex process regulated by
resistant to chemically induced hepatic neoplasms, indicating many factors, such as vascular endothelial growth factor
that TGF-α upregulation was closely linked to the oncogenic (VEGF), platelet-derived growth factor (PDGF), and basic
process (Jhappan et al, 1990; Russell et al, 1996). HB-EGF fibroblast growth factor (bFGF). VEGF and its receptors
protein was a potent mitogen for hepatocytes, detected in 59% VEGF-R1, -2, and -3 are overexpressed in HCC, and overex-
to 100% of HCC (Inui et al, 1994). Moreover, overexpression pression of VEGF is associated with poor prognosis (Dhar
of HB-EGF was found at an earlier stage of HCC with mod- et al, 2002; Tseng et al, 2008). PDGF recruits pericytes and
erately or well-differentiated features (Ito et al, 2001b), suggest- smooth muscle cells around new vessels, which are important
ing that HB-EGF is an important ligand in initiation of this components of arteries. bFGF is involved in endothelial cell
disease. As a consequence, MAPK activity is increased in HCC migration, capillary branching, and the activity of proteases,
compared with adjacent nontumorous tissues (Osada et al, essential for angiogenesis as well. These growth factors bind
2005; Schmidt et al, 1997). and stimulate their corresponding receptors on angiogenic
The IGF signaling cascade is a well-defined pathway that cells, and activate the subsequent signal pathways including
regulates energy metabolism and cell growth. It consists of MAPK, PI3K/AKT, SRC, and phospholipase C (PLC)-γ path-
ligands, receptors, adaptors, and subsequent MAPK and PI3K/ ways. Sorafenib is a small-molecule multikinase inhibitor that
AKT pathways. The ligands are IGF-1, IGF-2, and insulin as inactivates VEGF receptors, PDGF receptors, v-Kit Hardy-
well. The activation of this pathway begins when the ligands Zuckerman 4 feline sarcoma viral oncogene homolog (KIT),
bind to receptors, including IGF-1R homodimers and heterodi- and B-Raf proto-oncogene (BRAF) (Wilhelm et al, 2006). It is
mers consisting of IGF-1R and insulin receptors, resulting in the first drug proven to be effective in advanced HCC (Llovet
the activation of their kinase domain. Activated receptors then et al, 2008). Thus the antiangiogenic approach may become a
phosphorylate adaptor proteins such as IRS-1 and IRS-2, mainstream of the systemic treatment of HCC, and in fact,
which are able to trigger the MAPK and PI3K/AKT signaling many clinical trials for use of such agents are currently
cascades. Negative regulators of this pathway are IGF-2R and underway.
LIVER CANCER STEM CELLS taurocholate cotransporting polypeptide was identified as a

candidate receptor of this virus (Ni et al, 2014; Yan et al, 2012),
There exists a hierarchy in a tumor in terms of the potential to which needs further confirmation. The partially double-
divide and of cell-specific functions that differentiated cells stranded DNA is repaired to form a covalently closed circular
exert. A small population of tumor cells with the highest poten- DNA (cccDNA) moiety in the nucleus to serve as a stable
tial and an undifferentiated state gives rise to a bulk tumor template for transcription of the pregenomic mRNA and other
population. Such undifferentiated tumor cells, called cancer species required for productive viral replication. This cccDNA
stem cells (CSCs) or tumor-initiating cells, are able to self- template remains in the nucleus during chronic viral infection
renew and produce differentiated cells by asymmetric division and may persist in the liver for the lifetime of the individual
(Sell & Leffert, 2008; Visvader & Lindeman, 2008). The clinical (Wands, 2004).
importance of cancer stem cell theory is that CSCs are gener- The DNA of the HBV integrates randomly into hepatocyte
ally resistant to conventional anticancer agents and radiother- chromosomes and acts as a nonselective insertional mutagenic
apy, and that when CSCs are brought to distant organs, along agent. Secondary chromosomal rearrangements involving
with a bloodstream or lymphatic flow, they can form metastatic duplications, translocations, and deletions reveal that the major
lesions. Therefore eradication of cancer cells is so difficult that oncogenic effect of HBV integration may be increased genomic
we should target a few, therapeutic-resistant CSCs, as well as instability of the host’s cellular DNA. In greater than 90% of
more differentiated bulky tumor cell population sensitive to patients with HBV-related HCC, fragments of viral DNA have
conventional treatment. In HCC, identification of CSC surface been found integrated into the host genome. Most HCCs
markers has moved toward a deeper understanding of liver contain integrated forms of a high molecular weight, as shown
CSCs because liver CSCs can be prospectively isolated and by Southern blot analysis of DNA extracted from tumor tissue.
analyzed. Liver CSC surface markers include CD133, CD90, This event occurs randomly, because no common chromo-
EpCAM, CD13, CD24, OV6, and CD44 (Ding & Wang, somal site of insertion has been detected (Bréchot, 2000).
2014). It is necessary for the effective targeting of liver CSCs However, a large-scale analysis of HBV DNA integration sites
to find specific pathways that are used for the expansion and revealed that, in some special instances, the integration event
maintenance of stem cell properties. Several studies identified can disrupt the function of specific regulatory genes (Murakami
TGF-β, Janus activating kinase (JAK)/signal-transducer and et al, 2005). Of note, out of 50 HBV-related tumors studied,
activator of transcription 3 (STAT3), NOTCH, and PI3K/ insertion of viral DNA into the TERT locus was observed in
AKT/mTOR pathways as regulatory networks essential for the three with HBV-induced HCC. Telomerase reactivation was
activation and functions of liver CSCs (Wen et al, 2013; Wu found, suggesting that the HBV integration event into the
et al, 2012; Zhao et al, 2013). TERT locus was an important pathologic event in these tumors.
Another gene family recurrently affected by HBV integra-
HEPATITIS B VIRUS tion includes those involved in calcium signaling. Studies indi-
cate that HBV DNA had inserted into the gene encoding for
HBV is the prototype member of the Hepadnaviridae family. SERCA (sarco/endoplasmic reticulum calcium adenosine tri-
Viral members of this group also infect ducks, ground squirrels, phosphatase), which plays a pivotal role in regulating intracel-
and woodchucks. These small, partially double-stranded DNA lular calcium levels and shows as a second messenger involved
viruses contain four overlapping open reading frames (ORFs), in cell proliferation and programmed cell-death pathways
including preC/core, preS/S, P, and X, except for the duck, in (Chami et al, 2000). Collectively, cellular genes involved in
which X is missing (Fig. 9D.3A and B). PreC/core ORF chromosomal integrity and in growth factor–mediated signaling
encodes the precore protein, a precursor of hepatitis B early pathways are occasionally targeted by HBV integration events,
antigen (HBeAg) and core protein (HBcAg), a component of but in the vast majority of tumors, the viral integration is
the nucleocapsid. PreS/S ORF encodes for three proteins, random throughout the host genome (Murakami et al, 2005;
including large (L), middle (M), and small (S; HBsAg) pro- Paterlini-Bréchot et al, 2003). Thus integration of HBV into
teins. The S accounts for approximately 90% of all protein hepatocyte DNA produces specific and nonspecific genetic
produced from preS/S transcripts. The P gene encodes for a alterations that contribute to hepatocarcinogenesis.
DNA-dependent DNA polymerase, which also has reverse Studies on murine models expressing HBV-related trans-
transcriptase and RNase H activities. The partially double- genes such as HBx, as shown in Figure 9D.4, as well as trun-
stranded HBV genome (approximately 3.2 kb in length) exists cated preS/S regions, provide evidence for their role in hepatic
within a nucleocapsid. The X region encodes for a multifunc- tumor development (Chisari et al, 1989; Kim et al, 1991). The
tional protein, HBx. Although HBx is not a component of HBV most well-established transgene in this regard encodes for the
particles, it is believed to play a crucial role in viral replication. HBx protein. This is a 154–amino-acid molecule highly con-
Acute and chronic infection of the liver with HBV appears not served among mammalian hepadnaviruses (Chen et al, 1993),
to produce cytopathic effects on hepatocytes; however, several and it has multifunctional and pleiotropic properties that mod-
components of the viral particles activate the host’s immune ulate cellular functions, including transcription, signaling cas-
response, and cytotoxic T cells (CTLs) eliminate HBV-infected cades, DNA repair, protein degradation, and cell-cycle control
hepatocytes (Bertoletti & Ferrari, 2003; Fattovich, 2003). Such (Murakami, 1999). During viral replication, HBx is localized
immune responses induce sustained cycles of hepatocyte injury in the cytosol with a minor fraction present in the nucleus.
and regeneration that contribute to cirrhosis and HCC tumor Cytosolic HBx activates the RAS/RAF/MEK/ERK, PI3K/AKT
formation. pathway, SRC kinase, and JAK /STAT cascades, leading to the
HBV enters the hepatocyte by means of one or more yet-to- increased cell proliferation (Bouchard & Schneider, 2004).
be-defined cell-surface receptors, and the envelope glycoprotein Constitutive expression of HBx also promotes hepatocarcino-
is subsequently removed (Fig. 9D.3C). More recently, sodium genesis, in combination with activation of the insulin/IGF-1/
Envelope
P
PreS/S
P Core
HBV genome
~ 3.2 kb
DNA
PreS/S
(+) strand
(-) strand X
Lipid bilayer
Nucleocapsid
A PreC/core B
HBV
Cell injury
and increased cell
turnover
Recycling
Core particle plus

strand synthesis
Core particle minus
Entry of HBV
strand synthesis
into cell
Vesicular transport
to cell membrane
Budding into
endoplasmic
Core particle reticulum
HBx protein
Translation
Core assembly and
RNA packaging HBx as a
transcriptional
transactivator
Cytoplasm
Transcription
cccDNA
Repair
Mutations that increase

viral replication
Nucleus
Viral integration
with cellular DNA
C
FIGURE 9D.3. A, Structure of the hepatitis B virus (HBV) DNA genome, showing the four open reading frames involved in generation of preC/core,
preS/S, P, and X proteins. B, Structure of infectious HBV particle, showing the nucleocapsid containing an HBV genome and polymerase (P) and
an envelope derived from the lipid bilayer, where preS/S proteins are embedded. C, Life cycle of HBV, showing viral entry through receptors, followed
by uncoating and translocation of HBV DNA to the nucleus, where it is repaired to generate a covalently closed circular DNA (cccDNA) form that
serves as the template for transcription of the pregenomic and other viral mRNA necessary for replication. (From Wands JR: Prevention of hepatocel-
lular carcinoma, N Engl J Med 351:1567–1570, 2004. Copyright © 2004 with permission from Massachusetts Medical Society. All rights reserved.)
JAK STAT
HBx Enhanced
IRS proliferation
p53
Mitochondria RAS RAF MEK ERK
P13K
Ca2+ PYK2 SRC
AKT
Reactive oxygen species Increased

Enhanced Androgen androgen
viability receptor sensitivity
DNA damage
Impaired p53
DNA repair RNA pol II
systems HBx complex Modulated
transcription
HBx
Transcr.
DNA factor
Cytoplasm Nucleus
HBV genome
Insertional mutagenesis
and chromosome instability
FIGURE 9D.4. Characteristics of the HBx protein and its involvement in tumor formation. HBx plays a crucial role in hepatitis B virus (HBV)-induced
carcinogenesis. HBx is located in the cytoplasm and activates cellular signaling cascades. This viral nonstructural protein also inhibits p53-mediated
apoptosis. Nuclear HBx modulates a set of transcription factors through interaction with a RNA polymerase complex. The HBV genome integrates
into the host genome during persistent viral infection and promotes chromosome instability. AKT, v-Akt murine thymoma viral oncogene homolog;
ERK, extracellular signal-regulated protein kinase; HBx, hepatitis B virus X protein; IRS, insulin receptor substrate; JAK, Janus activating kinase; MEK,
mitogen-activated protein kinase; PI3K, phosphatidylinositol-3-kinase; PYK, proline-rich tyrosine kinase; RAF, Raf-1 proto-oncogene; RAS, rat
sarcoma (oncoprotein); SRC, sarcoma; STAT, signal-transducer and activator of transcription; transcr., transcription.
IRS-1/MEK/ERK cascade (Longato et al, 2009). In addition, expression has been shown to block p53-mediated apoptosis,
nuclear HBx has been reported to act as a transcriptional coact- and it provides a clonal selective advantage to HBV-infected
ivator, although it does not directly bind to DNA. Growth hepatocytes (Elmore et al, 1997; Huo et al, 2001; Wang et al,
molecules known to be influenced by HBx expression include 1995).
cyclic adenosine monophosphate (cAMP) response element– Interestingly, HBx can trigger the release of calcium ion
binding protein (CREB), activating transcription factor 2 from mitochondria, leading to the enhanced replication of HBV
(ATF2), activating enhancer binding protein 2 (AP-2), and DNA through interaction with a proline-rich tyrosine kinase 2
CREB-binding protein/p300 (Bouchard & Schneider, 2004). (PYK2) and SRC kinase (Bouchard et al, 2001). Calcium
Several reports indicate an interaction between HBx and release and localization of HBx in mitochondrial membranes
p53 tumor suppressor protein. In this context, HBx binds cause oxidative stress, in which reactive oxygen species (ROS)
to p53 and may suppress its functions. For example, HBx are produced (Bouchard et al, 2003; Yang & Bouchard, 2012).
binds to p53 in the nucleus and inhibits expression of p53 ROS directly damage DNA, leading to aberrant DNA replica-
responsive genes. Nuclear HBx also alters the association of tion. HBx also influences the androgen signaling pathway,
p53 with transcription factors, such as excision repair cross- which may explain, in part, the known male predominance of
complementation group 3 (ERCC3) and transcription factor HBV-induced HCC (Chiu et al, 2007). Taken together, these
IIH (TFIIH), which are involved in nucleotide excision repair findings indicate that HBx plays a complex and pleiotropic role
(Jia et al, 1999; Wang et al, 1994, 1995). Moreover, HBx in the multistep process of tumor development.
Evidence has been accumulating that the risk of HCC is HEPATITIS C VIRUS
substantially increased by viral factors, such as the level of
HBV replication produced by naturally occurring mutations in HCV is a member of the Flaviviridae family and is the only
the core and precore promoter regions (Baptista et al, 1999; member of the genus Hepacivirus (Tellinghuisen & Rice, 2002).
Kao et al, 2003; Kuang et al, 2004). A high viral replication The HCV genome consists of a single positive-strand RNA of
phenotype places the infected liver at greater risk for transfor- approximately 9.6 kb in length (Fig. 9D.5A). HCV RNA con-
mation, as shown in Figure 9D.3C. Finally, with the develop- tains a large (approximately 9.0 kb) ORF in which structural and
ment of diagnostic techniques sensitive enough to detect very nonstructural coding regions are located near the 5′ and 3′ ends
low levels of serum HBV DNA (<100 copies/mL), it has of the viral genome, respectively. Both 5′ and 3′ untranslated
become increasingly apparent that many patients with chronic (UTR) domains have been found to be essential for viral RNA
HCV infection also are infected with low levels of HBV. In replication. Translation of the HCV RNA can be initiated by an
this setting, HBV maintains its oncogenic properties, and evi- internal ribosome entry site (IRES) located in the 5′ UTR. The
dence has accumulated that occult HBV infection, defined as single ORF encodes for a polyprotein precursor, consisting of
less than 10,000 virions per mL of serum, may be associated about 3000 amino acids; it is cleaved into 10 smaller proteins by
with chronic hepatitis and cirrhosis of heretofore unknown the action of several proteases derived from both host and virus.
origin (Wands, 2004). These 10 viral-related molecules include three structural (core
HCV genome ~ 9.6 kb

Open reading frame ~ 9.0 kb
5´ UTR 3´ UTR
Core E1 E2 p7 NS2 NS3 NS4A NS4B NS5A NS5B
IRES
After translation and processing
p22 gp35 gp70 p21 p70 p4 p27 p56 p66
Core E1 E2 p7 NS2 NS3 NS4A NS4B NS5A NS5B
Nucleo- ? Metallo- Serine-protease Anchor for Replicase cofactor RNA-dependent

capsid protease NTPase RNA replicase? Immune evasion RNA polymerase
RNA helicase
Envelope Cofactor of
A NS3 protease
Attachment
HCV particle
Hepatocyte
Receptor Release
Replication
Lipid
Entry droplet
(endocytosis)
n
io
Assembly
t
la
ns
a
Tr
Ribosome
HCV
polyprotein
RNA replication Nucleus

complex
Structural
proteins Endoplasmic reticulum
Cytoplasm
B
FIGURE 9D.5. Diagram of hepatitis C virus (HCV) and the replication process. A, Components of the HCV genome showing the structure of HCV
RNA. HCV proteins, including their role in HCV replication, are presented. B, Life cycle of HCV showing the known steps of viral replication, including
attachment/entry, translation, replication, processing, assembly, and release of virus into the blood. IRES, Internal ribosome entry site; NS, nonstruc-
tural; NTPase, nucleoside triphosphatase; UTR, untranslated region.
[C], E1, and E2) and 7 nonstructural (NS; p7, NS2, NS3, NS4A, a complex and initiate replication of the RNA genome in col-
NS4B, NS5A, NS5B) proteins. The core forms a nucleocapsid laboration with some host proteins. Viral particles are assem-
that contains the HCV genome. The nucleocapsid is covered by bled from the structural proteins, and the RNA genome
an envelope composed of a lipid bilayer, in which the two struc- becomes encased between membranes derived from lipid drop-
tural proteins E1 and E2 are embedded. The function of p7 is lets and the ER. Assembled particles are delivered to the plasma
not well understood, but it appears to be a transmembrane membrane and released into the blood by exocytosis. Impor-
protein with ion channel activity. NS2 is a metalloprotease that tantly, the entire process related to HCV replication is restricted
cleaves between NS2 and NS3. The NS3 is a serine protease that to the cytoplasm; unlike HBV, the viral RNA does not form a
produces NS4A, NS4B, NS5A, and NS5B viral proteins; it also DNA intermediate, and thus the HCV genome does not inte-
has RNA helicase and nucleoside triphosphatase (NTPase) grate into the host’s cellular DNA.
activity. NS4A acts as a cofactor for NS3 activity, and NS4B is The evolution of chronic hepatitis to cirrhosis during per-
an integral membrane protein located on the cytoplasmic side sistent HCV infection may be mediated by the host immune
of endoplasmic reticulum (ER); it has been implicated in the response directed against epitopes that reside on the various
assembly of replicase complex. The NS5A protein is believed to viral proteins. Chronic HCV infection may develop as a result
act as a component of RNA replicase complex, and it plays a role of immune evasion as a function of quasi-species formation
in evasion of the host’s cellular immune response. The NS5B is during active viral replication. The RNA polymerase of HCV
an RNA-dependent RNA polymerase essential for the replica- lacks proofreading capacity and contributes to rapid diversifi-
tion of the HCV genome. cation of the viral population; therefore mutant HCV can
The life cycle of HCV consists of at least six different stages: escape from the adaptive immune response because of selec-
attachment/entry, translation, processing, genome replication, assem- tion pressure (Rehermann & Nascimbeni, 2005). Moreover,
bly, and release into the circulation (Fig. 9D.5B). HCV particles several viral proteins, including NS3 and NS5A, may contrib-
attach to the hepatocyte cell-surface membrane and enter the ute to viral persistence as a result of attenuation of cellular
cell via key proteins such as CD81, SR-BI, claudin-1, and interferon activity, which appears essential for virus elimina-
occludin. After entry into the hepatocyte, the nucleocapsid is tion (Foy et al, 2003; Polyak et al, 2001). These immune-
delivered to the cytoplasm, and the HCV RNA is released and evasion mechanisms promote chronicity of HCV infection,
immediately translated. The large polyprotein is processed on which promotes hepatic inflammation, cirrhosis, and HCC
the cytoplasmic side of the ER, and nonstructural proteins form (Fig. 9D.6).
Stellate cell
TGF-β
Promotion of
fibrosis
Reactive oxygen species Increased

production DNA injury
NS3
Core RAF MEK ERK
Enhanced p53 Immune cell
proliferation
WNT1 β-Catenin p21WAF1
GSK-3β
RNA pol Interferon

complex response
AKT PI3K NS5A pathway
NS5B
Enhanced Attenuated
Replication with viability GRB2 viral
low fidelity elimination
CDK2 p53
p21WAF1
Hepatocyte
FIGURE 9D.6. Pathogenic role of hepatitis C virus (HCV) proteins. This diagram illustrates how HCV proteins, including core, NS3, NS5A, and
NS5B, may modulate cellular function. Such proteins promote hepatocarcinogenesis in diverse ways via activation of signaling pathways and stellate
cells, and they suppress immune responses to the virus and generate oxidative stress in the liver. AKT, v-Akt murine thymoma viral oncogene homolog;
CDK2, cyclin-dependent kinase 2; ERK, extracellular signal-regulated protein kinase; GRB2, growth factor receptor-bound protein 2; GSK-3β, gly-
cogen synthetase kinase-3β; MEK, mitogen-activated protein kinase; NS, nonstructural; PI3K, phosphatidylinositol-3-kinase; pol, polymerase; TGF-β,
transforming growth factor-β; WNT, wingless type.
At least four HCV proteins, including core, NS3, NS5A, transgenic mice (Moriya et al, 1998). Moreover, iron loading
and NS5B, have been proposed to have cellular transforming in transgenic mice expressing the full-length HCV genome
activity both in vitro and in vivo as depicted in Figure 9D.6. promotes HCC formation, again implicating chronic oxidative
Transgenic mice overexpressing HCV core in the liver have stress in the pathogenesis of HCV-mediated HCC (Furutani
been shown to develop steatosis as well as HCC (Moriya et al, et al, 2006). The mechanisms underlying oxidative stress–
1998). In addition, the core protein has been shown to have induced hepatocarcinogenesis involve increased chromosome
transforming potential in vitro (Ray et al, 1996). HCV core instability and mutations in the host cellular DNA produced by
binds numerous cellular proteins and modulates the RAF/ the action of ROS (Machida et al, 2006; Naganuma et al,
MEK /ERK signal transduction pathway (Levrero, 2006). The 2004). Moreover, HCV directly induces lipid accumulation in
core protein also augments the TGF-β pathway by upregulating hepatocytes as a result of ER stress, because viral replication
TGF-β expression in hepatic stellate cells to promote fibrosis occurs on the ER membranes, where the core impedes ER
(Bataller et al, 2004). Enhanced proliferation of HCC cells has functions (Lonard et al, 2004). Hepatic steatosis induced by
been demonstrated by overexpression of the HCV core protein HCV infection promotes oxidative stress and ROS production
in vitro. This phenomenon was due to upregulation of WNT1 (Koike, 2005). Collectively, many of the HCV proteins are
expression, suggesting a functional link between HCV core believed to contribute to hepatocyte transformation during per-
expression during active viral replication and subsequent activa- sistent viral infection through stimulation of cell proliferation,
tion of the WNT/β-catenin signaling cascade (Fukutomi et al, increased cell survival, induction of genomic instability, and
2005). promotion of immune evasion.
NS3 has been shown to complex with the wild-type p53
protein (Ishido & Hotta, 1998). By modulating the activity of FUTURE DIRECTIONS
p53, NS3 inhibits transcription of the leucine carboxyl methyl-
transferase 2 (LCMT2) gene, which encodes a cell-cycle regula- In the postgenomic era, comprehensive examination of tumors
tor, p21WAF1/CIP1. In addition, it has been found to repress via next-generation sequencing and microarray technologies
LCMT2 promoter activity in a dose-dependent manner and makes it highly likely that molecular genetic changes evolving
stimulates cell growth (Kwun et al, 2001). from normal liver to dysplasia to HCC will be more precisely
The NS5A protein may act as a transcriptional modulator defined. Moreover, rapid advances in epigenetics analysis,
through interactions with other cellular proteins, such as GRB2, miRNA function, and cancer stem cell biology will greatly
p53, p21WAF1/CIP1, and CDK2. The functional conse- enhance our understanding of the molecular pathogenesis of
quences have been inhibition of hepatocyte apoptosis, leading HCC. Traditional anticancer drugs, which target the DNA
to persistent HCV infection (Reyes, 2002). One potential replication machinery, have marginal therapeutic effect on
mechanism by which NS5A may be able to exert an effect on HCC growth, and new molecular targets will need to be
gene expression and cellular growth is through functional asso- defined. To date, unlike other cancer types, including breast,
ciations with p53 and TATA box–binding proteins (TBPs). In colon, and lung, most clinical trials of molecularly targeted
addition, NS5A can bind directly to the p85 regulatory subunit therapeutics against HCC, except for sorafenib, have not
of PI3K and phosphorylate AKT, resulting in activation of yielded expected results (see Chapter 101). Reasons for failure
pro–cell survival signals (Street et al, 2004). More recently, it are heterogeneous, but we should learn from each trial and
has been observed that NS5A-mediated activation of PI3K further deepen the understanding of altered properties of
resulted in the stabilization and accumulation of β-catenin in tumor cells (Llovet & Hemandez-Gea, 2014). Such investiga-
the cytoplasm and nucleus through inactivation of GSK-3β tions emphasize the importance of unraveling the molecular
(Milward et al, 2010; Street et al, 2005). mechanisms of liver carcinogenesis, which may ultimately
Another possible pathway activated in HCV-induced HCC result in “personalized” medical approaches for this devastat-
involves oxidative stress generated by increased production of ing disease.
ROS during persistent viral infection. Indeed, ROS production
in the liver has been found to be increased in HCV core References are available at expertconsult.com.
Chapter 9D Molecular biology of liver carcinogenesis and hepatitis 172.e1
REFERENCES Donato F, et al: Alcohol and hepatocellular carcinoma: the effect of

lifetime intake and hepatitis virus infections in men and women, Am
American Cancer Society: Global cancer facts & figures, ed 2, Atlanta, J Epidemiol 155:323–331, 2002.
Ga, 2011, American Cancer Society. Elmore LW, et al: Hepatitis B virus X protein and p53 tumor suppres-
Baptista M, et al: High prevalence of 1762T 1764A mutations in the sor interactions in the modulation of apoptosis, Proc Natl Acad Sci
basic core promoter of hepatitis B virus isolated from black Africans U S A 94:14707–14712, 1997.
with hepatocellular carcinoma compared with asymptomatic carri- El-Serag HB: Epidemiology of viral hepatitis and hepatocellular carci-
ers, Hepatology 29:946–953, 1999. noma, Gastroenterology 142:1264–1273, 2012.
Bataller R, Brenner DA: Liver fibrosis, J Clin Invest 115:209–218, 2005. El-Serag HB, Kanwal F: Epidemiology of hepatocellular carcinoma in
Bataller R, et al: Hepatitis C virus core and nonstructural proteins the United States: Where are we? Where do we go?, Hepatology
induce fibrogenic effects in hepatic stellate cells, Gastroenterology 60:1767–1775, 2014.
126:529–540, 2004. El-Serag HB, Rudolph KL: Hepatocellular carcinoma: epidemiology
Bertoletti A, Ferrari C: Kinetics of the immune response during HBV and molecular carcinogenesis, Gastroenterology 132:2557–2576,
and HCV infection, Hepatology 38:4–13, 2003. 2007.
Boige V, et al: Concerted nonsyntenic allelic losses in hyperploid hepa- Farazi PA, DePinho RA: Hepatocellular carcinoma pathogenesis: from
tocellular carcinoma as determined by a high-resolution allelotype, genes to environment, Nat Rev Cancer 6:674–687, 2006.
Cancer Res 57:1986–1990, 1997. Farazi PA, et al: Differential impact of telomere dysfunction on initia-
Boissan M, et al: Overexpression of insulin receptor substrate-2 in tion and progression of hepatocellular carcinoma, Cancer Res 63:
human and murine hepatocellular carcinoma, Am J Pathol 167:869– 5021–5027, 2003.
877, 2005. Fattovich G: Natural history and prognosis of hepatitis B, Semin Liver
Bouchard MJ, et al: Calcium signaling by HBx protein in hepatitis B Dis 23:47–58, 2003.
virus DNA replication, Science 294:2376–2378, 2001. Fattovich G, et al: Hepatocellular carcinoma in cirrhosis: incidence and
Bouchard MJ, Schneider RJ: The enigmatic X gene of hepatitis B virus, risk factors, Gastroenterology 127:S35–S50, 2004.
J Virol 78:12725–12734, 2004. Fattovich G, et al: Natural history of chronic hepatitis B: special
Bouchard MJ, et al: Activation and inhibition of cellular calcium emphasis on disease progression and prognostic factors, J Hepatol
and tyrosine kinase signaling pathways identify targets of the HBx 48:335–352, 2008.
protein involved in hepatitis B virus replication, J Virol 77:7713– Ferber MJ, et al: Integrations of the hepatitis B virus (HBV) and human
7719, 2003. papillomavirus (HPV) into the human telomerase reverse transcrip-
Branda M, Wands JR: Signal transduction cascades and hepatitis B tase (hTERT) gene in liver and cervical cancers, Oncogene 22:3813–
and C-related hepatocellular carcinoma, Hepatology 43:891–902, 3820, 2003.
2006. Ferlay J, et al: Cancer incidence and mortality worldwide: Sources,
Bréchot C: Molecular bases for the development of hepatitis B virus methods and major patterns in GLOBOCAN 2012, Int J Cancer
(HBV)-related hepatocellular carcinoma (HCC), Semin Cancer Biol 136:E359–E386, 2015.
10:211–231, 2000. Foy E, et al: Regulation of interferon regulatory factor-3 by the hepatitis
Bressac B, et al: Selective G to T mutations of p53 gene in hepato C virus serine protease, Science 300:1145–1148, 2003.
cellular carcinoma from southern Africa, Nature 350:429–431, Fujimoto A, et al: Whole-genome sequencing of liver cancers identifies
1991. etiological influences on mutation patterns and recurrent mutations
Breuhahn K, et al: Dysregulation of growth factor signaling in human in chromatin regulators, Nat Genet 44:760–764, 2012.
hepatocellular carcinoma, Oncogene 25:3787–3800, 2006. Fujito T, et al: Prognostic significance of β-catenin nuclear expression
Calvisi DF, et al: Activation of β-catenin during hepatocarcinogenesis in hepatocellular carcinoma, Hepatogastroenterology 51:921–924,
in transgenic mouse models: relationship to phenotype and tumor 2004.
grade, Cancer Res 61:2085–2091, 2001. Fukutomi T, et al: Hepatitis C virus core protein stimulates hepatocyte
Calvisi DF, et al: Disruption of β-catenin pathway or genomic instabil- growth: correlation with upregulation of Wnt-1 expression, Hepatol-
ity defines two distinct categories of liver cancer in transgenic mice, ogy 41:1096–1105, 2005.
Gastroenterology 126:1374–1386, 2004. Furutani T, et al: Hepatic iron overload induces hepatocellular carci-
Cantarini MC, et al: Aspartyl-asparagyl β hydroxylase overexpression noma in transgenic mice expressing the hepatitis C virus polyprotein,
in human hepatoma is linked to activation of insulin-like growth Gastroenterology 130:2087–2098, 2006.
factor and Notch signaling mechanisms, Hepatology 44:446–457, Giannelli G, et al: Laminin-5 with transforming growth factor-β1
2006. induces epithelial to mesenchymal transition in hepatocellular carci-
Chami M, et al: Hepatitis B virus-related insertional mutagenesis impli- noma, Gastroenterology 129:1375–1383, 2005.
cates SERCA1 gene in the control of apoptosis, Oncogene 19:2877– Gouas D, et al: The aflatoxin-induced TP53 mutation at codon 249
2886, 2000. (R249S): biomarker of exposure, early detection and target for
Chang MH, et al: Decreased incidence of hepatocellular carcinoma in therapy, Cancer Lett 286:29–37, 2009.
hepatitis B vaccinees: a 20-year follow-up study, J Natl Cancer Inst Gramantieri L, et al: Cyclin G1 is a target of miR-122a, a microRNA
101:1348–1355, 2009. frequently down-regulated in human hepatocellular carcinoma,
Chen HS, et al: The woodchuck hepatitis virus X gene is important for Cancer Res 67:6092–6099, 2007.
establishment of virus infection in woodchucks, J Virol 67:1218– Gressner OA: Less Smad2 is good for you! A scientific update on cof-
1226, 1993. fee’s liver benefits, Hepatology 50:970–978, 2009.
Chisari F, et al: Molecular pathogenesis of hepatocellular carcinoma in Guichard C, et al: Integrated analysis of somatic mutations and focal
hepatitis B virus transgenic mice, Cell 59:1145–1156, 1989. copy-number changes identifies key genes and pathways in hepato-
Chiu CM, et al: Hepatitis B virus X protein enhances androgen cellular carcinoma, Nat Genet 44:694–698, 2012.
receptor-responsive gene expression depending on androgen level, Hassan MM, et al: Risk factors for hepatocellular carcinoma: synergism
Proc Natl Acad Sci U S A 104:2571–2578, 2007. of alcohol with viral hepatitis and diabetes mellitus, Hepatology
Cieply B, et al: Unique phenotype of hepatocellular cancers with 36:1206–1213, 2002.
exon-3 mutations in β-catenin gene, Hepatology 49:821–831, 2009. Higashitsuji H, et al: Reduced stability of retinoblastoma protein by
Davila JA, et al: Hepatitis C infection and the increasing incidence of gankyrin, an oncogenic ankyrin-repeat protein overexpressed in
hepatocellular carcinoma: a population-based study, Gastroenterology hepatomas, Nat Med 6:96–99, 2000.
127:1372–1380, 2004. Higashitsuji H, et al: The oncoprotein gankyrin binds to MDM2/
de la Coste A, et al: Somatic mutations of the β-catenin gene are fre- HDM2, enhancing ubiquitylation and degradation of p53, Cancer
quent in mouse and human hepatocellular carcinomas, Proc Natl Cell 8:75–87, 2005.
Acad Sci U S A 95:8847–8851, 1998. Higgs MR, et al: Downregulation of Gadd45beta expression by hepa-
Dhar DK, et al: Requisite role of VEGF receptors in angiogenesis of titis C virus leads to defective cell cycle arrest, Cancer Res 70:4901–
hepatocellular carcinoma: a comparison with angiopoietin/Tie 4911, 2010.
pathway, Anticancer Res 22:379–386, 2002. Huo TI, et al: Hepatitis B virus X mutants derived from human hepa-
Ding J, Wang H: Multiple interactive factors in hepatocarcinogenesis, tocellular carcinoma retain the ability to abrogate p53-induced apop-
Cancer Lett 346:17–23, 2014. tosis, Oncogene 20:3620–3628, 2001.
Paterlini-Bréchot P, et al: Hepatitis B virus-related insertional muta- Llovet JM, Hemandez-Gea V: Hepatocellular carcinoma: reasons for
genesis occurs frequently in human liver cancers and recurrently phase III failure and novel perspectives on trial design, Clin Cancer
targets human telomerase gene, Oncogene 22:3911–3916, 2003. Res 20:2072–2079, 2014.
Inagawa S, et al: Expression and prognostic roles of β-catenin in hepa- Llovet JM, et al: Sorafenib in advanced hepatocellular carcinoma, N
tocellular carcinoma: correlation with tumor progression and post- Engl J Med 359:378–390, 2008.
operative survival, Clin Cancer Res 8:450–456, 2002. Lonard A, et al: Steatosis and hepatitis C virus: mechanisms and sig-
Inoue M, et al: Influence of coffee drinking on subsequent risk of nificance for hepatic and extrahepatic disease, Gastroenterology
hepatocellular carcinoma: a prospective study in Japan, J Natl Cancer 126:586–597, 2004.
Inst 97:293–300, 2005. Longato L, et al: Overexpression of insulin receptor substrate-1 and
Inui Y, et al: Expression of heparin-binding epidermal growth factor in hepatitis Bx genes causes premalignant alterations in the liver, Hepa-
human hepatocellular carcinoma, Gastroenterology 107:1799–1804, tology 49:1935–1943, 2009.
1994. Ma WL, et al: Androgen receptor is a new potential therapeutic target
Ishido S, Hotta H: Complex formation of the nonstructural protein 3 for the treatment of hepatocellular carcinoma, Gastroenterology
of hepatitis C virus with the p53 tumor suppressor, FEBS Lett 135:947–955, 2008.
438:258–262, 1998. Machida K, et al: Hepatitis C virus triggers mitochondrial permeability
Ishizaki Y, et al: Immunohistochemical analysis and mutational analyses transition with production of reactive oxygen species, leading to
of β-catenin, AXIN family and APC genes in hepatocellular carcino- DNA damage and STAT3 activation, J Virol 80:7199–7207, 2006.
mas, Int J Oncol 24:1077–1083, 2004. Mao TL, et al: Expression of mutant nuclear β-catenin correlates with
Ito Y, et al: Expression and clinical significance of Erb-B receptor family non-invasive hepatocellular carcinoma, absence of portal vein spread,
in hepatocellular carcinoma, Br J Cancer 84:1377–1383, 2001a. and good prognosis, J Pathol 193:95–101, 2001.
Ito Y, et al: Expression of heparin-binding epidermal growth factor-like Matsuda Y, et al: pl6INK4 is inactivated by extensive CpG methylation in
growth factor in hepatocellular carcinoma: an immunohistochemical human hepatocellular carcinoma, Gastroenterology 116:394–400,
study, Oncol Rep 8:903–907, 2001b. 1999.
Jee SH, et al: Cigarette smoking, alcohol drinking, hepatitis B, and risk McGlynn KA, et al: International trends and patterns of primary liver
for hepatocellular carcinoma in Korea, J Natl Cancer Inst 96:1851– cancer, Int J Cancer 94:290–296, 2001.
1856, 2004. Meng F, et al: MicroRNA-21 regulates expression of the PTEN tumor
Jhappan C, et al: TGF alpha overexpression in transgenic mice induces suppressor gene in human hepatocellular cancer, Gastroenterology
liver neoplasia and abnormal development of the mammary gland 133:647–658, 2007.
and pancreas, Cell 15:1137–1146, 1990. Merle P, et al: Functional consequences of Frizzled-7 receptor overex-
Jia L, et al: Hepatitis B virus X protein inhibits nucleotide excision pression in human hepatocellular carcinoma, Gastroenterology
repair, Int J Cancer 80:875–879, 1999. 127:1110–1122, 2004.
Jin M, et al: p16 is a major inactivation target in hepatocellular carci- Merle P, et al: Oncogenic role of the Frizzled-7/β-catenin pathway in
noma, Cancer 89:60–68, 2000. hepatocellular carcinoma, J Hepatol 43:854–862, 2005.
Kanai Y, et al: DNA hypermethylation at the D17S5 locus and reduced Moriya K, et al: The core protein of hepatitis C virus induces hepatocel-
HIC-1 mRNA expression are associated with hepatocarcinogenesis, lular carcinoma in transgenic mice, Nat Med 4:1065–1067, 1998.
Hepatology 29:703–709, 1999. Murakami S: Hepatitis B virus X protein: structure, function and
Kao JH, et al: Basal core promoter mutations of hepatitis B virus biology, Intervirology 42:81–99, 1999.
increase the risk of hepatocellular carcinoma in hepatitis B carriers, Murakami Y, et al: Large-scaled analysis of hepatitis B virus (HBV)
Gastroenterology 124:327–334, 2003. DNA integration in HBV-related hepatocellular carcinomas, Gut
Kew MC: Synergistic interaction between anatoxin B1 and hepatitis B 54:1162–1168, 2005.
virus in hepatocarcinogenesis, Liver Int 23:405–499, 2003. Murata H, et al: Promoter hypermethylation silences cyclooxygenase-2
Kim CM, et al: HBx gene of hepatitis B virus induces liver cancer in (Cox-2) and regulates growth of human hepatocellular carcinoma
transgenic mice, Nature 351:317–320, 1991. cells, Lab Invest 84:1050–1059, 2004.
Kim M, et al: Functional interaction between Wnt3 and Frizzled-7 Nagai H, et al: Comprehensive allelotyping of human hepatocellular
leads to activation of the Wnt/β-catenin signaling pathway in hepa- carcinoma, Oncogene 14:2927–2933, 1997.
tocellular carcinoma cells, J Hepatol 48:780–791, 2008. Naganuma A, et al: Promotion of microsatellite instability by hepatitis
Koike K: Hepatitis C as a metabolic disease: Implication for the patho- C virus core protein in human non-neoplastic hepatocyte cells,
genesis of NASH, Hepatol Res 33:145–150, 2005. Cancer Res 64:1307–1314, 2004.
Kojima K, et al: MicroRNA122 is a key regulator of α-fetoprotein Nagao K, et al: Telomerase reverse transcriptase mRNA expression and
expression and influences the aggressiveness of hepatocellular carci- telomerase activity in hepatocellular carcinoma, J Gastroenterol
noma, Nat Commun 2:338, 2011. 34:83–87, 1999.
Kota J, et al: Therapeutic microRNA delivery suppresses tumorigenesis Nakagawa H, Shibata T: Comprehensive genome sequencing of the
in a murine liver cancer model, Cell 137:1005–1017, 2009. liver cancer genome, Cancer Lett 340:234–240, 2013.
Kuang SY, et al: Specific mutations of hepatitis B virus in plasma Nault JC, et al: High frequency of telomerase reverse-transcriptase
predict liver cancer development, Proc Natl Acad Sci U S A 101:3575– promoter somatic mutations in hepatocellular carcinoma and pre-
3580, 2004. neoplastic lesions, Nat Commun 4:2218, 2013.
Kwun HF, et al: p53-dependent transcriptional repression of p21(waf1) Ni Y, et al: Hepatitis B and D viruses exploit sodium taurocholate
by hepatitis C virus NS3, J Gen Virol 82:2235–2241, 2001. co-transporting polypeptide for species-specific entry into hepato-
Laurent-Puig P, Zucman-Rossi J: Genetics of hepatocellular tumors, cytes, Gastroenterology 146:1070–1083, 2014.
Oncogene 25:3778–3786, 2006. Nishimoto A, et al: Functional evidence for a telomerase repressor gene
Laurent-Puig P, et al: Genetic alterations associated with hepatocellular on human chromosome 10p15.1, Oncogene 20:828–835, 2001.
carcinomas define distinct pathways of hepatocarcinogenesis, Gastro- Nishiyama M, Wands JR: Cloning and increased expression of an
enterology 120:1763–1773, 2001. insulin receptor substrate-1-like gene in human hepatocellular car-
Lee S, et al: Aberrant CpG island hypermethylation along multistep cinoma, Biochem Biophys Res Commun 183:280–285, 1992.
hepatocarcinogenesis, Am J Pathol 163:1371–1378, 2003. Ogata H, et al: Deletion of the SOCS3 gene in liver parenchymal cells
Lee CM, et al: Telomerase activity and telomerase catalytic subunit in promotes hepatitis-induced hepatocarcinogenesis, Gastroenterology
hepatocellular carcinoma, Hepatogastroenterology 51:796–800, 2004. 131:179–193, 2006.
Lee HC, et al: Loss of Raf kinase inhibitor protein promotes cell pro- Osada S, et al: Evaluation of extracellular signal regulated kinase
liferation and migration of human hepatoma cells, Gastroenterology expression and its relation to treatment of hepatocellular carcinoma,
131:1208–1217, 2006. J Am Coll Surg 201:405–411, 2005.
Levrero M: Viral hepatitis and liver cancer: the case of hepatitis C, Parkin DM, et al: Cancer incidence in five continents (vol VIII), IARC
Oncogene 25:3834–3847, 2006. Scientific Publications, No. 155, Lyon, France, 2002, IARC
Li M, et al: Inactivating mutations of the chromatin remodeling gene Publications.
ARID2 in hepatocellular carcinoma, Nat Genet 43:828–829, 2011. Perz JF, et al: The contributions of hepatitis B virus and hepatitis C
Liew CT, et al: High frequency of p16INK4A gene alterations in hepato- virus infections to cirrhosis and primary liver cancer worldwide,
cellular carcinoma, Oncogene 18:789–795, 1999. J Hepatol 45:529–538, 2006.
Chapter 9D Molecular biology of liver carcinogenesis and hepatitis 172.e3
Pez F, et al: Wnt signaling and hepatocarcinogenesis: molecular targets Tornesello ML, et al: Mutations in TP53, CTNNB1 and PIK3CA
for the development of innovative anticancer drugs, J Hepatol genes in hepatocellular carcinoma associated with hepatitis B and
59:1107–1117, 2013. hepatitis C virus infections, Genomics 102:74–83, 2013.
Plentz RR, et al: Hepatocellular telomere shortening correlates with Tovar V, et al: IGF activation in a molecular subclass of hepatocellular
chromosomal instability and the development of human hepatoma, carcinoma and pre-clinical efficacy of IGF-1R blockage, J Hepatol
Hepatology 40:80–86, 2004. 52:550–559, 2010.
Pollak M: Insulin and insulin-like growth factor signalling in neoplasia, Tsai WC, et al: MicroRNA-122 plays a critical role in liver homeostasis
Nat Rev Cancer 8:915–928, 2008. and hepatocarcinogenesis, J Clin Invest 122:2884–2897, 2012.
Polyak SJ, et al: Hepatitis C virus nonstructural 5A protein induces Tseng PL, et al: Overexpression of VEGF is associated with positive
interleukin-8, leading to partial inhibition of the interferon-induced p53 immunostaining in hepatocellular carcinoma (HCC) and
antiviral response, J Virol 75:6095–6106, 2001. adverse outcome of HCC patients, J Surg Oncol 98:349–357, 2008.
Ray RB, et al: Hepatitis C virus core protein cooperates with Ras and Urabe Y, et al: Telomere length in human liver diseases, Liver 16:293–
transforms primary rat embryo fibroblasts to tumorigenic phenotype, 297, 1996.
J Virol 70:4438–4443, 1996. Visvader JE, Lindeman GJ: Cancer stem cells in solid tumours: accu-
Rehermann B, Nascimbeni M: Immunology of hepatitis B virus and mulating evidence and unresolved questions, Nat Rev Cancer 8:755–
hepatitis C virus infection, Nat Rev Immunol 5:215–229, 2005. 768, 2008.
Reyes GR: The nonstructural NS5A protein of hepatitis C virus: an Wands JR: Prevention of hepatocellular carcinoma, N Engl J Med
expanding, multifunctional role in enhancing hepatitis C virus 351:1567–1570, 2004.
pathogenesis, J Biomed Sci 9:187–197, 2002. Wang XW, et al: Hepatitis B virus X protein inhibits p53 sequence-
Russell WE, et al: Liver regeneration and hepatocarcinogenesis in trans- specific DNA binding, transcriptional activity, and association with
forming growth factor-α-targeted mice, Mol Carcinog 15:183–189, transcription factor ERCC3, Proc Natl Acad Sci U S A 91:2230–
1996. 2234, 1994.
Satoh S, et al: AXIN1 mutations in hepatocellular carcinomas and Wang XW, et al: Abrogation of p53-induced apoptosis by the hepatitis
growth suppression in cancer cells by virus-mediated transfer of B virus X gene, Cancer Res 55:6012-6016, 1995
AXIN1, Nat Genet 24:245–250, 2000. Weihrauch M, et al: Frequent k-ras-2 mutations and p16INK4A meth-
Schaff Z, et al: Overexpression of transforming growth factor-alpha in ylation in hepatocellular carcinomas in workers exposed to vinyl
hepatocellular carcinoma and focal nodular hyperplasia from Euro- chloride, Br J Cancer 84:982–989, 2001.
pean patients, Hum Pathol 25:644–651, 1994. Wen W, et al: Cyclin G1 expands liver tumor-initiating cells by Sox2
Schmidt CM, et al: Increased MAPK expression and activity in primary induction via Akt/mTOR signaling, Mol Cancer Ther 12:1796–1804,
human hepatocellular carcinoma, Biochem Biophys Res Commun 2013.
236:54–58, 1997. Wilhelm S, et al: Discovery and development of sorafenib: a multiki-
SEER Cancer Statistics Factsheets: Liver and intrahepatic bile duct cancer. nase inhibitor for treating cancer, Nat Rev Drug Discov 5:835–844,
National Cancer Institute. Bethesda, MD, 2007-2011. Available at 2006.
http://seer.cancer.gov/statfacts/html/livibd.html. Wong CM, et al: β-Catenin mutation and overexpression in hepatocel-
Sell S, Leffert HL: Liver cancer stem cells, J Clin Oncol 26:2800–2805, lular carcinoma: clinicopathologic and prognostic significance,
2008. Cancer 92:136–145, 2001.
Shimojima M, et al: Detection of telomerase activity, telomerase RNA Wu JN, Roberts CW: ARID1A mutations in cancer: another epigenetic
component, and telomerase reverse transcriptase in human hepato- tumor suppressor?, Cancer Discov 3:35–43, 2013.
cellular carcinoma, Hepatol Res 29:31–38, 2004. Wu MH, et al: Androgen receptor promotes hepatitis B virus–induced
Starley BQ, et al: Nonalcoholic fatty liver disease and hepatocellular hepatocarcinogenesis through modulation of hepatitis B virus RNA
carcinoma: a weighty connection, Hepatology 51:1820–1832, 2010. transcription, Sci Transl Med 32:32–35, 2010.
Street A, et al: The hepatitis C virus NS5A protein activates a phos- Wu K, et al: Hepatic transforming growth factor beta gives rise to
phoinositide 3-kinase-dependent survival signaling cascade, J Biol tumor-initiating cells and promotes liver cancer development, Hepa-
Chem 279:12232–12241, 2004. tology 56:2255–2267, 2012.
Street A, et al: Hepatitis C virus NS5A-mediated activation of phos- Yarden Y, Sliwkowski MX: Untangling the ErbB signalling network, Nat
phoinositide 3-kinase results in stabilization of cellular β-catenin and Rev Mol Cell Biol 2:127–137, 2001.
stimulation of β-catenin-responsive transcription, J Virol 79:5006– Yan H, et al: Sodium taurocholate cotransporting polypeptide is a
5016, 2005. functional receptor for human hepatitis B and D virus, Elife 1:e00049,
Takeo S, et al: Examination of oncogene amplification by genomic 2012.
DNA microarray in hepatocellular carcinomas: comparison with Yang B, Bouchard MJ: The hepatitis B virus X protein elevates cytosolic
comparative genomic hybridization analysis, Cancer Genet Cytogenet calcium signals by modulating mitochondrial calcium uptake, J Virol
130:127–132, 2001. 86:313–327, 2012.
Tanaka S, et al: Biological effects of human insulin receptor substrate-1 Yeh YC, et al: Elevation of transforming growth factor α and its rela-
overexpression in hepatocytes, Hepatology 26:598–604, 1997. tionship to the epidermal growth factor and α-fetoprotein levels in
Tannapfel A, et al: INK4a-ARF alterations and p53 mutations in hepa- patients with hepatocellular carcinoma, Cancer Res 47:896–901,
tocellular carcinomas, Oncogene 20:7104–7109, 2001. 1987.
Tellinghuisen TL, Rice CM: Interaction between hepatitis C virus Yu J, et al: Methylation profiling of twenty-four genes and the concor-
proteins and host cell factors, Curr Opin Microbiol 5:419–427, dant methylation behaviours of nineteen genes that may contribute
2002. to hepatocellular carcinogenesis, Cell Res 13:319–333, 2003.
Theise ND, et al: Dysplastic nodules and hepatocarcinogenesis, Clin Zhang J, et al: Expression of transforming growth factor-α and hepatitis
Liver Dis 6:497–512, 2002. B surface antigen in human hepatocellular carcinoma tissues and its
Thompson MD, Monga SPS: Wnt/β-catenin signaling in liver health significance, World J Gastroenterol 15:830–833, 2004.
and disease, Hepatology 45:1298–1305, 2007. Zhao W, et al: 1B50-1, a mAb raised against recurrent tumor cells,
Thorgeirsson SS, Grisham JW: Molecular pathogenesis of human hepa- targets liver tumor-initiating cells by binding to the calcium channel
tocellular carcinoma, Nat Genet 31:339–346, 2002. α2δ1 subunit, Cancer Cell 23:541–556, 2013.
CHAPTER 10
Liver immunology
Steven C. Katz, Zubin M. Bamboat, Venu G. Pillarisetty, and Ronald P. DeMatteo
FUNDAMENTALS OF IMMUNOLOGY suppressor cells (MDSCs) regulate immune responses in the

periphery and prevent autoimmunity. Immunoinhibitory recep-
The immune system is composed of two major components: tors, including programmed death-1 (PD-1) and cytotoxic
innate and adaptive immunity (Fig. 10.1; Janeway, 2001). Innate T-lymphocyte–associated antigen-4 (CTLA-4), are crucial
immunity refers to the nonspecific first line of defense against modulators of T-cell function, working in concert with suppres-
danger signals from pathogens or tumor cells. The repertoire sor cells to limit and control immune responses. As will be
of innate immune cells includes natural killer (NK) cells, mac- discussed later, Tregs, MDSCs, and immunoinhibitory path-
rophages, and dendritic cells (DCs). Innate immune cells are ways in the liver cooperate to create a highly tolerogenic milieu.
armed with receptors for danger signals found on pathogens or Intrahepatic tolerance is a fundamental aspect of liver immu-
host cells. Such pattern recognition receptors (PRRs) enable nology that is central to its position as an immune organ at the
the rapid response of innate immune cells to infection or host interface between ingested exogenous antigens and secondary
injury. PRRs bind to well-conserved molecules from microbes, lymphoid organs.
including lipopolysaccharide, other bacterial cell wall moieties, The balance between tolerance and immunity in the liver is
and pathogen nucleic acids. PRR signaling and the ensuing tightly regulated by intrahepatic immune cells and associated
response leads to destruction of the invading pathogen or tumor signaling pathways. Several features of liver immunology point
via phagocytosis or release of various cytotoxic or inflammatory to the skewing of this balance toward tolerance under normal
agents. Activation of innate immunity can ultimately lead to physiologic conditions (Fig. 10.3). First, the fact that the liver
stimulation of adaptive immune cells such as T and B cells, via is one of the most common sites for metastatic disease suggests
cytokine secretion and antigen presentation. Crosstalk between that malignant cells are able to exploit intrahepatic immuno-
the innate and adaptive immune systems bridges the nonspe- suppression. Second, compared with other solid-organ trans-
cific initial response to a highly specialized system capable of plants, liver allografts do not require immunosuppression in
long-lasting immunologic memory. certain strains of mice and require less immunosuppression in
Adaptive immunity involves antigen-specific responses, humans. Third, the immune system is often unable to clear
which occur de novo during an initial immune response or chronic hepatitis B and C viral infections. Additionally, oral
rapidly upon repeat exposure to a particular pathogen. The ingestion or portal vein injection of foreign proteins can lead to
adaptive immune system comprises T and B lymphocytes that tolerance in animal models. Conversely, the liver is the site of
circulate within the blood, lymphatic tissues, and nonlymphoid several autoimmune processes, including primary sclerosing
organs, including the liver. T and B cells express highly specific cholangitis and primary biliary cirrhosis. Despite the prominent
cell-surface receptors capable of recognizing particular anti- role that the intrahepatic immune system plays in modulating
gens. T-cell activation requires presentation of antigen by benign and malignant disorders, the study of liver immunology
antigen-presenting cells (APCs) such as DCs. APCs mediate remains underdeveloped. We review our current understanding
antigen presentation to T cells within the context of major of liver immune cell function and their role in disease.
histocompatibility complex molecules (MHC I or MHC II). In
addition, APCs provide a critical “second signal” through Anatomic Considerations
costimulatory molecules, and the response is further modulated Because the vascular supply of the liver derives principally from
by secreted cytokines (Fig. 10.2). Classically, CD4+ helper T the portal venous system draining the gut, a heavy antigen
cells recognize antigen in the context of MHC II, whereas load is delivered to the liver, particularly after meals. Portal
CD8+ cytotoxic T cells engage antigen loaded onto MHC I venous blood flows slowly through the vast network of hepatic
molecules. Several subsets of CD4+ cells (T helper, or Th cells) sinusoids, which are discontinuously lined by fenestrated
orchestrate and polarize the immune response to address par- endothelium lacking a basement membrane (Fig. 10.4). The
ticular situations. sluggish flow of blood allows for the efficient capture of antigens
Although activation of innate and adaptive immunity is by leukocytes traveling in the blood within the sinusoids and
essential for combating pathogens and malignant cells, overly by the endothelial cells lining the sinusoids. The liver reticulo-
exuberant immune responses can result in severe tissue damage. endothelial system, comprising liver sinusoidal endothelial cells
The immune system is normally able to distinguish self from (LSECs) and Kupffer cells (KCs), is very efficient at extracting
nonself and is controlled by numerous regulatory mechanisms. antigens from portal blood. LSECs capture and process antigen
During T-cell development, autoreactive cells are deleted at levels comparable to professional APCs, such as DCs (Katz
through an elegant process of selection within the thymus. Cell et al, 2004). The microscopic anatomy of the liver also favors
types such as regulatory T cells (Tregs) and myeloid-derived the ability of bloodborne leukocytes to interact with hepatic
173
INNATE IMMUNITY ADAPTIVE IMMUNITY
B
PMN
B
MAC B B
NK DC CD4 CD4
CD4
CD4
CD4 CD4
MAC
PMN
CD8
CD8 CD8
DC
CD8
CD8 CD8
Tumor or microbe antigen

Cytokine Antibody
HOURS 0-12 DAYS 1-5
FIGURE 10.1 Innate and adaptive immunity. A, Innate immunity includes the body’s initial defenses against infection. It includes certain comple-
ment proteins, epithelial barriers, natural killer (NK) cells, neutrophils (polymorphonuclear [PMN]), phagocytes such as macrophages (MACs), and
antigen-presenting cells such as dendritic cells (DCs). Innate immune cells may directly kill tumor or infected cells, and then present antigen to adap-
tive immune cells. B, In contrast, adaptive immunity refers to the precisely targeted immune mechanisms that occur later in the immune response.
The adaptive immune system has B-cell–mediated humoral (dissolved) and T-cell–mediated cellular components. The innate and adaptive immune
systems overlap extensively, communicating by direct cellular contact or cytokine secretion. B-cell secretion of antibodies functions to block infections
and trigger the destruction of pathogenic organisms. T-cell–mediated immunity occurs after antigen presentation and leads to direct cellular lysis by
CD8 T cells with support from the CD4 arm.
parenchymal cells and resident immune cells of the liver. The release large amounts of soluble MHC class I molecules that
microanatomic and rheologic features of the hepatic sinusoids potentially could function to neutralize cytolytic T cells. KCs
facilitate highly efficient antigen presentation and interactions and NK T cells have been implicated as important in the devel-
among immune cells. opment of hepatic tolerance, because the depletion of either
has been associated with a loss of oral tolerance. Liver DCs and
Tolerance and Immunosuppression T cells also have been suggested to play a role in liver tolerance
One of the most intriguing aspects of liver immunology is the because, as a group, they have been shown to be less immuno-
strong tendency to develop tolerance to orally ingested antigens genic than their counterparts from lymphoid organs, such as
and liver allografts. From a teleologic perspective, tolerance to the spleen and peripheral blood (Bamboat et al, 2009a; Katz
oral antigens is clearly advantageous. In experimental animal et al, 2005). LSECs have been found to promote tolerance, and
models and clinical liver transplantation studies, a greater prohepatic B cells are suppressed in certain models (Thorn et al,
pensity for graft acceptance has been noted compared with 2014). Controversy exists, however, as to whether these differ-
transplantation of other solid organs. A liver transplant protects ences are developmental, due to conditioning within the liver
a kidney allograft transplanted simultaneously from the same or as a result of chemotactic signals in the liver that attract a
donor (Creput et al, 2003). Unfortunately, primary and meta- unique population of cells. It is likely that both intrinsic differ-
static tumors in the liver exploit intrahepatic immunosuppres- ences in liver immune cells and environmental factors are criti-
sion to evade destruction by the immune system. A deeper cal in defining the nature of liver immunology.
understanding of liver tolerance may support therapeutic inter-
ventions to stimulate immunity for cancer or control liver LIVER IMMUNE CELLS
immune cell function for inflammatory conditions.
Multiple competing theories to explain the cause of liver Although the liver performs myriad nonlymphoid functions, it
tolerance have been advanced. One postulate is that, as in contains numerous nonparenchymal cells (NPCs), one quarter
central thymic tolerance, clonal deletion of antigen-specific T of which are leukocytes. The composition of the intrahepatic
cells occurs in the liver. Another theory is that liver allografts leukocyte population is markedly different from that seen in
Chapter 10 Liver immunology 175
Effector function
Treg
Tolerance CD8
Immunity CD4
Proinflammatory
-γ Intracellular pathogens
IFN Treg
CD8 T
GF- Antiinflammatory
β, IL
-1 0 Tolerance CD8 Intrahepatic tumor
Treg
PD-1
IFN-γ Proinflammatory PD-L1
Th1 Intracellular pathogens CD4
CD8
Naïve T TNF, IL-1
or B cell CD8
2 Th2
IL-4 Allergy, asthma CD8
APC IL-5, 13 Helminth infections CD4
CD8
1 IL-17 Extracellular pathogens
Th17
TNF, IL-21 Autoimmunity Treg
CD4
Signal 3 CD4
Treg IL-10 Immunosuppression
CD4
TGF-β Tolerance
IL-10 Antibody production FIGURE 10.3 The balance between tolerance and immunity in the
Bcell Autoimmune disease
TGF-β, IFN liver. The liver exhibits a fine balance between the induction of tolerance
FIGURE 10.2 Antigen-presenting cell (APC) instruction of T and and excessive immune responses. Examples of the liver as a tolerogenic
B cells. Signals 1 (antigen presentation), 2 (costimulation), and 3 (cyto- organ include its role in oral tolerance, chronic hepatitis infection, and
kine production) between APCs and naïve T or B cells govern the transplantation. In addition, the liver is among the most common sites
subsequent adaptive immune response. Naïve CD4+ T cells can differ- of distant metastatic disease in patients with cancer. In contrast, the liver
entiate into four types of T cells—Th1, Th2, Th17, and Treg—each with is also the site of many autoimmune diseases, such as autoimmune
a distinct cytokine profile and specific effector function. In turn, pro- hepatitis, primary sclerosing cholangitis, and primary biliary cirrhosis.
grammed CD4+ T cells modulate CD8+ T cells that can differentiate into Immune responses to intrahepatic antigens occur when effector CD4+
either cytotoxic or regulatory cells capable of killing foreign pathogens and CD8+ T cells are not suppressed by regulatory T cells (Treg) or
or suppressing immune responses, respectively. B cells can also dif- immunoinhibitory pathways such as the programmed death-1/pro-
ferentiate into cytokine and antibody-producing cells, which play a role grammed death ligand-1 (PD-1/PD-L1) axis (left). In contrast, when
in responses against tumors, pathogens, and autoimmune diseases. Tregs expand or immunoinhibitory pathways are upregulated, tolerance
APCs, Antigen-presenting cells; IFN-γ, interferon-γ; IL, interleukin; TGF- occurs (right). Hepatocytes may also present antigens and express
β, transforming growth factor-β; TNF, tumor necrosis factor. immunomodulatory molecules. (From Saied A, et al: Immunotherapy for
solid tumors—a review for surgeons, J Surg Res 187:525–535, 2014.)
other organs (Fig. 10.5). The liver contains most of the cellular
components of innate and adaptive immunity. Importantly,
liver immune cells demonstrate unique functional properties Dendritic Cells
that tend to promote a tolerogenic milieu. Substantial hetero- DCs are a heterogeneous population of leukocytes that are
geneity among liver immune cells precludes simple generaliza- primarily responsible for the capture of antigens in the periph-
tions concerning their function. The pleiotropic roles attributed ery and subsequent presentation to immune effector cells. DCs
to liver immune cells are likely to be highly contextual, render- are now recognized as the most potent APCs of the immune
ing their study both fascinating and challenging. system. Murine DCs are typically isolated based on their high
expression of the integrin CD11c. Immature DCs are special-
Antigen-Presenting Cells ized to capture antigens and then migrate to lymph nodes,
Experimental and clinical observations that antigens passing where they can interact with T cells. After an encounter with
through the liver can lead to tolerance make the understanding a pathogenic stimulus, such as bacterial lipopolysaccharide,
of intrahepatic antigen presentation particularly relevant (Li DCs undergo phenotypic and functional changes, whereby
et al, 2004). APCs play a crucial role in driving adaptive their ability to capture antigens is diminished, but they increase
immune responses and bridging innate to adaptive immunity. their expression of MHC class II and T cell costimulatory
DCs, LSECs, KCs, and hepatic B cells all play a role in antigen molecules. Costimulatory molecule expression is essential in
presentation within the liver. The context in which an APC facilitating antigen presentation and efficient T-cell activation.
presents an antigen can dramatically alter the response of Inflammatory conditions often promote a process of matura-
antigen-specific T cells. Specifically, when antigen presentation tion, whereby DCs increase expression of MHC and costimula-
occurs in conjunction with the appropriate costimulatory mol- tory molecules, enabling efficient antigen presentation and
ecules, T cells proliferate and develop an immunogenic pheno- T-cell activation. However, liver DC phenotype and function
typic and functional profile (Fig. 10.6). In contrast, antigens is somewhat unique, as we discuss later.
presented in the absence of costimulation or presence of immu- The body of literature addressing the function of DCs grown
noinhibitory signals lead to anergy or activation-induced T-cell in vitro from murine bone marrow progenitors or human
death, two of the mechanisms of peripheral tolerance induction peripheral blood mononuclear cells is vast. DCs isolated directly
and maintenance. from the spleen, lymph nodes, or thymus also have been well
Dendritic FIGURE 10.4 Microvascular hepatic anatomy. A, The liver is orga-

Plate of
hepatocytes cell nized as lobules defined by their relation to portal vascular bundles and
Bile duct
Arteriole Portal central veins. B, Liver sinusoidal endothelial cells (LSECs) line the hepatic
vein
sinusoids and have fenestrated membranes. A variety of immune cells
Sinusoid exist within the sinusoids and are able to traverse the sinusoidal mem-
brane to enter and exit the space of Disse, which is in contact with
hepatocytes ( From Crispe IN: Hepatic T cells and liver tolerance, Nat
Rev Immunol 3:51–62, 2003.) C, Scanning electron micrograph of LSEC
(×20,000) demonstrated the classic fenestrated cell membrane. (From
Katz SC: Liver sinusoidal endothelial cells are insufficient to activate T
cells, J Immunol 173:230–235, 2004.)
Lymphocyte
Central
vein Kupffer cell
A
To gain a better understanding of liver DCs as they function
in situ, recent focus has been on freshly isolated liver DCs. In
our initial studies, we expanded the number of liver DCs to
facilitate their isolation. Because granulocyte-macrophage
LSEC Hepatocyte
colony-stimulatory factor (GM-CSF) had been used to grow
DC progenitors in culture from liver nonparenchymal cells, and
because we had previously shown that in vivo GM-CSF over-
Space of Disse Hepatic expression increases the number of myeloid DCs in the spleen
sinusoid
Kupffer Lymphocyte (Miller et al, 2002a), we suspected that it might expand liver
cell Immature DCs as well. We found that overexpression of GM-CSF using
DC
an adenoviral vector led to a dramatic increase in the number
of highly immunostimulatory liver DCs (Pillarisetty et al,
2003). We also showed that a large proportion of the cells
recruited to the liver by GM-CSF were DC precursor cells that
Blood from develop into CD11c+ DCs in culture, which suggests that DC
portal vein
Blood to development can occur in the liver. We used immunohisto-
central vein
chemistry to locate both types of recruited cells. The liver DCs
Lymph drains from space of Disse and enters were concentrated around the central veins, whereas the DC
lymphatic capillaries in the portal area
precursors were distributed diffusely throughout the liver
B parenchyma.
Our experience with expanding liver DCs sheds light on in
situ liver DC development. We found that compared with the
relatively well-studied DCs from the spleen, CD11c+ liver DCs
were immature and only weakly immunostimulatory (Pillari-
setty et al, 2004). In contrast to spleen DCs, liver DCs are
heterogeneous in their expression of MHC class II and costim-
ulatory molecules. Myeloid (CD11b+) and lymphoid (CD8α+)
liver DCs, which each comprise approximately 10% of the total
population of DCs in the liver, were as able to activate T cells,
as were their splenic counterparts. The bulk of the remaining
cells, which had low-to-no expression of CD11b and CD8α,
were poor T-cell stimulators. This study showed that the pres-
ence of these atypical DCs accounted for the weakly activating
nature of liver DCs on the whole. More recently, we discovered
that the CD11chi subset of DCs within the liver is required for
effective presentation of soluble protein (Plitas et al, 2008).
Using a transgenic mouse in which CD11chi DCs can be
depleted selectively, we found that activation of antigen-specific
C
CD8+ T cells in the liver only occurred in the presence of
CD11chi DCs.
Our investigation into liver DCs has now been extended into
humans. We found that freshly isolated DCs from human liver
studied. In contrast, because of the rarity of DCs in the liver, exhibit tolerogenic properties when compared with autologous
few studies have shown the phenotype or function of liver DCs. blood DCs. Human liver DCs are weaker stimulators of T cells
Initial studies showed that liver DC progenitors are relatively and produce the antiinflammatory cytokine interleukin-10 (IL-
immature with poor immunostimulatory ability that can 10), which induces the differentiation of naïve CD4+ T cells
promote immunologic tolerance (Lau & Thomson, 2003; Lu into regulatory T cells with suppressive function (Fig. 10.7;
et al, 1994, 2001). Bamboat et al, 2009a).
4%
2%
6%
5%
5%
LSEC
DC
3% Kupffer
T Cells
NKT Cells
B Cells
NK Cells
75%
LSEC
T CELL KC
DC
B
FIGURE 10.5 Liver immune cell composition. A, After mechanical and enzymatic digestion of freshly harvested murine liver specimens, hepa-
tocytes were removed by centrifugation. Flow cytometry analysis was then performed on hepatic nonparenchymal cells (NPCs). Liver sinusoidal
endothelial cells (LSECs) constitute most (~75%) of the liver NPCs. Liver leukocytes contain a particularly high proportion of natural killer T (NKT)
cells and Kupffer cells. Liver dendritic cells (DC) are unique in their subset composition and function. B, LSECs line the hepatic sinusoidal spaces,
where liver T cells, KCs, and DCs interact to orchestrate immune responses.
Kupffer Cells Techniques that allow the more precise delineation of cel-
Liver macrophages, referred to as Kupffer cells, have long been lular subtypes have stimulated reassessment of previously held
believed to be the primary phagocytic cells of the liver. KCs notions concerning KCs and other immune cells. Specifically,
represent the largest pool of macrophages in the body, derived the advent of flow cytometry has allowed us to define cells by
from monocytic precursors in the blood. They are typically their surface marker expression, rather than simply by size or
found in the hepatic sinusoids; however, they also can migrate other crude physical characteristics shared by many cell types.
through the space of Disse to interact with hepatocytes (see Fig. This point is illustrated by the fact that many liver DCs express
10.4). KCs have been thought to play a major role in antigen the cell-surface marker CD11b, which has been used to define
presentation and have been implicated in portal venous toler- KCs in the liver. Much of the antigen-presenting ability previ-
ance. Evidence also suggests that KCs may regulate T-cell ously attributed to KCs may be a result of DC function. KCs
responses to antigens and induce immune tolerance to liver and MDSCs also have similar cell-surface marker expression.
allografts (Sun et al, 2003). The lack of consistent definitions of cells by surface marker
independently capable of triggering a T-cell–mediated immune

Teff
response does not, however, exclude the possibility that LSECs,
in concert with DCs or KCs, play an important role in antigen
IL-2
IFN-γ Carcinoma -1 presentation in the liver.
PD
PD-L1
Effector Cells
Teff
PD-L1 PD-1 Teff T Cells
Like other liver immune cell populations, intrahepatic T cells
have unique properties enabling them to contribute to mainte-
Carcinoma
MDSC nance of a tolerogenic milieu. T cells are a heterogeneous
DC population of adaptive immune cells with both effector and
suppressor subtypes. CD4 helper T cells (Th1 and Th2 cells)
CTLA4 orchestrate immune responses, CD8 cytotoxic T cells destroy
B7
-1
/2
infected host or malignant cells, and Tregs play an immuno-
modulatory role. The liver also has a rich supply of nonclassic
Treg Teff T-cell subsets, including NKT and γδ T cells. The nature of
the interactions between APCs and T cells, and hence T-cell
IL-10
TGF-β
functional polarization, determine the outcome of a particular
immune response.
FIGURE 10.6 Context of antigen presentation determines T-cell T cells are classified based on the types of antigen recogni-
activation status. The three main APCs are dendritic cells (DCs), mac- tion receptors present on their surface. The liver contains a full
rophages, and B cells. The immune response depends on the context complement of T-cell subsets, although the relative proportions
in which antigens are presented to T cells. The type of APC and the of each population are different when compared with lymphoid
presence or absence of costimulatory molecules are important in deter-
organs. The broadest way to classify T cells is as either con-
mining whether a T cell has no response (anergy) or is activated. Pre-
sentation of antigen by an APC to a T cell typically result in triggering of
ventional or unconventional. Conventional T cells express the
an adaptive immune response (left), unless suppressor cells or immu- αβ T-cell receptor in association with either CD4 or CD8.
noinhibitory signals intervene (right). APC, Antigen-presenting cell; IFN, These are the most prevalent T cells in the body and account
interferon; IL, interleukin; MDSC, myeloid-derived suppressor cells; PD- for about one third of the murine liver Tcell population. In
L1, programmed death ligand-1; Teff, effector T cell; TGF-β, transform- contrast, unconventional T cells expressing NK markers or the
ing growth factor-β; Treg, regulatory T cells. (From Khan H, et al: The γδ T-cell receptor comprise a greater proportion of liver T cells,
prognostic value of liver tumor T cell infiltrates, J Surg Res 191:189– approximately 50% and 10%, respectively. Immunosuppressive
195, 2014.) Tregs expressing FOXP3 are heavily represented in the liver.
In contrast, the liver also contains T cells known as Th17 cells,
which are capable of producing the highly inflammatory cyto-
expression and the continued use of less precise methods of cell kine IL-17. Th17 cells play an important role in the promotion
isolation have led to continued ambiguity regarding KCs. An of inflammatory and fibrotic disorders affecting the liver.
alternative explanation for the overlap in functional definitions The diversity of conventional T cells is based on their rec-
for DCs, KCs, and MDSCs is that all of these cell types rep- ognition of specific peptide antigen motifs within the context of
resent alternative liver myeloid cell differentiation programs, MHC class I or II molecules expressed by APCs. The αβ T-cell
developing from a common precursor cell. receptor is highly variable, and numerous T cells, each recogniz-
ing a different antigen presented by APCs, are present in the
Liver Sinusoidal Endothelial Cells immune system. CD8+ T cells respond to peptides presented
Liver sinusoidal endothelial cells are highly specialized cells that on MHC class I molecules, which are expressed by nearly every
line the hepatic sinusoids. LSECs are distinguished by the pres- cell in the body, excluding erythrocytes. Activated CD8+ T cells
ence of fenestrations in their cellular membranes (see Fig 10.4). become cytotoxic T lymphocytes. CD4+ helper T cells recognize
The fenestrations are believed to facilitate the selective passage antigens presented on MHC class II molecules on the surface
of antigens between the sinusoid and the hepatic parenchyma, of professional APCs. CD4+ T-cell subsets, Th1 and Th2 cells,
and may also increase the surface area available for antigen then regulate and amplify the immune response by secreting
presentation. This strategic placement puts LSECs in the ideal cytokines, which affect nearby effector cells.
position to interact with antigens and immune cells passing The interaction between CD4+ and CD8+ cells is clearly
between the liver and the portal venous system. important, and some studies have suggested that the ratio
Several studies have shown that, in addition to serving as a between these two cell types, which is approximately 2 : 1 in the
structural component of the hepatic sinusoids, LSECs are spleen, lymph nodes, and peripheral blood, is reversed in the
immune cells with the ability to capture and present antigen to liver (Crispe, 2003). Murine data from our laboratory, using
T cells (Knolle & Limmer, 2001). As with KCs, considerable more inclusive isolation techniques, have shown that the liver
controversy surrounds the immunologic function of LSECs. In has a similar CD4+ to CD8+ T-cell ratio of approximately
contrast to earlier work, we have shown that although LSECs 1.5 : 1, as found in the spleen and node. Prior studies have also
are highly capable of capturing various antigens in vivo and in supported the notion that intrahepatic T cells are destined for
vitro, they lack the ability to activate T cells in the absence of apoptosis or death. In contrast, we have found that liver T-cell
exogenous costimulation (Katz et al, 2004). The differences in function is suppressed within the intrahepatic space but can be
results may derive from the use of more specific methods of cell rescued upon transfer to the spleen or stimulation ex vivo (Katz
isolation in the latter study. The finding that LSECs are not et al, 2005). This finding suggests that intrahepatic T-cell
**
**
** **
3000 Sham 5000 *
Sham
I/R
Serum ALT (IU/L)

2500 4000 I/R
ALT (IU/L)
2000
3000
1500
2000
1000
500 1000
0 0
WT WT TLR9–/–TLR9–/– Treatment PBS DT PBS DT
A Isotype 1A8 Isotype 1A8 B Donor cells PBS PBS cDCs cDCs
70% Hepatic I/R

Bone marrow
monocytes
CCR2-mediated
Liver cDC Inflammatory recruitment
Monocyte
TLR9
↓TNF
Hepatocyte ↓IL-6 ↓ALT

death
↓ROS
IL-10 ROS
Neutrophil
↑TNF
TLR9 ↑ALT
↑ROS
C
FIGURE 10.7 The role of Toll-like receptor 9 (TLR9) in neutrophils and dendritic cells (DCs) during liver ischemia/reperfusion (I/R). A,
Wild-type (WT) and TLR9−/− mice were depleted of neutrophils using an antibody (1A8) or isotype control (Isotype). Mice then underwent sham lapa-
rotomy (Sham) or 1 hour of segmental ischemia, followed by 12 hours of reperfusion (I/R). Serum alanine aminotransferase (ALT) was then measured.
B, Conventional DCs (cDCs) were depleted selectively in transgenic mice via a single injection of diphtheria toxin (DT). Control mice were injected
with phosphate-buffered saline (PBS). Twelve hours later, mice were injected intravenously with cDCs or PBS just prior to sham or I/R. Serum ALT
was measured 12 hours later. C, After segmental liver I/R (70% ischemia), the resultant hepatocyte death causes release of host DNA, which activates
cDCs and neutrophils via TLR9. cDCs produce interleukin-10 (IL-10), which suppresses the function of inflammatory monocytes. As a result, less
inflammatory cytokines (tumor necrosis factor [TNF], IL-6) and reactive oxygen species (ROS) are produced, causing less liver injury. In contrast,
neutrophil TLR9 activation causes an increase in TNF and ROS production, which increases ALT. *P < .05; **P < .01; CCR2, C-C chemokine recep-
tor. (A, From Wiley InterScience. Bamboat ZM, et al: Toll-like receptor 9 inhibition confers protection from liver ischemia-reperfusion injury, Hepatology
51:621–632, 2009; B and C, modified from the American Society for Clinical Investigation. Bamboat ZM, et al: Conventional dendritic cells reduce
ischemia/reperfusion injury in mice via IL-10, J Clin Invest 120:559–569, 2010.)
dysfunction is reversible, which points to potential therapeutic immunosuppressive properties as well (Katz et al, 2005). The
opportunities. high proportion of γδ T cells in the liver suggests that they have
an important immunologic role, but further investigation is
γδ T Cells required. As with most lymphocyte populations, heterogeneity
The γδ T-cell receptor is relatively invariant, and can recognize among liver γδ T cells precludes simple generalizations con-
multiple nonpeptide antigens without the need for MHC pre- cerning their functions.
sentation. γδ T cells represent 10% of liver T cells, whereas
they comprise only a small proportion (< 5%) of T cells in the Natural Killer T Cells
blood or lymphoid organs. γδ T cells also are abundant at other NKT cells share characteristics of conventional T cells and NK
environmental interfaces, including the skin and mucosal sur- cells and are defined by the presence of several T-cell and NK
faces. Through secretion of activating and modulatory cyto- cell-surface markers. Most NKT cells react against glycolipid
kines, γδ T cells help orchestrate early responses to atypical antigens in the context of CD1d, which is a MHC class I–like
bacterial and viral pathogens. γδ T cells can also promote anti- glycoprotein. CD1d is expressed on APCs and hepatocytes.
tumor immunity through their early secretion of interferon-γ NKT express invariant T-cell receptor chains that are con-
(IFN-γ) (Gao et al, 2003). A study from our group focused on served across species, suggesting an important role for NKT
liver γδT cells and demonstrated that this cell type has cells in the innate immune response to pathogens. Activated
NKT cells are capable of producing IFN-γ and IL-4, which are T-cell activity that can be usurped by tumors to induce T-cell
the prototypical Th1 and Th2 cytokines, respectively. suppression (see Fig. 10.6). CTLA-4 is expressed on acti-
NKT cells constitute a relatively large proportion of T cells vated T cells and constitutively on Tregs, and blocking the
found in the liver compared with other organs. In addition, an activity of CTLA-4 enables T-cell functional rescue. Like
expansion of NKT cells is seen in the liver in several models of CTLA-4, PD-1 is a coinhibitory receptor but with distinct
liver injury, such as partial hepatectomy. NKT cells have been biologic properties. PD-1 has two known ligands: programmed
shown to play a role in the mechanisms of inflammatory dis- death ligand-1 (PD-L1) and PD-L2. Many tumors and sup-
eases and clearance of infection in the liver. Depletion of NKT pressive immune cells express PD-L1, and PD-1 engagement
cells has been shown to abrogate the effects of experimentally by PD-L1 results in T-cell functional exhaustion (Curiel et al,
induced hepatitis in a mouse model, and mice lacking NKT 2003). Exhausted T cells have a markedly diminished capac-
cells are susceptible to viral and bacterial infections. NKT cells ity for cytokine production, proliferation, and tumor lysis.
are now also thought to play a part in tumor surveillance in the Although blockade of CTLA-4 or PD-1 may result in autoim-
liver. Data from murine primary and metastatic tumor models mune or inflammatory toxicity, targeting PD-1 in the liver
have shown that NKT cells can mediate tumor rejection, in may be associated with protection of normal parenchyma. In
part because of their ability to secrete IFN-γ (Teng et al, 2007). an animal model, we have recently demonstrated that PD-1
Work in other murine models suggests that liver NKTs have blockade can resuscitate T cells within the liver while actually
the capacity to suppress T-cell proliferation and hence contrib- minimizing inflammatory injury to normal parenchyma (Licata
ute to the immunosuppressive nature of the intrahepatic space et al, 2013).
(Katz et al, 2005). Tregs are a subset of CD4 T cells and mediate tolerance by
suppressing antigen-specific T cells (Shevach, 2002). Differen-
Natural Killer Cells tiation of Tregs is programmed by the FOXP3 transcription
NK cells are innate responders and, unlike T cells, do not factor, which is also useful for identifying Tregs experimentally.
possess specific antigen receptors. By expressing a variety of Immunosuppression by Tregs is mediated through numerous
activating and inhibitory receptors, NK cells can bind ligands mechanisms, including secretion of tolerogenic cytokines and
on their target cells. The resulting activation of NK cells causes expression of PD-L1. MDSCs work in concert with Tregs in
the release of lytic granules, or cytokines such as IFN-γ, which promoting an immunosuppressive environment. MDSCs are a
kill the infected host of tumor cell in an MHC-unrestricted heterogeneous group of cells derived from a myeloid lineage
fashion. NK cells are a major component of murine and human pathway. Phenotypically, MDSCs have features of immature
liver lymphocytes (see Fig. 10.5) and mediate inflammatory neutrophils, monocytes, or NK cells. MDSCs express PD-L1
reactions seen in viral and autoimmune hepatitis. Bulk human and also promote T-cell suppression through production of
liver NK cells possess weaker lytic capabilities when compared suppressive cytokines and enzymes. Tregs and MDSCs are
with autologous blood NK cells (Burt et al, 2009) because the likely both important contributors to baseline liver tolerance
liver has a greater proportion of NK-cell subtypes with weaker and suppression of immune responses to intrahepatic infection
cytolytic function when compared with blood NK cells. or cancer.
B Cells
CYTOKINES
B cells, along with T cells, mediate adaptive immune responses.
B cells are responsible for humoral immunity through antibody Cytokines (see Chapter 11) are small proteins that enable com-
production. B cells may also function as APCs under certain munication among immune cells, providing signals that govern
conditions. Although hepatic B cells have received very little and shape responses to tumor or pathogen. Cytokines can be
attention, they make up a significant proportion of liver lym- roughly divided into two categories, those of innate and those
phocytes. They play prominent roles in viral and autoimmune of adaptive immunity (Table 10.1). Cytokines also can be dis-
disease affecting the liver (Novobrantseva et al, 2005). In con- tinguished based on whether they are inflammatory or antiin-
trast, the contribution of hepatic B cells to the biology of liver flammatory. However, individual cytokine function is highly
tumors is poorly defined. Our group has recently reported that contextual, depending on environmental cues and target cell
MDSCs in the liver suppress hepatic B cells through downregu- features. Although hundreds of cytokines have been described,
lation of CD80, which is a costimulatory molecule involved in we limit our discussion here to a select few that have been
T-cell activation (Thorn, 2014). This may be a mechanism shown to be produced in, and have important effects on, the
through which hepatic B-cell function is altered to contribute liver.
toward intrahepatic tolerance. Further work is required to
develop our understanding of how hepatic B cells are affected Tumor Necrosis Factor-α
by the liver microenvironment and in turn influence intrahe- Tumor necrosis factor-α (TNF)-α was originally isolated from
patic diseases. animals treated with lipopolysaccharide, and its name is derived
from the fact that it was found to lyse tumors in vitro and in
IMMUNOINHIBITORY PATHWAYS AND vivo. TNF-α is one of the most important inflammatory cyto-
SUPPRESSOR CELLS kines. It directly regulates innate immunity and often has sys-
temic effects. TNF-α is primarily produced by macrophages in
Suppressive or coinhibitory signaling pathways are important response to lipopolysaccharide from gram-negative bacteria. It
mediators of intrahepatic tolerance. The immunoinhibitory also is secreted by T cells, mast cells, DCs, and NK cells.
receptors receiving the most attention in laboratories and clin- TNF-α has consistently been found to play a pathogenic role
ical trials are CTLA-4 and PD-1 (Brahmer et al, 2012; Finn, in acute liver failure in multiple animal models and in humans.
2008). The PD-1 and CTLA-4 axes are pivotal regulators of TNF receptors are present on most cell types, and the TNF
TABLE 10.1 Cytokines Involved in Innate and Adaptive Immunity

Cytokine Principal Source Principal Target: Biologic Effects
Innate Immunity
TNF-α Macrophages, T cells, DCs Endothelial cells: activation
Neutrophils: activation
Hypothalamus: fever
Liver: synthesis of acute-phase proteins
Muscle and fat: catabolism
Multiple cell types: apoptosis
IL-1 Macrophages, endothelial cells Endothelial cells: activation
Hypothalamus: fever
Liver: synthesis of acute-phase proteins
Chemokines Macrophages, endothelial cells, T Leukocytes: chemotaxis, activation, migration into tissues
cells, fibroblasts, platelets
IL-12 Macrophages, DCs, NK DCs T cells: Th1 differentiation
NK DCs, NK cells, and T cells: IFN-γsynthesis, increased cytologic activity
IFN-α, IFN-β IFN-α: macrophages, plasmacytoid All cells: antiviral state, increased inhibition of IL-12 production and
DCs IFN-β: fibroblasts expression of costimulatory and MHC class II molecules
IL-10 Macrophages, T cells (mainly Th2) Macrophages, DCs: inhibition of IL-12 production and expression of
costimulatory and MHC class II molecules
IL-6 Macrophages, DCs, endothelial Liver: synthesis of acute-phase proteins
cells, T cells
IL-15 Macrophages, others NK cells: proliferation
T cells: proliferation (memory CD8+ cells)
IL-18 Macrophages NK cells and T cells: IFN-γ synthesis
Adaptive Immunity
IL-2 T cells T cells; proliferation, increased cytokine synthesis; potentiates Fas-
mediated apoptosis
IL-4 CD4+ T cells (Th2), mast cells B cells: isotype switching to IgE
T cells: Th2 differentiation, proliferation
NK DCs and NK cells: increased IFN-γ production
IL-5 CD4+ T cells (Th2) Eosinophils: activation, increased production
IFN-γ CD8+ T cells (Th1), NK DCs, NK Macrophages: activation
cells
B cells: isotype switching to opsonizing and complement-fixing IgG
subclasses
T cells: Th1 differentiation
Various cells: increased MHC class I and II expression, increased antigen
processing and presentation to T cells
TGF-β T cells (regulatory), macrophages, T cells: inhibition of proliferation and effector functions
other cell types
B cells: inhibition of proliferation; IgA production
Macrophages: inhibition
Lymphotoxin T cells Neutrophils: recruitment and activation
Lymphoid organogenesis
IL-13 CD4+ T cells (Th2) B cells: isotype switching to IgE
Epithelial cells: increased mucus production
Macrophages: inhibition
IL-17 CD4+ T cells (Th17) Bone marrow progenitors: neutrophil development
Fibroblasts and endothelium: IL-6 and G-CSF production
DCs, Dendritic cells; G-CSF, granulocyte colony-stimulating factor; IFN, interferon; IgA, IgG, immunoglobulin A and G, respectively; IL, interleukin; NK, natural killer; MHC, major histocompatability
complex; TGF-β, transforming growth factor-β; Th, T helper; TNF, tumor necrosis factor.
Modified from Abbas AK: Cellular and molecular immunology, Philadelphia, 2003, Saunders Elsevier.
receptor family is a large group of proteins with myriad roles crosstalk among intrahepatic immune cell subsets has greatly
in inflammation and immunity. informed our appreciation of liver disease pathogenesis. The
following sections highlight the role of the hepatic immune
Interleukin-6 system in a variety of diseases affecting the liver.
IL-6 is a pleiotropic cytokine that is highly expressed by the
liver and has effects on innate and adaptive immunity. It is Transplantation
synthesized by DCs, KCs, T cells, B cells, endothelial cells, Transplant immunology has shed significant insight into the
fibroblasts, and keratinocytes. IL-6 is produced as an early unique properties of the liver as an immunologic organ. Unlike
response to TNF-α and is thought to amplify the inflamma- other organs, the liver can be accepted across MHC barriers in
tory response in the liver through the induction of acute-phase animal models of transplantation (Qian et al, 1994). In addi-
protein production by hepatocytes. As part of its promotion of tion, systemic, donor-specific tolerance often develops in liver
the early immune response, IL-6 also stimulates the differen- allograft recipients (Kamada et al, 1981). In humans, liver
tiation of neutrophils from bone marrow precursors. In transplants confer protection from other organs from the same
coöperation with TGF-β, IL-6 promotes expansion of proin- donor (Rasmussen et al, 1995), and recipients can often be
flammatory Th17 cells. Additionally, experimental models weaned off immunosuppression altogether (Mazariegos et al,
have shown that absence of IL-6 in cultures impairs DC 1997). The unique ability of allografts to resist rejection and
T-cell stimulatory function; however, IL-6–knockout mice promote extrahepatic tolerance speaks to the immunosuppres-
have normal numbers and types of DCs in their spleens and sive properties liver immune cells.
livers (Bleier et al, 2004). IL-6 also is known to play an The development of mouse models of orthotopic liver trans-
important role in liver regeneration after trauma or hepatec- plantation has provided an opportunity to gain mechanistic
tomy. We recently found that serum IL-6 levels correlated insight into the regulation of immune responses in the liver
with activity of intrahepatic T-cell immunotherapy for liver (Qian et al, 1991). Of the factors that are thought to contribute
metastases (Saied et al, 2014). to hepatic tolerance (Box 10-1), microchimerism and induction
of activated T-cell apoptosis have been the most intensively
Type I Interferons (Interferon-α and Interferon-β) studied and widely accepted. Defined as the persistence of
Type I IFNs refers to a family of structurally similar cytokines donor cells in allograft recipients, microchimerism has been
that are crucial in the immune response to viral infection. postulated to be a prerequisite for organ allograft acceptance
IFN-α is secreted by immune cells, whereas IFN-β is secreted (Starzl et al, 1992, 1996). The high antigenic load derived from
by multiple cell types, such as fibroblasts, in response to viral the donor liver coupled with persistence of donor APCs within
infection. IFN-α was previously thought to be produced by lymphoid tissues of the recipient may provide a continuing
macrophages; however, more recent evidence has shown that source of allostimulation (Sriwatanawongsa et al, 1995; Suzuki
plasmacytoid DCs are its primary producers (Siegal et al, et al, 1988). The persistent activation of alloreactive recipient
1999). Type I IFNs have direct antiviral action by triggering T cells is thought to result in exhaustive deletion and induction
virally infected cells to produce enzymes that interfere with of systemic tolerance. The role of specific donor-derived liver
viral RNA or DNA replication. They also lead to the increased APCs, such as DCs and KCs, in contributing to chimerism-
expression of MHC class I molecules on the surface of virally induced tolerance remains unclear. Although freshly isolated
infected cells, increasing their likelihood of being killed by liver DCs have been shown to exhibit tolerogenic potential via
cytolytic CD8+ T cells. Additionally, type I IFNs inhibit cel- the production of IL-10 (Bamboat et al, 2009a), others have
lular proliferation. IFN-α has been used clinically in the treat- shown that donor-derived liver DCs prime recipient T cells to
ment of viral hepatitis and as an adjuvant therapy for differentiate into proinflammatory subtypes that promote rejec-
melanoma. tion (Bosma et al, 2010).
The immunoregulatory role of KCs has also been exten-
Type II Interferons (Interferon-γ) sively investigated. KCs have been implicated in the induction
IFN-γ is a type II IFN. It is the archetypal proinflammatory
cytokine associated with antitumor immune responses. IFN-γ
activates macrophages and plays a crucial role in bridging
innate and adaptive immunity by increasing antigen presenta- BOX 10.1 Factors Contributing to Hepatic Tolerance
tion through the MHC class I and II pathways (Schroder et al, • Immunosuppression via release of soluble MHC class I antigens
2004). It is produced by T cells, B cells, NKT cells, DCs, and • Distinct lymphocyte population with distinct functions (KCs,
macrophages. IFN-γ promotes Th1-driven cytotoxic T-cell DCs, NK, and NKT cells, LSECs)
responses, which are essential for clearance of virally infected • Constant antigen load from the portal circulation
cells or tumors. In addition, IFN-γ promotes MHC expression • Induction of activated T-cell apoptosis
• Microchimerism
on APCs and enhances NK-cell cytotoxic activity. Increases in
• Immature phenotype and tolerogenic function of resident liver
serum IFN-γ levels have been correlated with favorable clinical
APCs (DCs, KCs, LSECs)
responses to immunotherapy treatments. • Altered Th1 versus Th2 profile within the liver (favoring Th2)
• Presence of suppressor cells (Tregs and MDSCs)
IMMUNE SYSTEM IN BENIGN • Immunoinhibitory pathways (CTLA-4, PD-1)
LIVER DISEASES APCs, Antigen-presenting cells; CTLA-4, cytotoxic T-lymphocyte–associated antigen-4; DCs,
dendritic cells; KCs, Kupffer cells; LSECs, liver sinusoidal endothelial cells; MDSC, myeloid-
Intrahepatic immune cells play prominent roles in myriad derived suppressor cells; MHC, major histocompatibility complex; NK, natural killer; NKT, natural
inflammatory, infectious, and neoplastic diseases affecting the killer T cell; PD-1, programmed death-1; Treg, regulatory T cells.
liver. Our understanding of liver immune cell biology and Modified from Gershwin ME, et al: Liver immunology. Philadelphia, 2003, Hanley and Belfus.
of oral tolerance and in lymphocyte apoptosis, which has been www.who.int/mediacentre/factsheets). The cost to society is
a proposed mechanism of immunosuppression in liver trans- magnified when the increased risk for cirrhosis and liver cancer
plantation (Callery et al, 1989). More recently, KCs have been in those infected with HBV and HCV are taken into account.
shown to play a crucial role in induction of antigen-specific Immune-mediated or inflammatory destruction of liver paren-
T-cell tolerance (Breous et al, 2009) and are thought to sup- chyma is a common final pathway in viral hepatitis. The pres-
press T-cell response via the production of prostaglandins (You ence of large numbers of KCs and DCs in the liver, which
et al, 2008). Despite evidence of the tolerogenic potential of produce TNF-α in response to viral antigens, leads to high local
KCs, the use of gadolinium chloride to suppress KC function levels of inflammatory cytokines and untoward host cell damage.
has not been shown to impact survival of liver allografts in Similarly, T cells, NKT cells, and NK cells in the liver produce
mouse models of liver transplantation. The disparate findings IFN-γ, which also serves to amplify the local immune response
concerning the role of KCs in liver transplantation may reflect by activating KCs and DCs. Consistent with this, chronic active
the opposing effects of tolerogenic MDSCs and proinflamma- hepatitis is morphologically identified by piecemeal necrosis
tory macrophages, both of which share considerable phenotypic and a predominantly mononuclear cell infiltrate (Ferrari et al,
overlap with KCs. 1999).
The overall bias of intrahepatic T-cell responses toward Despite immune cell activation and inflammatory cytokine
tolerance and apoptosis might account for the survival of liver production, the host is frequently unable to clear HBV or HCV.
allografts (Crispe, 2003). Activated T cells and their subse- Therefore, to achieve successful viral control and limit collat-
quent apoptosis within the liver are thought to be another criti- eral damage, a sustained, antigen-specific immune response is
cal component in the induction of hepatic tolerance and the necessary (Gerlach et al, 1999; Maini et al, 2000). Recognition
acceptance of liver allografts. Recently, hepatic stellate cells of multiple viral epitopes by the host is also advantageous, as
(HSCs) have been found to mediate liver T-cell apoptosis via it theoretically minimizes the emergence of mutants that escape
the PD-L1/PD-1 axis (Charles et al, 2013). HSCs may also the immune system. Unfortunately, effective T-cell responses
induce expansion of Tregs as another mechanism of influencing against HBV and HCV are susceptible to suppression by the
T-cell tolerance (Dangi et al, 2012). PD-1/PD-L1 immunoinhibitory access and Treg (Fuller et al,
From a more clinical perspective, patients transplanted for 2013). The tolerogenic predisposition of the liver may suppress
end-stage liver disease secondary to hepatitis B or C often adaptive immunity to HBC and HCV, accounting in part for
experience higher viral loads after transplantation secondary to the persistence of these infections.
the antirejection immunosuppressive regimen. Recurrent viral A number of additional host parameters have been shown
disease in the transplanted liver is inevitable and is therefore in experimental models to alter the immune response to viral
one of the most common reasons for allograft failure (Brown, hepatitis infection. Age can influence the outcome of infection,
2005). Recent work using adoptive immunotherapy of donor as vertical transmission to neonates generally results in persis-
lymphocytes provides some encouraging insight into addressing tent infection throughout life. Chronic viral infection in animal
recurrent viral disease in transplant recipients. Lymphocytes models of hepatitis occurs more often when the initial insult
enriched for NK and NKT cells harvested from donor livers at fails to elicit an acute inflammatory response. In woodchucks,
the time of explantation were primed ex vivo for 2 days prior the absence of an acute infection following viral inoculation
to reinjection into allograft recipients. Hepatitis C viral loads correlated with an indolent and chronic course of disease
were found to be significantly lower at 1 month posttransplan- (Nakamura et al, 2001).
tation in patients treated with adoptive therapy versus controls Emerging research tools that will undoubtedly shed more
(Ohira et al, 2009). Using human-hepatocyte chimeric mice, light on the immune mechanisms underlying viral persistence
the authors showed that adoptive transfer of activated donor and resistance to treatment include the development of chime-
lymphocytes conferred protection via the production of IFN-γ. ric mice that are reconstituted with human hepatocytes and
Other groups have also shown that antigen-specific CD8+ T immune cells (Manz & Di Santo, 2009). The transplantation
cells are thought to be important in preventing recurrence of of human hepatocytes into immunodeficient mice will allow for
hepatitis in transplant recipients via the production of IFN-γ large-scale screening therapeutics against human hepatitis
(Jo et al, 2009). viruses. Recently, an immunocompetent humanized murine
Although the hepatic immune system has been shown to HCV infection model has been developed (Chen et al, 2014).
clearly play an important role in liver and systemic tolerance, Chimeric murine models will enable researchers to study the
definitive mechanisms in humans remain elusive. It is likely that in vivo effects of human hepatotropic viruses on human immune
multiple suppressive cell types and several overlapping immu- responses to infection for the first time.
noinhibitory pathways interact to form a network of tolerance
within the intrahepatic space. Understanding the mechanisms Autoimmune Hepatitis
controlling the tolerogenic propensity of the liver is of great Autoimmune hepatitis (AIH) is an idiopathic disorder that
significance in transplantation, as it will permit the design of leads to cirrhosis. In contrast to the preceding sections that
novel approaches to reduce immunosuppressive-related mor- highlighted disease states associated with hepatic tolerance,
bidity and mortality. AIH is a manifestation of an overactive hepatic immune
system. AIH likely reflects an imbalance between proinflam-
Hepatitis matory immune responses and immunosuppressive factors,
The viruses hepatitis B (HBV) and hepatitis C (HCV) are the latter including PD-1/PD-L1, Treg, and MDSC. It has a
noncytopathic and hepatotropic, and cause acute and chronic female predominance and is characterized by elevated levels
liver disease (see Chapter 70). The worldwide burden of disease of immunoglobulin G (IgG) autoantibodies (Obermayer-
is immense, with an estimated 500 million people infected with Straub et al, 2000). An experimental model of autoimmune
either of these two viruses (World Health Organization, http:// hepatitis is based on the treatment of mice with concanavalin
A (ConA), a plant lectin known to activate T cells in vitro. A of postoperative allograft dysfunction and morbidity (Clavien
single intravenous injection of ConA leads to severe liver et al, 1992). During liver I/R, the release of endogenous mol-
damage and is associated with the activation of CD4+ T cells ecules signal danger to the host by activating innate immune
and the production of TNF-α and IFN-γ. Further studies cells through their interaction with Toll-like receptors (TLRs)
showed that NKT cells are the most important subset of (Lotze et al, 2007). Such “danger signals,” or danger-associated
CD4+ T cells involved in mediating ConA hepatitis. The molecular patterns (DAMPs), can be classified as intracellular
importance of NKT cells in mediating autoimmune liver proteins, nucleic acids, or components of the extracellular
damage also is supported by the observation that injection of matrix that are released following host cell injury (see Fig.
α-galactosylceramide, an activator of NKT cells through 10.7). The ensuing host innate immune response results in
CD1d binding, leads to a similar form of injury as that seen untoward collateral tissue damage that culminates in hepato-
in ConA hepatitis. AIH has also been linked with dysfunction cyte death and a systemic inflammatory response. The injury
of Tregs, leading to a hyperfunctional Th17 compartment and stems from the inability of TLRs on innate immune cells to
excessive inflammation (Grant et al, 2014). The pathogenesis distinguish infectious ligands from DAMPs released by autolo-
of AIH highlights the critical importance of balance between gous tissue. The need for effective approaches to manage
proinflammatory and antiinflammatory immune processes patients with I/R-induced organ damage is highlighted by the
within the liver. fact that current treatment is merely supportive care.
Using a murine model of segmental liver I/R, we reported
Primary Biliary Cirrhosis that hepatic DCs and TLR9 play critical roles in modulating
Primary biliary cirrhosis (PBC) is an autoimmune liver disease the host immune response during liver I/R (Bamboat et al,
that leads to the destruction of intrahepatic bile ducts and 2009b). Hepatocyte DNA released during ischemia binds
subsequent cholestasis and cirrhosis (Bergasa et al, 2004; TLR9 in a variety of liver immune cells that then cause liver
Lindor et al, 2009). Most cases of PBC are associated with damage. Using TLR9-knockout mice, we found that the global
antimitochondrial (95%) and antinuclear antibodies (50%) absence of TLR9 signaling confers protection from liver I/R
antibodies directed against self-antigens. Dysregulated CD4+ injury. Interestingly, the immune effects of TLR9 activation
helper T cells and CD8+ cytotoxic T lymphocytes are thought appear to be cell specific. In neutrophils, host DNA binds TLR9
to be important in the pathogenesis of PBC. This is supported and exacerbates injury through production of inflammatory
by the findings of lymphoid infiltration of the portal tracts and cytokines and reactive oxygen species (ROS). In contrast, DCs
aberrant expression of MHC class II on biliary epithelial cells. respond to host DNA by curtailing injury via production of
Because biliary epithelial cells are the primary target of injury antiinflammatory IL-10, which confers protection by suppress-
in PBC, their expression of MHC molecules and cytokines ing the function of inflammatory monocytes that migrate to the
likely plays an important role in the pathogenesis of this disease. liver from the bone marrow (see Fig 10.7). Human liver DCs
The degree of Th17 activity within the liver of PBC patients also have a propensity to secrete large amounts of IL-10 at
has been associated with severity of disease, and IL-17 pro- baseline and after TLR activation (Bamboat et al, 2009a). T
motes excessive fibrosis (Yang et al, 2014). In addition to lymphoctes also play a role in I/R tissue injury, with Th17 cells
pathogenic T-cell function, biliary epithelial cells from patients promoting influx of neutrophils (Kono et al, 2011). Improved
with primary biliary cirrhosis attract mononuclear cells via the understanding of the complex interactions between DAMPs
chemokine CX3CL-1 and overexpress the inflammatory cyto- and PRRs on immune cells will promote the development of
kines IL-6 and TNF-α (Shimoda et al, 2010). rational, more effective approaches to limit liver I/R injury.
Primary Sclerosing Cholangitis Obstructive Jaundice

Primary sclerosing cholangitis (PSC) results from fibrosis of the Obstructive jaundice leads to intrahepatic inflammation, and
intrahepatic and extrahepatic bile ducts (Lee & Jonas, 2007; jaundiced patients are at increased risk for complications due
see Chapter 41). The pathogenesis is thought to result from to immunologic derangements (Armstrong et al, 1984). Our
immune activation within the liver after bacteria gain access to group reported that bile duct ligation is associated with expan-
the portal circulation via a diseased intestinal epithelium. The sion of immunosuppressive liver Tregs, which inhibit T-cell
disease is associated with T cells infiltrating the portal tracts. responses to stimulation by DCs (Katz et al, 2011). Subse-
In addition, an increase in production of circulating proinflam- quently, we reported that suppression of liver T-cell function
matory cytokines has been shown to correlate with disease in the setting of bile duct ligation is also due to increased
progression. Pathologic Th17 cell activation in response to expression of PD-1. As noted above, PD-1 is an immunoinhibi-
bacteria and PRR activity contributes to inflammation in PSC tory molecule known to suppress T-cell function. Inhibition of
patients (Katt et al, 2014). The end result is biliary cirrhosis PD-1 is an attractive strategy for enhancing liver T-cell function
and a markedly elevated risk for cholangiocarcinoma. Most as PD-1 also mediates paradoxic proinflammatory effects in the
cases of PSC are associated with underlying inflammatory liver (Licata et al, 2013).
bowel disease, specifically ulcerative colitis. In contrast to most
autoimmune diseases, which predominantly affect women, Partial Hepatectomy
PSC affects men twice as often as it does women. Unlike most other mammalian organs, the liver has the capacity
to regenerate after resection or traumatic injury (see Chapters
Ischemia/Reperfusion Injury 6, 103, and 108C). This unique property allows for partial liver
Liver ischemia/reperfusion (I/R) injury is a well-recognized transplantation from live donors and major hepatic resections
consequence of trauma, circulatory shock, partial hepatectomy, for malignancy. The early response to partial hepatectomy (PH)
and liver transplantation. It contributes to the shortage of involves the activation of NF-κB in liver NPCs and hepatocytes,
organs available for transplantation and is a major determinant which causes additional injury through the production of ROS,
TNF-α, and other inflammatory cytokines. During time, IL-6 understanding the relationship between the hepatic immune
and IL-10 production increases, favoring a more antiinflamma- system and the parasite. Eggs that are laid by adult worms in
tory state that limits immune-mediated damage and promotes the portal vein become trapped within the hepatic sinusoids and
hepatocyte regeneration. Liver DCs have been reported to trigger an immune response that results in granuloma forma-
expand following PH in mouse models (Castellaneta et al, tion. The early immune response mediated by eosinophils and
2006). After PH, liver DCs produce IL-10 and promote the neutrophils is proinflammatory. Later, the cytokine profile
development of antiinflammatory T cells. These data suggest within the liver becomes antiinflammatory and continues for
that inherent tolerogenic properties of hepatic DCs may serve the course of disease, ultimately resulting in liver fibrosis (Burke
to promote hepatocyte regeneration after hepatectomy. et al, 2009). KCs, IL-4, and IL-13 are thought to play a major
Like liver I/R, PH results in the release of a variety of danger role in skewing the immune response during the course of
signals that bind PRRs on innate immune cells, triggering a schistosomiasis infection (Hayashi et al, 1999).
cascade of signaling events that modulate the rate of parenchy-
mal regeneration. Deficiency in the intracellular signaling mol- IMMUNE SYSTEM IN MALIGNANT
ecule myeloid differentiation primary response 88 (MyD88), LIVER DISEASES
which acts downstream of almost all TLRs (except TLR3),
results in poorer outcome in mouse models of PH (Seki et al, Evidence on the importance of the hepatic immune system in
2005). In contrast, the absence of TLR3, which signals in a malignancy comes from data which reveal that (1) when com-
MyD88-independent manner, or blocking signaling of another pared with autologous blood, human liver specimens from
PRR, the receptor for advanced glycation end products patients with malignancy contain a disproportionately higher
(RAGE), promotes liver regeneration and survival after massive percentage of NK cells with reduced antitumor function (Burt
hepatectomy (Cataldegirmen et al, 2005; Zorde-Khvalevsky et al, 2009) and (2) the hepatic T-cell infiltrate predicts survival
et al, 2009). These data suggest that a balance exists between following surgery in patients with metastatic colorectal liver
various TLRs and RAGE that governs the net effect of partial cancer (Katz et al, 2009). Tumor-infiltrating lymphocytes
hepatectomy. Identification of mechanisms that limit regenera- (TILs) indicate a specific host response to tumor antigens.
tion after massive injury holds the key to expanding the limits TILs may be utilized with therapeutic intent or studied for
of liver transplantation and salvaging livers and hosts over- prognostic information (Rosenberg et al, 1986). TILs have
whelmed by carcinoma and toxic insults. been demonstrated to predict outcomes in a wide variety of
solid tumors, with the magnitude of this effect being dependent
Drug-Induced Liver Disease on tumor site and disease stage (Gooden et al, 2011). An
The prevalence of liver failure as a result of drug-induced toxic- increasing number of studies have focused on TILs as predic-
ity is increasing at an alarming rate and is the most common tors of outcome for primary and metastatic liver tumors. Our
cause of acute liver failure in the developed world (Lee, 2007). evolving understanding of intrahepatic antitumor immunity
The injury is characterized by toxin-induced hepatocyte death promises to enable development of novel therapeutic approaches
and activation of the innate immune system. This triggers a using immunomodulatory agents and adoptive cell therapy.
cascade of proinflammatory signals that result in further hepa-
tocyte death and, ultimately, in liver failure. Through the use Immune Response to Primary Liver Cancer
of mouse models of acetaminophen-induced liver injury, the Hepatocellular carcinoma (HCC) is one of the most common
roles of various components of the hepatic immune system have malignant tumors worldwide. The immune response to neo-
been well characterized. LSECs are thought to mediate liver plastic cells has been shown to be a potential determinant of
injury by detecting hepatocyte DNA via TLR9, resulting in survival in HCC patients (Fu et al, 2007; Gao et al, 2007;
production of pro–IL-1 and pro–IL-18, which are cleaved into Sasaki et al, 2008). High densities of DCs, T-cell subsets, and
their active forms by an intracellular protein complex called the tumor-associated macrophages have all been reported to be
inflammasome (Imaeda et al, 2009). Widespread inflammation favorable prognostic factors in HCC (Cai et al, 2005). TIL
ensues, resulting in additional parenchymal injury. responses to HCC occur in a more complex biologic context
than metastatic tumors given the underlying inflammatory pro-
Bacterial and Parasitic Liver Disease cesses that drive HCC carcinogenesis and progression.
Intrahepatic immunosuppression not only appears to be HCC TILs have been compared with T cells in adjacent
involved in tolerance to liver allografts but may also play a role normal liver tissue, and HCC TILs reflect a more immunosup-
in the vulnerability of the liver to various bacterial and parasitic pressed state. Mathai and coworkers (2012) demonstrated that
infections (see Chapters 12, 45, 72, 73, and 74). The liver is HCC TIL populations have a higher number of Tregs com-
susceptible to pyogenic abscesses, particularly in patients with pared with normal surrounding liver. Increased ratios of Tregs
malignant biliary tract disease undergoing complex endoscopic, (FOXP3+) to total T cells (CD3+) were independently associ-
percutaneous, and surgical interventions (Mezhir et al, 2010). ated with poorer differentiation. Furthermore, high FOXP3+ to
The host response to bacteria within the liver results in forma- CD8+ TIL ratios were associated with shorter overall and
tion of an abscess in an effort to curtail the spread of infection. disease-free survival times in HCC patients on both univariate
Neutrophils are recruited to the site of infection via chemotactic and multivariate analysis (Mathai et al, 2012). Immune sup-
signals released by the invading bacteria and activation of compression within HCC may also be attributed to increased
plement. Neutrophils then release cytokines and reactive expression of PD-1 on TILs. As noted earlier, PD-1 is an
oxygen species that promote inflammation and kill bacteria. immunoinhibitory receptor that suppresses T-cell division and
Helminthic infections of the liver, such as schistosomiasis, cytokine production (Wu et al, 2009). PD-1 is engaged by
echinococcosis, and ascariasis have a hepatic phase in their life PD-L1, which has been detected on HCC tumor cells. The
cycles. Animal models of schistosomiasis have helped in presence of Tregs and PD-1+ TILs is reflective of an
immunosuppressive milieu, but may represent a therapeutic outcome following resection of neuroendocrine tumor liver
opportunity as anti–PD-1 antibodies are now in clinical use for metastases (Katz et al, 2010). Tumor MHC I expression has
a variety of solid tumors. also been correlated to immune infiltrates and outcome (Tur-
TIL responses have also been studied in patients with biliary cotte et al, 2014).
tract cancers. Cholangiocarcinoma (CCA) is the second most A critical finding from studies of CRLM is that although
common primary intrahepatic malignancy, and most patients patients with favorable TIL profiles are more likely to have a
are not candidates for curative surgical intervention. A study of good outcome, the majority of individuals do not demonstrate
123 patients with intrahepatic CCA who underwent curative an effective intratumoral immune response. We speculate that
surgical resection revealed that IL-17+ (Th17) and FOXP3+ immunosuppressive intrahepatic immune cells and immunoin-
(Treg) TILs were enriched predominantly within CCA tumors, hibitory pathways limit the function of effector T cells. This
whereas CD8+ TILs were most abundant at the tumor margin. presents a therapeutic opportunity to deliver effective adoptive
IL-17+ TIL counts were an independent predictor of decreased cellular immunotherapy in conjunction with suppressive
patient survival, reflecting the importance of the Th17/Treg pathway inhibition to overcome the factors curtailing endoge-
balance in liver diseases (Gu et al, 2012). Similar to HCC, nous antitumor immunity.
tumor expression of PD-L1 was correlated with CCA stage and
tumor differentiation (Ye et al, 2009). Improving our under- Immunotherapy for Liver Cancer
standing of CCA immune responses and tumor-driven immu- Although many investigators have attempted to manipulate the
nosuppression may offer insights into novel therapeutic options immune system for the treatment of cancer (Hunder et al,
and improve risk stratification. 2008), few attempts have been made to directly target intrahe-
patic immune cells. The primary goal of cancer immunother-
Immune Response to Metastatic Liver Cancer apy, particularly for liver tumors, is to deliver or induce potent
The high prevalence of metastatic disease to the liver is likely antitumor immunity while reversing intrahepatic suppression.
due to multiple factors. Evidence suggests that particular Reversal of intrahepatic tolerance has been accomplished in
characteristics of the liver immune system may play an impor- animal models by activating liver immune cells, depleting sup-
tant role in this propensity. As delineated earlier, the primary pressor cells, or blocking immunoinhibitory pathways such as
mediators of innate and adaptive immunity found in the liver, the PD-1/PD-L1 axis. We previously demonstrated the impor-
DCs, and T cells, are unique in their distribution of subtypes tance of IL-12 production by DCs in activating NK cells toward
and skew toward tolerogenic function (Katz et al, 2005; Pil- protection against tumor in a mouse melanoma liver metastasis
larisetty et al, 2004). In addition, the liver contains suppres- model (Miller et al, 2002b). Despite having less lytic potential
sive cells such as Tregs and MDSCs that may also promote than blood NK cells, we found that human liver NK cells do
the establishment of metastatic colonies within the intrahe- have the capacity to become potent antitumor cells when acti-
patic space. vated in the presence of KCs and TLR3 ligands (Burt et al,
Several studies suggest that intrahepatic adaptive immune 2009).
responses to colorectal liver metastases (CRLM) are predictive In addition to DCs, recent work on the role of peritumoral
of recurrence and death following resection. Both CD4+ and monocytes in patients with HCC has garnered much attention
CD8+ T cells are activated within CRLM, and activated CD4+ among immunologists and hepatologists alike. Activated mono-
T helper cells may promote tumor-selective activity of cytotoxic cytes in the peritumoral stroma of HCC have been shown to
CD8+ T lymphocytes (Wagner et al, 2008). Following hepatic foster a state of tolerance and promote tumor progression via
resection, patients with high numbers of CD8+ intratumoral T the expression of an immunosuppressive molecule, PD-L1
cells were more likely to survive 10 years or longer. Among (Kuang et al, 2009). In addition, tumor expression of PD-L1
patients who survived for 10 years or more, 31% had high levels was recently shown to serve as a predictor of recurrence in
of CD8 T cells. In contrast, only 8% of those who survived less patients with resected HCC (Gao et al, 2009). Blocking anti-
than 2 years had a high level of CD8+ TIL (Katz et al, 2009). bodies to molecules such as PD-1 and PD-L1 are currently
Ratios of TIL subsets have been studied as well and may approved for human use.
provide more biologic insight than individual TIL counts. In As described above, most patients fail to mount effective
patients with CRLM, high CD8+ to CD4+ TIL ratios were an antitumor immunity, likely due to the immunosuppressive
independent predictor of long-term survival following resection nature of the intrahepatic space. Genetically modified chime-
after adjusting for multiple variables, including clinical risk ric antigen receptor T cells (CAR-Ts) offer a potential solu-
score. In a more recent study, high CD4+ to CD3+ and CD8+ tion to providing effective antitumor immunity for CRLM.
to CD3+ TIL ratios were a significant correlate of improved and CAR-Ts are produced from patient autologous T cells, using
recurrence-free survival (Katz et al, 2013). a retroviral system to engineer expression of an immune
As reported in the setting of primary liver cancer, TIL ratios, receptor that is activated upon engagement of tumor antigens
including FOXP3+ Treg, are predictive of outcome following such as carcinoembryonic antigen (CEA). In our recent phase
CRLM resection. High FOXP3+ to CD4+ and FOXP3+ to I Hepatic Immunotherapy for Metastases (HITM) trial, we
CD8+ TIL ratios were independent predictors of shorter overall demonstrated the safety and encouraging signals of clinical
survival. Overall survival at 5 years for patients with a high efficacy of hepatic artery anti-CEA CAR-T infusions in
FOXP3+ to CD4+ TIL ratio was 34% compared with 51% for patients with CRLM (Saied et al, 2014). Future HITM trials
patients with a low ratio (odds ratio = 1.6; P = .03). Similarly, will combine regional CAR-T infusions with immunomodula-
5 year survival was 35% in those with a high FOXP3+ to CD8+ tory agents designed to overcome intrahepatic suppression. It
ratio compared with 46% in those with a low FOXP3+ to CD8+ is likely that innovative combinatorial immunotherapeutic
TIL ratio (hazard ratio = 1.5; P = .05) (Katz et al, 2013). approaches will achieve greater clinical success than single-
FOXP3+ TIL counts were also found to be predictors of agent strategies.
SUMMARY understanding of the liver’s immune system and how it relates

to the immune system as a whole continues to evolve. Deepen-
The liver contains a diverse population of immune cells that ing our insight into liver immune cell biology and immunoin-
demonstrate a strong tendency toward tolerance. Although the hibitory pathway signaling will provide exciting opportunities
propensity toward tolerance is advantageous when responding for manipulating the intrahepatic immune system for treatment
to oral antigens and in the setting of transplantation, intrahe- of benign and malignant conditions.
patic immunosuppression may be exploited by pathogens and
malignant cells to evade detection and destruction. Our References are available at expertconsult.com.
Chapter 10 Liver immunology 187.e1
Gerlach JT, et al: Recurrence of hepatitis C virus after loss of virus-

REFERENCES specific CD4(+) T cell response in acute hepatitis C, Gastroenterology
117:933–941, 1999.
Armstrong CP, et al: Wound healing in obstructive jaundice, Br J Surg Gooden MJ, et al: The prognostic influence of tumour-infiltrating lym-
71:267–270, 1984. phocytes in cancer: a systematic review with meta-analysis, Br J
Bamboat ZM, et al: Human liver dendritic cells promote T cell hypo- Cancer 105:93–103, 2011.
responsiveness, J Immunol 182:1901–1911, 2009a. Grant CR, et al: Dysfunctional CD39(POS) regulatory T cells and
Bamboat ZM, et al: Toll-like receptor 9 inhibition confers protection aberrant control of T-helper type 17 cells in autoimmune hepatitis,
from liver ischemia-reperfusion injury, Hepatology 51:621–632, Hepatology 59:1007–1015, 2014.
2009b. Gu F-M, et al: Intratumoral IL-17+ cells and neutrophils show strong
Bergasa NV, et al: Primary biliary cirrhosis: report of a focus study prognostic significance in intrahepatic cholangiocarcinoma, Ann
group, Hepatology 40:1013–1020, 2004. Surg Oncol 19:2506–2514, 2012.
Bleier JI, et al: Increased and long-term generation of dendritic cells Hayashi N, et al: Kupffer cells from Schistosoma mansoni-infected mice
with reduced function from IL-6-deficient bone marrow, J Immunol participate in the prompt type 2 differentiation of hepatic T cells in
172:7408–7416, 2004. response to worm antigens, J Immunol 163:6702–6711, 1999.
Bosma BM, et al: Migration of allosensitizing donor myeloid dendritic Hunder NN, et al: Treatment of metastatic melanoma with autologous
cells into recipients after liver transplantation, Liver Transpl 16:12– CD4+ T cells against NY-ESO-1, N Engl J Med 358:2698–2703,
22, 2010. 2008.
Brahmer JR, et al: Safety and activity of anti-PD-L1 antibody in Imaeda AB, et al: Acetaminophen-induced hepatotoxicity in mice is
patients with advanced cancer, N Engl J Med 366:2455–2465, 2012. dependent on Tlr9 and the Nalp3 inflammasome, J Clin Invest
Breous ES, et al: Hepatic regulatory T cells and Kupffer cells are 119:305–314, 2009.
crucial mediators of systemic T cell tolerance to antigens targeting Janeway CA Jr: How the immune system works to protect the host from
murine liver, Hepatology 50:612–621, 2009. infection: a personal view, Proc Natl Acad Sci U S A 98:7461–7468,
Brown RS: Hepatitis C and liver transplantation, Nature 436:973–978, 2001.
2005. Jo J, et al: Analysis of CD8+ T cell-mediated inhibition of hepatitis C
Burke ML, et al: Immunopathogenesis of human schistosomiasis, virus replication using a novel immunological model, Gastroenterol-
Parasite Immunol 31:163–176, 2009. ogy 136:1391–1401, 2009.
Burt BM, et al: The lytic potential of human liver NK cells is restricted Kamada N, et al: Reversal of transplantation immunity by liver graft-
by their limited expression of inhibitory killer Ig-like receptors, J ing, Nature 292:840–842, 1981.
Immunol 183:1789–1796, 2009. Katt J, et al: Increased T helper type 17 response to pathogen stimula-
Cai XY, et al: Relationship between dendritic cells and memory T tion in patients with primary sclerosing cholangitis, Hepatology
lymphocytes in tumor site and prognosis of hepatocellular carci- 58:1084–1093, 2014.
noma, Zhonghua Yi Xue Za Zhi 85:671–675, 2005. Katz SC, et al: Liver sinusoidal endothelial cells are insufficient to
Callery MP, et al: Kupffer cell blockade inhibits induction of tolerance activate T cells, J Immunol 173:230–235, 2004.
by the portal venous route, Transplantation 47:1092–1094, 1989. Katz SC, et al: Conventional liver CD4 T cells are functionally distinct
Castellaneta AA, et al: Functional modification of CD11c+ liver den- and suppressed by environmental factors, Hepatology 42:293–300,
dritic cells during liver regeneration after partial hepatectomy in 2005.
mice, Hepatology 43:807–816, 2006. Katz SC, et al: T cell infiltrate predicts long-term survival following
Cataldegirmen G, et al: RAGE limits regeneration after massive liver resection of colorectal cancer liver metastases, Ann Surg Oncol
injury by coordinated suppression of TNF-alpha and NF-kappaB, 16:2524–2530, 2009.
J Exp Med 201:473–484, 2005. Katz SC, et al: T cell infiltrate and outcome following resection of
Charles R, et al: Human hepatic stellate cells inhibit T-cell response intermediate-grade primary neuroendocrine tumours and liver
through B7-H1 pathway, Transplantation 96:17–24, 2013. metastases, HPB 12:674–683, 2010.
Chen J, et al: Persistent hepatitis C virus infections and hepatopatho- Katz SC, et al: Obstructive jaundice expands intrahepatic regulatory T
logical manifestations in immune-competent humanized mice, Cell cells, which impair liver T lymphocyte function but modulate liver
Res 24:1050–1066, 2014. cholestasis and fibrosis, J Immunol 187:1150–1156, 2011.
Clavien PA, et al: Preservation and reperfusion injuries in liver Katz SC, et al: Regulatory T Cell infiltration predicts outcome follow-
allografts: an overview and synthesis of current studies, Transplanta- ing resection of colorectal cancer liver metastases, Ann Surg Oncol
tion 53:957–978, 1992. 20:946–955, 2013.
Creput C, et al: Incidence of renal and liver rejection and patient Knolle PA, Limmer A: Neighborhood politics: the immunoregulatory
survival rate following combined liver and kidney transplantation, function of organ-resident liver endothelial cells, Trends Immunol
Am J Transplant 3:348–356, 2003. 22:432–437, 2001.
Crispe IN: Hepatic T cells and liver tolerance, Nat Rev Immunol 3:51– Kono H, et al: Role of IL-17A in neutrophil recruitment and hepatic
62, 2003. injury after warm ischemia-reperfusion mice, J Immunol 187:4818–
Curiel TJ, et al: Blockade of B7-H1 improves myeloid dendritic cell- 4825, 2011.
mediated antitumor immunity, Nat Med 9:562–567, 2003. Kuang DM, et al: Activated monocytes in peritumoral stroma of hepa-
Dangi A, et al: Selective expansion of allogeneic regulatory T cells by tocellular carcinoma foster immune privilege and disease progression
hepatic stellate cells: role of endotoxin and implications for allograft through PD-L1, J Exp Med 206:1327–1337, 2009.
tolerance, J Immunol 188:3667–3677, 2012. Lau AH, Thomson AW: Dendritic cells and immune regulation in the
Ferrari C, et al: Immunopathogenesis of hepatitis C virus infection, liver, Gut 52:307–314, 2003.
J Hepatol 31(Suppl 1):31–38, 1999. Lee WM: Acetaminophen toxicity: changing perceptions on a social/
Finn OJ: Cancer immunology, N Engl J Med 358:2704–2715, 2008. medical issue, Hepatology 46:966–970, 2007.
Fu J, et al: Increased regulatory T cells correlate with CD8 T cell Lee CK, Jonas MM: Pediatric hepatobiliary disease, Curr Opin Gastro-
impairment and poor survival in hepatocellular carcinoma patients, enterol 23:306–309, 2007.
Gastroenterology 132:2328–2339, 2007. Li W, et al: Role of the liver in peripheral tolerance: induction through
Fuller MJ, et al: Immunotherapy of chronic hepatitis C virus infection oral antigen feeding, Am J Transplant 4:1574–1582, 2004.
with antibodies against programmed cell death-1 (PD-1), Proc Natl Licata L, et al: Biliary obstruction results in PD-1-dependent liver T
Acad Sci U S A 110:15001–15006, 2013. cell dysfunction and acute inflammation mediated by Th17 cells and
Gao Y, et al: Gamma delta T cells provide an early source of interferon neutrophils, J Leukoc Biol 94:813–823, 2013.
gamma in tumor immunity, J Exp Med 198:433–442, 2003. Lindor KD, et al: Primary biliary cirrhosis, Hepatology 50:291–308,
Gao Q, et al: Intratumoral balance of regulatory and cytotoxic T cells 2009.
is associated with prognosis of hepatocellular carcinoma after resec- Lotze MT, et al: The grateful dead: damage-associated molecular
tion, J Clin Oncol 25:2586–2593, 2007. pattern molecules and reduction/oxidation regulate immunity,
Gao Q, et al: Overexpression of PD-L1 significantly associates with Immunol Rev 220:60–81, 2007.
tumor aggressiveness and postoperative recurrence in human hepa- Lu L, et al: Propagation of dendritic cell progenitors from normal
tocellular carcinoma, Clin Cancer Res 15:971–979, 2009. mouse liver using granulocyte/macrophage colony-stimulating factor
and their matu-rational development in the presence of type-1 col- Saied A, et al: Neutrophil:lymphocyte ratios and serum cytokine
lagen, J Exp Med 179:1823–1834, 1994. changes after hepatic artery chimeric antigen receptor-modified
Lu L, et al: A novel subset of dendritic cells propagated from the liver T-cell infusions for liver metastases, Cancer Gene Ther 21:457–462,
promotes differentiation of T regulatory cells and enhances allograft 2014.
survival, Transplant Proc 33:229, 2001. Sasaki A, et al: Prognostic value of tumor-infiltrating FOXP3+ regula-
Maini MK, et al: The role of virus-specific CD8(+) cells in liver damage tory T cells in patients with hepatocellular carcinoma, Eur J Surg
and viral control during persistent hepatitis B virus infection, J Exp Oncol 34:173–179, 2008.
Med 191:1269–1280, 2000. Schroder K, et al: Interferon-gamma: an overview of signals, mecha-
Manz MG, Di Santo JP: Renaissance for mouse models of human nisms and functions, J Leukoc Biol 75:163–189, 2004.
hematopoiesis and immunobiology, Nat Immunol 10:1039–1042, Seki E, et al: Contribution of Toll-like receptor/myeloid differentiation
2009. factor 88 signaling to murine liver regeneration, Hepatology 41:443–
Mathai AM, et al: Role of Foxp3-positive tumor-infiltrating lympho- 450, 2005.
cytes in the histologic features and clinical outcomes of hepatocel- Shevach EM: CD4+ CD25+ suppressor T cells: more questions than
lular carcinoma, Am J Surg Pathol 36:980–986, 2012. answers, Nat Rev Immunol 2:389–400, 2002.
Mazariegos GV, et al: Weaning of immunosuppression in liver trans- Shimoda S, et al: CX3CL1 (fractalkine): a signpost for biliary inflam-
plant recipients, Transplantation 63:243–249, 1997. mation in primary biliary cirrhosis, Hepatology 51:567–575, 2010.
Mezhir JJ, et al: Current management of pyogenic liver abscess: surgery Siegal FP, et al: The nature of the principal type 1 interferon-producing
is now second-line treatment, J Am Coll Surg 210:975–983, 2010. cells in human blood, Science 284:1835–1837, 1999.
Miller G, et al: Endogenous granulocyte-macrophage colony- Sriwatanawongsa V, et al: The essential roles of parenchymal tissues
stimulating factor overexpression in vivo results in the long-term and passenger leukocytes in the tolerance induced by liver grafting
recruitment of a distinct dendritic cell population with enhanced in rats, Nat Med 1:428–432, 1995.
immunostimulatory function, J Immunol 169:2875–2885, 2002a. Starzl TE, et al: Cell migration, chimerism, and graft acceptance,
Miller G, et al: Adenovirus infection enhances dendritic cell immuno- Lancet 339:1579–1582, 1992.
stimulatory properties and induces natural killer and T cell-mediated Starzl TE, et al: Liver transplants contribute to their own success, Nat
tumor protection, Cancer Res 62:5260–5266, 2002b. Med 2:163–165, 1996.
Nakamura I, et al: Pathogenesis of experimental neonatal woodchuck Sun Z, et al: Hepatic allograft-derived Kupffer cells regulate T cell
hepatitis virus infection: chronicity as an outcome of infection is response in rats, Liver Transpl 9:489–497, 2003.
associated with a diminished acute hepatitis that is temporally defi- Suzuki G, et al: Antigen-induced suppression of the proliferative
cient for the expression of interferon gamma and tumor necrosis response of T cell clones, J Immunol 140:1359–1365, 1988.
factor-alpha messenger RNAs, Hepatology 33:439–447, 2001. Teng M, et al: Combined natural killer t-cell based immunotherapy
Novobrantseva TI, et al: Attenuated liver fibrosis in the absence of B eradicates established tumors in mice, Cancer Res 67(15):7495–
cells, J Clin Invest 115:3072–3082, 2005. 7504, 2007.
Obermayer-Straub P, et al: Autoimmune hepatitis, J Hepatol 32(Suppl Thorn M, et al: Liver metastases induce reversible hepatic B cell dys-
1):181–197, 2000. function mediated by Gr-1+CD11b+ myeloid cells, J Leukoc Biol
Ohira M, et al: Adoptive immunotherapy with liver allograft-derived 96:883–894, 2014.
lymphocytes induces anti-HCV activity after liver transplantation in Turcotte S, et al: Tumor MHC class I expression improves the prog-
humans and humanized mice, J Clin Invest 119:3226–3235, 2009. nostic value of T cell density in resected colorectal liver metastases,
Pillarisetty VG, et al: GM-CSF expands dendritic cells and their pro- Cancer Immunol Res 2:530–537, 2014.
genitors in mouse liver, Hepatology 37:641–652, 2003. Wagner P, et al: Detection and functional analysis of tumor infiltrating
Pillarisetty VG, et al: Liver dendritic cells are less immunogenic than T-lymphocytes (TIL) in liver metastases from colorectal cancer, Ann
spleen dendritic cells because of differences in subtype composition, Surg Oncol 15:2310–2317, 2008.
J Immunol 172:1009–1017, 2004. Wu K, et al: Kupffer cell suppression of CD8+ T cells in human hepa-
Plitas G, et al: Dendritic cells are required for effective cross- tocellular carcinoma is mediated by B7-H1/programmed death-1
presentation in the murine liver, Hepatology 47:1343–1351, 2008. interactions, Cancer Res 69:8067–8075, 2009.
Qian SG, et al: Orthotopic liver transplantation in the mouse, Trans- Yang CY, et al: IL-12/Th1 and IL-23/Th17 biliary microenvironment
plantation 52:562–564, 1991. in primary biliary cirrhosis: implications for therapy, Hepatology
Qian S, et al: Murine liver allograft transplantation: tolerance and 59:1944–1953, 2014.
donor cell chimerism, Hepatology 19:916–924, 1994. Ye Y, et al: Interaction of B7-H1 on intrahepatic cholangiocarcinoma
Rasmussen A, et al: Combined transplantation of liver and kidney from cells with PD-1 on tumor-infiltrating T cells as a mechanism of
the same donor protects the kidney from rejection and improves immune evasion, J Surg Oncol 100:500–504, 2009.
kidney graft survival, Transplantation 59:919–921, 1995. You Q, et al: Mechanism of T cell tolerance induction by murine
Rosenberg SA, et al: A new approach to the adoptive immunotherapy hepatic Kupffer cells, Hepatology 48:978–990, 2008.
of cancer with tumor-infiltrating lymphocytes, Science 233:1318– Zorde-Khvalevsky E, et al: Toll-like receptor 3 signaling attenuates
1321, 1986. liver regeneration, Hepatology 50:198–206, 2009.
CHAPTER 11
Cytokines in liver, biliary, and pancreatic disease
Jason Maggi and George Miller
OVERVIEW and continue to be investigated as potential therapeutic targets

(Table 11.1).
Cytokines are small proteins that play a primary role in the
human response to a variety of stimuli. These nonstructural Endotoxin and the Immune Response
proteins have an effect directly on cells and in the communica- Exclusive to gram-negative bacteria, lipopolysaccharide (LPS) is
tion between cells. Pleiotropic in nature, many cytokines can one of the most studied pathogen-associated activators of inflam-
produce both proinflammatory and antiinflammatory effects, mation. An activator of proinflammatory cytokines, LPS plays a
depending upon the timing, levels, and context in which they large role in tissue damage and development of septic shock,
are secreted. The liver possesses extensive immunologic func- classifying it as an endotoxin (Dinarello, 2004). Although the
tion, and considerable research has implicated cytokines as exact mechanism of LPS pathogenesis in human disease remains
crucial mediators in the development of hepatic diseases, as undefined, its role in activation of the host innate immune
well as regeneration and repair. Although not as well defined, response is widely known (Table 11.2). Additionally, elucidation
it would correspond that pancreatic pathophysiology derives of the molecular signaling process initiated by LPS continues,
from a similar relationship. Elucidation of the mechanisms of opening potential areas for targeted treatment (Fig 11.1).
these mediators allows an enhanced understanding of the Endotoxins occupy much of the outer leaflet of the outer
natural history of liver, biliary, and pancreatic surgical diseases. membrane of gram-negative bacteria. These endotoxins are
Although produced by all nucleated cells, constitutive produc- composed of three major components: a conserved lipid A
tion of cytokines is all but absent without noxious stimuli. This region, a core oligosaccharide, and an outer O-specific oligosac-
chapter reviews these stimuli, as well as the specific mediators charide chain, which is specific to each bacterial strain and
involved, in the pathophysiology of liver, biliary, and pancreatic elicits host production of different antibodies. The lipid A
disease. Specific attention is paid to surgical diseases, as well moiety produces the inflammatory biologic effects of LPS
as potential therapeutic targets. through its interactions with host serum proteins and cellular
receptors (Ulevitch & Tobias, 1999). Although not intrinsically
CYTOKINES, STIMULI, AND SIGNALING harmful itself, the combined interaction of these LPS compo-
nents with host proteins and lipoproteins produces the proin-
Immune Recognition and Toll-Like Receptors flammatory cytokines that give rise to symptoms seen in the
Immune response to microbial pathogens and biologic insults acute septic response (Gioannini & Weiss, 2007).
requires quick identification of the threat to minimize damage The host innate immune recognition and response to LPS
to the host. The innate immune system is a first-line defense is highly ordered and is crucial to preventing overwhelming
against microbial pathogens, providing the ability to distinguish infection and resultant sepsis. Although it is crucial that the
self from nonself via cell-surface receptors and antigen- high sensitivity of the system allows rapid mobilization of host
presenting cells. Additionally, central to this defense mecha- protective mechanisms, self-limitation of this response is just as
nism is the ability to recognize cellular damage. vital, preventing irreversible tissue damage and allowing return
The liver, biliary system, and pancreas play a vital role in to homeostasis. Extensive study has recognized a strictly
development of immunologic responses because of their close ordered sequence of interactions with multiple extracellular and
physiologic relationship with the gut. Portal return of blood to cell-surface host proteins, including LPS-binding protein
the liver contains a tremendous amount of antigens and micro- (LBP), cluster of differentiation (CD)14, myeloid differentia-
bial products, which functions to shape immunologic tolerance tion protein-2 (MD-2), and Toll-like receptor 4 (TLR4)
as well as an organized response of inflammatory mediators (Gioannini & Weiss, 2007; Miyake, 2007).
(Carvalho et al, 2012). Recognition of cellular damage and Initiation of the host innate immune response to LPS is
microbial products is accomplished via intracellular and cell- binding of it to LBP to form the LPS-LBP complex (Thomas
surface–expressed pattern recognition receptors (PRRs) et al, 2002). The importance of LBP in beginning this inflam-
(Michelsen & Arditi, 2007). matory cascade is noted by the rise in LBP ribonucleic acid
An example of essential PRRs in the liver is the family of (RNA) and protein synthesis during the initial acute phase of
Toll-like receptors (TLRs). TLRs are a family of at least 10 endotoxic shock (Hallatschek et al, 2004). LBP, a 58 kDa gly-
transmembrane receptors found in cells throughout the human coprotein secreted by hepatocytes into the bloodstream, binds
body. Although essential for protective immunity in the normal the lipid A portion of LPS with high affinity and facilitates
state, aberrant or prolonged responses can produce catastrophic transfer to, and the subsequent effects of, CD14-positive cells.
effects on the host. Accordingly, TLRs have been demonstrated Although several cellular receptors have been found to
to be essential in the development of multiple hepatic diseases bind LPS, the CD14 molecule functions specifically to facilitate
188
Chapter 11 Cytokines in liver, biliary, and pancreatic disease 189
TABLE 11.1 Toll-Like Receptors in Hepatic, Biliary, and Pancreatic Disease

Receptor Known Ligands Location Associated Disease
TLR2 Peptidoglycan Extracellular Acute pancreatitis

Hemeglutinin Pancreatic cancer
Porins Ischemia/reperfusion
TLR3 DS-DNA (viral) Intracellular Chronic pancreatitis
Pancreatic cancer
Hepatitis B
Hepatitis C
Hepatocellular carcinoma
TLR4 Lipopolysaccharide Extracellular Acute pancreatitis
Nonalcoholic steatohepatosis
Cirrhosis
Hepatic fibrosis
Ischemia/reperfusion
TLR7/8 SS-RNA (viral) Intracellular Hepatitis B
RNA-immunoglobulin complexes
TLR9 DNA–immunoglobulin complexes Intracellular Acute pancreatitis
Unmethylated “CpG” DNA Pancreatic cancer
Hepatitis B
Hepatitis C
Hepatic fibrosis
Cirrhosis
CpG, Cytosine-guanine dinucleotide; DS, double stranded; SS, single stranded; TLR, Toll-like receptor.
that purified endotoxin aggregates alone cannot produce the

TABLE 11.2 Host Defense Responses to Lipopolysaccharide
same cellular activation effect, even at doses 1000 times higher
and Endotoxemia
than those required with the CD14 complex (Weiss, 2003).
Acute-Phase Response Host Defense The initial signal-transducing component of this pathway is
TLR4. Although expressed at low levels in healthy hepatic cells,
Thermoregulation Elevated core temperature
Antimicrobial response increasing evidence points to altered LPS-TLR4 complex sig-
Immunologic Complement activation naling in the pathogenesis of chronic liver disease. This complex,
Leukocytosis (neutrophilia) with the addition of the extracellular protein MD-2, functions
Proinflammatory response (IL-1α, to catalyze the intracellular signaling cascade (Nagai et al,
IL-1B, TNF-α, IL-1R, TGF-β, 2002). Although MD-2 can directly bind to LPS in the absence
protease inhibitors)
of TLR4, it appears that TLR4 stimulation, and the subsequent
B-cell stimulation, antibody production
intracellular signaling cascade, can only be achieved by their
Coagulation Decreased protein C
Decreased antithrombin III coordinated binding of the LPS-LBP complex. TLR4 intracel-
Increased cell adhesion lular signaling proceeds through the recruitment of adaptor
Activation of coagulation cascade molecules interacting through the Toll-like and interleukin
Increased tissue factor production (IL)-1 receptor (TIR) domain. Myeloid differentiation primary
Prostaglandin production
response 88 (MyD88), a signal transducer of the MyD88-
Platelet aggregation, platelet-activating
factor dependent interleukin-1 receptor (IL-1R) pathway, is essential
Fibrinogen for LPS-induced proinflammatory cytokine production. This
Metabolic Lipolysis has been demonstrated in the complete deficit of hepatic acute-
Mobilized amino acids phase protein production in MyD88-deficient mice (Kawai
Altered glucose metabolism et al, 1999; Yamamoto et al, 2004). Concurrently, TLR4 sig-
Increased corticosteroid production
naling proceeds via a MyD88-independent pathway utilizing
IL, Interleukin; IL-1R, interleukin-1 receptor; TGF-β, transforming growth factor-β; TNF-α, tumor the TIR domain–containing adaptor molecule-inducing
necrosis factor-α. interferon-β (IFN-β) (TRIF). This pathway appears to be
essential to the production of IFN-β and IFN-inducible genes
(Yamamoto et al, 2003).
the innate immune cellular response to LPS. CD14, a Downstream signaling of the TLR4-mediated pathway
glycosylphosphatidylinositol-anchored protein, functions as a results in activation of the nuclear factor -κB (NF-κB)–inducing
ligand-binding protein for the LPS-LBP complex. Lacking an kinase (NIK), causing NF-κB translocation from the cytoplasm
intracellular domain capable of signal transduction, the CD14 to the nucleus (Yang et al, 1999). This process, with resultant
complex functions to promote enhanced engagement of the transcription of NF-κB–dependent genes and mitogen-activated
signal transducing mechanism. This is demonstrated by the fact protein kinases (MAPKs), produces the majority of the
downregulation of NF-κB (Chuang & Ulevitch, 2004). These

LPS examples, along with multiple additional proteins that function
MD-2 to regulate TLR4 signaling, can prove to be potential targets
for counteraction of destructive inflammatory responses.
In direct defense against gram-negative bacteria and the
TLR4 harmful effects of LPS, polymorphonuclear leukocytes (PMNs)
possess the ability to produce an LPS antagonistic protein,
TIRAP/MAL bactericidal/permeability-increasing protein (BPI). BPI is a
TIR domain TRAM
member of a family of proteins named lipid transfer/LBP, which
MyD88 have the ability to bind to lipid substrates, such as the lipid A
Death TRIF moiety in LPS. Stored in the primary granules of PMNs, serum
domain TICAM-1
IRAK1 concentrations of BPI in the healthy state are very low. However,
IRAK4 levels rise sharply alongside LBP in response to the onset of
acute inflammation (Froon et al, 1995). As a homolog to LBP,
TRAF6
BPI can bind endotoxin aggregates and gram-negative bacteria,
but with a much higher affinity than LBP. The effects of BPI
MAP3K
IRF3 IKK function to neutralize endotoxin at multiple levels. Binding to
complex gram-negative bacteria causes a cytotoxic effect, and allows
gram-negative bacteria–BPI complexes to be taken up by PMNs
without CD14 activation (Fierer et al, 2002). Additionally, BPI
p38 has the ability to directly inhibit LBP-endotoxin interaction with
Interferon JNK CD14 and subsequent cell activation. In vitro studies of recom-
induction NF-κB activation ERK binant BPI have demonstrated direct inhibition of LPS-mediated
FIGURE 11.1 Lipopolysaccharide (LPS) signal transduction. LPS- monocyte stimulation (Heumann et al, 1993). This ability of
binding protein (LBP)-endotoxin complexes bind to CD14 and interact BPI to act as an endogenous inhibitor of LPS has spurred inves-
with Toll-like receptor 4 (TLR4) and MD-2 membrane proteins, which tigation of its use as a potential therapeutic for endotoxin-
activates intracellular myeloid differentiation response gene 88 (MyD88)- related inflammatory conditions. Recombinant BPI has been
dependent and -independent pathways that ultimately lead to nuclear demonstrated in human trials to significantly reduce levels of
factor (NF)-κB activation and type I interferon production. ERK, extracel- endotoxin and circulating tumor necrosis factor-α (TNF-α),
lular signal-regulated kinase; IκB, inhibitor of NF-κB; IKK, inhibitor of and has shown modest success in decreasing morbidity related
NF-κB kinase; IRAK, interleukin-1 receptor–associated kinase; IRF3, to severe meningococcal sepsis (Levin et al, 2000). Additional
interferon regulatory factor 3; JNK, Jun kinase; MAL, MyD88 adaptor-
work continues in determining BPI’s role as a potential antian-
like; MAP3K, mitogen-activated kinase kinase kinase; MD-2, myeloid
differentiation protein-2; NF-κB–inducing kinase; TICAM-1, Toll-like
giogenesis therapeutic agent (Wu et al, 2003).
receptor adaptor molecule-1 TIRAP, Toll-like/interleukin 1 receptor (TIR)
domain-containing adaptor protein; TRAF, tumor necrosis factor
Tumor Necrosis Superfamily
receptor–associated factor; TRIF, TIR domain–containing adaptor– The members of the TNF superfamily have continued to be
inducing interferon (IFN)-β; TRAM, Toll receptor–associated molecule; extensively studied since the initial discovery of TNF-α nearly
TRIF, Toll receptor–associated activator of interferon. (From Miyake K: 40 years ago. Composed of more than 20 primarily type II
Innate recognition of lipopolysaccharide by Toll-like receptor 4-MD-2, transmembrane proteins with more than 30 receptors, this
Trends Microbiol 12:186–192, 2004.) family of ligands has a profound effect on the immune response,
cell proliferation, and apoptosis (Bazzoni & Beutler, 1996;
Grewal, 2009). Additionally, several of these members have
proinflammatory cytokines responsible for the host responses profound proinflammatory effects.
seen with endotoxemia (Alexopoulou et al, 2001). TNF-α was initially isolated from LPS-challenged mice in
This system of signaling and enhanced response to LPS to an attempt to define a soluble factor causing necrosis of trans-
activate the innate immune response is advantageous to the plantable tumors (Carswell et al, 1975). It is initially synthe-
host in that it can provide enhanced protection from gram- sized as a 26 kDa cell-surface–associated molecule; it is then
negative induced infection. However, an unregulated response cleaved by the matrix metalloproteinase (MMP) TNF-α-
can cause an extremely damaging host response, as is seen with converting enzyme to produce a soluble 17 kDa form (Moss
overwhelming sepsis. Specific regulatory mechanisms are et al, 1997). Secreted primarily by monocytes and macro-
needed to prevent the potential shock, multisystem organ phages, TNF-α is an acute-phase protein that, through its
failure, or possible death that can result without modulation of cell-associated and soluble forms, can act in juxtacrine and
the previously described signaling pathways. endocrine signaling. In addition to its role in necrosis and
Macrophages previously exposed to LPS demonstrate apoptosis, it has also been defined as a significant factor in the
reduced responses to repeat stimulation; this LPS tolerance is physiologic effects seen with severe inflammatory reactions
important for the negative regulation of the systemic inflamma- (Box 11.1).
tory response (Medvedev et al, 2006). At the cell-surface level, Induction of the inflammatory response by TNF-α causes a
RP 105 (radioprotective 105), a homolog of TLR4, can func- profound effect on vascular endothelial cells, producing vaso-
tion to competitively inhibit the interaction between LPS and dilation and capillary permeability that function to increase
TLR4 (Divanovic et al, 2005). At the TIR domain of TLR4, transendothelial passage of fluid via alterations of adhesion
overexpression of TRIAD3A, an E3 ubiquitin-protein ligase molecules, increased production of nitric oxide, and cyclooxy-
inhibitor, functions to promote degradation of TLR4 and genase (COX). Further cellular damage and deleterious
production of oxygen- and nitrogen-reactive species. This has

BOX 11.1 Cellular and Systemic Effects of Tumor Necrosis
been demonstrated in mouse models of nonalcoholic steato-
Factor-α
hepatosis, in which LPS-induced TNF-α upregulation pro-
Cellular duced accelerated pathologic hepatocyte destruction (Leist
Lymphocytes et al, 1995).
Activation of cytotoxic T-cells TNF-α has long been investigated as a potential chemo-
T-cell colony formation therapeutic agent, due to recognized antiproliferation proper-
Polymorphonuclear Lymphocytes ties and effects on vascular permeability (Roberts et al, 2011).
Superoxide production However, systemic use has been limited by the significant side
Increased phagocytosis effects. Low-dose TNF-α administration in healthy volunteers
Increased cellular adhesion produced an acute activation of neutrophils, reflected by a rapid
Monocytes/Macrophages
increase in plasma concentrations of IL-6, IL-8, and acute-
Induced cytokine production
phase proteins. Additionally, participants demonstrated a sus-
Chemotaxis
Increased metabolism tained decrease in lymphocytes, basophils, and eosinophils
Endothelial Cells (Corssmit et al, 1997). Due to the increased doses of TNF-α
Increased coagulation needed in chemotherapeutic regimens, such as with isolated
Increased vascular permeability limb perfusions, continuous monitoring of systemic leakage and
Induce nitric oxide synthetase infusion rate adjustment is necessary (Sorkin et al, 1995).
Prevent fibrinolysis
Induce endothelial adhesion molecules Interleukin-1
Induce prostacyclin In similar fashion to the TNF superfamily, IL-1 is a very potent
Promote angiogenesis
factor in activation of the innate immune system and inflam-
Fibroblasts
matory response. Due to the discovery that cytoplasmic domain
Induced collagenase production
Increased matrix metalloproteinase production of the IL-1 receptor type I is quite homologous to those of
TLRs, similar activation of proinflammatory immune responses
Systemic would be expected (Dinarello, 2009). However, unlike most
Central Nervous System other cytokine families, the IL-1 family includes members that
Increased core temperature actively function in suppression of the innate immune response
Anorexia (Garlanda et al, 2004).
Pituitary dysfunction The IL-1 family is composed of 11 ligands, including
Increased production of astroglial and microglial cells IL-1F1, IL-1F2, and IL-1F3, which function as a specific
receptor antagonist for the IL-1 family (Weber et al, 2010).
Gastrointestinal
IL-1 primarily produces an immune response via the ability to
Ischemia
Hemorrhage induce gene expression of multiple inflammatory mediators.
Acute-phase protein release This is observed with the IL-1–mediated production of COX-2
Hepatosplenomegaly and inducible nitric oxide synthase, as well as increased expres-
sion of intercellular adhesion molecule-1 and vascular cell
Cardiovascular adhesion molecule-1 (True et al, 2000). These products, as
Shock and tissue injury well as interaction with and potentiation of other cytokines,
Increased vascular permeability result in the symptoms of inflammation, vasodilation, and
Tachycardia movement of activated inflammatory cells to tissues, which are
Cardiovascular collapse
seen with activation of the innate immune response. Stimula-
Endocrine
tion of IL-1 can occur via LPS, as is seen with other cytokine
Insulin resistance
Protein catabolism families. However, other factors, such as TNF-α, complement
Stress hormone release C5a, hypoxia, or IL-1 ligands per se, can induce increased IL-1
Adrenal hemorrhage messenger RNA transcription.
Although the IL-1F1 precursor is found in constitutive levels
in nearly all nonhematopoietic cells, it is not actively secreted
or found in the circulation. It is upregulated in hypoxia, facili-
inflammatory effects are produced by the triggering effect of tating increased transcription of proinflammatory mediators
TNF-α on secondary mediators, such as IL-1, IL-6, IL-10, when its intracellular concentrations are increased (Werman
IFN-γ, platelet-activating factor, epinephrine, cortisol, and et al, 2004). Fully active as an intracellular precursor molecule,
growth hormone. This activation contributes to the wide range IL-1F1 is only released from necrotic cells, producing a sterile
of physiologic effects seen with overwhelming septic shock inflammation and inducing other proinflammatory molecules,
(Tracey & Cerami, 1994). including IL-1F2 (Chen et al, 2007).
In addition to the inflammatory effects of TNF-α, increasing In contrast to the IL-1F1 precursor, the IL-1F2 precursor
evidence implicates its role in the process of apoptosis. Although is not constitutively active and is found primarily in monocytes,
a natural and protective host mechanism, dysregulated cell macrophages, and dendritic cells (Unlu et al, 2007). Processing
death can occur with overexpression of TNF. Several members and secretion requires cleavage by the intracellular cysteine
of the TNF receptor group possess “death domains,” which are protease caspase-1, which is the rate-limiting step in IL-1F2
protein interaction domains that activate caspases and other secretion (Netea et al, 2009). Inhibition of this process for
signaling molecules that result in gene transcription and IL-1F2, as well as other members of the IL-1 family, has been
an area of intense study to reduce the severity of IL-1–based Epithelial/

inflammation and disease (Martinon & Tschopp, 2004). Block- Immune cells Malignant cells Fibroblasts
ade of IL-1F1 and IL-1F2 effects remains an intense area of
focus in the treatment of autoinflammatory diseases. Multiple IL-6 IL-11
experimental models have demonstrated reduced hepatic fibro-
sis in mouse models (Kamari et al, 2011) (see Chapter 7). gp130
Additionally, with the recognition of the role of chronic inflam-
mation in the development of tumors, investigation into the RAS/ERK JAK/STAT3 PI3K/AKT
Tumor cell
role of IL-1 in carcinogenesis has continued to grow. IL-1 has proliferation
been implicated in angiogenesis promotion in pancreatic carci-
noma (Matsuo et al, 2009), whereas inhibition of IL-1F1 was
demonstrated to reduce hepatocellular carcinoma development Tumor cell
(Sakurai et al, 2008). Future work into both caspase and direct survival
IL-1 inhibition may open new insights into halting tumor EMT/ Metastasis Angiogenesis Inflammation
invasion
progression.
Interleukin-6 Malignant cells Microenvironment
Although also involved in inflammation through effects on cell FIGURE 11.2 Interleukin (IL)-6 and IL-11 signaling in cancer. IL-6
differentiation, proliferation, and apoptosis, IL-6 has also been and IL-11, produced by immune cells, fibroblasts, and epithelial and
found to play an important role in other functions, including malignant cells, activate the JAK/STAT3, SHP-2-RAS-ERK, and PI3K-
AKT pathways, through which they induce cell proliferation, survival,
immune regulation and oncogenesis (Elinav et al, 2013).
EMT/invasion, metastasis, angiogenesis and inflammation. AKT, Protein
IL-6 belongs to a nine member superfamily of cytokines,
kinase B; EMT, epithelial-to-mesenchymal transition; ERK, extracellular
which includes leukemia inhibitory factor, oncostatin M, signal-regulated kinase; gp130, glycoprotein 130; JAK, Janus activating
cardiotrophin-1, neurotrophic factor, as well as IL-11. Signal kinase; PI3K, phosphatidylinositol-3-kinase; RAS, rat sarcoma (protein);
transduction for all members of this family utilizes the signal- STAT, signal transducer and activator of transcription. (From Taniguchi
transducing receptor glycoprotein 130 kDa (gp130), making K, Karin M: IL-6 and related cytokines as the critical lynchpins between
this a potential target for future therapies. Upon formation of inflammation and cancer, Semin Immunol 26:54–74, 2014.)
gp130 homodimers by ligand binding, signaling is carried
forward primarily by the Janus activating kinase/signal trans-
ducer and activator of transcription pathway. downregulating the membrane-bound FAS apoptotic signaling
IL-6 has a wide range of effects, both in the hematopoietic pathway (Schneider et al, 1998).
system and in the innate immune response. It has been dem- The function of FAS in processes separate from apoptosis
onstrated that IL-6 has a substantial role in the balance between continues to be investigated. Although FAS was thought to be
IL-17–producing T cells and regulatory T cells; IL-6 overpro- found in cells of lymphoid or myeloid lineage, it is now known
duction and upregulation of T helper (Th)17 cells is thought to be present in a wide variety of nonlymphoid cells. Soluble
to be a critical factor in development of multiple autoimmune FASL has been demonstrated to induce inflammation through
disorders (Kimura & Kishimoto, 2010). IL-6 and IL-11, with both NIK signaling as well as phosphatidylinositol-3-kinase
known effects on cell proliferation, are felt to play a crucial role (O’Reilly et al, 2009). Ligation of FAS in these cases produces
in the development of hepatocellular carcinoma (Fig. 11.2) (see chemotactic factors, such as IL-8 and monocyte chemoattrac-
Chapter 91). Current studies are underway investigating inhibi- tant protein-1, recruiting neutrophils, although not activating
tion of these cytokines in solid tumors. them. This is contrary to, and separate from, the apoptotic
properties of FAS. Nonapoptotic function becomes vitally
FAS Ligand important in liver cells, which have the highest constitutive
Identification of the FAS ligand (FASL or CD95L) through expression of FAS. It has been demonstrated in mice that FAS
expression cloning confirmed that this membrane-associated is intimately involved in liver regeneration after partial hepatec-
protein is also a member of the TNF superfamily (Suda et al, tomy (see Chapter 6), rather than apoptosis, and functions in
1993). Initially thought to function exclusively as an inducer neutrophil recruitment in viral hepatitis (Desbarats & Newell,
of apoptosis, continued study has implicated FAS as a factor 2000). FAS also plays a role in inflammatory processes of
in inflammation, carcinogenesis, and cellular proliferation intestinal epithelial cells via interaction with TLR signaling
(Peter et al, 2007). Activation of FAS has certainly been pathways. These properties of FAS more closely resemble the
proved to trigger apoptosis, observed by the presence of FASL functions of the TNF/tumor necrosis factor receptor 1 (TNFR1)
at the surface of lymphocytes and natural killer (NK) cells, system.
where it functions to promote the elimination of transformed Additionally, links between FAS and carcinogenesis con-
and infected cells. Additionally, rapidly fatal hepatitis has tinue to be explored. Studies have demonstrated increased
been observed in murine models with upregulated FAS FAS-induced production of chemotactic factors and promotion
signaling pathway (Huang et al, 1999) (see Chapters 9D of tumor invasiveness and metastasis via MAPK pathways
and 70). (Barnhart et al, 2004). Additionally, increased serum levels of
A five stage model of activation of membrane-bound FAS soluble FAS have been correlated with decreased survival in
has been proposed, detailing a rapid caspase-dependent apop- multiple tumor types (Konno et al, 2000; Ugurel et al, 2001).
tosis of FAS-bearing cells (Lee et al, 2006). FAS also exists in This supports the view that FAS ligation, in certain physiologic
a soluble form, the result of cleavage via metalloproteases. This situations, can signal not only nonapoptotic pathways, but pos-
soluble FAS has been proposed as a FAS receptor antagonist, sibly prooncogenic transmission as well.
Transforming Growth Factor-β hypoalbuminemia found in patients with chronic liver disease
Transforming growth factor-β (TGF-β) is another pleiotropic and have been predictive of a negative clinical course.
cytokine involved in many aspects of cellular function, includ- Acute-phase protein production has been found to occur as
ing differentiation, migration, angiogenesis, and apoptosis two separate groups, influenced by the presence of the specific
(Massague, 2008). It is stored as an intracellular latent form cytokine and signaling mechanism. Type I proteins, such as
until it is activated by several factors, including MMPs, integ- C-reactive protein, serum amyloid A, and complement C3, are
rins, and thrombospondin 1 (Ge & Greenspan, 2006). Of sig- primarily influenced by IL-1 and TNF-α, whereas type II pro-
nificant importance to TGF-β, the downstream effects of this teins, such as fibrinogen, α1 antitrypsinase, and ceruloplasmin,
cytokine are quite context and cell dependent. The primary are primarily stimulated by IL-6 (Moshage, 1997). These pro-
signaling mechanism is through the phosphorylation of SMAD2 teins have a wide range of effects in the acute response, both
and SMAD3, allowing translocation into the nucleus and resul- proinflammatory and antiinflammatory (Table 11.3). For
tant gene activation (Shi & Massague, 2003). Additionally, example, both haptoglobin and hemopexin act as antioxidants,
activation can occur via SMAD-independent pathways, primar- protecting against reactive oxygen species, and haptoglobin aids
ily through MAPKs. in angiogenesis, critical to hepatic regeneration (Cid et al,
The function of TGF-β is often contradictory, as it can 1993) (see Chapter 6).
function to arrest growth, as well as act as a tumor promoter. Prolonged and unregulated response can result in irrevers-
The functional switch can occur at many points in the signaling ible tissue damage in the host, and counterregulatory mecha-
pathway. In the SMAD-dependent pathway, location of forma- nisms are necessary. An IL-1 receptor antagonist has been
tion of the active complex appears to play a role in conversion implicated in attenuating the acute response; studies of pre-
to oncogenic properties, as increased phosphorylation of the treatment with IL-1 receptor antagonist (IL-1ra), a recom
linker region of SMAD3, in preference to the C-terminal binantly synthesised endogenous inhibitor of IL-1, have
region, is seen in many advanced tumors (Wrighton et al, demonstrated decreased progression of inflammation. IL-6,
2009). Although less common, mutations can also play a role which is an extremely strong inflammatory stimulus early in the
in tumor development, which is seen with increased SMAD4 response, has been implicated to counteract this effect via a
inactivation in pancreatic carcinoma (Goggins et al, 1998). negative feedback loop, decreasing production of multiple pro-
Additionally, aberrant TGF-β signaling has been implicated in inflammatory cytokines (Xing et al, 1998). Additionally, activa-
the progression of hepatic fibrosis through the increased pro- tion of the hypothalamic-pituitary-adrenal axis, primarily by
duction of connective tissue growth factor by hepatic progenitor IL-1 and IL-6, induces production of steroid hormones.
cells (Gressner et al, 2002) (see Chapter 7). Increased release of glucocorticoids facilitates transport of free
amino acids to the liver, providing the substrates necessary for
Cytokines in the Liver and Pancreas Regulation of production of antiinflammatory acute-phase products.
Cytokine Expression in the Liver
Hepatic response to inflammation and parenchymal injury pro- Cytokines and the Pancreas
duces a variety of effects in the host, both locally and systemi- Defining the effects of specific cytokines in pancreatic dysfunc-
cally. Because as many as 80% of all macrophages in the body tion is complicated by the pleiotropic effect of these molecules.
are Kupffer cells, the liver has a major impact on production Pancreatic islet cells produce multiple cytokines in the well
of inflammatory mediators. Although cytokine expression in the state, regulating B-cell function and replication. Increasing
liver is low at baseline, as a nonspecific first line of defense from IL-1 levels stimulates B-cell proliferation to respond to exog-
bacterial products from the hepatic circulation, a rapid, orches- enous stressors. However, prolonged stress has been implicated
trated response is needed for host protection. in islet cell dysfunction and destruction. Mouse models have
Hepatic cytokine contribution has been demonstrated in demonstrated the role of TGF-β in the progression of type 1
human studies, where cannulation of the hepatic vein after diabetes, and elevated concentrations of this cytokine have been
intravenous endotoxin administration revealed that large implicated in increased microvascular complications in diabetic
amounts of both TNF-α and IL-6 in the systemic circulation children (Zorena et al, 2013).
were derived from the splanchnic bed (Fong et al, 1990). TGF-β has also been studied as a contributing factor in
Increasing evidence has also defined the role of hepatic stellate the development of chronic pancreatic inflammation (see
cells in recruitment and induction of infiltrating leukocytes, Chapters 54 and 55) and pancreatic adenocarcinoma (see
amplifying the inflammatory response. Infiltration of T cells Chapter 59). Elevated levels are found in patients with chronic
and NK cells can enhance production of TNF-α, FASL, and pancreatitis, correlating with progression of fibrosis and
TGF-β, as described in studies of virally induced hepatic injury insulin resistance. It has been postulated that insensitivity to
(Melhem et al, 2006). Increasing levels of these cytokines, TGF-β stimulation, with resultant elevated levels, results in a
through varied mechanisms of insult to the host, has been switch from inhibitory effects to increased proliferation of
demonstrated in both hepatic injury and regeneration. pancreatic adenocarcinomas, in concert with additional cyto-
kine influence (Table 11.4). Accordingly, elevated TGF-β1
Cytokines and the Hepatic Acute-Phase Response and TGF-β2 receptor levels have been associated with
In response to hepatic injury, release of IL-1, IL-6, and TNF-α advanced cancers, with a trend toward worse overall survival
accounts for the stimulus to upregulate production of acute- (Javle et al, 2014).
phase reactant proteins. Whereas cytokine-stimulated tran-
scription factors produce increased amounts of acute-phase Cytokines and Apoptosis
proteins, levels of constitutively active proteins, such as albumin, The process of hepatic inflammation exists to protect the hepa-
transferrin, and blood factor XII, decrease proportionately tocytes and the host from excessive injury, while promoting
(Baumann & Gauldie, 1994). These findings account for the repair of tissue damage. However, with overwhelming or
TABLE 11.3 Hepatic Acute-Phase Proteins and Their Major Biologic Functions
Acute-Phase Reactant Major Biologic Functions
Secreted Pathogen Recognition Receptors (PRRs)

C-Reactive protein Small pentraxin; recognizes phosphocholine and polysaccharides of bacteria. Mediates complement
activation and phagocytosis. Induces release of ROS and cytokines from macrophages
Serum amyloid P Small pentrxin; Ca2+-dependent binding of lipoprotein ligands of bacteria. Interaction with complement
and matrix components
Peptidoglycan-recognition N-Acetylmuramoyl-L-Ala amidase; hydrolyzes the lactyl bond between the MurNAc and L-alanine in
protein-2 bacterial peptidoglycan; elimination of microbial peptidoglycans to prevent excessive inflammation
Proteinase Inhibitors
α1-Antichymotrypsin Serine proteinase inhibitor; inhibitor of cathepsin G
α1-Antitrypsin Trypsin inhibitor. Inactivator of neutrophil elastase. Modulation of neutrophil chemotaxis and inhibition of
neutrophil superoxide production
α2-Macroglobulin Plasma protein. Inhibits serine, cysteine, aspartate and metal proteases. Exhibits cytokine, growth factor,
and hormone binding functions. Modulation of mitogen and antigen-driven T-cell response
Pancreatic secretory trypsin Specific trypsin inhibitor that prevents autoactivation of trypsinogen
inhibitor
α1-Cysteine proteinase Endogenous inhibitor of lysosomal cysteine proteinases
inhibitor
Components of the Complement System
C3 Central component of the complement system, activated by classical, alternative, and lectin pathways
C4 Classical complement pathway. Functions as C3 convertase.
C9 Part of lytic membrane attack complex. Formation of transmembrane pore on invading pathogen
C4b-binding protein Human complement inhibitor. Accelerates decay of C3 convertase
Factor B Formation of C3 convertase. Amplifier of alternative pathway
C1 inhibitor Serine protease inhibitor. Inactivator of complement serine esterases. Inactivator of factors XI and XII of
coagulation system
Iron homeostasis
Hepcidin Hepatic bactericidal protein. Main regulator of body iron homeostasis. Acts as an antimicrobial peptide
by limiting iron availability
Ferritin Member of ferritin-like superfamily. Provides iron homeostasis
Metal-Chelating Proteins or Transport Proteins
Haptoglobin Crucial role in heme-iron recovery. Binds free hemoglobin, exerting bacteriostatic and antioxidant
function. Reduces loss of free hemoglobin through glomerular filtration
Ceruloplasmin Plasma protein with a broad oxidase specificity. Regulation of iron and copper transport and homeostasis
Hemopexin Heme scavenging at systemic level. Highest binding affinity to heme, exerting antioxidant functions
Coagulation and Fibrinolytic System
Fibrinogen Clotting factor. Central regulator of the inflammatory response
Plasminogen Proenzyme. Converted into active serine protease plasmin. Modulator of the inflammatory response by
plasmin-mediated directional cell migration of hematopoietic cells
Tissue plasminogen activator Highly specific serine proteinase. One of two principal plasminogen activators. Acts as fibrinogen
cytokine. Modulator of extracellular matrix remodeling. Supports cell migration and invasion by multiple
interactions with integrins, endocytosis receptors, and growth factors
Plasminogen-activator Physiologic inhibitor of the two plasminogen activators
inhibitor 1
Protein S Vitamin K–dependent plasma glycoprotein. Acts as cofactor to activated protein C in the degradation of
factors Va and VIIa
Other Acute-Phase Proteins
Angiotensinogen Protease inhibitor. Converted into angiotensin; part of the renin-angiotensin-aldosterone system
IL-1 receptor antagonist Member of IL-1 family. Inhibits activities of IL-1α and IL-1β
α1-Acid glycoprotein Inhibits activation of neutrophils and upregulates the expression of IL-1 receptor antagonist by
macrophages. Inhibitor of leukocyte extravasation
Acute-phase serum amyloid Apolipoprotein. Influences cholesterol metabolism. Downregulates fever and inhibits PGE2 as well as
A oxidative burst
IL, Interleukin; PGE2, prostaglandin E2, MurNAc, N-acetylmuramic acid; ROS, reactive oxygen species.
Modified from Bode JG, et al: Hepatic acute phase proteins-regulation by IL-6- and IL-1-type cytokines involving STAT3 and its crosstalk with NF-κB-dependent signaling, Eur J Cell Biol 91:
496–505, 2012.
transduction of the apoptotic pathway (Fig. 11.4). Multiple

TABLE 11.4 Cytokines and Their Function in Pancreatic Cancer
studies have demonstrated increased FAS expression and
Cytokine Function number of FASL-bearing cells in the elimination of infected
Proinflammatory cells of patients with hepatitis B and C (Mochizuki et al, 1996;
Yoneyama et al, 2002). Additionally, increased levels of bile
IL-6 Promotion of angiogenesis
Association with cachexia salts, as seen with hepatic injury, promote increased transloca-
Association with poor prognosis tion of FAS to the plasma membrane, furthering the apoptotic
IL-8 Promotion of angiogenesis cascade (Sodeman et al, 2000).
Promotion of migration/metastasis Recognition of these effectors of apoptosis opens the doorway
Association with aggressiveness and invasiveness for potential targets for treatment. Caspase inhibition has been
TNF-α Promotion of migration/metastasis demonstrated to increase survival in mice undergoing massive
Association with cachexia hepatectomy (Yoshida et al, 2007). Additionally, targeting of
Association with poor prognosis
Used in Phase I/II clinical trials the FAS pathway led to decreased progression to fulminant
hepatic failure. Further elucidation of these pathways will
IL-1β Association with invasiveness
Promotion of migration/metastasis provide more targets for therapeutic treatments.
Antiinflammatory
Nitric Oxide
TGF-β Promotion of immune evasion
Association with poor prognosis Altered redox states and oxidative stress have been extensively
Used in Phase I/II clinical trials studied in a multitude of disease states due to their participation
Il-10 Promotion of immune evasion in processes such as inflammation and apoptosis (Wink et al,
Association with poor prognosis 1999). Oxidative stress implies an imbalance between oxidant
IL, Interleukin; TGF-β, transforming growth factor-β; TNF-α, tumor necrosis factor-α. and antioxidant agents; when the capacity of the antioxidant
system is overwhelmed, a harmful level of redox state is
Modified from Roshani R, et al: Inflammatory cytokines in human pancreatic cancer, Cancer Lett
345:157–163, 2014. achieved, and the potential for cellular damage arises.
Central to the balance of redox states in hepatocytes is nitric
oxide (NO). NO is a small, hydrophobic molecule with an
extremely short half-life. Created by nitric oxide synthase
prolonged stimulation, the protective mechanisms shift to a (NOS), NO presence can be greatly increased via production
regulated process of cell death, apoptosis. by inducible NO synthase (iNOS). iNOS is expressed in all
TNF-α would appear to be a primary cytokine responsible hepatic cells, and its production can be induced by IL-1, TNF-
for the initiation of apoptosis; in many cases, such as endotox- α, and LPS (Sass et al, 2001). NO plays a complicated role in
emia, it certainly is an integral factor (Xaus et al, 2000). hepatocyte cellular function. In steady-state low-redox condi-
However, due to its extremely pleiotropic nature, the exact tions, NO plays a protective role in the liver, having been
mechanism of contribution remains unclear. This is demon- demonstrated to inhibit apoptosis and abate mitochondrial dys-
strated by the fact that cultured hepatocyte cell lines do not function (Chen et al, 2003; Wang et al, 2002). NO also func-
demonstrate massive apoptosis with isolated TNF-α adminis- tions as a cytoprotective molecule by acting as an electron
tration; sensitization with hepatotoxic agents, such as acceptor for S-nitrosylation, causing inhibition of caspase activ-
D-galactosamine, is necessary. Additionally, TNF-α has a role ity (Kim et al, 2000). Upregulation of iNOS has also been
in cellular survival through its activation of NF-κB. Activation found to be an important factor in the propagation of the
of NF-κB has been demonstrated to play a role in hepatic inflammatory response, which, as previously discussed, can
inflammation and protection from apoptosis (Bohlinger et al, play a protective role for the host. NO has been demonstrated
1996; Reinhard et al, 1997). Thus a balance between the pro- to induce the expression of heat-shock protein 70 (Kim et al,
inflammatory and apoptotic signaling effects of TNF-α deter- 1997). These proteins may function by refolding damaged pro-
mines the path taken. teins in the liver, modulating caspase activation, and regulating
TNF-α apoptotic signaling signals the activation of caspase-8 expression of heme oxygenase-1, conferring a protective mech-
and caspase-3 to begin the process of ligand-mediated apopto- anism from apoptosis. NO also can bind soluble guanyl cyclase,
sis. The importance of these intracellular proteases in the apop- which can increase intracellular cyclic guanosine monophos-
tosis cascade is evident in the fact that caspase-3 knockout mice phate, suppressing apoptosis and caspase activity (Rodrigo
are much more resistant to the induction of lethal hepatitis and et al, 2004).
massive cellular damage than their wild-type counterparts In conditions of oxidative stress and altered redox states,
(Nicholson & Thornberry, 1997; Woo et al, 1999). Activation upregulation of iNOS has been associated with deleterious
of these caspases is signaled by ligand binding of a group of effects due to continued oxidative damage and propagation of
receptors termed death receptors. These receptors belong to inflammation. Studies in mice have demonstrated reduced
the tumor necrosis superfamily and include FAS, TNFR1, and amounts of hepatic damage with inhibition of NO synthesis,
TNF-related apoptosis-inducing ligand receptor 1. Binding by and iNOS has been implicated as a requirement for the forma-
their respective ligands activates these caspases and initiates the tion of ethanol-induced hepatic injury (McKim et al, 2003).
process of apoptosis. Alternatively, inhibition of procaspase-8 Viral hepatitis has also been associated with increased iNOS
is thought to direct signaling to cellular necrosis (Fig. 11.3). expression, with several studies suggesting an association
Central to hepatic apoptosis is upregulation of the FAS/ between levels of iNOS induction and disease severity (Atik
FASL system, which has been implicated in a variety of dis- et al, 2008). Thus it appears that the NO functioning in either
eases. Given the similarities between FAS and TNF-α, it would a protective or destructive role can vary depending upon cel-
follow that follow that there is much overlap in their signal lular levels and redox conditions (Table 11.5).
Complex II
Complex I (DISC)
TNF TNF TNF TNF
Procaspase-8
TNFR1 TNFR1 TNFR1 FADD TNFR1

TRADD TRADD TRADD
TRAF3 TRADD
RIP1 TRAF2=K63 RIP1-K63 TRAF2=K63 RIP1-K63 TNFR2 RIP1
TRAF2 RIP3
cIAP1 cIAP1 cIAP1 cIAP1
TRAF3
LUBAC
cIAP2 cIAP2 complex cIAP2 cIAP2
RIP1
Procaspase-8 RIP3
Inhibition Inactivaqtion
of caspases of RIP1
MAP3Ks activation Activation NECROPTOSOME
of IKK complex
IKK (MEMO)
IKK Procaspase-8
IKK activation
FADD
TRADD
JNK TNFR2 RIP1 P
κBa degradation RIP3
Activation of
cIAP1 cIAP2 effective caspases
AP-1 ACTIVATION NF-κB ACTIVATION NECROPTOSIS APOPTOSIS

FIGURE 11.3 TNFR1 activation leads to the formation of different signaling complexes. The binding of TNF to TNFR1 initially involves the
constitution of a transmembrane signaling complex (complex I) in which the receptor interacts, through the adaptor protein TRADD, with RIP1 and
also with TRAF2, TRAF3, cIAP1, and cIAP2. Complex I is also able to activate the transcriptional factor NF-κB through the IKK complex. Both TNFR1
and the adaptor proteins can undergo conformational changes that lead to the internalization of the signaling complex and to changes in the adaptor
proteins that bind to the receptor. This new signaling complex (complex II or DISC) differs from complex I in that TRADD, procaspase-8 and RIP3
also interact with TNFR1. In this complex the inactivation of RIP1 and RIP3 leads to the activation of caspase-8 and thus to the triggering of the
apoptotic process. Under conditions in which caspases activation is inhibited, RIP1 and RIP3 could be phosphorylated, which leads to cell death
through a caspase-independent process known as necroptosis. AP-1, Activator protein-1; cIAP, cellular inhibitor of apoptosis; DISC, death-inducing
signaling complex; FADD, FAS death domain; IKK, inhibitor of NF-κB kinase; IκB, inhibitor of NF-κB; JNK, N-terminal jun kinase; K63, lysine (position)
63; LUBAC, linear ubiquitin chain assembly complex; MAP3K, mitogen-activated protein kinase kinase kinase; MEMO, mediator of ERBB2-driven
cell motility; NF-κB, nuclear factor κB; NF-κB inductor kinase; RIP, receptor-interacting serine-threonine kinase; TNF, tumor necrosis factor; TNFR1,
TNF receptor 1; TRADD, TNF receptor–associated protein with death domain; TRAF, TNF receptor adaptor factor. (From Cabal-Hierro L, Lazo PS:
Signal transduction by tumor necrosis factor receptors, Cell Signal 24:1297–1305, 2012.)
Liver Regeneration mice and specific cytokine inhibitors have demonstrated the
The liver possesses a unique capacity for repair after injury; it importance of TNF-α and IL-6 in the initiation phase of regen-
is the only solid organ with the ability to fully regenerate eration (Michalopoulos & DeFrances, 1997). This is supported
damaged or lost tissue (see Chapters 6, 103, and 108). This by the finding that expression of inflammatory cytokines is
ability differs significantly from other organs, which heal by scar upregulated soon after hepatic resection, with serum levels of
formation (Fausto et al, 2006). However, in certain disease TNF-α and IL-6 detectable for up to 5 days (Abshagen et al,
states, cellular damage outpaces the regenerative capacity, and 2007). Progression to the proliferative phase is primarily regu-
healthy functioning epithelial tissue is replaced by connective lated by mitogenic factors such as hepatocyte growth factor,
tissue; such is the case with hepatic fibrosis and the develop- and the termination has been associated with members of the
ment of cirrhosis. Insights into the mechanisms involved in TGF-β superfamily, such as TGF-β1, follistatin, and activin.
hepatic regeneration open the door for potential therapeutic Inhibition of these TGF-β members in combination has been
targets to prevent chronic liver diseases. demonstrated to disrupt normal cessation of the proliferative
The process of liver regeneration has been described to phase (Oe et al, 2004).
occur in three phases: (1) initiation, (2) proliferation, and (3) An increasing body of evidence suggests that TNF-α, and,
termination (Zimmermann, 2004). Studies utilizing transgenic possibly more important, signaling through TNFR1, are crucial
in initiating the process of hepatic regeneration. Expression of

TABLE 11.5 Roles of iNOS in Hepatic Injury
TNF-α is upregulated as soon as 30 minutes after hepatic
Condition/Inducers NO Effect Mechanism injury via NF-κB pathways (Yang et al, 2005). Injecting mice
In Vivo with anti–TNF-α antibodies was demonstrated to delay onset
of regeneration and DNA synthesis, whereas TNFR1 knockout
Endotoxemia Protective Inhibition of apoptosis
Toxic Oxidative stress mice had significantly delayed proliferation and decreased cell-
Circulatory failure cycle gene expression (Shimizu et al, 2009). Interestingly,
TNF-α Protective Inhibits apoptosis injection of IL-6 into these TNFR1–deficient mice corrected
CCl4 Protective Decreases oxidative stress the delay in regeneration (Yamada et al, 1997). This may
Liver Regeneration Protective Inhibits apoptosis
suggest that not only are factors other than TNF-α essential
for the initiation of regeneration, but also that a major role of
Ischemia/ Toxic Oxidative damage
reperfusion TNF-α and TNFR1 is to upregulate IL-6 production.
Hemorrhagic shock Toxic Direct toxicity, activates
IL-6 has been firmly established as a critical factor in hepatic
inflammation regeneration. Similar to TNF-α, IL-6 levels are found to be
Alcoholic liver injury Protective Unclear significantly elevated shortly after partial hepatectomy. IL-6–
In Vitro deficient mice demonstrate decreased hepatocyte DNA synthe-
sis and delayed proliferation; injecting these mice with IL-6
TNF-α, FAS Protective Inhibits caspase/apoptosis
antibody HSP70 upregulation before partial hepatectomy produces regeneration rates equiva-
H2O2 Protective Heme oxygenase-1
lent to their wild-type littermates (Cressman et al, 1996).
upregulation However, injection via the subcutaneous or intravenous route
Acetaminophen Protective Modulates GSH levels resulted in significantly different outcomes, with the subcutane-
ously injected mice showing higher survival rates (Blinden-
CCl4, Carbon tetrachloride; GSH, reduced glutathione; H2O2, peroxide; HSP70, 70-kDa heat-shock
bacher et al, 2003). This suggests that both IL-6 level and
protein, iNOS, inducible nitric oxide synthase; NO, nitric oxide; TNF-α, tumor necrosis factor-α.
duration of exposure play a critical role in the process of regen-
Modified from Li J, Billiar TR: Nitric oxide. IV. Determinants of nitric oxide protection and toxicity in
the liver, Am J Physiol 276:G1069–G1073, 1999.
eration. Additionally, IL-6 has been demonstrated to exert a
protective effect in the regenerating liver. Treatment with IL-6
TNF-α
FASL
Cell membrane
TNFR
TRAIL FAS FADD
FLIP Mitochondria
TRAIL-FADD
R1/R2 Smac/Diablo
Caspase-8 Omi/HtrA2
Caspase-10 Cyto c
IAPs Apoptosome
Caspase-3 BCL-2
Caspase-7 Caspase-9
p53
NF-κB/survivin-XIAP
Apoptosis
Proliferation
Nucleus
FIGURE 11.4 Regulation of apoptosis-signaling pathways. The apoptotic cell death requires the interplay of a multitude of factors. These related
factors are organized in a tight and efficient manner to mediate apoptotic signaling. Both FAS and TNF-α demonstrate significant overlap in the
apoptotic pathway via activation of caspases 3 and 8 to signal cell death. The activation of apoptosis causes cell death instantaneously unless the
cell death is inhibited by potent antiapoptotic signals. Transcription factors such as NF-κB and p53 have the ability to mediate apoptosis through
upregulation or downregulation of apoptosis-related gene expression in nuclei. BCL-2, B-cell leukemia/lymphoma (apoptosis regulator)-2; BIR, bacu-
lovirus inhibitor of apoptosis protein (IAP) repeat; FADD, FAS death domain; FLICE, FADD-like IL-1β–converting enzyme, FLIP, FLICE-inhibitory protein;
HtrA2, HtrA serine peptidase 2; IAPs, inhibitor of apoptosis proteins;NF-κB, nuclear factor κB; Smac/Diablo, second mitochondria-derived activator
of caspase/direct inhibitor of apoptosis-binding protein with low pI; TNF, tumor necrosis factor; TNFR, TNF receptor; TRAF2, TNF receptor-associated
factor 2; TRAIL, TNF-related apoptosis-inducing ligand; TRAIL-R1, TRAIL receptor-1 XIAP, X-linked inhibitor of apoptosis protein. (From Wang K, Lin
B: Inhibitor of apoptosis proteins (IAPs) as regulatory factors of hepatic apoptosis, Cell Signal 25:1970–1980, 2013.)
before partial hepatic resection in mice reduced the oxidative Nitric oxide has also been demonstrated to contribute to the
dysfunction seen after such an insult (Jin et al, 2007). Protec- integrity of the gut mucosal barrier. LPS-induced iNOS pro-
tion has also been demonstrated through IL-6–mediated cell- duction increased permeability of an experimental intestinal
cycle arrest and activation of DNA mismatch repair genes, monolayer, resulting in a 10-fold increase in bacterial transloca-
ensuring a correct course of regeneration (Tachibana et al, tion (Forsythe et al, 2002). Supporting these findings, a
2014). study of 44 cirrhotic patients demonstrated that the level of
Adequate production of IL-6 and subsequent proliferation is disease progression corresponded to an increased presence of
linked to TNF-α and TNFR1 activation. However, as discussed NO-secreting CD14+ macrophages, implicating these cells as a
previously, TNF-α can also be a potent stimulator of apoptosis. potential source of permeability increasing cytokines (Du
The decision to trigger the antiapoptotic pathway appears to be Plessis et al, 2013).
regulated by the activation of NF-κB. After inhibition of NF-κB,
administration of TNF-α to mouse livers upregulated caspases-3 Pathogenesis of Proinflammatory Cytokines
and caspase-9 and produced massive hepatocyte apoptosis (Xu Although it is impossible to establish a direct link between an
et al, 1998). Similarly, inhibition of NF-κB using an adenoviral individual cytokine and disease progression, mounting circum-
vector produced rapid onset of liver failure after partial hepatec- stantial evidence continues to demonstrate their contributions
tomy (Iimuro et al, 1998). These findings provide the link to, and potential targets for, multiple diseases.
between TNF-α signaling and IL-6 production, as well as the Deregulation of multiple cytokines has been linked to out-
necessity of NF-κB–dependent gene transcription for appropri- comes in hepatitis (see Chapters 9D and 70). Patients infected
ate induction of hepatic regeneration. with hepatitis C have demonstrated increased hepatocyte apop-
tosis (McPartland et al, 2005). TNF-α has been shown to be
PATHOGENESIS OF ENDOTOXIN AND an integral factor in the development of fulminant hepatic
CYTOKINES IN LIVER, BILIARY, AND failure in humans and mice, and elevated serum TNF-α levels
have been correlated with mortality in alcoholic hepatitis
PANCREATIC DISEASE (Rachakonda et al, 2014). Elevated TNF-α levels correlated
In patients with advanced liver disease, there is an enhanced with level of fibrosis in hepatitis B and C patients (Akcam et al,
susceptibility to systemic endotoxemia; rarely does this occur 2012), and patients infected with hepatitis C demonstrate a
in patients without advanced disease or end-organ dysfunction. correlation between increased disease severity and elevated
Elevated levels of plasma endotoxin have been found in patients levels of soluble FAS (Hayashi & Mita, 1997). Elevated soluble
with a variety of hepatic disorders, and the degree of elevation FAS and TNF-α levels corresponded with elevated alanine
has been correlated with severity of disease (Lin et al, 1995). aminotransferase levels and increased severity of disease
The concept of the passage of viable intestinal bacteria through (Raghuraman et ak, 2005). Additionally, elevated IL-6 levels
the gut mucosa into mesenteric lymph nodes and sterile tissue strongly associated with liver-related mortality in hepatitis C
has been defined as bacterial translocation (Berg & Garlington, patients (Nakagawa et al, 2015). Studies of hepatitis B–infected
1979). mice have implicated IL-6 overexpression to altered cell-cycle
Intestinal bacterial overgrowth, as has been observed in regulation and impaired hepatic regeneration (Quetier et al,
patients with advanced liver disease, has been closely linked 2013).
with development of bacterial translocation (Chang et al, 1998; However, in advanced viral hepatitis, elevated TNF-α levels
Llovet et al, 1998). However, other factors also likely contribute often do not correspond with apoptosis of infected cells and
to translocation. Structural changes of intestinal mucosa in cir- hepatic regeneration. This is presumably due to increased
rhotic patients has been observed (Norman et al, 1980), as well NF-κB activation, preventing the apoptotic functions of TNF-α
as increased permeability allowing tight junction translocation (Collins et al, 2014). The finding that administration of anti–
(Pascual et al, 2003; Zuckerman et al, 2004). Use of nonselec- TNF-α functions in an antiviral manner, upregulating apopto-
tive β-blockers to decrease portal venous pressure gradient sis of infected cells and inhibiting regeneration, supports this
resulted in decreased translocation, as well as reduced serum line of thought (Brenndorfer et al, 2010). These findings
IL-6 and LBP levels (Reiberger et al, 2013). Additionally, mul- suggest a complex interaction between these cytokines and the
tiple disease states, such as hemorrhagic shock, have been iden- ability to remove virally infected cells.
tified to produce similarly increased permeability (Kompan Similarly, in patients with alcoholic hepatitis, elevated serum
et al, 1999). levels of TNF-α and IL-6 appear to play a significant role in
Disease states and increased translocation stimulate the gut- disease severity. In patients with acute alcoholic hepatitis, levels
associated lymphoid tissue to respond with increased produc- of these cytokines were strongly associated with 180 day mor-
tion of multiple cytokines, producing an inflammatory state, tality (Rachakonda et al, 2014). Endotoxin levels are also sig-
possibly potentiating bacterial translocation. Bile duct–ligated nificantly increased in these patients compared with healthy
mice demonstrated elevated levels of TNF-α in the lamina control participants, and concentrations have been correlated
propria of the gut, as well as accelerated hepatic fibrosis (Hart- with mortality (Fujimoto et al, 2000). Elevated endotoxin
mann et al, 2012). Similarly, elevated serum TNF-α levels are levels are likely due to increased hepatic exposure to LPS prod-
also found in mesenteric lymph nodes of cirrhotic patients ucts via translocation from the gut (Pinzone et al, 2012). The
(Genesca et al, 2003), and administration of anti-TNF anti- enhanced inflammatory reaction seen is likely mediated through
bodies has been demonstrated to significantly reduce bacterial activation of Kupffer cells via the LPS/LBP pathway, providing
translocation (Frances et al, 2007). Additionally, elevation of a potential therapeutic target of therapy for alcoholic hepatitis
serum IL-6 levels has been implicated as a contributing factor (Sandahl et al, 2014).
for increased gut permeability in hemorrhagic shock models Aberrant signaling pathways have also been implicated in
(Yang et al, 2003). the progression to malignancy. Reduced FAS expression has
been implicated in the loss of normal apoptotic mechanisms reperfusion. eNOS knockout mice demonstrated increased
seen with the development of hepatocellular carcinoma (Fuku- severity of hepatic injury after ischemia compared with their
zawa et al, 2001; Shin et al, 1998). IL-6 levels were found to wild-type counterparts. Additionally, TNF-α expression was
correlate with poor prognosis in hepatocellular carcinoma (Jang upregulated to a level five times that in the wild-type mice
et al, 2012; Ohishi et al, 2014). Additionally, although no (Hines et al, 2002), suggesting a protective effect from cytokine
direct link between NO and development of carcinoma has injury. This protection has demonstrated effects beyond the
been defined, in patients with hepatocellular carcinoma, ele- liver, as NO has also been shown to reduce the incidence of
vated levels of iNOS in conjunction with COX-2 expression onset of hepatopulmonary syndrome in transplanted rats (Diao
had a significant negative effect on patient survival (Rahman et al, 2012).
et al, 2002). Despite these findings, it is often difficult to establish a link
Experimental models of obstructive jaundice demonstrate between elevated levels of proinflammatory mediators and a
similar alterations in cytokine levels, and TNF-α and IL-6 have negative clinical outcome. Insight into this link comes from
been identified as the dominant cytokines (see Chapter 8). studies of patients undergoing thoracoabdominal aortic aneu-
Endotoxin-mediated stimulation of the innate immune system rysm repair; intraoperatively, supraceliac aortic clamping pro-
produced a markedly exaggerated production of TNF-α and duces a demonstrable I/R injury, reflected by enhanced release
IL-6 in common bile duct–ligated mice compared with control of proinflammatory cytokines (Welborn et al, 2000). Microar-
subjects (Badger et al, 2012; O’Neil et al, 1997). This finding ray analysis of peripheral blood leukocytes found changes in
suggests a process of sensitization of macrophages to the effects 146 genes that were correlated with progression to multiorgan
of LPS. Support for these results is demonstrated by altered dysfunction, measured at 2 and 24 hours postreperfusion.
Kupffer cell function in obstructive jaundice, with increased Forty-one genes demonstrated increased expression at 2 hours;
responsiveness to the effects of LBP (Minter et al, 2005). several of these are involved in the innate immune response
Although IL-6 levels decreased after drainage and relief of (Feezor et al, 2004). Similarly, in a study of 25 patients under-
obstruction (Akiyama et al, 1998), slower resolution of IL-6 going thoracoabdominal aortic aneurysm repair, elevated levels
elevation has been linked with lower albumin levels and of neutrophil NF-κB strongly predicted onset of multisystem
increased frequency of positive bile cultures. Although no direct organ failure and mortality (Foulds et al, 2001).
link between IL-6 levels and outcomes can be concluded, pro-
longed elevation of IL-6 after drainage has been associated with Cytokines in Hepatic Tumor Ablation
elevated mortality in malignant obstructions (Yilmaz et al, For hepatic tumors not amenable to resection, multiple ablative
2013). techniques have been developed for local control with the goal
of improving overall survival (see Chapter 98). Although the
Cytokines in Ischemia-Reperfusion Injury mechanisms of ablation differ, all techniques aim for controlled
Ischemia-reperfusion (I/R) injury continues to be an important tumor cell destruction while sparing functional remnant liver.
cause of organ dysfunction after restoration of adequate blood However, unlike resection, ablative techniques leave the remain-
flow. This type of injury is seen in a variety of surgical condi- ing debris of devitalized tumor in situ, with large amounts of
tions, including transplantation, aortic surgery, elective hepatic tumor antigen resulting in immunologic reactions and cytokine
resections, as well as trauma-related disorders. Although the alterations.
exact mechanisms of I/R injury remain unclear, considerable Radiofrequency ablation (RFA) heats intratumoral cells to
evidence exists indicating the release of several cytokines fol- a temperature of 60° C to 100° C to create a coagulative necro-
lowing an ischemic insult (Liu et al, 2000). sis (see Chapter 98B). For optimal control, a goal of a 1 cm
TNF-α appears to play a central role in I/R injury. Murine margin of tissue surrounding the tumor should be achieved
models have demonstrated a significant increase in TNF-α (Dodd et al, 2001). Although tumor size limitations of approxi-
levels after 90 minutes of ischemia, followed by 60 minutes of mately 3 cm were required to obtain an adequate treatment
reperfusion (Colletti et al, 1990). Although the exact mecha- margin (Berber et al, 2004), advances in the technology of
nism stimulating TNF-α release is not yet defined, it appears thermal delivery are expanding this limit. In a human study of
that production is significantly influenced by Kupffer cell acti- 17 RFA-treated patients, no appreciable changes in cytokines
vation. In rats, Kupffer cells isolated from the liver after I/R (TNF-α, IL-1, IL-6, IL-1R, IL-10) were noted within the 48
demonstrated a several hundred fold increase in TNF-α pro- hours following treatment; median hepatic volume treated was
duction compared with controls. Treatment with anti–TNF-α 36.8%. Additionally, stimulation with LPS of whole-blood
antibodies significantly decreased both TNF-α and IL- 6 pro- samples demonstrated retained the ability for normal cytokine
duction (Wanner et al, 1999). Mediation of the effects of production, indicating adjacent Kupffer cells are inactivated by
TNF-α occurs primarily through activation of NF-κB. Trans- coagulative necrosis (Schell et al, 2002). However, in rat
location to the nucleus and binding activity of NF-κB models of large-volume RFA, 50% to 60% liver-volume treat-
increased as much as 66% after 5 minutes of ischemia (Li ment produced a significant rise in TNF-α and IL-6 indications
et al, 1999). Additionally, inhibition of NF-κB activation has (Ng et al, 2006). This may indicate that both tumor size and
demonstrated significantly decreased hepatic apoptosis and overall liver volume treated can be limitations for effective, safe
caspase-3 activity (Giakoustidis et al, 2010). This is similar to treatment.
findings with induction of heme oxygenase-1, which functions A more recent technique for tumor ablation is irreversible
to suppress TNF-α/TNFR1-directed hepatic apoptosis (Kim electroporation (see Chapter 98C). This therapy utilizes short,
et al, 2013). intense electric fields across cell membranes, increasing perme-
NOS also appears to be a factor in hepatic ischemia- ability. This allows for ablation of large amounts of tissue
reperfusion injury. The endothelial isoform of NOS (eNOS) without thermal effect. Preliminary animal studies have dem-
has most clearly been defined as a protective molecule during onstrated elevated percentages of CD4+ T lymphocytes and
IL-10 levels compared with hepatic resection groups (Li et al, agonists are currently demonstrating promise in murine pan-
2012), opening the possibility for future adjuvant immuno- creatic carcinoma models as tumor vaccine adjuvants (Huynh
therapy. Further studies are underway to fully elucidate the et al, 2012). Given the known importance of this receptor
relationship between electroporation and the inflammatory family in cellular signaling, TLR4 is also a potential target for
response. therapy. Despite some initial success, much work remains
before TLR antagonists can be considered as therapeutic
FUTURE DIRECTIONS agents. Multiple clinical trials continue to investigate TLRs as
potential targets for therapy (Aranda et al, 2014).
With increased understanding of cytokine functions and their Although evaluated in patients for the treatment of sepsis,
role in certain molecular pathways, intensive investigation into use of antiendotoxin as a therapy to abate the inflammatory
focused targeting of these molecules has progressed greatly over response would appear to be applicable to a variety of disorders.
the past decade. Agents are currently available as both inhibi- However, results from numerous studies investigating the clini-
tors and inducers of certain cytokines for conditions such as cal utility of antiendotoxin antibodies have been disappointing.
chronic autoimmune diseases, inflammatory disorders, and Multiple possibilities exist as potential sources for this lack of
infectious complications. Additionally, inhibition of endotoxin efficacy. Variability in patient selection, timing of treatment,
and the inflammatory response continues to show progress. and nonstandardization of preparations all have been investi-
Utilizing these advances, progression to targeted cytokine gated as potential sources. Additionally, variability in lipid A
therapy and inhibition for liver, biliary, and pancreatic diseases structure among gram-negative bacteria adds to the difficulty
certainly appears to be a logical direction for future research. of developing an effective treatment. This is supported in find-
The utility of cytokine treatment for a variety of hepatic ings that predicted mortality risk appears to be species depen-
diseases has been known for some time. Pegylated IFN-α, with dent (Hurley & Opal, 2013). Despite early setbacks, further
an elimination half-life of 75 hours, has been used in combina- investigation into delivery of antiendotoxin as subunit vaccines
tion treatment of chronic hepatitis C for more than 15 years. would appear to be an attractive option, given the relative lack
This breakthrough demonstrated the applicability of cytokine of new antimicrobials coming to market (Cross, 2014).
therapy in hepatic diseases. However, due to the pleiotropic Increased success has been observed with modulators of
nature of these cytokines, inhibition or induction of individual downstream mediators of the inflammatory process, particu-
targets proves difficult as a treatment for hepatic, biliary, and larly in the treatment of rheumatoid arthritis. Currently, five
pancreatic diseases. Interactions between multiple networks TNF-blocking agents have received U.S. Food and Drug
and pathways appear to be the driving force behind disease Administration approval, with promising results. TNF-α block-
progression, rather than alterations by an individual mediator. ers infliximab (Remicade) and adalimumab (Humira) have also
However, utilizing the known effects of these cytokines to been utilized in the treatment of complicated Crohn disease.
enhance immune responses is an area of current study. Additionally, a synthetic IL-1R antagonist, anakinra (Kineret),
Cytokine-induced killer cells—autologous lymphocytes incu- is currently in use for treatment of rheumatoid arthritis.
bated with cytokines such as IL-1 and IL-2—have shown favor- As further knowledge of the mechanisms of these agents is
able effects in the treatment of hepatocellular carcinoma in gained in the future, their applicability in hepatic, biliary, and
phase 1 and 2 studies (Ma et al, 2012). Current phase 3 studies pancreatic disorders will likely become increasingly evident.
are now underway. Although the complex interactions of multiple cytokines and
Targeting of Toll-like receptors as a means to achieve anti- pathways make targeting specific actors in hepatobiliary and
neoplastic effects in patients with pancreatic ductal adenocar- pancreatic diseases quite difficult, it is exactly this complexity
cinoma is also a current area of study. Phase 1/2 trials of that provides many untapped potential opportunities for inter-
macrophage-activating lipopeptide-2, a synthetic agonist of vention in the future.
TLR2 and TLR6, have demonstrated promising results in
R2-resected tumors (Schmidt et al, 2007). TLR2 synthetic References are available at expertconsult.com.
Chapter 11 Cytokines in liver, biliary, and pancreatic disease 200.e1
Diao TJ, et al: Protective effect of nitric oxide on hepatopulmonary

REFERENCES syndrome from ischemia-reperfusion injury, World J Gastroenterol
18(25):3310–3316, 2012.
Abshagen K, et al: Loss of NF-kappaB activation in Kupffer cell- Dinarello CA: Infection, fever, and exogenous and endogenous pyro-
depleted mice impairs liver regeneration after partial hepatectomy, gens: some concepts have changed, J Endotoxin Res 10(4):201–222,
Am J Physiol Gastrointest Liver Physiol 292(6):G1570–G1577, 2007. 2004.
Akcam FZ, et al: Cytokine levels and histopathology in chronic hepa- Dinarello CA: Immunological and inflammatory functions of the
titis B and chronic hepatitis C, J Interferon Cytokine Res 32(12):570– interleukin-1 family, Annu Rev Immunol 27:519–550, 2009.
574, 2012. Divanovic S, et al: Negative regulation of Toll-like receptor 4 signaling
Akiyama T, et al: Serum and bile interleukin 6 after percutaneous by the Toll-like receptor homolog RP105, Nat Immunol 6(6):571–
transhepatic cholangio-drainage, Hepatogastroenterology 45(21):665– 578, 2005.
671, 1998. Dodd GD 3rd, et al: Radiofrequency thermal ablation: computer anal-
Alexopoulou L, et al: Recognition of double-stranded RNA and activa- ysis of the size of the thermal injury created by overlapping ablations,
tion of NF-kappaB by Toll-like receptor 3, Nature 413(6857):732– AJR Am J Roentgenol 177(4):777–782, 2001.
738, 2001. Du Plessis J, et al: Activated intestinal macrophages in patients with
Aranda F, et al: Trial watch: Toll-like receptor agonists in oncological cirrhosis release NO and IL-6 that may disrupt intestinal barrier
indications, Oncoimmunology 3:e29179, 2014. function, J Hepatol 58(6):1125–1132, 2013.
Atik E, et al: Inducible nitric oxide synthase and histopathological cor- Elinav E, et al: Inflammation-induced cancer: crosstalk between
relation in chronic viral hepatitis, Int J Infect Dis 12(1):12–15, 2008. tumours, immune cells and microorganisms, Nat Rev Cancer
Badger SA, et al: Cytokine response to portal endotoxaemia and neu- 13(11):759–771, 2013.
trophil stimulation in obstructive jaundice, Eur J Gastroenterol Fausto N, et al: Liver regeneration, Hepatology 43(2 Suppl 1):S45–S53,
Hepatol 24(1):25–32, 2012. 2006.
Barnhart BC, et al: CD95 ligand induces motility and invasiveness of Feezor RJ, et al: Genomic and proteomic determinants of outcome in
apoptosis-resistant tumor cells, EMBO J 23(15):3175–3185, 2004. patients undergoing thoracoabdominal aortic aneurysm repair, J
Baumann H, Gauldie J: The acute phase response, Immunol Today Immunol 172(11):7103–7109, 2004.
15(2):74–80, 1994. Fierer J, et al: The role of lipopolysaccharide binding protein in resis-
Bazzoni F, Beutler B: The tumor necrosis factor ligand and receptor tance to Salmonella infections in mice, J Immunol 168(12):6396–
families, N Engl J Med 334(26):1717–1725, 1996. 6403, 2002.
Berber E, et al: Laparoscopic radiofrequency ablation of hepatic Fong YM, et al: The acute splanchnic and peripheral tissue metabolic
tumors: prospective clinical evaluation of ablation size comparing response to endotoxin in humans, J Clin Invest 85(6):1896–1904,
two treatment algorithms, Surg Endosc 18(3):390–396, 2004. 1990.
Berg RD, Garlington AW: Translocation of certain indigenous bacteria Forsythe RM, et al: Lipopolysaccharide-induced enterocyte-derived
from the gastrointestinal tract to the mesenteric lymph nodes and nitric oxide induces intestinal monolayer permeability in an auto-
other organs in a gnotobiotic mouse model, Infect Immun 23(2):403– crine fashion, Shock 17(3):180–184, 2002.
411, 1979. Foulds S, et al: Transcription factor NF kappa B expression and post-
Blindenbacher A, et al: Interleukin 6 is important for survival after surgical organ dysfunction, Ann Surg 233(1):70–78, 2001.
partial hepatectomy in mice, Hepatology 38(3):674–682, 2003. Frances R, et al: Bacterial translocation is downregulated by anti-TNF-
Bohlinger I, et al: DNA fragmentation in mouse organs during endo- alpha monoclonal antibody administration in rats with cirrhosis and
toxic shock, Am J Pathol 149(4):1381–1393, 1996. ascites, J Hepatol 46(5):797–803, 2007.
Brenndorfer ED, et al: Anti-tumor necrosis factor alpha treatment pro- Froon AH, et al: Lipopolysaccharide toxicity-regulating proteins in
motes apoptosis and prevents liver regeneration in a transgenic bacteremia, J Infect Dis 171(5):1250–1257, 1995.
mouse model of chronic hepatitis C, Hepatology 52(5):1553–1563, Fujimoto M, et al: Plasma endotoxin and serum cytokine levels in
2010. patients with alcoholic hepatitis: relation to severity of liver distur-
Carswell EA, et al: An endotoxin-induced serum factor that causes bance, Alcohol Clin Exp Res 24(Suppl 4):48s–54s, 2000.
necrosis of tumors, Proc Natl Acad Sci U S A 72(9):3666–3670, Fukuzawa K, et al: Expression of fas/fas ligand (fasL) and its involve-
1975. ment in infiltrating lymphocytes in hepatocellular carcinoma (HCC),
Carvalho FA, et al: Toll-like receptor-gut microbiota interactions: J Gastroenterol 36(10):681–688, 2001.
perturb at your own risk! Annu Rev Physiol 74:177–198, 2012. Garlanda C, et al: Intestinal inflammation in mice deficient in Tir8, an
Chang CS, et al: Small intestine dysmotility and bacterial overgrowth inhibitory member of the IL-1 receptor family, Proc Natl Acad Sci U
in cirrhotic patients with spontaneous bacterial peritonitis, Hepatol- S A 101(10):3522–3526, 2004.
ogy 28(5):1187–1190, 1998. Ge G, Greenspan DS: BMP1 controls TGFbeta1 activation via cleav-
Chen CJ, et al: Identification of a key pathway required for the sterile age of latent TGFbeta-binding protein, J Cell Biol 175(1):111–120,
inflammatory response triggered by dying cells, Nat Med 13(7):851– 2006.
856, 2007. Genesca J, et al: Increased tumour necrosis factor alpha production in
Chen T, et al: Role of nitric oxide in liver injury, Curr Mol Med 3(6): mesenteric lymph nodes of cirrhotic patients with ascites, Gut
519–526, 2003. 52(7):1054–1059, 2003.
Chuang TH, Ulevitch RJ: Triad3A, an E3 ubiquitin-protein ligase Giakoustidis DE, et al: Attenuation of liver ischemia/reperfusion
regulating Toll-like receptors, Nat Immunol 5(5):495–502, 2004. induced apoptosis by epigallocatechin-3-gallate via down-regulation
Cid MC, et al: Identification of haptoglobin as an angiogenic factor in of NF-kappaB and c-Jun expression, J Surg Res 159(2):720–728,
sera from patients with systemic vasculitis, J Clin Invest 91(3):977– 2010.
985, 1993. Gioannini TL, Weiss JP: Regulation of interactions of gram-negative
Colletti LM, et al: Role of tumor necrosis factor-alpha in the patho- bacterial endotoxins with mammalian cells, Immunol Res 39(1–
physiologic alterations after hepatic ischemia/reperfusion injury in 3):249–260, 2007.
the rat, J Clin Invest 85(6):1936–1943, 1990. Goggins M, et al: Genetic alterations of the transforming growth factor
Collins AS, et al: Hepatitis C virus (HCV)-induced suppressor of cyto- beta receptor genes in pancreatic and biliary adenocarcinomas,
kine signaling (SOCS) 3 regulates proinflammatory TNF-alpha Cancer Res 58(23):5329–5332, 1998.
responses, J Leukoc Biol 96(2):255–263, 2014. Gressner AM, et al: Roles of TGF-beta in hepatic fibrosis, Front Biosci
Corssmit EP, et al: Effects of interferon-alpha (IFN-alpha) administra- 7:d793–d807, 2002.
tion on leucocytes in healthy humans, Clin Exp Immunol 107(2):359– Grewal IS: Overview of TNF superfamily: a chest full of potential
363, 1997. therapeutic targets, Adv Exp Med Biol 647:1–7, 2009.
Cressman DE, et al: Liver failure and defective hepatocyte regeneration Hallatschek W, et al: Inhibition of hepatic transcriptional induction of
in interleukin-6-deficient mice, Science 274(5291):1379–1383, 1996. lipopolysaccharide-binding protein by transforming-growth-factor
Cross AS: Anti-endotoxin vaccines: back to the future, Virulence beta 1, Eur J Immunol 34(5):1441–1450, 2004.
5(1):219–225, 2014. Hartmann P, et al: Toll-like receptor 2-mediated intestinal injury and
Desbarats J, Newell MK: Fas engagement accelerates liver regeneration enteric tumor necrosis factor receptor I contribute to liver fibrosis in
after partial hepatectomy, Nat Med 6(8):920–923, 2000. mice, Gastroenterology 143(5):1330–1340.e1331, 2012.
Hayashi N, Mita E: Fas system and apoptosis in viral hepatitis, J Gas- Ma Y, et al: Cytokine-induced killer (CIK) cell therapy for patients
troenterol Hepatol 12(9–10):S223–S226, 1997. with hepatocellular carcinoma: efficacy and safety, Exp Hematol
Heumann D, et al: Competition between bactericidal/permeability- Oncol 1(1):11, 2012.
increasing protein and lipopolysaccharide-binding protein for lipo- Martinon F, Tschopp J: Inflammatory caspases: linking an intracellular
polysaccharide binding to monocytes, J Infect Dis 167(6):1351–1357, innate immune system to autoinflammatory diseases, Cell
1993. 117(5):561–574, 2004.
Hines IN, et al: Role of nitric oxide in liver ischemia and reperfusion Massague J: TGFbeta in Cancer, Cell 134(2):215–230, 2008.
injury, Mol Cell Biochem 234-235(1–2):229–237, 2002. Matsuo Y, et al: Interleukin-1alpha secreted by pancreatic cancer cells
Huang DC, et al: Activation of Fas by FasL induces apoptosis by a promotes angiogenesis and its therapeutic implications, J Surg Res
mechanism that cannot be blocked by Bcl-2 or Bcl-x(L), Proc Natl 153(2):274–281, 2009.
Acad Sci U S A 96(26):14871–14876, 1999. McKim SE, et al: Inducible nitric oxide synthase is required in alcohol-
Hurley JC, Opal SM: Prognostic value of endotoxemia in patients with induced liver injury: studies with knockout mice, Gastroenterology
gram-negative bacteremia is bacterial species dependent, J Innate 125(6):1834–1844, 2003.
Immun 5(6):555–564, 2013. McPartland JL, et al: Apoptosis in chronic viral hepatitis parallels his-
Huynh AS, et al: Novel toll-like receptor 2 ligands for targeted pancre- tological activity: an immunohistochemical investigation using anti-
atic cancer imaging and immunotherapy, J Med Chem 55(22):9751– activated caspase-3 and M30 cytodeath antibody, Int J Exp Pathol
9762, 2012. 86(1):19–24, 2005.
Iimuro Y, et al: NFkappaB prevents apoptosis and liver dysfunction Medvedev AE, et al: Tolerance to microbial TLR ligands: molecular
during liver regeneration, J Clin Invest 101(4):802–811, 1998. mechanisms and relevance to disease, J Endotoxin Res 12(3):133–
Jang JW, et al: Serum interleukin-6 and C-reactive protein as a prog- 150, 2006.
nostic indicator in hepatocellular carcinoma, Cytokine 60(3):686– Melhem A, et al: Anti-fibrotic activity of NK cells in experimental liver
693, 2012. injury through killing of activated HSC, J Hepatol 45(1):60–71,
Javle M, et al: Biomarkers of TGF-beta signaling pathway and prog- 2006.
nosis of pancreatic cancer, PLoS ONE 9(1):e85942, 2014. Michalopoulos GK, DeFrances MC: Liver regeneration, Science
Jin X, et al: Interleukin-6 inhibits oxidative injury and necrosis after 276(5309):60–66, 1997.
extreme liver resection, Hepatology 46(3):802–812, 2007. Michelsen KS, Arditi M: Toll-like receptors and innate immunity in
Kamari Y, et al: Lack of interleukin-1alpha or interleukin-1beta inhibits gut homeostasis and pathology, Curr Opin Hematol 14(1):48–54,
transformation of steatosis to steatohepatitis and liver fibrosis in 2007.
hypercholesterolemic mice, J Hepatol 55(5):1086–1094, 2011. Minter RM, et al: Altered Kupffer cell function in biliary obstruction,
Kawai T, et al: Unresponsiveness of MyD88-deficient mice to endo- Surgery 138(2):236–245, 2005.
toxin, Immunity 11(1):115–122, 1999. Miyake K: Innate immune sensing of pathogens and danger signals
Kim YM, et al: Nitric oxide protects cultured rat hepatocytes from by cell surface Toll-like receptors, Semin Immunol 19(1):3–10,
tumor necrosis factor-alpha-induced apoptosis by inducing heat 2007.
shock protein 70 expression, J Biol Chem 272(2):1402–1411, Mochizuki K, et al: Fas antigen expression in liver tissues of patients
1997. with chronic hepatitis B, J Hepatol 24(1):1–7, 1996.
Kim SJ, et al: Role of heme oxygenase 1 in TNF/TNF receptor- Moshage H: Cytokines and the hepatic acute phase response, J Pathol
mediated apoptosis after hepatic ischemia/reperfusion in rats, Shock 181(3):257–266, 1997.
39(4):380–388, 2013. Moss ML, et al: Cloning of a disintegrin metalloproteinase that pro-
Kim YM, et al: Nitric oxide prevents tumor necrosis factor alpha- cesses precursor tumour-necrosis factor-alpha, Nature 385(6618):
induced rat hepatocyte apoptosis by the interruption of mitochon- 733–736, 1997.
drial apoptotic signaling through S-nitrosylation of caspase-8, Nagai Y, et al: Essential role of MD-2 in LPS responsiveness and TLR4
Hepatology 32(4 Pt 1):770–778, 2000. distribution, Nat Immunol 3(7):667–672, 2002.
Kimura A, Kishimoto T: IL-6: regulator of Treg/Th17 balance, Eur J Nakagawa H, et al: Impact of serum levels of interleukin-6 and adipo-
Immunol 40(7):1830–1835, 2010. nectin on all-cause, liver-related, and liver-unrelated mortality in
Kompan L, et al: Effects of early enteral nutrition on intestinal perme- chronic hepatitis C patients, J Gastroenterol Hepatol 30(2):379–388,
ability and the development of multiple organ failure after multiple 2015.
injury, Intensive Care Med 25(2):157–161, 1999. Netea MG, et al: Differential requirement for the activation of the
Konno R, et al: Serum soluble fas level as a prognostic factor in patients inflammasome for processing and release of IL-1beta in monocytes
with gynecological malignancies, Clin Cancer Res 6(9):3576–3580, and macrophages, Blood 113(10):2324–2335, 2009.
2000. Ng KK, et al: Safety limit of large-volume hepatic radiofrequency abla-
O’Reilly LA, et al: Membrane-bound Fas ligand only is essential for tion in a rat model, Arch Surg 141(3):252–258, 2006.
Fas-induced apoptosis, Nature 461(7264):659–663, 2009. Nicholson DW, Thornberry NA: Caspases: killer proteases, Trends
Lee KH, et al: The role of receptor internalization in CD95 signaling, Biochem Sci 22(8):299–306, 1997.
EMBO J 25(5):1009–1023, 2006. Norman DA, et al: Water and electrolyte movement and mucosal mor-
Leist M, et al: Tumor necrosis factor-induced hepatocyte apoptosis phology in the jejunum of patients with portal hypertension, Gastro-
precedes liver failure in experimental murine shock models, Am J enterology 79(4):707–715, 1980.
Pathol 146(5):1220–1234, 1995. Oe S, et al: Intact signaling by transforming growth factor beta is not
Levin M, et al: Recombinant bactericidal/permeability-increasing required for termination of liver regeneration in mice, Hepatology
protein (rBPI21) as adjunctive treatment for children with severe 40(5):1098–1105, 2004.
meningococcal sepsis: a randomised trial. rBPI21 Meningococcal Ohishi W, et al: Serum interleukin-6 associated with hepatocellular
Sepsis Study Group, Lancet 356(9234):961–967, 2000. carcinoma risk: a nested case-control study, Int J Cancer 134(1):154–
Li C, et al: Early activation of transcription factor NF-kappaB during 163, 2014.
ischemia in perfused rat heart, Am J Physiol 276(2 Pt 2):H543– O’Neil S, et al: Obstructive jaundice in rats results in exaggerated
H552, 1999. hepatic production of tumor necrosis factor-alpha and systemic and
Li X, et al: Immunologic response to tumor ablation with irreversible tissue tumor necrosis factor-alpha levels after endotoxin, Surgery
electroporation, PLoS ONE 7(11):e48749, 2012. 122(2):281–286, discussion 286–287, 1997.
Lin RS, et al: Endotoxemia in patients with chronic liver diseases: Pascual S, et al: Intestinal permeability is increased in patients with
relationship to severity of liver diseases, presence of esophageal advanced cirrhosis, Hepatogastroenterology 50(53):1482–1486,
varices, and hyperdynamic circulation, J Hepatol 22(2):165–172, 2003.
1995. Peter ME, et al: The CD95 receptor: apoptosis revisited, Cell
Liu P, et al: Role of endogenous nitric oxide in TNF-alpha and IL-1beta 129(3):447–450, 2007.
generation in hepatic ischemia-repefusion, Shock 13(3):217–223, Pinzone MR, et al: Microbial translocation in chronic liver diseases, Int
2000. J Microbiol 2012:694629, 2012.
Llovet JM, et al: Translocated intestinal bacteria cause spontaneous Quetier I, et al: Hepatitis B virus HBx protein impairs liver regenera-
bacterial peritonitis in cirrhotic rats: molecular epidemiologic evi- tion through enhanced expression of IL-6 in transgenic mice, J
dence, J Hepatol 28(2):307–313, 1998. Hepatol 59(2):285–291, 2013.
Chapter 11 Cytokines in liver, biliary, and pancreatic disease 200.e3
Rachakonda V, et al: Stratification of risk of death in severe acute Unlu S, et al: Phosphorylation of IRF8 in a pre-associated complex
alcoholic hepatitis using a panel of adipokines and cytokines, Alcohol with Spi-1/PU.1 and non-phosphorylated Stat1 is critical for LPS
Clin Exp Res 38(11):2712–2721, 2014. induction of the IL1B gene, Mol Immunol 44(13):3364–3379, 2007.
Raghuraman S, et al: Characterization of soluble FAS, FAS ligand and Wang Y, et al: Mechanisms of hepatoprotection by nitric oxide, Ann N
tumour necrosis factor-alpha in patients with chronic HCV infection, Y Acad Sci 962:415–422, 2002.
J Clin Virol 34(1):63–70, 2005. Wanner GA, et al: Differential effect of anti-TNF-alpha antibody on
Rahman MA, et al: Impact of PTEN expression on the outcome of proinflammatory cytokine release by Kupffer cells following liver
hepatitis C virus-positive cirrhotic hepatocellular carcinoma patients: ischemia and reperfusion, Shock 11(6):391–395, 1999.
possible relationship with COX II and inducible nitric oxide syn- Weber A, et al: Interleukin-1 (IL-1) pathway, Sci Signal 3(105):cm1,
thase, Int J Cancer 100(2):152–157, 2002. 2010.
Reiberger T, et al: Non-selective betablocker therapy decreases intes- Weiss J: Bactericidal/permeability-increasing protein (BPI) and
tinal permeability and serum levels of LBP and IL-6 in patients with lipopolysaccharide-binding protein (LBP): structure, function and
cirrhosis, J Hepatol 58(5):911–921, 2013. regulation in host defence against Gram-negative bacteria, Biochem
Reinhard C, et al: Tumor necrosis factor alpha-induced activation of Soc Trans 31(Pt 4):785–790, 2003.
c-jun N-terminal kinase is mediated by TRAF2, EMBO J 16(5): Welborn MB, et al: The relationship between visceral ischemia, proin-
1080–1092, 1997. flammatory cytokines, and organ injury in patients undergoing
Roberts NJ, et al: Systemic use of tumor necrosis factor alpha as an thoracoabdominal aortic aneurysm repair, Crit Care Med 28(9):3191–
anticancer agent, Oncotarget 2(10):739–751, 2011. 3197, 2000.
Rodrigo R, et al: Alterations in soluble guanylate cyclase content and Werman A, et al: The precursor form of IL-1alpha is an intracrine
modulation by nitric oxide in liver disease, Neurochem Int 45(6):947– proinflammatory activator of transcription, Proc Natl Acad Sci U S
953, 2004. A 101(8):2434–2439, 2004.
Sakurai T, et al: Hepatocyte necrosis induced by oxidative stress and Wink DA, et al: Antioxidant effects of nitric oxide, Methods Enzymol
IL-1 alpha release mediate carcinogen-induced compensatory pro- 301:413–424, 1999.
liferation and liver tumorigenesis, Cancer Cell 14(2):156–165, 2008. Woo M, et al: In vivo evidence that caspase-3 is required for Fas-
Sandahl TD, et al: Hepatic macrophage activation and the LPS mediated apoptosis of hepatocytes, J Immunol 163(9):4909–4916,
pathway in patients with alcoholic hepatitis: a prospective cohort 1999.
study, Am J Gastroenterol 109(11):1749–1756, 2014. Wrighton KH, et al: Phospho-control of TGF-beta superfamily signal-
Sass G, et al: Inducible nitric oxide synthase is critical for immune- ing, Cell Res 19(1):8–20, 2009.
mediated liver injury in mice, J Clin Invest 107(4):439–447, 2001. Wu FP, et al: Effects of major liver resection, with or without recom-
Schell SR, et al: Pro- and antiinflammatory cytokine production after binant bactericidal/permeability-increasing protein (rBPI21), on the
radiofrequency ablation of unresectable hepatic tumors, J Am Coll angiogenic profile of patients with metastatic colorectal carcinoma,
Surg 195(6):774–781, 2002. J Surg Oncol 84(3):137–142, 2003.
Schmidt J, et al: Intratumoural injection of the toll-like receptor-2/6 Xaus J, et al: LPS induces apoptosis in macrophages mostly through
agonist ‘macrophage-activating lipopeptide-2’ in patients with pan- the autocrine production of TNF-alpha, Blood 95(12):3823–3831,
creatic carcinoma: a phase I/II trial, Br J Cancer 97(5):598–604, 2000.
2007. Xing Z, et al: IL-6 is an antiinflammatory cytokine required for control-
Schneider P, et al: Conversion of membrane-bound Fas(CD95) ligand ling local or systemic acute inflammatory responses, J Clin Invest
to its soluble form is associated with downregulation of its proapop- 101(2):311–320, 1998.
totic activity and loss of liver toxicity, J Exp Med 187(8):1205–1213, Xu Y, et al: NF-kappaB inactivation converts a hepatocyte cell line
1998. TNF-alpha response from proliferation to apoptosis, Am J Physiol
Shi Y, Massague J: Mechanisms of TGF-beta signaling from cell mem- 275(4 Pt 1):C1058–C1066, 1998.
brane to the nucleus, Cell 113(6):685–700, 2003. Yamada Y, et al: Initiation of liver growth by tumor necrosis factor:
Shimizu T, et al: Gene expression during liver regeneration after partial deficient liver regeneration in mice lacking type I tumor necrosis
hepatectomy in mice lacking type 1 tumor necrosis factor receptor, factor receptor, Proc Natl Acad Sci U S A 94(4):1441–1446, 1997.
J Surg Res 152(2):178–188, 2009. Yamamoto M, et al: Role of adaptor TRIF in the MyD88-independent
Shin EC, et al: Expression of Fas-related genes in human hepatocel- toll-like receptor signaling pathway, Science 301(5633):640–643,
lular carcinomas, Cancer Lett 134(2):155–162, 1998. 2003.
Sodeman T, et al: Bile salts mediate hepatocyte apoptosis by increasing Yamamoto M, et al: TIR domain-containing adaptors define the speci-
cell surface trafficking of Fas, Am J Physiol Gastrointest Liver Physiol ficity of TLR signaling, Mol Immunol 40(12):861–868, 2004.
278(6):G992–G999, 2000. Yang R, et al: IL-6 is essential for development of gut barrier dysfunc-
Sorkin P, et al: Systemic leakage and side effects of tumor necrosis tion after hemorrhagic shock and resuscitation in mice, Am J Physiol
factor alpha administered via isolated limb perfusion can be manipu- Gastrointest Liver Physiol 285(3):G621–G629, 2003.
lated by flow rate adjustment, Arch Surg 130(10):1079–1084, 1995. Yang L, et al: NF-kappaB activation in Kupffer cells after partial hepa-
Suda T, et al: Molecular cloning and expression of the Fas ligand, a tectomy, Am J Physiol Gastrointest Liver Physiol 289(3):G530–G538,
novel member of the tumor necrosis factor family, Cell 75(6):1169– 2005.
1178, 1993. Yang RB, et al: Signaling events induced by lipopolysaccharide-
Tachibana S, et al: Interleukin-6 is required for cell cycle arrest and activated toll-like receptor 2, J Immunol 163(2):639–643, 1999.
activation of DNA repair enzymes after partial hepatectomy in mice, Yilmaz B, et al: Prognostic utility of serum neopterin in obstructive
Cell Biosci 4(1):6, 2014. jaundice secondary to malignant lesions treated by percutaneous
Thomas CJ, et al: Evidence of a trimolecular complex involving LPS, transhepatic biliary drainage, J Vasc Interv Radiol 24(6):865–872,
LPS binding protein and soluble CD14 as an effector of LPS 2013.
response, FEBS Lett 531(2):184–188, 2002. Yoneyama K, et al: The expression of Fas and Fas ligand, and the
Tracey KJ, Cerami A: Tumor necrosis factor: a pleiotropic cytokine effects of interferon in chronic liver diseases with hepatitis C virus,
and therapeutic target, Annu Rev Med 45:491–503, 1994. Hepatol Res 24(4):327–337, 2002.
True AL, et al: Activation of NF-kappaB induced by H(2)O(2) and Yoshida N, et al: Improvement of the survival rate after rat massive
TNF-alpha and its effects on ICAM-1 expression in endothelial cells, hepatectomy due to the reduction of apoptosis by caspase inhibitor,
Am J Physiol Lung Cell Mol Physiol 279(2):L302–L311, 2000. J Gastroenterol Hepatol 22(11):2015–2021, 2007.
Ugurel S, et al: Increased soluble CD95 (sFas/CD95) serum level cor- Zimmermann A: Regulation of liver regeneration, Nephrol Dial Trans-
relates with poor prognosis in melanoma patients, Clin Cancer Res plant 19(Suppl 4):iv6–iv10, 2004.
7(5):1282–1286, 2001. Zorena K, et al: Relationship between serum transforming growth
Ulevitch RJ, Tobias PS: Recognition of gram-negative bacteria and factor beta 1 concentrations and the duration of type 1 diabetes
endotoxin by the innate immune system, Curr Opin Immunol mellitus in children and adolescents, Mediators Inflamm 2013:849457,
11(1):19–22, 1999. 2013.
CHAPTER 12
Infections in hepatic, biliary, and pancreatic surgery
Nicholas Spinelli, Matthew S. Strand, and Ryan C. Fields
Infections are a significant cause of morbidity and mortality in Chapter 43, 72, and 73). These same organisms may also pass
the surgical patient. Furthermore, infectious complications through the liver to access the systemic circulation, potentially
occurring in the postoperative setting have become a topic of leading to sepsis and septic shock. These pathologic states do
scrutiny in both the medical and public arenas. This increased not normally occur in the healthy human because of defense
awareness about infections and other perioperative complica- mechanisms that exist at multiple organ-system levels. In this
tions has been driven by major quality outcome initiatives, such section, we are interested in those mechanisms specific to the
as the National Surgical Quality Improvement Program hepatobiliary system and pancreas.
(NSQIP) sponsored by the American College of Surgeons Pathogenic organisms may enter the liver by two major
(ACS) (http://site.acsnsqip.org). It is necessary for physicians pathways. The first is the portal venous system, which delivers
and other health care providers to understand the nature of blood from the GI tract. In addition to absorbed nutrients,
infection in the surgical patient and how best to prevent it. portal venous blood also contains enteric bacteria and toxins
Caring for patients with hepatobiliary and/or pancreatic that may result in illness if not appropriately cleared by the liver.
disease can be quite challenging with regard to infection, The second pathway into the liver is via retrograde entry
especially in patients undergoing major hepatic or pancreatic through the biliary system. Organisms already present within
resection after preoperative biliary instrumentation. Although the systemic circulation may also enter the liver via a third
the hepatobiliary and pancreatic (HPB) surgical subspecialty pathway—the arterial inflow. Intravenous drug use and endo-
has evolved greatly during the past several decades, the morbid- carditis are two important mechanisms by which this can occur
ity following these procedures continues to be significant. (Fig. 12.1).
Infectious complications, in particular, are an important con- Both the liver and the biliary system contain important
tributor to this morbidity. Modern 30 day surgical-site infection defense mechanisms that prevent enteric organisms from
(SSI) rates after HPB procedures can be as high as 20% to 40% establishing infection within the liver (i.e., hepatic abscess) and/
(Ceppa et al, 2013). Infection is also associated with increased or reaching the systemic circulation. Generally, defense mecha-
hospital stay, operative times, transfusions, blood loss, intensive nisms are categorized into three types: physical, chemical, and
care unit use, and readmission rates (Kent et al, 2013). In immunologic (Table 12.1). The single most important mecha-
addition to these short-term sequelae, long-term sequelae nism within the liver is the immunologic defense provided by
include possible delay in adjuvant therapy. Kupffer cells. Kupffer cells are phagocytic cells located in the
The goal of this chapter is to understand the range of infec- sinusoids of the liver. They are derived from circulating mono-
tious complications that may accompany resections of the liver, cytes and represent almost 90% of the tissue macrophages
biliary tree, and pancreas and to gain an appreciation for the present in the human body. Kupffer cells are constantly exposed
important surgery-specific risk factors for infection. We also to enteric organisms and endotoxins, leading to their activation
review potential ways of mitigating infection risk at the preop- as phagocytes. Once activated, Kupffer cells also release various
erative, operative, and postoperative levels of patient care. With cytokines, prostanoids, nitric oxide, and reactive oxygen species
a general lack of level I evidence to guide clinical decisions, that contribute to their immune function (see Chapters 7, 10,
there continues to be controversy surrounding some of these and 11). Additionally, Kupffer cells indirectly regulate the
putative risk-lowering methods. phenotype of surrounding hepatocytes, stellate cells, endothelial
We begin this chapter with a brief review of the normal host cells, and other immune cells present in the liver (Bilzer et al,
defenses preventing the entry of infectious organisms into the 2006). Without Kupffer cells, enteric organisms and their toxins
liver, biliary system, and pancreas. We then specifically define entering the liver through portal venous blood flow would freely
postoperative infection, which can be divided into SSIs and pass into the systemic circulation. Reduction of the Kupffer cell
remote-site infections. A short discussion of the general risk mass via liver resection is an important consideration when
factors for the development of SSIs and the general measures evaluating a patient’s potential for postoperative infectious
that can be used to prevent them is also warranted before morbidity.
delving into the surgery-specific risks that have been identified Portal hypertension, whether from presinusoidal, sinusoidal,
in the HPB surgery literature. or postsinusoidal causes, undermines the ability of the Kupffer
cells to clear portal venous inflow of enteric organisms and
NORMAL HOST DEFENSES toxins. Portal venous obstruction results in the shunting of
portal venous blood to the systemic venous circulation through
Bacteria that are normally present in the gastrointestinal (GI) the development of collateral pathways. Depending on the
tract can invade the liver and biliary tree directly, resulting in degree of shunting and portal hypertension, this may allow a
hepatic abscess and ascending cholangitis, respectively (see significant portion of portal venous blood to bypass Kupffer
201
Exogenous nonenteric Enteric organisms also exist within the biliary system. Tight junctions prevent
organisms in the Gl tract mixing of blood and bile within the liver by forming a physical
barrier between the bile canaliculi and hepatic sinusoids. Biliary
ductule epithelial cells each contain a single cilium that helps
Hepatic artery Portal vein via Biliary system move bile toward the extrahepatic ducts (Gilroy et al, 1995;
via systemic portal venous via sphincter Itoshima et al, 1977). Mucus produced by the extrahepatic bile
circulation system of Oddi ducts prevents prolonged contact between the ductal epithelium
and bacteria. Both of these features promote forward flow of
bile out of the liver, thus preventing retrograde ascent of any
Liver bacteria that have entered the biliary system.
Bile salts are a chemical barrier to infection and have several
important properties (see Chapter 8). First, they have both
Hepatic veins bacteriostatic and bactericidal properties (Stewart et al, 1986;
Sung et al, 1993) capable of maintaining the normal balance
of intestinal flora. Second, they have a trophic effect on the
Systemic circulation intestinal epithelium, allowing it to maintain its own effective
Portosystemic barrier against bacterial translocation (Stewart et al, 1986).
shunting in
Last, bile salts have an antiendotoxin effect (Stewart et al,
portal HTN
1986).
FIGURE 12.1. There are multiple routes by which microorganisms can
The biliary system also contains immunologically active
access the liver. Exogenous bacteria present in the bloodstream from
substances that combat infection, including immunoglobulin A
intravenous drug abuse or endocarditis travel to the liver through the
hepatic arterial system, whereas enteric bacteria typically enter the liver (Emmrich et al, 1998; Scott-Conner & Grogan, 1994), fibro-
via the portal venous system or through the biliary system. These enteric nectin (Wilton et al, 1987), and complement factors (Sumiyoshi
organisms can then invade the systemic circulation if not eradicated by et al, 1997).
host defense mechanisms present in the liver. Patients with portosys- As seen in Table 12.1, the pancreas contains similar defense
temic shunting are at increased risk for enteric bacteremia because the mechanisms that prevent its invasion by enteric organisms. An
liver’s defense mechanisms are bypassed as portal blood is diverted important difference is that there are no Kupffer cell analogues
directly into the systemic circulation. GI, Gastrointestinal; HTN, in the pancreas. The reason for this lies in the fact that the
hypertension. pancreas does not represent a “window” to the systemic circula-
tion. Its venous outflow enters the portal system that ultimately
enters the liver. Thus, unlike the liver, it does not represent a
TABLE 12.1 Normal Host Defense Mechanisms in the defense against enteric microorganisms attempting to enter the
Hepatobiliary System and Pancreas systemic circulation.
Hepatobiliary Pancreas
DEFINING INFECTIOUS COMPLICATIONS
Physical Biliary sphincter Pancreatic sphincter
Hepatic tight junctions Pancreatic tight Defining postoperative infectious complications can be accom-
junctions plished by considering two broad categories of infection: SSI
Bile flow Pancreatic juice flow and remote-site infection.
Mucus Mucus
Cilia Cilia
Surgical-Site Infection
Chemical Bile salts Pancreatic fluid The Centers for Disease Control and Prevention (CDC)
Immunologic Kupffer cells — further defines SSI based on the physical level of infection
(Mangram et al, 1999). Superficial incisional SSIs involve
Immunoglobulin A Immunoglobulin A
infections of the skin and/or subcutaneous tissue, whereas deep
Fibronectin —Complement
Complement incisional SSIs refer to those infections located in the muscle/
fascial layer of the abdominal wall. Infections located within the
abdominal cavity define organ/space SSIs (Fig. 12.2). A detailed
definition of each type of SSI is included in Box 12.1. Note
cells, thus increasing the possibility of sepsis (see Fig. 12.1) (see that these definitions are also used by the American College of
Chapters 76 and 81). Surgeons in defining SSI for the NSQIP program.
The biliary system has a number of ways to prevent retro-
grade invasion into the liver by pathogenic microorganisms. Remote-Site Infections
These defense mechanisms span all categories listed earlier. Remote-site infections include those infections distant from the
From a physical barrier standpoint, the sphincter of Oddi is incision and operative field: respiratory tract infections, urinary
effective at preventing bacterial ascent into the liver. However, tract infections, and catheter-related bloodstream infections.
removal of this mechanism frequently occurs in HPB operations
and other procedures (e.g., endoscopic retrograde cholangio- GENERAL RISK FACTORS FOR THE
pancreatography with sphincterotomy and/or stenting) and is an DEVELOPMENT OF SURGICAL-SITE INFECTION
important contributor to postoperative and periprocedural
infectious morbidity. Cholangitis with septic shock is the most There are two important sources of risk regarding the develop-
serious complication. Microscopic physical defense mechanisms ment of a postoperative SSI. Both patient-specific factors and
Chapter 12 Infections in hepatic, biliary, and pancreatic surgery 203
BOX 12.1 Defining Surgical-Site Infection (SSI)

Superficial Incisional SSI
An infection that occurs within 30 days after the operation AND
involves only the skin or subcutaneous tissue of the incision, AND
Skin
the patient has at least ONE of the following:
Superficial
a. Purulent drainage, with or without laboratory confirmation,
incisional
SSI from the superficial incision.
Subcutaneous
b. Organisms isolated from an aseptically obtained culture of
tissue
fluid or tissue from the superficial incision.
c. At least ONE of the following signs or symptoms: pain or
Deep tenderness, localized swelling, redness, or heat, AND
Deep soft tissue incisional superficial incision is deliberately opened by surgeon,
(fascia and SSI
muscle) UNLESS incision is culture negative.
d. Diagnosis of a superficial incisional SSI by the surgeon or
attending physician.
Organ/space Organ/space
SSI
Deep Incisional SSI
An infection that occurs within 30 days after the operation AND
FIGURE 12.2. Surgical-site infections (SSIs), as defined by the Centers involves deep soft tissues of the incision (e.g., fascial and muscle
for Disease Control and Prevention, are classified according to the layers), AND the patient has at least ONE of the following:
physical level of infection. Superficial incisional SSIs involve infections of a. Purulent drainage from the deep incision but not from the
the skin and/or subcutaneous tissue, whereas deep incisional SSIs refer organ/space component of the surgical site.
to those infections located in the muscle/fascial layer of the abdominal b. A deep incision that spontaneously dehisces or is
wall. Infections located within the actual abdominal cavity define organ/ deliberately opened by a surgeon when the patient has at
space SSIs. (Modified from Mangram AJ, et al: Guideline for prevention least ONE of the following signs or symptoms: fever
of SSI, 1999. Hospital Infection Control Practices Advisory Committee. (>38° C), localized pain, or tenderness, UNLESS site is
Infect Control Hosp Epidemiol 20(4):250-278, 1999; quiz 279-280.) culture negative.
c. An abscess or other evidence of infection involving the
deep incision that is detected on direct examination, during
reoperation, or by histopathologic examination or imaging
test.
procedure-related factors combine to yield varying degrees of d. Diagnosis of a deep incisional SSI by the surgeon or
risk. attending physician.
Patient-Related Risk Organ/Space SSI

A recent review discusses the important risk factors generated An infection that occurs within 30 days after the operation AND
involves any part of the anatomy (e.g., organs or spaces), other than
by the patient (Kirby et al, 2009). Some of these risk factors
the incision, which was opened or manipulated during the operation,
may not be modifiable at the time of operation, but it is impor- AND the patient has at least ONE of the following:
tant to be aware of them while the patient receives care. Specifi- a. Purulent drainage from a drain that is placed into the organ/
cally, patient-related risk factors for SSI include age, nutritional space.
status, diabetes, smoking, obesity, coexisting infections at a b. Organisms isolated from an aseptically obtained culture of
remote body site, colonization with microorganisms, altered fluid or tissue in the organ/space.
immune response, and length of preoperative stay. Many of c. An abscess or other evidence of infection involving the
these patient-related risk factors are taken into account with organ/space that is detected on direct examination, during
use of the ACS-NSQIP Surgical Risk Calculator (http:// reoperation, or by histopathologic examination or imaging
www.riskcalculator.facs.org). This online tool sponsored by the test.
d. Diagnosis of an organ/space incisional SSI by the surgeon
ACS is designed to estimate the probability of a postoperative
or attending physician.
complication, including infectious morbidity, based on a par-
ticular procedure. It may be an important clinical adjunct for Modified from Mangram AJ, et al: Guideline for prevention of SSI, 1999. Hospital Infection
both surgeon and patient alike in the preoperative decision- Control Practices Advisory Committee. Infect Control Hosp Epidemiol 20(4):250-278, 1999; quiz
279-280.
making process.
The authors of the aforementioned review also include a
comprehensive discussion detailing the general measures that
can be used to prevent SSI. It cannot be overemphasized that Procedure-Related Risk
the mitigation of infectious risk occurs at all phases of patient The ultimate source for a SSI is the patient’s endogenous
care (Tables 12.2 to 12.4). microbial flora. It must also be kept in mind that any patient
Many of these general preventive measures were incorpo- can experience an infectious complication from the introduc-
rated into a perioperative surgical care bundle and applied to tion of outside microbes into the wound through a break in
patients undergoing pancreaticoduodenectomy in a retrospec- sterile technique. However, for the purposes of this section, we
tive study conducted at Thomas Jefferson University. The are interested in how the nature of the operation itself might
results showed a lower rate of wound infections in patients contribute to the risk of infection. In the most general sense,
subjected to the surgical care bundle (7.7%) versus those who this is perhaps best understood by examining the CDC wound
were not (15%). This difference was statistically significant classification system (Mangram et al, 1999). Wounds are cate-
(P =.01) (Lavu et al, 2012). gorized into four different classes (Box 12.2). With an increase
TABLE 12.2 General Preoperative Interventions to Prevent Surgical-Site Infection

Intervention Evidence References
Reduce hemoglobin A1c levels to <7% before operation Class II data Anderson et al, 2008
Smoking cessation 30 days before operation Class II data Anderson et al, 2008; Mangram et al, 1999
Administer specialized nutritional supplements or enteral nutrition Class I and class II Anderson et al, 2008; Anonymous, 1991;
to patients at severe nutritional risk for 7-14 days preoperatively; data with Mangram et al, 1999; Weimann et al, 2006
preoperative parenteral nutrition should not be routinely used, significant
except selectively in patients with severe underlying malnutrition heterogeneity
Adequately treat preoperative infections, such as urinary tract Class II data Anderson et al, 2008; Mangram et al, 1999
infections
Decolonization of unselected patients with mupirocin is not Class I data Anderson et al, 2008; Kalmeijer et al, 2002;
currently recommended Konvalinka et al, 2006; Laupland & Conly,
2003; Mangram et al, 1999; Perl et al, 2002;
Suzuki et al, 2003
Identification and decolonization of Staphylococcus aureus carriers Limited class I data Hacek et al, 2008; Rao et al, 2008
may be a potentially useful intervention, but requires further
investigation
Preoperative showering with chlorhexidine. Not currently Class I data Anderson et al, 2008; Mangram et al, 1999;
recommended Webster & Osborne, 2007, 2012
Modified from Teresa C. Horan, Robert P. Gaynes, William J. Martone, William R. Jarvis and T. Grace Emori: CDC definitions of nosocomial surgical site infections: a modification of CDC definitions of
surgical wound infections, Infection Control 13: 606-608, 1992.
TABLE 12.3 General Perioperative Interventions to Prevent Surgical-Site Infection

Remove hair only if it will interfere with the operation; hair Class I data Anderson et al, 2008; Bratzler & Hunt, 2006;
removal by clipping immediately prior to the operation or Kjonniksen et al, 2002; Mangram et al,
with depilatories; no preoperative or perioperative shaving 1999; Springer, 2007
of surgical site
Use an antiseptic surgical scrub or alcohol-based hand Class II data Anderson et al, 2008; Mangram et al, 1999
antiseptic for preoperative cleansing of the operative team
members’ hands and forearms
Prepare the skin around the operative site with an Class II data Anderson et al, 2008; Digison, 2007;
appropriate antiseptic agent, including preparations based Mangram et al, 1999
on alcohol, chlorhexidine, or iodine/iodophors
Administer prophylactic antibiotics for most clean- Strong class I data Anonymous, 1999; Bratzler & Hunt, 2006;
contaminated and contaminated procedures, and selected Classen et al, 1992; Mangram et al, 1999;
clean procedures; use antibiotics appropriate for the Springer, 2007
potential pathogens
Administer prophylactic antibiotics within 1 hr before incision Strong class II data Anonymous, 1999; Bratzler & Hunt, 2006;
(2 hr for vancomycin and fluoroquinolones Classen et al, 1992; Mangram et al, 1999,
Springer, 2007
Use higher dosages of prophylactic antibiotics for morbidly Limited class II data Forse et al, 1989; Mangram et al, 1999
obese patients
Use vancomycin as a prophylactic agent only when there is a Class I data Anderson et al, 2008; Anonymous, 1999;
significant risk of MRSA infection Bolon et al, 2004; Finkelstein et al, 2002;
Mangram et al, 1999
Provide adequate ventilation, minimize operating room traffic, Class II and class III Anderson et al, 2008; Mangram et al, 1999
and clean instruments and surfaces with approved data
disinfectants
Avoid “flash” sterilization Class II data Anderson et al, 2008; Mangram et al, 1999
Carefully handle tissue, eradicate dead space, and adhere to Class III data Anderson et al, 2008; Mangram et al, 1999
standard principles of asepsis
Leave contaminated or dirty infected wounds open, with the Limited class I, class II Brasel et al, 1997; Cohn et al, 2001;
possible exception of wounds following operations for data Mangram et al, 1999
perforated appendicitis
Redose prophylactic antibiotics with short half-lives Limited class I, class II Mangram et al, 1999; Scher, 1997;
intraoperatively if operation is prolonged (for cefazolin if data Swoboda et al, 1996
operation > 3 hr) or if there is extensive blood loss
Maintain intraoperative normothermia Class I data, some Anderson et al, 2008; Barone et al, 1999;
contradictory class Bratzler & Hunt, 2006; Kurz et al, 1996;
II data Mangram et al, 1999; Sessler & Akca,
2002; Springer, 2007; Walz et al, 2006
MRSA, Methicillin-resistant Staphylococcus aureus.
Modified from Kirby JP, Mazuski JE: Prevention of SSI. Surg Clin North Am 89(2):365-389, 2009, viii.
TABLE 12.4 General Postoperative Interventions to Prevent Surgical-Site Infection

Discontinue prophylactic antibiotics within 24 hours after the Class I data Anonymous, 1999; Barie, 2000; Bratzler & Hunt,
procedure (48 hr for cardiac surgery and liver transplant 2006; DiPiro et al, 1986; Mangram et al, 1999,
procedures); preferably, discontinue prophylactic McDonald et al, 1998; Springer, 2007
antibiotics after skin closure
Maintain serum glucose levels < 200 mg/dL on postoperative Class II data Anderson et al, 2008; Bratzler & Hunt, 2006; Carr
days 1 and 2 et al, 2005; Furnary et al, 1999; Lazar et al, 2004;
Springer, 2007; Zerr et al, 1997
Monitor wound for the development of surgical-site infection Class III data Anderson et al, 2008; Mangram et al, 1999
Modified from Kirby JP, Mazuski JE: Prevention of SSI. Surg Clin North Am 89(2):365-389, 2009, viii.
organ/space SSI. Organ/space infection can be more specifically

BOX 12.2 Surgical Wound Classification
categorized into (1) intraabdominal perihepatic abscess and
Class I Wounds (2) intraabdominal extrahepatic abscess (see Chapter 27).
Considered clean wounds. They are defined by an uninfected opera- Furthermore, these patients are also at risk for remote-site
tive wound in which no inflammation is encountered and the respira-
infections, including bloodstream infections, respiratory tract
tory, alimentary, genital, or uninfected urinary tract is not entered.
infections, and urinary tract infections. Risk of infection
Therefore the only microorganisms that are likely to be introduced
into the wound are those from the skin. exists at the preoperative, operative, and postoperative levels
(Table 12.5).
Class II Wounds
Known as clean-contaminated wounds. They include wounds in Preoperative Risk Mitigation
which the respiratory, alimentary, genital, or urinary tracts are entered Postoperative infectious complications after hepatic resection
under controlled conditions and without unusual contamination. have been linked to the preoperative state of the liver, the
Specifically, operations involving the biliary tract, appendix, vagina, amount of liver parenchyma resected, and the development of
and oropharynx are included in this category, provided no evidence postoperative liver dysfunction. Yang and colleagues (2014)
of infection or major break in technique is encountered. With class
found cirrhosis and hepatolithiasis to be independent preopera-
II wounds and major abdominal operations, in addition to the skin
tive risk factors for the development of postoperative SSIs (see
flora, there is also the added contamination from the organisms that
are colonizing the luminal structures of the gastrointestinal tract. Chapter 44). Hepatolithiasis as a risk factor for SSI has also
been confirmed by Uchiyama and colleagues (2011). Garwood
Class III Wounds and colleagues (2004) gleaned that the extent of hepatic resec-
These are contaminated wounds. They involve open, fresh, acciden- tion is associated with postoperative infectious complications.
tal wounds. Operations with gross spillage from the gastrointestinal This study also identified advanced age and multiple comorbid
tract and incisions in which acute, nonpurulent inflammation is conditions as important preoperative contributors to infectious
encountered are included here. morbidity. Schindl and colleagues (2005) established a relation-
ship between the extent of resection, residual liver volume, and
Class IV Wounds
the development of infection. Although a precise residual liver
Known as dirty infected wounds. They are defined by old traumatic
volume to predict postoperative infection could not be found,
wounds with retained devitalized tissue and those that involve exist-
ing clinical infection or perforated viscera. there was a significant relationship linking severe hepatic dys-
function and postoperative infection. Furthermore, severe
Modified from Mangram AJ, et al: Guideline for prevention of SSI, 1999. Hospital Infection hepatic dysfunction could be predicted by small residual liver
Control Practices Advisory Committee. Infect Control Hosp Epidemiol 20(4):250-278, 1999; quiz volume and high body mass index (BMI). Nanashima and
279-280.
colleagues (2014) examined factors associated with SSI in a
retrospective analysis of posthepatectomy patients. Liver failure
in wound classification level, there is an increase in the risk of was significantly associated with deep SSIs.
infectious complications, which is explained by the increased Hepatic resection removes Kupffer cell mass, which is the
exposure of the surgical field to the patient’s endogenous liver’s principal mechanism for clearing the portal inflow of
microorganisms. enteric microorganisms and their associated toxins. Hepatic
In the next section, we more closely examine the risk of resection also results in a decrease in bile production and the
postoperative infection specific to the HPB surgery patient associated beneficial chemical and immunologic effects that
undergoing a particular procedure. come from bile salts. Resection with biliary reconstruction will
also negate the physical barrier of the sphincter of Oddi, further
contributing to infectious risk. Depending on the extent of
DISCUSSION OF SURGERY-SPECIFIC RISK resection, a normal healthy liver in a reasonable surgical can-
FACTORS FOR POSTOPERATIVE INFECTIOUS didate may be able to compensate for the compromise in these
COMPLICATIONS defenses. However, this may not be the case for the patient with
diseased liver parenchyma. The preoperative mitigation of the
Hepatic Resection risk of postoperative infectious complications after hepatic
The range of SSI following hepatic resection includes the previ- resection therefore begins with a full appreciation of the preex-
ously defined superficial incisional SSI, deep incisional SSI, and isting condition of the patient’s liver. Adjunctive assessments
TABLE 12.5 Risk Factors for Infectious Complications After Hepatectomy
206
Aim/Objective/End
Author, Year Title Study Type Patients Point Population Results and Conclusions
Yanaga et al, Intraperitoneal Retrospective 149 To examine Elective hepatectomy The rate of intraperitoneal septic complications after
1986 septic perioperative risk hepatectomy was 12.8%. Chi-square tests showed that
complications factors for significant risk factors for intraperitoneal sepsis after
after intraperitoneal hepatectomy were right or extended right lobectomy, age
hepatectomy septic > 65, operative duration > 5 hr, blood loss > 3 L, and
complications postoperative bleeding requiring laparotomy. The authors
after hepatectomy concluded that limiting operative time and blood loss
were important to prevent postoperative intraperitoneal
sepsis after hepatectomy.
Garwood et al, Infectious Retrospective 207 To assess the Hepatectomy The overall rate of infection after hepatectomy was 3.3%;
2004 complications characteristics of however, mortality among this subset was 33% (3 of 9
after hepatic surgical infections patients). All deaths from infection after hepatectomy had
resection after hepatic MRSA isolated from cultures. Operative duration was
resection to significantly longer in patients suffering infection (7.9 vs.
identify factors 5.6 hours, P = .02) compared with those without.
accounting for Although operative duration was the only statistically
increased significant risk factor for postoperative infection, the
postoperative authors concluded that advanced age, comorbid
mortality conditions, and extent of hepatic resection contribute to
infectious risk after hepatectomy.
Schindl et al, The value of Prospective 104 To examine the Elective hepatectomy The rate of postoperative infection was 31.7%. Twenty-two
2005 residual liver relationship patients had a residual liver volume <26.6%; 11
volume as a between residual developed severe hepatic dysfunction postoperatively,
PART 1 LIVER, BILIARY, AND PANCREATIC ANATOMY AND PHYSIOLOGY
predictor of liver volume and and 11 did not. Eight of 11 (72.7%) patients with
hepatic postoperative postoperative severe hepatic dysfunction developed
dysfunction after complications infection, whereas only 2 of 11 (18.2%) without severe
major liver after hepatectomy hepatic dysfunction developed infection (P = .030). The
resection authors concluded that the risk of postoperative infection
is related more strongly to the incidence of postoperative
liver dysfunction rather than residual liver volume.
Togo et al, 2007 Perioperative Retrospective 535 To assess risk Hepatectomy with or The overall rate of posthepatectomy infection declined over
infection control factors for SSI without biliary or four periods of time, from 44.7% to 24.1% to 15.0% to
and its after hepatectomy intestinal 9.2%. Independent risk factors for SSI after hepatectomy
effectiveness in over four different reconstruction included use of silk suture (ARR, 4.56) and bile leak
hepatectomy time periods (ARR, 22.99). The authors advocated careful operative
patients procedure, performance of cholangiography, checking for
bile leak using indigo carmine, and the use of synthetic
absorbable sutures.
Pessaux et al, Randomized clinical Prospective 200 To assess the value Elective hepatectomy Pneumonia occurred more frequently in the patients
2007 trial evaluating of routine receiving nasogastric tube decompression (13.0% vs.
the need for nasogastric tube 5.0%, P = .047). There were no differences in mortality,
routine decompression length of stay, or return of bowel function. The authors
nasogastric after elective concluded that routine nasogastric decompression after
decompression hepatic resection hepatectomy had no advantages and that it increased
after elective the risk of pulmonary complications.
hepatic resection
Togo et al, 2008 Usefulness of Prospective rat 313 To assess the Hepatic resection In the animal model, the number of bacteria in silk used
absorbable model, usefulness of without biliary during hepatectomy was 1000 times that of Vicryl. In the
sutures in retrospective synthetic reconstruction patient analysis, the respective incidences of SSI and
preventing SSI in human absorbable infection on the cut surface of the liver in the Vicryl group
hepatectomy analysis sutures (Vicryl) in (3.2%, 1.6%) were significantly lower than in the silk
preventing SSI group (11.2%, 8.8%; P = .0045). The adjusted risk ratio
after hepatectomy (ARR) for SSI in the silk group was 3.4. Operative length
(ARR, 7.31) and bile fistula (ARR, 17.5) were also
independent predictors of SSI. The authors concluded
that the use of synthetic absorbable sutures, instead of
silk sutures, can prevent the development of SSI in
patients undergoing hepatectomy.
Okabayashi et al, Risk factors and Retrospective 152 To identify risk Hepatectomy The overall rate of postoperative infection was 14.5%.
2009b predictors for SSI factors and the Independent risk factors for SSI were BMI > 23.6 (OR,
after hepatic predictors for the 3.7), EBL > 810 mL (OR, 4.4), postoperative bile leak
resection prevention of SSI (OR, 5.2), and postoperative sliding scale insulin use (OR,
after hepatectomy 1.2) as opposed to artificial pancreas. The authors
concluded that postoperative hyperglycemia, bile leak,
obesity, and high intraoperative blood loss correlated with
postoperative infectious risk, and that postoperative
artificial pancreas use is a safe and beneficial device for
maintaining glycemic control.
Kobayashi et al, Risk factors of SSI Retrospective 405 To compare the Hepatectomy without The overall incidence of SSI was 5.8%. Independent risk
2009 after clinicopathologic biliary reconstruction factors for SSI were intraoperative bowel injury (OR,
hepatectomy for factors between 20.08), blood loss > 2000 mL (OR, 4.4), and age > 65 yr
liver cancers patients who (OR, 2.4). The authors concluded that administering
developed SSI prophylactic antibiotics to patients with risk factors for
and those who infection after hepatectomy might reduce infectious
did not after morbidity.
hepatectomy for
liver cancer
Uchiyama et al, Risk factors for Prospective 308 To clarify the Hepatectomy The overall rate of SSI was 11.0%. SSI was significantly
2011 postoperative collection, incidence of SSIs more common after resection for hepatolithiasis compared
infectious retrospective after hepatectomy to resection for HCC or metastasis (23.8% vs. 11.3% vs.
complications analysis 2.7% respectively, P < .001). The authors concluded that
after hepatectomy postoperative SSIs correlate with infected bile.
Moreno Predictors of SSI Prospective 2332 To identify predictors Segmental or The overall rate of postoperative infection ranged from 9.7%
Elola-Olaso after liver resection: of SSI after liver hemihepatectomy, for segmentectomy to 18.3% for trisectionectomy.
et al, 2012 a multicentre resection trisectionectomy Independent predictors for SSI included serum Na > 145
analysis using (OR, 6.05), hypoalbuminemia per mg/dL from mean (OR,
National Surgical 0.67), serum bilirubin > 1.0 (OR, 1.7), need for dialysis
Quality (OR, 4.46 for deep space infection), longer operative time
Improvement (OR, 1.003/min over mean) and smoking status within
Program data 1 yr of procedure (OR, 1.66 for deep space SSI).
Arikawa et al, Risk factors for SSI Retrospective 171 To elucidate risk Hepatectomy for HCC The overall rate of SSI was 21%. Independent risk factors
2011 after factors for SSI for the development of SSI after hepatectomy were bile
hepatectomy for after hepatectomy leakage and excessive blood loss. The authors
hepatocellular for cancer concluded that preventing bile leaks was paramount to
carcinoma limiting infectious morbidity after hepatectomy.
Harimoto et al, Prospective Prospective, 125 To assess the effect Hepatic resection The overall rate of SSI was 13.6%. Independent risk factors
2011 randomized randomized of suture material without biliary for SSI were blood loss > 1000 mL (OR, 7.6). Although
controlled trial on SSI after reconstruction there was no statistically significant difference in SSI rate
Chapter 12 Infections in hepatic, biliary, and pancreatic surgery
investigating the hepatectomy for silk versus Vicryl use (15.8% vs. 11.3%), recovery
type of sutures from infection was faster in the Vicryl group (28 vs. 54
used during days), P < .05.
hepatectomy
207
Continued
TABLE 12.5 Risk Factors for Infectious Complications After Hepatectomy—cont’d
208
Aim/Objective/End
Author, Year Title Study Type Patients Point Population Results and Conclusions
Pessaux et al, Identification and Prospective 555 To determine the Elective hepatectomy The overall rate of postoperative infectious complications
2013 validation of risk collection, risk factors for was 24.3%. Independent risk factors for postoperative
factors for retrospective postoperative infectious complications were nasogastric tube (OR, 1.8),
postoperative analysis infectious blood transfusion (OR, 1.9), and diabetes (OR, 2.4). The
infectious complications only independent risk factor for a surgical postoperative
complications following partial infection was portal vein resection (OR, 5.5). Risk factors
following hepatectomy for a medical postoperative infectious complication were
hepatectomy preop biliary drainage (OR, 1.9), blood transfusion (OR,
2.1), diabetes mellitus (OR, 2.9), and atrial fibrillation (OR,
3.6). The authors concluded that addressing the above
risk factors could reduce the incidence of postoperative
infectious complications after hepatectomy.
Nakahira et al, Proposal for a Prospective 1926 To evaluate Hepatopancreatobiliary The overall rate of infectious complications after
2013 subclassification collection, differences procedures, hepatopancreatobiliary procedures was 23.2%.
of hepatobiliary- retrospective among excluding Independent risk factors for SSI after hepatectomy were
pancreatic analysis hepatobiliary cholecystectomy drain placement (OR, 2.8) and operative duration > 5 hr
operations for procedures to (OR, 2.9). The authors concluded that hepatectomy and
SSI surveillance determine the nonhepatectomy hepatopancreatobiliary procedures differ
following optimal with respect to the incidence of SSI and should be
assessment of subdivision for assessed separately for the purposes of SSI surveillance.
results of SSI surveillance
prospective
multicenter data
Sadamori et al, Risk factors for Retrospective 359 To investigate the Hepatectomy for HCC The overall rate of SSI after hepatectomy was 14.5%,
2013 organ/space SSI causative without including 6.7% incisional SSI and 8.6% organ/space SSI.
after bacteria, reconstruction Independent risk factors for SSI after hepatectomy were
hepatectomy for management, operative time > 280 minutes (OR, 2.32), repeat
HCC in 359 outcome, and hepatectomy (OR, 3.43), blood transfusion (OR, 7.56 for
recent cases characteristics of incisional SSI), and bile leakage (OR, 3.01 for organ/
SSIs after space SSI).
hepatectomy
Nanashima et al, Associated factors Retrospective 526 To clarify the factors Hepatectomy The overall incidence of SSI after hepatectomy ranged from
2014, with SSIs after associated with 5% to 8% depending on the time period. Independent
hepatectomy: posthepatectomy risk factors for superficial SSI included not using a
predictions and SSI vessel-sealing device (OR, 3.1). Independent risk factors
countermeasures for deep SSI were male gender (OR, 2.0), hepatic failure
by a retrospective (OR, 3.3), and bile leakage (OR, 4.8).
cohort study
Kurmann et al, Hepatic steatosis is Retrospective 231 To investigate the Liver or colorectal The overall rate of SSI was 29.3%. The SSI rate in steatotic
2014 associated with impact of hepatic resection patients was 47.5% compared with nonsteatotics 26.6%
SSI after hepatic steatosis on SSI (P < .05). Hepatic steatosis was an independent risk
and colorectal in patients that factor for SSI in patients undergoing hepatectomy (OR,
surgery underwent open 10.33) or colectomy (OR, 6.67).
abdominal
surgery
Yang et al, 2014 Risk factors of SSI Retrospective 7388 To assess risk Hepatectomy The overall rate of SSI was 9.4%. On multivariate analysis,
after hepatic factors for SSI the only independent risk factors for both incisional and
resection after hepatectomy organ/space infections were hepatolithiasis (OR, 1.58
and 1.66, respectively), cirrhosis (OR, 1.61 and 1.70,
respectively), and blood transfusion (OR, 1.32 and 1.71,
respectively).
BMI, Body mass index; EBL, estimated blood loss; HCC, hepatocellular carcinoma; OR, odds ratio; SSI, surgical-site infection.
such as liver biopsy and measurement of portal venous pres- developed tool allows the surgeon to estimate the probability
sures may be necessary where there is uncertainty concerning of unfavorable outcomes, including postoperative infectious
the health of the liver. complications.
The extent of parenchymal resection is addressed in this
section on preoperative risk mitigation because the amount of Operative Risk Mitigation
liver to be resected is generally addressed before the patient is There are several operative risk factors that are associated with
brought to the operating room (see Chapters 100 and 108). A postoperative infectious complications. Factors that have come
full assessment will allow the surgeon to carefully tailor the to light in the recent literature include bile leakage, length of
appropriate procedure to a particular patient. Should the risk surgery, increased blood loss, and intraoperative bowel injury
of postoperative hepatic dysfunction and its related complica- (Arikawa et al, 2011; Kobayashi et al, 2009; Moreno Elola-
tion of postoperative infection be deemed too high for formal Olaso et al, 2012; Nakahira et al, 2013; Nanashima et al, 2014;
resection, then thoughtful consideration of other treatment Okabayashi et al, 2009b; Togo et al, 2007). Therefore, in an
modalities may be needed. For example, parenchymal-sparing attempt to mitigate these risk factors, it is important to expedi-
techniques, such as segmental hepatectomy, ablation, or even tiously progress through the operation by using meticulous
arterial-based modalities, may be required. Preoperative portal technique to avoid excess blood loss and iatrogenic injury to
vein embolization can be considered in certain patients in surrounding structures. It is also essential to pay particular
whom the anticipated residual liver is questionable. When the attention to the identification and treatment of intraoperative
risk of postoperative complications is prohibitive, then not bile leaks once the liver is transected. For example, some sur-
operating or ablating may be the prudent course of action. geons routinely coat the cut surface of the liver with hydrogen
The use of systemic chemotherapy should also be taken peroxide. Small bile leaks then become more obvious at the site
into account in the overall treatment plan for patients undergo- of green bubbling.
ing hepatic resection, especially for an indication of colorectal An intraoperative air leak test was recently found to be very
liver metastasis. Neoadjuvant chemotherapy, in particular, can effective in the detection of bile leaks, thus decreasing the rate
theoretically increase the risk of infection due to its negative of postoperative biliary complications (Zimmitti et al, 2013).
effects on the liver, including steatosis, steatohepatitis, and This maneuver involves the placement of a transcystic catheter
sinusoidal obstruction syndrome, making it less than ideal from that is used to inject air into the biliary tree after the upper
the standpoint of postoperative infection. Its use continues to abdomen is submerged in saline and the distal common bile
be debated; however, a recent retrospective study conducted by duct is occluded. Bile leaks are identified at the site of streaming
Scilletta and colleagues (2014) suggested that neoadjuvant air bubbles and are directly repaired. The authors compared
chemotherapy was not a significant risk factor for SSIs in the rates of postoperative biliary complications between 103
patients undergoing liver resection for colorectal hepatic metas- patients who underwent air leak testing and 120 matched
tases (see Chapters 71 and 100). patients who underwent hepatic resection before air leak testing
Nordlinger and colleagues (2008) conducted a randomized was used. None of the hepatic resections in either group were
controlled trial comparing liver resection for resectable colorec- accompanied by biliary reconstruction. The authors noted a
tal liver metastases in patients with and without perioperative significantly lower rate of postoperative bile leaks in the air
chemotherapy. Perioperative chemotherapy was defined as six leak–tested group (1.9% vs. 10.8%, P = .008). This adjunctive
cycles of FOLFOX4 (5-fluororuracil plus leucovorin and maneuver seems easy to use and appears effective in the preven-
oxaliplatin) before and after surgery. There were 182 patients tion of bile leaks. This may have a positive impact on postopera-
within each arm of the study. Infectious complications that were tive infectious complications after hepatic resection but requires
analyzed included wound infection, intraabdominal infection, further validation.
and urinary infection. There was a trend toward higher rates of Regarding parenchymal transection techniques, no one
these complications in the perioperative chemotherapy group, method or combination of methods has been shown to be
but it was not statistically significant. Therefore it would seem superior. Therefore it is recommended that the surgeon use the
that for resectable colorectal cancer liver metastasis, chemo- technique that is most familiar, while limiting the amount of
therapy before hepatectomy is safe from a postoperative infec- necrotic liver parenchyma left behind (Yanaga et al, 1986). We
tious standpoint. also believe that it is important to suction any pooled blood and
Other preoperative contributors to postoperative infectious bile at the end of the operation (see Chapter 103).
complications after hepatic resection include advanced age, The use of absorbable suture material versus silk during
presence of diabetes mellitus, obesity, presence of an open hepatectomy has also been studied with regard to infectious
wound, hypernatremia, hypoalbuminemia, elevated serum bili- complication. The data are conflicting. In a retrospective analy-
rubin, dialysis, comorbid conditions, repeat hepatectomy, and sis, silk suture was found to be significantly associated with SSI
hepatic steatosis (Garwood et al, 2004; Kurmann et al, 2014; (Togo et al, 2007). This was confirmed in a study conducted
Moreno Elola-Olaso et al, 2012; Okabayashi et al, 2009b; on rats (Togo et al, 2008). However, a prospective randomized
Pessaux et al, 2013; Sadamori et al, 2013; Togo et al, 2007). controlled trial found no difference between silk and Vicryl
Note that many of these preexisting conditions may not be suture with regard to infection (Harimoto et al, 2011).
modifiable before the time of operation. Therefore it cannot be
overemphasized that the nature of the planned liver resection Perioperative Antibiotics
be considered in light of the general condition of the patient. The Clinical Practice Guidelines for Antimicrobial Prophylaxis
The ACS’s Risk Calculator (http://www.riskcalculator.facs.org) published in the American Journal of Health-System Pharmacy
can also provide some valuable insight into the expected post- (Bratzler et al, 2013) represent the most current recommenda-
operative recovery of an individual patient and may help with tions on preoperative antimicrobial prophylaxis. They do not
preoperative decision making. As discussed earlier, this recently include specific recommendations for hepatic resection with or
without biliary-enteric reconstruction, or cases in which the of prophylactic drainage in GI surgery in a systematic review
patient has undergone preoperative biliary drainage. The guide- and meta-analysis, concluding that many GI operations can be
lines, however, do provide recommendations on biliary tract safely performed without the use of drains. Regarding liver
procedures, including cholecystectomy, exploration of the surgery specifically, this paper suggests that surgical drains do
common bile duct, and choledochoenterostomy. The guidelines not necessarily prevent biloma formation and do not always
state that the most common organisms associated with infection prevent the need for percutaneous drainage. A grade A recom-
after these particular procedures include Escherichia coli, Kleb- mendation was given against prophylactic drainage in elective
siella species, and enterococci. The recommended antibiotic for hepatic resection. This is supported by several randomized
open biliary tract surgery is cefazolin; alternative agents are studies (Belghiti et al, 1993, Fong et al, 1996, Liu et al, 2004)
ampicillin-sulbactam, cefotetan, cefoxitin, and ceftriaxone. It is and a systematic review (Gurusamy et al, 2007). A more recent
important to note, however, that none of the cephalosporins multicenter international prospective study also concluded that
cover enterococci (see Chapters 12, 29, and 30). intraoperatively placed surgical drains do not prevent the need
A reasonable approach to the prevention of SSI in the setting for additional percutaneous drainage (Brooke-Smith et al,
of hepatic resection would be to administer antimicrobial pro- 2015).
phylaxis for all elective hepatic resections regardless of antici-
pated biliary-enteric reconstruction. As mentioned earlier, the Postoperative Risk Mitigation
recommended antimicrobial agent for biliary tract procedures Nasogastric Decompression
from the Clinical Practice Guidelines for Antimicrobial Prophy- Pessaux and colleagues (2007) conducted a randomized clinical
laxis is cefazolin. This agent can theoretically be used for hepatic trial examining the utility of postoperative nasogastric decom-
resection as well. However, we typically broaden our preopera- pression after elective hepatectomy. The authors randomized
tive coverage with a second-generation cephalosporin or a 200 patients to nasogastric tube use versus no nasogastric tube.
carbapenem. We currently favor ertapenem because it is a single The use of a nasogastric tube was significantly associated with
agent with broad coverage, dosed daily. Whatever agent is used, an increased rate of pneumonia and atelectasis but did not
it should be given within 60 minutes of skin incision and reduce overall surgical complications, medical morbidity,
redosed appropriately intraoperatively to maintain adequate in-hospital mortality, duration of ileus, or length of hospital
tissue levels. stay. The authors concluded that routine nasogastric decom-
The Clinical Practice Guidelines for Antimicrobial Prophy- pression offers no benefit. We do not routinely use nasogastric
laxis state that, in general, the “shortest effective duration of tubes in our hepatectomy patients.
antimicrobial administration for preventing SSI is not known,”
but that “postoperative antimicrobial administration is not Early Enteral Nutrition and Synbiotics
necessary for most procedures.” This notion is further sup- The concept of early enteral nutrition has been studied in
ported by a randomized controlled trial recently conducted in patients undergoing liver resection. Richter and colleagues
Japan that explored the merits of postoperative antibiotic pro- (2006) conducted a systematic review of early enteral nutrition
phylaxis after liver resection (Hirokawa et al, 2013). A total of following open liver resection, concluding that it is safe and that
241 hepatic resection patients were studied. Half of the patients it decreases the incidence of postoperative complications in
received postoperative antibiotics, whereas the other half did comparison to parenteral nutrition. The authors noted a statis-
not. There was no difference between the two patient groups tically significant lower rate of wound infections and catheter-
with regard to signs of infection, incidence of the systemic related infections with early enteral feeding versus parenteral
inflammatory response syndrome, infectious complications, nutrition. However, pneumonias and intraabdominal abscesses
SSI, or remote-site infections. The authors concluded that were not significantly decreased. It should be noted that enteral
postoperative antibiotic prophylaxis cannot prevent postopera- nutrition in this review was, in general, started on the second
tive infections after liver resection. We do not routinely continue postoperative day via an operative jejunal tube. We do not
antimicrobial prophylaxis postoperatively. advocate placement of jejunal feeding catheters in the otherwise
healthy liver resection patient. However, for the malnourished
Drains patient, early enteral nutrition perhaps via a planned jejunal
Two types of drains have been studied in the literature. tube is recommended over parenteral nutrition.
Nakayama and colleagues (2014) evaluated the utility of sub- Enteral nutrition supports gut-barrier function, mitigating
cutaneous drains in the prevention of wound infection. After the risk of bacteremia originating from the enteric system
fascial closure, a 10 Fr drain connected to negative pressure (Mizuno et al, 2010). However, a perturbation in the microbial
was placed in the subcutaneous tissue. The authors were not balance within the intestinal lumen can also pave the way for
able to significantly lower the wound infection rate. infections after liver surgery. This is particularly evident in cir-
Intraabdominal drains have been analyzed more systemati- rhotic patients (Yeh et al, 2003). Synbiotic therapy, which is a
cally. The rationale for leaving an intraabdominal drain is to combination of probiotic and prebiotic therapy, has been
detect and prevent biloma formation in the event of bile leakage studied as a method of potentially reducing infectious morbid-
after hepatic resection. Bile leakage and subsequent biloma ity after liver surgery. Probiotics are defined as live bacteria that
formation is an important contributor to infectious complica- are capable of improving the microbial balance within the
tions after hepatic resection, as stated earlier. However, the intestinal lumen. Prebiotics are chemicals that enhance the
overall trend reflected in the literature does not support the use growth of beneficial bacteria. Several studies have documented
of drains in elective hepatic surgery. Foregoing prophylactic encouraging results in the ability of synbiotic therapy to help
drainage after elective hepatic surgery is consistent with the prevent postoperative infectious complications after liver
general notion that drainage may be unnecessary in most GI surgery, especially in patients with biliary cancer (Kanazawa
operations. Petrowsky and colleagues (2004) studied the value et al, 2005; Sugawara et al, 2006; Usami et al, 2011).
Enhanced recovery after surgery (ERAS) pathways have (Farges et al, 2013). The most recent National Comprehensive
shown promise in improving postoperative outcomes in patients Cancer Network (NCCN) guidelines recommend biliary drain-
undergoing liver surgery (Hughes et al, 2014). Important age via endoscopic retrograde cholangiography or percutaneous
components of ERAS pathways include continuing nutrition as transhepatic cholangiography and portal vein embolization in
long as 2 hours before hepatic resection, avoidance of a naso- patients with a very small FLR. Routine drainage is not
gastric tube, and early postoperative diet resumption. Although recommended.
there have been no studies specifically examining the effects of
ERAS pathways on infectious complications, avoidance of Pancreatic Resection
perioperative starvation is theoretically beneficial from an infec- The location and nature of disease within the pancreas dictates
tious standpoint and may one day be proven. ERAS pathways the specific type of pancreatic resection used. With modern
should therefore be considered for those patients undergoing imaging modalities, resection type can often be determined
routine hepatectomy. preoperatively. The most commonly performed resections
include pancreaticoduodenectomy and distal pancreatectomy
Blood Glucose Control (typically with splenectomy when performed for cancer); hence
Since the landmark paper by van Den Berghe and colleagues the studies of infectious complications after pancreatectomy
(2001) demonstrating improved outcomes with intensive focus on these two procedures. Enucleation and central pan-
insulin therapy, there has been much focus on tight blood createctomy are less common procedures.
glucose control in the surgical patient. Dysglycemia has been The range of SSI following pancreaticoduodenectomy or
studied in patients undergoing hepatectomy specifically. Huo distal pancreatectomy includes the previously defined superfi-
and colleagues (2003) demonstrated increased hepatic decom- cial incisional SSI, deep incisional SSI, and organ/space SSI
pensation in diabetic patients undergoing hepatic resection for (see Chapters 27 and 66). For pancreaticoduodenectomy,
hepatocellular carcinoma. Little and colleagues (2002) showed organ/space infection can be further categorized by the contrib-
an association with increased mortality in diabetic patients uting anastomosis: intraabdominal abscesses, or infected fluid
undergoing hepatectomy for colorectal cancer metastasis. collections can be related to an infected bile, pancreatic, or
Regarding postoperative infectious complications in particular, enteric leak, or a combination of these. Organ/space infection
Ambiru and colleagues (2008) demonstrated an increase in SSI after distal pancreatectomy is typically related to pancreatic
in HPB surgery patients with poor postoperative blood glucose leak. As with any major abdominal operation, these patients are
control. Therefore tight glycemic control is paramount after also at risk for the development of abscesses not related to
hepatectomy. Methods of glycemic control include sliding anastomotic leakage or secondary to an infected hematoma.
scales and continuous insulin infusions; however, other less Uncontained intraabdominal sepsis requiring reexploration is
conventional approaches have been used. A closed-loop artifi- also a possibility. Furthermore, these patients can develop
cial pancreas system has been shown to reduce SSI (Okabayashi remote-site infections, including bloodstream infections, chol-
et al, 2009a) as well as liver dysfunction (Okabayashi et al, angitis, respiratory tract infections, urinary tract infections, and
2011) in patients undergoing hepatic resection. Promising Clostridium difficile infections.
results have also been obtained with the hyperinsulinemic- General risk factors for SSI discussed earlier apply to patients
normoglycemic clamp technique. This method was shown to undergoing pancreatic resection; however, there are risk factors
reduce complications, including infections after hepatectomy in specific to pancreatic resection that warrant discussion.
a study by Fisette and colleagues (2012). These are presented according to each stage of patient care:
preoperative, intraoperative, and postoperative. In many
Preoperative Biliary Drainage in the Hilar instances, it may not be possible to significantly modify risk;
Cholangiocarcinoma Patient (See Chapters 27 and 51) nonetheless, it is important to understand potential risk factors
Preoperative biliary drainage before hepatic resection for extra- so that the patient can be monitored with the appropriate
hepatic hilar cholangiocarcinoma is a crucial consideration degree of vigilance in the postoperative period.
when anticipating the surgical approach and postoperative
outcomes. Postoperative outcomes after liver resection tend to Preoperative Risk Mitigation
be worse in patients with obstructive jaundice (Belghiti et al, Body Mass Index and Nutritional Status
2000). Preoperative biliary drainage was therefore initially com- Elevated BMI is a risk factor for infectious complications fol-
monplace before the performance of elective liver resection for lowing pancreatic resection (Table 12.6). House and colleagues
obstructed patients with hilar cholangiocarcinoma. However, (2008) studied postoperative complications in 356 patients who
the procedure is associated with preoperative cholangitis, and underwent pancreaticoduodenectomy for pancreatic adenocar-
relatively recent literature has confirmed that it is also associ- cinoma with the goal of identifying preoperative patient and
ated with increased postoperative complications, infectious radiographic factors associated with postoperative morbidity.
complications in particular (Ferrero et al, 2009; Hochwald Complications developed in 38% of this patient population,
et al, 1999). This has called into question whether or not with the most common pancreatic fistula/abscess, wound infec-
routine preoperative biliary drainage should be performed in tion, and delayed gastric emptying. Wound infection rates were
hilar cholangiocarcinoma patients undergoing hepatic resec- significantly higher in those patients with a BMI greater than
tion. In fact, a recent review by Liu and colleagues (2011) or equal to 30 (P = .03). The authors also determined that the
concluded that preoperative drainage should not be routinely degree of visceral fat as seen on preoperative axial imaging
performed. However, there are data to suggest that there is a correlated with higher rates of overall complications and pan-
role for preoperative biliary decompression in certain scenarios, creatic fistula.
including very small functional liver remnant (FLR) patients Su and colleagues (2010) analyzed 101 pancreaticoduode-
(Kennedy et al, 2009) and perhaps for right-sided resections nectomy patients with regard to postoperative infectious
212
TABLE 12.6 Risk Factors for Infectious Complications After Pancreatectomy: Preoperative Risk Factors
Author, Year Title Study Type Patients Aim/Objective/End Point Population Results and Conclusions
House et al, Preoperative predictors for Prospective 356 Postoperative Pancreaticoduodenectomy The overall complication rate was 38%, most
2008 complications after collection, complications commonly pancreatic fistula or abscess, wound
pancreaticoduodenectomy: retrospective infection, and delayed gastric emptying. Multivariate
impact of BMI and body fat review analysis revealed that absence of pancreatic
distribution atrophy and extent of central obesity (as measured
by visceral fat) were associated with postoperative
complications. Patients with retrorenal visceral fat
>2 cm developed significantly more complications
(51% vs. 31%) and pancreatic fistulas (24% vs
10%). Patients with BMI > 30 kg/m2 had
significantly more wound infections (21% vs. 12%).
Su et al, Factors influencing infectious Prospective 101 Postoperative Pancreaticoduodenectomy Nineteen preoperative variables were correlated with
2010 complications after collection, infectious postoperative infectious morbidity. Infectious
pancreatoduodenectomy retrospective complications complications developed in 56% of patients.
review Serious infection (bacteremia, intraabdominal
infection, pneumonia) occurred significantly more
often in patients with BMI > 25 (67% vs. 24%). BMI
significantly increased operative time and was the
only independent risk factor for postoperative
infection.
Greenblatt Preoperative factors predict Prospective 4945 30 day complication Pancreaticoduodenectomy The overall complication rate was 27.1%. Sepsis
et al, 2011 perioperative morbidity and collection, and mortality rates (15.3%), SSI (13.1%), and respiratory complications
mortality after retrospective including pneumonia (9.5%) were the most

pancreaticoduodenectomy review common complications. Preoperative factors
associated with a significant increase in 30 day
morbidity were BMI > 25, age > 80 yr, male gender,
dependent functional status, COPD, steroid use,
bleeding disorder, leukocytosis, serum creatinine >
1.0, and albumin < 3.4.
Kelly et al, Risk stratification for distal Prospective 2322 30 day complication Distal pancreatectomy The overall complication rate was 28.1%. Sepsis
2011 pancreatectomy using collection, and mortality rates (8.7%), SSI (5.9%), and pneumonia (4.7%) were the
ACS-NSQIP: preoperative retrospective most common complications. Independent
factors predict morbidity review predictors of morbidity included male gender, BMI
and mortality > 30, smoking, steroid use, neurologic disease,
preoperative SIRS/sepsis, serum albumin < 3.4,
serum creatinine > 1.4, and abnormal platelet
count (<50,000 or > 400,000).
La Torre et al, Malnutrition and pancreatic Retrospective 143 SSI, morbidity, Pancreaticoduodenectomy Nutritional status as assessed by the malnutrition
2013 surgery: prevalence and mortality, hospital or distal pancreatectomy universal screening tool (MUST) and nutritional risk
outcomes stay length index (NRI) were independent risk factors for
morbidity after pancreatectomy.
Shinkawa NRI as an independent Retrospective 64 SSI Pancreaticoduodenectomy The overall SSI rate was 33%. Pancreatic fistula (OR
et al, 2013 predictive factor for the 26.8) and NRI < 97.5 (OR, 6.5) were independent
development of SSI after risk factors for SSI.
pancreaticoduodenectomy
ACS-NSQIP, American College of Surgeons–National Surgical Quality Improvement Program; BMI, body mass index; COPD, chronic obstructive pulmonary disease; OR, odds ratio; SIRS, systemic inflammatory response syndrome; SSI, surgical-site infection.
complications. Nineteen perioperative variables were studied in as independent risk factors for SSI. The authors then conducted
an effort to determine which predicted infectious morbidity. a multivariate regression analysis on just preoperative and
The authors noted infectious complications in 55% of the intraoperative risk factors (excluding the postoperative develop-
patients. They concluded that high BMI (defined as BMI > 25) ment of pancreatic fistula). A NRI of less than 97.5 remained
significantly increased operative time and was the only indepen- an independent predictor of SSI.
dent risk factor for serious postoperative infectious complica- As noted above, modification of the aforementioned preop-
tions, including bacteremia, intraabdominal infection, and erative risk factors may be difficult or even impossible prior to
pneumonia. pancreatectomy. This is particularly true if the indication for
Larger multiinstitutional data have yielded similar conclu- resection is cancer or suspicion of cancer, which is common. In
sions about the association between BMI and morbidity follow- these instances, proceeding to the operating room expeditiously
ing pancreatic resection. Greenblatt and colleagues (2011) used may be the prudent course of action, especially in the clearly
the ACS-NSQIP database in an attempt to formulate a predic- resectable and otherwise healthy operative candidate. However,
tion tool for patients undergoing pancreaticoduodenectomy. more than one-third of patients about to undergo pancreatico-
The authors examined preoperative factors that might predict duodenectomy can be considered borderline candidates from a
perioperative morbidity and mortality. Although this study was medical standpoint (Tzeng et al, 2014). These patients are at
not designed to predict who would incur an infectious compli- significant risk for postoperative morbidity (including infectious
cation specifically, the authors found that the most frequent complications) as well as mortality. Therefore, as suggested by
complications after pancreaticoduodenectomy included sepsis Tzeng and colleagues, surgeons should strongly consider
(15.3%), SSI (13.1%), and respiratory complications (9.5%). improving the condition of the patient to mitigate infectious/
The overall complication rate in 1342 patients was 27.1%. overall morbidity and mortality in these “borderline resectable
Elevated BMI was a significant predictor of morbidity (after type C” patients before surgery. Those patients receiving
adjusting for confounding variables), and morbidity increased neoadjuvant therapy should take advantage of this time and use
incrementally with BMI. Other predictors included older age, it as a “window of opportunity” to modify BMI, improve
male gender, dependent functional status, chronic obstructive nutritional/functional status, control hypertension, and/or quit
pulmonary disease, steroid use, bleeding disorder, leukocytosis, smoking. For patients seen with surgically resectable tumors
elevated serum creatinine, and hypoalbuminemia. but significant reversible functional deficits, it may be worth-
Kelly and colleagues (2011) attempted to identify preopera- while to administer neoadjuvant therapy while the patient
tive and operative risk factors for the development of complica- is medically optimized. Regardless of whether neoadjuvant
tions after distal pancreatectomy. The authors also used the therapy consists of chemoradiation or chemotherapy alone,
multiinstitutional prospective ACS-NSQIP database. Their either type of preoperative therapy is considered safe with
efforts concerned the development of a risk score for patients regard to postoperative complications (Araujo et al, 2013;
undergoing distal pancreatectomy. The study population Cheng et al, 2006; Cho et al, 2014; Heinrich et al, 2008).
included 2322 patients. The overall 30 day complication and
mortality rates were 28.1% and 1.2%, respectively. Similar to Preoperative Biliary Drainage
the analysis conducted by Greenblatt and colleagues (2011), Preoperative biliary drainage in the setting of an obstructing
this study was not designed to specifically address infectious pancreatic head mass continues to be debated. Earlier studies
complications. However, the most common complications were suggested that perioperative mortality is higher when pancre-
sepsis, SSI, and pneumonia. Multivariate analysis determined aticoduodenectomy is performed on the hyperbilirubinemic
that high BMI was a preoperative predictor of postoperative patient (Bottger et al, 1999; Braasch et al, 1977; Lerut et al,
morbidity. The other preoperative variables associated with 1984) (see Chapters 29, 30, and 66). More recent work has
postoperative complications included male gender, smoking, also shown preoperative jaundice to be a poor prognostic factor
steroid use, neurologic disease, preoperative systemic inflam- with regard to overall survival for patients undergoing resection
matory response syndrome /sepsis, hypoalbuminemia, elevated of the head of the pancreas for adenocarcinoma (Strasberg
creatinine, and abnormal platelet count. et al, 2014). However, the literature continues to suggest that
Poor preoperative nutritional status is another important attempts at normalizing the bilirubin preoperatively may have
risk factor for SSI and other postoperative morbidity in patients detrimental effects that manifest in the postoperative period.
undergoing pancreatic resection (see Table 12.6) (see Chapter Healthy patients with an intact sphincter of Oddi and a
26). La Torre and colleagues (2013) noticed a relationship normal biliary system have sterile bile for all of the reasons
between malnutrition and morbidity after pancreatic surgery in discussed previously. However, obstructive jaundice in the
their retrospective evaluation of data collected from 143 patients setting of a mass in the head of the pancreas results in bile stasis.
undergoing pancreatic resection for cancer. Malnutrition was This in turn promotes colonization of the biliary system, espe-
defined by using several different validated screening tools. It cially after the bile ducts are interrogated and drained via stents
was gleaned from multivariate analysis that malnutrition, as (Limongelli et al, 2007). The presence of bacteria in the biliary
defined by the malnutrition universal screening tool and the system is known as bacterobilia. When normal host defense
nutritional risk index (NRI), was an independent risk factor mechanisms present in the liver and biliary tree are overwhelmed
for overall morbidity, which included SSI. Shinkawa and col- by a critical level of bacterobilia and the biliary tree is not
leagues (2013) confirmed these findings in an examination of adequately drained, then pathogenic enteric organisms may
64 patients with pancreaticoduodecectomy with regard to reach the systemic circulation through the liver, causing sepsis
potential perioperative risk factors for SSI. SSI, as defined by (i.e., cholangitis).
the CDC, developed in 33% of these patients. Using multivari- It is rare for a patient with pancreatic cancer to be seen with
ate logistic regression analysis on perioperative factors, the cholangitis without having undergone attempts at biliary
authors identified pancreatic fistula and a NRI of 97.5 or less decompression. However, should this happen, the required
treatment is antibiotics with biliary drainage as described in the

Tokyo Guidelines (Gomi et al, 2013; Miura et al, 2013). In the Operative Risk Mitigation
United States, typical drainage procedures include decompres- Preoperative Antibiotics
sion via the percutaneous transhepatic approach or via endo- The current Clinical Practice Guidelines for Antimicrobial
scopic retrograde cholangiography. More commonly, cholangitis Prophylaxis include pancreaticoduodenectomy as a gastroduo-
develops during or soon after attempted biliary decompression denal procedure, where the most common organisms cultured
in the patient who is undergoing elective preoperative biliary from SSIs are E. coli, Proteus, Klebsiella, staphylococci, strepto-
drainage. Unfortunately, these patients who develop cholangitis cocci, enterococci, and sometimes Bacteroides species. The
preoperatively are at increased risk for postoperative complica- recommended antimicrobial prophylaxis is a single preoperative
tions, especially those related to infection (Kitahata et al, 2014; dose of cefazolin. The guidelines do not recommend continuing
Kondo et al, 2013) (Table 12.7). antimicrobial coverage beyond 24 hours postoperatively.
The mere presence of bacterobilia, often related to preopera- Despite these guidelines, antimicrobial prophylaxis specifically
tive biliary drainage, increases the risk of infectious complica- for pancreatic resections has not been well evaluated in terms
tions in the postoperative setting (Cortes et al, 2006; di Mola of the specific agent to use and its duration. Donald and col-
et al, 2014; Howard et al, 2006; Jagannath et al, 2005; Lermite leagues (2013) suggest that guideline-recommended antimicro-
et al, 2008; Limongelli et al, 2007; Povoski et al, 1999a, 1999b; bial prophylaxis may not be appropriate for patients undergoing
Sivaraj et al, 2010) (see Table 12.7). Therefore preoperative pancreaticoduodenectomy. They present an argument for
biliary drainage in the resectable patient should be given broadening perioperative antibiotic coverage with the use of
thoughtful consideration, especially in light of a recent multi- piperacillin-tazobactam. Other authors have advocated for the
center, randomized trial that showed routine preoperative selection of perioperative antimicrobial prophylaxis that is
biliary drainage increases the rate of postoperative complica- based on preoperative bile cultures obtained at the time of
tions in general (van der Gaag et al, 2010). preoperative biliary drainage (Sudo et al, 2007, 2014). Sour-
There are patients who do routinely undergo preoperative rouille and colleagues (2013) recommend a 5 day course of
biliary drainage. One group includes those patients with bor- postoperative antimicrobial prophylaxis for patients undergoing
derline resectable pancreatic head cancers who receive several pancreaticoduodenectomy. This is clearly an area of HPB
months of neoadjuvant therapy before surgery. Although these surgery that warrants study so that perioperative antimicrobial
patients may experience an increase in postoperative infec- therapy can be standardized. Our practice is to administer one
tious complications with prolonged preoperative biliary drain- dose of ertapenem preoperatively for reasons previously men-
age, the procedure appears to be relatively safe, and its tioned in the section “Hepatic Resection.”
associated risks are not prohibitive (Gerke et al, 2004; Pisters
et al, 2001). Other Operative Risk Factors
Another group of patients in whom preoperative biliary Pancreatic fistula is one of the most important risk factors
drainage may be routinely indicated includes symptomatic determining postoperative morbidity following pancreatic
patients with significant jaundice who are expected to wait more resection (Behrman et al, 2008; Shinkawa et al, 2013; Sugiura
than 1 week for surgical referral. Preoperative biliary drainage et al, 2012; Watanabe et al, 2012) (Table 12.9) (see Chapter
in this cohort also appears to be relatively safe, as demonstrated 66). Much of that morbidity stems from the strong relationship
by Howard and colleagues (2006). The authors of this study between intraabdominal infectious complications and the pres-
examined the relationship between bacterobilia (based on ence of a pancreatic fistula. Behrman and colleagues (2008)
intraoperative bile cultures) and infectious complications in 138 retrospectively studied 196 pancreatectomy patients with an
patients undergoing an operation (including a biliary enteric aim to identify risk factors for intraabdominal sepsis. Approxi-
anastomosis) for obstructive jaundice. Eighty-six (62%) patients mately 16% of these patients developed an infected intraab-
had preoperative biliary stenting, whereas 52 (38%) did not. dominal fluid collection, and overt pancreatic fistula as well as
Ninety-one patients had bacterobilia, 69 from the stented group soft pancreatic remnant were found to be statistically significant
and 22 from the other group. Overall infectious complications factors associated with its development. The authors also
occurred in 31 patients (22.4%), with the majority occurring observed that infected fluid collections may occur relatively
in the stented group (23 vs. 8). However, this difference was early in the postoperative course, and surgeons should have a
not statistically significant. Stented patients did have a signifi- low threshold to image and drain these collections, many of
cantly higher rate of wound infection (P = .03) and bacteremia which were polymicrobial in nature.
(P = .04) on subset analysis. The authors concluded that pre- Sugiura and colleagues (2012) retrospectively examined risk
operative stenting increases the number of patients with positive factors for SSI in 408 patients who underwent pancreaticoduo-
intraoperative bile cultures, bacteremia, and wound infection. denectomy. An incisional SSI developed in 61 patients, whereas
However, they also noted that preoperative stenting does not an organ/space infection developed in 195 patients. The follow-
increase overall infectious morbidity, noninfectious morbidity, ing were identified as significant risk factors for incisional SSI
mortality, or hospital length of stay. The authors ultimately upon multivariate analysis: length of operation greater than 480
state that preoperative biliary drainage is not unreasonable in minutes (odds ratio [OR], 3.22), main pancreatic diameter less
the jaundiced patient awaiting referral to an appropriate surgi- than or equal to 3 mm (OR, 2.18), and abdominal wall thick-
cal center. Table 12.8 is a compilation of the various organisms ness greater than 10 mm (OR, 2.16). Also, the following were
obtained during intraoperative bile collection. Because Howard significant risk factors for the development of organ/space SSI:
and colleagues (2006) conducted their study in Indianapolis, pancreatic fistula (OR, 7.56), use of semiclosed drainage system
the bile bacteriology results listed are probably representative (OR, 3.68), BMI greater than 23.5 (OR, 3.04), main pancreatic
of bacteria existing in the typical American patient undergoing duct diameter less than or equal to 3 mm (OR, 2.21), and
an HPB surgery. Text continued on p. 219

TABLE 12.7 Risk Factors for Infectious Complications After Pancreatectomy: Preoperative Biliary Drainage
Kondo et al, Selection of prophylactic Retrospective 116 SSI Pancreaticoduodenectomy Patients were divided into two groups based
2013 antibiotics according to the on when their surgery was performed.
microorganisms isolated Patients in the “early” group received
from surgical site infections standard preoperative antibiotics, whereas
(SSIs) in a previous series of patients in the “later” group received
surgeries reduces SSI prophylactic antibiotics selected on the
incidence after basis of the SSIs that occurred in the
pancreaticoduodenectomy “early” group. Independent risk factors for
SSI included preoperative cholangitis (OR,
2.8) and being in the “early” group (OR,
2.9).
Kitahata Preoperative cholangitis during Retrospective 127 Postoperative complications Pancreaticoduodenectomy The effect of external vs. internal biliary
et al, 2014 biliary drainage increases drainage on postoperative complications
the incidence of for patients undergoing
postoperative severe pancreaticoduodenectomy was assessed.
complications after The overall complication rate was 33%.
pancreaticoduodenectomy Preoperative cholangitis was significantly
more common in internally vs. externally
drained patients (22.4% vs. 1.7%).
Preoperative cholangitis was the only
independent risk factor for severe
complication after
pancreaticoduodenectomy (OR, 4.61).
Povoski et al, Preoperative biliary drainage: Retrospective 161 To determine the impact of Pancreaticoduodenectomy The overall rate of postoperative infectious
1999a impact on intraoperative bile preoperative biliary complications was 29%, most commonly
cultures and infectious drainage on intraoperative wound infection (14%) and intraabdominal
morbidity and mortality after bile cultures and abscess (12%). Preoperative biliary
pancreaticoduodenectomy postoperative infectious drainage was independently and
morbidity significantly associated with positive
intraoperative bile cultures (P < .001),
postoperative infectious complications (P
= .022), wound infections (P = .045), and
death (P = .021). There was also an
association with intraabdominal abscess
(P = .061). Authors concluded that
preoperative biliary drainage should be
avoided in operative candidates.
Povoski et al, Association of preoperative Retrospective 240 To determine whether Pancreaticoduodenectomy The overall rate of postoperative infectious
1999b biliary drainage with preoperative biliary complications was 34%, most commonly
postoperative outcome instrumentation and/or wound infection (16%) and intraabdominal
following drainage are associated abscess (14%). Preoperative biliary
pancreaticoduodenectomy with increased morbidity drainage but not biliary instrumentation
and mortality rates after alone was significantly associated with
pancreaticoduodenectomy overall complications (P < .025), infectious
complications (P < .014), intraabdominal
abscess (P < .022), and death (P < .037).
Again, the authors concluded that
preoperative biliary drainage should be

avoided in otherwise resectable
candidates.
215
Continued
216
TABLE 12.7 Risk Factors for Infectious Complications After Pancreatectomy: Preoperative Biliary Drainage—cont’d
Jagannath Effect of preoperative biliary Prospective 144 To examine the effects of Pancreaticoduodenectomy The overall rate of postoperative infectious
et al, 2005 stenting on immediate collection, preoperative biliary stenting complications was 26.4%, including
outcome after retrospective on early outcome after wound infection (21.5%) and sepsis
pancreaticoduodenectomy analysis pancreaticoduodenectomy (13.9%). Positive intraoperative bile
cultures were significantly associated with
postoperative morbidity (0.001) and
mortality (P = .019). Significantly
associated morbidity included bleeding,
pancreatic leak, bile leak, wound infection,
sepsis, and infectious complications.
Preoperative biliary stenting was not
significantly associated with positive
intraoperative bile cultures unless stenting
was complicated by pancreatitis,
cholangitis, bleeding, or distal stent
migration.
Cortes et al, Effect of bile contamination on Retrospective 79 To compare outcomes after Pancreaticoduodenectomy The overall rate of postoperative infectious
2006 immediate outcomes after pancreaticoduodenectomy complications was 49%. Sixty-five percent
pancreaticoduodenectomy in patients with sterile and of patients had positive-bile cultures, of
for tumor those with infected bile which 80% had endoscopic biliary
drainage. In the culture-negative group,
14% had preoperative drainage. Overall
morbidity was higher in the culture-positive
group (77% vs. 59%, P = .05), especially
with respect to infectious complications
(65% vs. 37%, P = .003). Among those
with infectious complications, there was
49% concordance with bile cultures.
Organisms were resistant to cefazolin in
97%.
Howard et al, Influence of bactibilia after Retrospective 138 To correlate bactibilia found Obstructive jaundice The overall rate of postoperative infectious
2006 preoperative biliary stenting after preoperative biliary complications was 22.4%. 62% of
on postoperative infectious stenting with that of the patients had preoperative biliary stenting,
complications bacteriology of whereas 38% did not. Stented patients
postoperative infectious had a higher rate of wound infection (P =
complications in patients .03) and bacteremia (P = .04). The authors
with obstructive jaundice concluded that preoperative biliary
stenting increases the incidence of
bactibilia, bacteremia, and wound infection
rates but does not increase morbidity,
mortality, or hospital length of stay.

Limongelli Correlation between Prospective 220 To examine the relationship Pancreaticoduodenectomy, The overall rate of postoperative infectious
et al, 2007 preoperative biliary drainage, collection, between preoperative total pancreatectomy, complications was 45%, including
bile duct contamination, and retrospective biliary drainage, and biliary bypass wound infection (29%), intraabdominal
postoperative outcomes for analysis intraoperative bile culture, collection (16%), and sepsis (13%).
pancreatic surgery and postoperative 51.4% had a positive intraoperative
morbidity and mortality in bile culture. Infectious complications
patients undergoing (OR, 1.8; P = .03) and wound infections
pancreatic surgery (OR, 2.8; P = .002) were significantly
associated with a positive intraoperative
bile culture. The authors concluded that
preoperative biliary drainage promotes
positive intraoperative bile cultures, which
may increase the risk of infectious
complications.
Lermite et al, Effect of preoperative Prospective 124 To report the influence of Pancreaticoduodenectomy The overall rate of postoperative infectious
2008 endoscopic biliary drainage collection, preoperative endoscopic complications was 28.6%; 89.3% of
on infectious morbidity after retrospective biliary drainage on patients who underwent preoperative
pancreatoduodenectomy: a analysis postoperative infectious biliary drainage had positive bile cultures
case-control study morbidity in patients as opposed to 19.4% in the undrained
undergoing group (P < .001). Infectious complications
pancreaticoduodenectomy occurred in 50% of drained patients, but
only in 21.4 % of undrained patients (P =
.05). The authors concluded that routine
preoperative drainage should be avoided
in surgical candidates with potential
exceptions in cases of cholangitis or
patients undergoing neoadjuvant treatment
or extensive preoperative assessment.
Sivaraj et al, Is bactibilia a predictor of poor Prospective 76 To examine the relationship Pancreaticoduodenectomy The overall rate of postoperative infectious
2010 outcome of collection, between bile infection and complications was 35.5%. Bile cultures
pancreaticoduodenectomy? retrospective infectious complications were positive in 46% and negative in 54%.
analysis following Culture-positive patients had a higher
pancreaticoduodenectomy incidence of postoperative infectious
complications and wound infection (P =
.015), intraabdominal abscess (P = .002),
bacteremia (P = .0043), renal insufficiency
(P = .037), and longer average hospital
stay (16 vs. 10 days; P = .0002). The
authors advocated preoperative biliary
drainage only in strictly selected cases.
Continued
217
218
TABLE 12.7 Risk Factors for Infectious Complications After Pancreatectomy: Preoperative Biliary Drainage—cont’d
van der Preoperative biliary drainage Multicenter 202 To compare preoperative Pancreaticoduodenectomy The overall rate of postoperative infectious
Gaag for cancer of the head of randomized biliary drainage with complications was 23%. Serious
et al, 2010 the pancreas trial surgery alone for patients complications related to either biliary
with cancer of the drainage or surgery occurred in 74% of
pancreatic head those who underwent drainage as
opposed to 39% of those who did not.
Serious complications included
pancreatitis, cholangitis, GI perforation,
hemorrhage, stent occlusion or need for
exchange, pancreatic leak, delayed gastric
emptying, bile leak, anastomotic leak,
intraabdominal abscess, wound infection,
portal vein thrombosis, pneumonia,
myocardial infarction, or need for repeat
laparotomy. Drainage did not increase
mortality or hospital length of stay. The
authors concluded that preoperative

drainage increases the rate of
complications.
di Mola et al, Influence of preoperative biliary Prospective 131 To determine whether Pancreaticoduodenectomy Overall morbidity and mortality rates were
2014 drainage on surgical collection, preoperative biliary 54.6% and 3%, respectively. Preoperative
outcome after retrospective drainage and/or biliary biliary drainage significantly increased
pancreaticoduodenectomy: analysis instrumentation are complications (OR, 10.18; P < .001).
single-center experience associated with Specific complications that were
postoperative morbidity significantly more frequent included wound
and mortality rates after infection (P < .001), postpancreatectomy
pancreaticoduodenectomy hemorrhage (P < .0185), hyperglycemia (P
< .001), and pancreatic fistula (P = .036).
Survival did not differ significantly between
the groups. The authors concluded that
except for classic indications, preoperative
biliary drainage should be used selectively
for patients with resectable pancreatic
head cancer.
GI, Gastrointestinal; OR, odds ratio.
TABLE 12.8 Bile Bacteriology

Patients With Bactibilia Who Patients With Bactibilia Who
All patients With Underwent Preoperative Did Not Undergo Preoperative
Microorganism Bactibilia (n = 91) Biliary Stenting (n = 69) Biliary Stenting (n = 22)
Enterococcus 52 (57) 47 (68) 5 (23)

Klebsiella species 27 (30) 18 (26) 9 (41)
Yeast 21 (23) 14 (20) 7 (32)
Escherichia coli 16 (18) 16 (23) 0
Staphylococcus aureus 13 (14) 10 (14) 3 (14)
Enterobacter species 13 (14) 8 (12) 5 (23)
Lactobacillus species 8 (9) 8 (12) 0
Veillonella species 7 (8) 6 (9) 1 (5)
Clostridium perfringens 6 (7) 6 (9) 0
Fusobacterium nucleatum 6 (7) 6 (9) 0
Alpha streptococci 6 (7) 4 (6) 2 (9)
Bacteroides species 4 (4) 2 (3) 2 (9)
Prevotella species 4 (4) 4 (6) 0
Torulopsis species 2 (2) 2 (3) 0
Citrobacter species 2 (2) 1 (1) 1 (5)
Others 10 (11) 8 (12) 2 (9)
Modified from Howard TJ, et al: Influence of bactibilia after preoperative biliary stenting on postoperative infectious complications. J Gastrointest Surg 10(4):523-531, 2006. In this report, 138 patients
undergoing an operation for obstructive jaundice had both their intraoperative bile and all postoperative infectious complications cultured. Ninety-one patients had bacteria-positive intraoperative bile. Of
those 91 patients with bactibilia, 69 had undergone preoperative biliary stenting and 22 had not. Values are number of patients (percentage) unless otherwise indicated.
operation longer than 480 minutes (OR, 1.78). Pancreatic cant postoperative fistula, leak, or abscess in patients undergo-
fistula was clearly the strongest risk factor for organ/space ing pancreatic resection.
infection. Regarding distal pancreatectomy specifically, Hamilton and
The prevention of pancreatic fistula after pancreatectomy colleagues (2012) conducted a randomized controlled trial
remains a central question within the general and HPB surgical examining the efficacy of mesh-reinforced stapled closure of the
communities. Schmidt and colleagues (2009) studied preopera- distal pancreas. The authors randomly assigned 54 patients to
tive and perioperative risk factors for the development of a mesh reinforcement and 46 patients to nonmesh reinforcement,
pancreatic fistula in pancreaticoduodenectomy patients. First, in which the primary outcome was clinically significant pancre-
they confirmed the notion that patients in whom a pancreatic atic leak. International Study Group of Pancreatic Fistula
fistula develops have a higher rate of postoperative wound infec- (ISGPF) grade B and C leaks occurred more frequently in the
tion and intraabdominal abscess. Second, they identified several patients without mesh reinforcement (20% vs. 1.9%, P .0007).
operative risk factors for the development of pancreatic fistula. Other operative risk factors contributing to postoperative
Their multivariate analysis showed that an invaginated pancre- infectious morbidity after pancreatic resection include longer
atic anastomosis and closed suction intraperitoneal drainage operative times (Ball et al, 2010; Procter et al, 2010; Sudo et al,
were predictive of a pancreatic fistula, whereas chronic pancre- 2014; Sugiura et al, 2012; Wang et al, 2007) and need for peri-
atitis and preoperative biliary stenting were protective of a operative blood transfusion (Ball et al, 2010) (see Table 12.9).
pancreatic fistula. Despite these findings, a recent review Procter and colleagues (2010) performed a retrospective analysis
(Schoellhammer et al, 2014) suggested that no one pancreatic of 299,359 general surgical procedures (including pancreatec-
anastomosis is superior and that more studies are needed to tomy), identified through the ACS-NSQIP database, looking for
identify the best anastomotic technique. Also, the authors risk factors associated with infectious complications. Their
concluded that there is no evidence to support routine use of multivariate analysis suggested that increased operative duration
stents or topical sealing agents in the prevention of pancreatic is an independent risk factor for infectious complications and
fistula. hospital length of stay.This was confirmed by Ball and colleagues
A recent randomized trial studied pasireotide as a possible (2010), who conducted a retrospective analysis on only pancre-
adjunct to prevent postoperative pancreatic fistula (Allen et al, aticoduodenectomy patients identified via the ACS-NSQIP
2014). Pasireotide is a somatostatin analogue with a longer database. Their study involved 4817 patients and determined
half-life than octreotide. The authors randomly assigned 300 that longer operative times were associated with both morbidity
patients undergoing either pancreaticoduodenectomy or distal and mortality. Also, there was a linear relationship between
pancreatectomy to either perioperative pasireotide or placebo. preoperative RBC transfusion and 30 day morbidity. This led the
The primary end point was the occurrence of pancreatic authors to suggest blood transfusion and operative time as
fistula, leak, or abscess of grade 3 or higher. This end point was quality indicators for pancreaticoduodenectomy.
significantly lower in those patients treated with pasireotide Another commonly discussed topic concerns the use
(9% vs. 21%, P = .006). The authors concluded that this of intraperitoneal drains. Despite much literature within the
perioperative medication decreases the rate of clinically signifi- recent past suggesting that intraperitoneal drainage may be
220
TABLE 12.9 Risk Factors for Infectious Complications After Pancreatectomy: Perioperative Risk Factors
Behrman Intraabdominal sepsis following Retrospective 196 To assess risk factors, Elective pancreatectomy Of the study population, 16.3%
et al, pancreatic resection: microbiology, and experienced intraabdominal sepsis.
2008 incidence, risk factors, management of Statistically significant risk factors for
diagnosis, microbiology, intraabdominal sepsis after intraabdominal sepsis were overt
management, and outcome pancreatectomy pancreatic fistula (18.8% vs. 5%) and
soft pancreatic remnant (74.2% vs.
42.3%). Length of stay was significantly
prolonged (28.5 vs. 15.2 days) as well
as mortality (15.6% vs. 1.8%). The
authors concluded that intraabdominal
sepsis after pancreatectomy was
associated with a soft pancreatic
remnant and overt pancreatic fistula, and
that intraabdominal sepsis increased
length of stay and mortality. They also
noted that this complication often occurs
early in the postoperative course and
frequently involves multiple microbes.

Ball et al, Perioperative blood transfusion Prospective 4817 To assess perioperative RBC Pancreaticoduodenectomy Overall morbidity and mortality rates were
2010 and operative time are quality collection, transfusion and operative 37% and 3%, respectively. Thirty-two
indicators for retrospective time as risk factors for percent received a transfusion. Morbidity
pancreatoduodenectomy review 30 day morbidity and (46% vs. 33%) and mortality (5.3% vs.
mortality after 1.9%) were increased in patients
pancreaticoduodenectomy receiving a transfusion (P < .05).
Morbidity increased stepwise with the
number of RBC units transfused (P <
.01). Longer operative times were
associated with increased 30-day
morbidity (P < .001) and mortality (P <
.01). The authors concluded that RBC
transfusion and operative time correlate
with morbidity, and are appropriate
quality indicators for
pancreaticoduodenectomy.
Procter General surgical operative Prospective 299, To assess the relationship General surgery patients Increased operative duration was
et al, duration is associated with collection, 359 between operative duration independently associated with increased
2010 increased risk-adjusted retrospective and infectious complications infectious morbidity and length of
infectious complication rates review in general surgery patients hospital stay. The OR for infectious
and length of hospital stay morbidity for procedures lasting 1 hr or
less was normalized to 1. The adjusted
OR for infectious morbidity increased by
1 for every 1.5 hr increase in operative
duration.

Sugiura Risk factor of SSI after Retrospective 408 To assess the risk factors for Pancreaticoduodenectomy The overall rate of postoperative infectious
et al, pancreaticoduodenectomy SSI and bacterial complications was 62.7%. Independent
2012 composition of SSIs in risk factors for incisional SSI were length
patients undergoing of operation > 480 min (OR, 3.22), main
pancreaticoduodenectomy pancreatic duct diameter < 3 mm (OR,
2.18), and abdominal wall thickness >
10 mm (OR, 2.16). Independent risk
factors for organ/space SSI were
pancreatic fistula (OR, 7.56), use of
semiclosed drainage system (OR, 3.68),
body mass index > 23.5 kg/m2 (OR,
3.04), main pancreatic duct diameter <
3 mm (OR, 2.21), and length of
operation > 480 min (OR, 1.78). The
authors concluded that efforts to reduce
pancreatic fistulas, to decrease length of
operation, and to use a closed drainage
system could lead to lower rates of SSI.
Watanabe Risk factors for intraabdominal Retrospective 206 To assess the risk factors for Pancreaticoduodenectomy The overall rate of intraabdominal infection
et al, infection after the development of was 21.4%. Multivariate analysis
2012 pancreaticoduodenectomy—a intraabdominal infection revealed that pancreatic fistula was an
retrospective analysis to after independent risk factor for
evaluate the significance of pancreaticoduodenectomy, intraabdominal infection (OR, 9.58).
preoperative biliary drainage focusing on preoperative Preoperative biliary drainage did not
and postoperative pancreatic biliary drainage and increase the rate of intraabdominal
fistula development of pancreatic infection. The authors concluded that
fistula preoperative biliary drainage might not
increase infectious morbidity.
Shinkawa Nutritional risk index (NRI) as an Retrospective 64 To determine the relationship Pancreaticoduodenectomy The overall SSI rate was 33%. Pancreatic
et al, independent predictive factor between preoperative fistula (OR, 26.8) and NRI < 97.5 (OR,
2013 for the development of SSI nutritional screening scores 6.5) were independent risk factors for
after and the development of SSI SSI. The authors concluded that using
pancreaticoduodenectomy after the NRI can help assess risk of infection
pancreaticoduodenectomy after pancreaticoduodenectomy.
Sudo Perioperative antibiotics Prospective 254 To evaluate a perioperative Pancreaticoduodenectomy Positive intraoperative bile cultures were
et al, covering bile contamination antibiotic strategy targeting significantly higher in the internally (85%)
2014 prevent abdominal infectious bile contamination and externally (90%) drained cases than
complications after associated with in nondrained cases (26%) (P < .001).
pancreatoduodenectomy in preoperative biliary drainage Overall morbidity rates (23%, 23%, and
patients with preoperative procedures for preventing 25%) and abdominal infectious
biliary drainage abdominal infectious complications (13%, 17%, and 14%) did
complications after not differ among the nondrained,
pancreatoduodenectomy internally drained, and externally drained
cases, respectively. Operative time
>360 min was associated with
intraabdominal infectious complications
(P = .045). The authors concluded that
perioperative antibiotics covering bile
contaminants may prevent abdominal
infections after pancreatoduodenectomy
in patients with or without preoperative
biliary drainage.
CI, Confidence interval; OR, odds ratio; RBC, red blood cell; SSI, surgical-site infection.
221
unnecessary and even harmful (Adham et al, 2013; Behrman preference but must be done meticulously. With regard to distal
et al, 2015; Conlon et al, 2001; Correa-Gallego et al, 2013; pancreatectomy, stapled closure of the pancreas with bioab-
Fisher et al, 2011; Mehta et al, 2013; Paulus et al, 2012; van sorbable mesh buttress appears promising in the prevention of
der Wilt et al, 2013), a recent randomized prospective multi- pancreatic fistula. Perioperative administration of pasireotide
center trial concluded that “elimination of intraperitoneal has demonstrated efficacy in reducing pancreatic fistula after
drainage in all cases of pancreaticoduodenectomy increases the pancreaticoduodenectomy and distal pancreatectomy. The use
frequency and severity of complications” (Van Buren et al, of intraperitoneal drains and the timing of their removal remain
2014). This most recent study randomized 137 patients under- controversial.
going pancreaticoduodenectomy. Half of the patients had an
intraperitoneal drain left in place, whereas the other half did Postoperative Risk Mitigation
not. These patients were followed prospectively for a range of Postoperative blood glucose control is important after surgical
complications. The study was stopped early due to the fact that procedures and has an impact on patient outcomes, as discussed
there was a substantial difference in mortality between the two earlier. Two hundred and sixty-five HPB surgery patients were
groups. The drained patients had a mortality of 3%, whereas studied by Ambiru and colleagues (2008) and were prospec-
the undrained group experienced a 12% mortality rate. Beyond tively evaluated for the development of SSI. Multivariate analy-
this, pancreaticoduodenectomy without intraperitoneal drain- sis showed that poor postoperative blood glucose was an
age was significantly associated with an increase in the number independent risk factor for SSI. The rate of SSI was 20% in
of complications per patient, an increase in the number of those patients with blood glucose levels below 200 versus 52%
patients who had at least one complication rated at grade 2 or in those without insulin infusion therapy (P < .01). Therefore
higher, and a higher average complications severity. From an blood glucose control in the postoperative setting is of particular
infectious standpoint, pancreaticoduodenectomy without intra- concern in the postpancreatectomy patient, especially because
peritoneal drainage was associated with a higher rate of intraab- some patients who were not previously diabetic may eventually
dominal abscess (25% vs. 10%, P = .027). This study therefore require insulin therapy.
provides strong evidence supporting the placement of intraperi- Synbiotic therapy has shown some promising results in
toneal drains at the time of pancreaticoduodenectomy. Taken mitigating infectious complications in hepatic resection patients
in the context of other contemporary literature on this subject, (see section “Hepatic Resection”). Similar results have been
the best approach to peritoneal drainage remains unclear and seen after resection of the head of the pancreas. Rayes and
should be individualized to the specific patient. Also, placing a colleagues (2007) conducted a randomized, double-blind trial
drain intraoperatively does not always obviate the need for a to evaluate the potential benefit of synbiotic therapy in patients
percutaneous drainage procedure in the early postoperative undergoing pancreaticoduodenectomy with preservation of the
period (Conlon et al, 2001). If placed at the time of pancreati- pylorus. The study included 80 patients total. All patients were
coduodenectomy, the timing of drain removal is also controver- started on enteral nutrition immediately postoperatively via a
sial. Recent prospective studies, including a randomized trial, nasojejunal tube placed at the time of the operation. The
have suggested that early drain removal based on drain amylase experimental group received enteral nutrition plus synbiotics,
levels can lower the rate of postoperative complications, includ- whereas the control group received enteral nutrition plus
ing infectious ones (Bassi et al, 2010; Kawai et al, 2006). placebo. Postoperative infectious complications occurred in
To summarize, risk mitigation at the operative level for a 12.5% of the patients receiving synbiotic versus 40% in the
pancreaticoduodenectomy or distal pancreatectomy consists of control group (P = .005). Although this evidence appears
the efficient performance of the operation using careful opera- compelling, we do not routinely use nasojejunal tubes. We also
tive technique in an effort to avoid unnecessary blood loss. do not advocate for the routine use of surgically placed feeding
There are no universally agreed upon techniques to reduce jejunostomy tubes, because we feel that their associated mor-
pancreatic fistula during the performance of pancreaticoduode- bidity outweighs their benefit (Padussis et al, 2014).
nectomy. The pancreaticoenteric anastomosis during pan
creaticoduodenectomy is still performed according to surgeon References are available at expertconsult.com.
Chapter 12 Infections in hepatic, biliary, and pancreatic surgery 222.e1
REFERENCES Cho SW, et al: Neoadjuvant radiation therapy and its impact on com-
plications after pancreaticoduodenectomy for pancreatic cancer:
Adham M, et al: Pancreatic resection: drain or no drain?, Surgery analysis of the American College of Surgeons National Surgical
154(5):1069–1077, 2013. Quality Improvement Program (ACS-NSQIP), HPB (Oxford) 16(4):
Allen PJ, et al: Pasireotide for postoperative pancreatic fistula, N Engl 350–356, 2014.
J Med 371(9):875–876, 2014. Classen DC, et al: The timing of prophylactic administration of antibi-
Ambiru S, et al: Poor postoperative blood glucose control increases otics and the risk of surgical-wound infection, N Engl J Med
surgical site infections after surgery for hepato-biliary-pancreatic 326(5):281–286, 1992.
cancer: a prospective study in a high-volume institute in Japan, Cohn SM, et al: Prospective randomized trial of two wound manage-
J Hosp Infect 68(3):230–233, 2008. ment strategies for dirty abdominal wounds, Ann Surg 233(3):409–
Anderson DJ, et al: Strategies to prevent surgical site infections in acute 413, 2001.
care hospitals, Infect Control Hosp Epidemiol 29(Suppl 1):S51–S61, Conlon KC, et al: Prospective randomized clinical trial of the value of
2008. intraperitoneal drainage after pancreatic resection, Ann Surg 234(4):
Anonymous: Perioperative total parenteral nutrition in surgical patients. 487–493, discussion 493–484, 2001.
The Veterans Affairs Total Parenteral Nutrition Cooperative Study Correa-Gallego C, et al: Operative drainage following pancreatic resec-
Group, N Engl J Med 325(8):525–532, 1991. tion: analysis of 1122 patients resected over 5 years at a single institu-
Anonymous: ASHP Therapeutic Guidelines on Antimicrobial Prophy- tion, Ann Surg 258(6):1051–1058, 2013.
laxis in Surgery. American Society of Health-System Pharmacists, Cortes A, et al: Effect of bile contamination on immediate outcomes
Am J Health Syst Pharm 56(18):1839–1888, 1999. after pancreaticoduodenectomy for tumor, J Am Coll Surg 202(1):
Araujo RL, et al: Does pre-operative chemoradiation for initially 93–99, 2006.
unresectable or borderline resectable pancreatic adenocarcinoma di Mola FF, et al: Influence of preoperative biliary drainage on surgical
increase post-operative morbidity? A case-matched analysis, HPB outcome after pancreaticoduodenectomy: single centre experience,
(Oxford) 15(8):574–580, 2013. Langenbecks Arch Surg 399(5):649–657, 2014.
Arikawa T, et al: Risk factors for surgical site infection after hepatec- Digison MB: A review of anti-septic agents for pre-operative skin
tomy for hepatocellular carcinoma, Hepatogastroenterology 58(105): preparation, Plast Surg Nurs 27(4):185–189, quiz 190–181, 2007.
143–146, 2011. DiPiro JT, et al: Single dose systemic antibiotic prophylaxis of surgical
Ball CG, et al: Peri-operative blood transfusion and operative time are wound infections, Am J Surg 152(5):552–559, 1986.
quality indicators for pancreatoduodenectomy, HPB (Oxford) Donald GW, et al: Perioperative antibiotics for surgical site infection in
12(7):465–471, 2010. pancreaticoduodenectomy: does the SCIP-approved regimen provide
Barie PS: Modern surgical antibiotic prophylaxis and therapy—less is adequate coverage?, Surgery 154(2):190–196, 2013.
more, Surg Infect (Larchmt) 1(1):23–29, 2000. Emmrich J, et al: Secretory immunoglobulin A in pancreatic juice and
Barone JE, et al: Hypothermia does not result in more complications pancreatic tissue of patients with chronic pancreatitis, Gut 42(3):
after colon surgery, Am Surg 65(4):356–359, 1999. 436–441, 1998.
Bassi C, et al: Early versus late drain removal after standard pancreatic Farges O, et al: Multicentre European study of preoperative biliary
resections: results of a prospective randomized trial, Ann Surg drainage for hilar cholangiocarcinoma, Br J Surg 100(2):274–283,
252(2):207–214, 2010. 2013.
Behrman SW, Zarzaur BL: Intra-abdominal sepsis following pancreatic Ferrero A, et al: Preoperative biliary drainage increases infectious
resection: incidence, risk factors, diagnosis, microbiology, manage- complications after hepatectomy for proximal bile duct tumor
ment, and outcome, Am Surg 74(7):572–578, discussion 578–579, obstruction, World J Surg 33(2):318–325, 2009.
2008. Finkelstein R, et al: Vancomycin versus cefazolin prophylaxis for cardiac
Behrman SW, et al: Routine drainage of the operative bed following surgery in the setting of a high prevalence of methicillin-resistant
elective distal pancreatectomy does not reduce the occurrence of staphylococcal infections, J Thorac Cardiovasc Surg 123(2):326–332,
complications, J Gastrointest Surg 19(1):72–79, discussion 79, 2015. 2002.
Belghiti J, et al: Seven hundred forty-seven hepatectomies in the 1990s: Fisette A, et al: High-dose insulin therapy reduces postoperative liver
an update to evaluate the actual risk of liver resection, J Am Coll Surg dysfunction and complications in liver resection patients through
191(1):38–46, 2000. reduced apoptosis and altered inflammation, J Clin Endocrinol Metab
Belghiti J, et al: Drainage after elective hepatic resection. A randomized 97(1):217–226, 2012.
trial, Ann Surg 218(6):748–753, 1993. Fisher WE, et al: Pancreatic resection without routine intraperitoneal
Bilzer M, et al: Role of Kupffer cells in host defense and liver disease, drainage, HPB (Oxford) 13(7):503–510, 2011.
Liver Int 26(10):1175–1186, 2006. Fong Y, et al: Drainage is unnecessary after elective liver resection, Am
Bolon MK, et al: Glycopeptides are no more effective than beta-lactam J Surg 171(1):158–162, 1996.
agents for prevention of surgical site infection after cardiac surgery: Forse RA, et al: Antibiotic prophylaxis for surgery in morbidly obese
a meta-analysis, Clin Infect Dis 38(10):1357–1363, 2004. patients, Surgery 106(4):750–756, discussion 756–757, 1989.
Bottger TC, Junginger T: Factors influencing morbidity and mortality Furnary AP, et al: Continuous intravenous insulin infusion reduces the
after pancreaticoduodenectomy: critical analysis of 221 resections, incidence of deep sternal wound infection in diabetic patients after
World J Surg 23(2):164–171, discussion 171–162, 1999. cardiac surgical procedures, Ann Thorac Surg 67(2):352–360, discus-
Braasch JW, Gray BN: Considerations that lower pancreatoduodenec- sion 360–352, 1999.
tomy mortality, Am J Surg 133(4):480–484, 1977. Garwood RA, et al: Infectious complications after hepatic resection,
Brasel KJ, et al: Cost-utility analysis of contaminated appendectomy Am Surg 70(9):787–792, 2004.
wounds, J Am Coll Surg 184(1):23–30, 1997. Gerke H, et al: Complications of pancreaticoduodenectomy after
Bratzler DW, et al: Clinical practice guidelines for antimicrobial pro- neoadjuvant chemoradiation in patients with and without preopera-
phylaxis in surgery, Am J Health Syst Pharm 70(3):195–283, 2013. tive biliary drainage, Dig Liver Dis 36(6):412–418, 2004.
Bratzler DW, Hunt DR: The surgical infection prevention and surgical Gilroy C, et al: Cilia in the porcine bile ductule: motile or sensory?,
care improvement projects: national initiatives to improve outcomes Histol Histopathol 10(2):301–304, 1995.
for patients having surgery, Clin Infect Dis 43(3):322–330, 2006. Gomi H, et al: TG13 antimicrobial therapy for acute cholangitis and
Brooke-Smith M, et al: Prospective evaluation of the International cholecystitis, J Hepatobiliary Pancreat Sci 20(1):60–70, 2013.
Study Group for Liver Surgery definition of bile leak after a liver Greenblatt DY, et al: Preoperative factors predict perioperative morbid-
resection and the role of routine operative drainage: an international ity and mortality after pancreaticoduodenectomy, Ann Surg Oncol
multicentre study, HPB (Oxford) 17(1):46–51, 2015. 18(8):2126–2135, 2011.
Carr JM, et al: Implementing tight glucose control after coronary artery Gurusamy KS, et al: Routine abdominal drainage for uncomplicated
bypass surgery, Ann Thorac Surg 80(3):902–909, 2005. liver resection, Cochrane Database Syst Rev (3):CD006232, 2007.
Ceppa EP, et al: Reducing surgical site infections in hepatopancreato- Hacek DM, et al: Staphylococcus aureus nasal decolonization in joint
biliary surgery, HPB (Oxford) 15(5):384–391, 2013. replacement surgery reduces infection, Clin Orthop Relat Res 466(6):
Cheng TY, et al: Effect of neoadjuvant chemoradiation on operative 1349–1355, 2008.
mortality and morbidity for pancreaticoduodenectomy, Ann Surg Hamilton NA, et al: Mesh reinforcement of pancreatic transection
Oncol 13(1):66–74, 2006. decreases incidence of pancreatic occlusion failure for left
pancreatectomy: a single-blinded, randomized controlled trial, Ann Lavu H, et al: Perioperative surgical care bundle reduces pancreatico-
Surg 255(6):1037–1042, 2012. duodenectomy wound infections, J Surg Res 174(2):215–221, 2012.
Harimoto N, et al: Prospective randomized controlled trial investigat- Lazar HL, et al: Tight glycemic control in diabetic coronary artery
ing the type of sutures used during hepatectomy, World J Gastroenterol bypass graft patients improves perioperative outcomes and decreases
17(18):2338–2342, 2011. recurrent ischemic events, Circulation 109(12):1497–1502, 2004.
Heinrich S, et al: Neoadjuvant chemotherapy generates a significant Lermite E, et al: Effect of preoperative endoscopic biliary drainage on
tumor response in resectable pancreatic cancer without increasing infectious morbidity after pancreatoduodenectomy: a case-control
morbidity: results of a prospective phase II trial, Ann Surg 248(6): study, Am J Surg 195(4):442–446, 2008.
1014–1022, 2008. Lerut JP, et al: Pancreaticoduodenal resection. Surgical experience and
Hirokawa F, et al: Evaluation of postoperative antibiotic prophylaxis evaluation of risk factors in 103 patients, Ann Surg 199(4):432–437,
after liver resection: a randomized controlled trial, Am J Surg 206(1): 1984.
8–15, 2013. Limongelli P, et al: Correlation between preoperative biliary drainage,
Hochwald SN, et al: Association of preoperative biliary stenting with bile duct contamination, and postoperative outcomes for pancreatic
increased postoperative infectious complications in proximal cholan- surgery, Surgery 142(3):313–318, 2007.
giocarcinoma, Arch Surg 134(3):261–266, 1999. Little SA, et al: Diabetes is associated with increased perioperative
House MG, et al: Preoperative predictors for complications after mortality but equivalent long-term outcome after hepatic resection
pancreaticoduodenectomy: impact of BMI and body fat distribution, for colorectal cancer, J Gastrointest Surg 6(1):88–94, 2002.
J Gastrointest Surg 12(2):270–278, 2008. Liu CL, et al: Abdominal drainage after hepatic resection is contrain-
Howard TJ, et al: Influence of bactibilia after preoperative biliary stent- dicated in patients with chronic liver diseases, Ann Surg 239(2):194–
ing on postoperative infectious complications, J Gastrointest Surg 201, 2004.
10(4):523–531, 2006. Liu F, et al: Preoperative biliary drainage before resection for hilar
Hughes MJ, et al: Enhanced recovery following liver surgery: a system- cholangiocarcinoma: whether or not? A systematic review, Dig Dis
atic review and meta-analysis, HPB (Oxford) 16(8):699–706, 2014. Sci 56(3):663–672, 2011.
Huo TI, et al: Diabetes mellitus is a risk factor for hepatic decompensa- Mangram AJ, et al: Guideline for prevention of surgical site infection,
tion in patients with hepatocellular carcinoma undergoing resection: 1999. Hospital Infection Control Practices Advisory Committee,
a longitudinal study, Am J Gastroenterol 98(10):2293–2298, 2003. Infect Control Hosp Epidemiol 20(4):250–278, quiz 279–280, 1999.
Itoshima T, et al: Scanning electron microscopy of the bile ductule, McDonald M, et al: Single- versus multiple-dose antimicrobial prophy-
Gastroenterol Jpn 12(6):476–482, 1977. laxis for major surgery: a systematic review, Aust N Z J Surg 68(6):
Jagannath P, et al: Effect of preoperative biliary stenting on immediate 388–396, 1998.
outcome after pancreaticoduodenectomy, Br J Surg 92(3):356–361, Mehta VV, et al: Is it time to abandon routine operative drain use? A
2005. single institution assessment of 709 consecutive pancreaticoduode-
Kalmeijer MD, et al: Surgical site infections in orthopedic surgery: the nectomies, J Am Coll Surg 216(4):635–642, discussion 642–634,
effect of mupirocin nasal ointment in a double-blind, randomized, 2013.
placebo-controlled study, Clin Infect Dis 35(4):353–358, 2002. Miura F, et al: TG13 flowchart for the management of acute cholangitis
Kanazawa H, et al: Synbiotics reduce postoperative infectious complica- and cholecystitis, J Hepatobiliary Pancreat Sci 20(1):47–54, 2013.
tions: a randomized controlled trial in biliary cancer patients undergo- Mizuno T, et al: Intraoperative bacterial translocation detected by
ing hepatectomy, Langenbecks Arch Surg 390(2):104–113, 2005. bacterium-specific ribosomal RNA-targeted reverse-transcriptase
Kawai M, et al: Early removal of prophylactic drains reduces the polymerase chain reaction for the mesenteric lymph node strongly
risk of intra-abdominal infections in patients with pancreatic head predicts postoperative infectious complications after major hepatec-
resection: prospective study for 104 consecutive patients, Ann Surg tomy for biliary malignancies, Ann Surg 252(6):1013–1019, 2010.
244(1):1–7, 2006. Moreno Elola-Olaso A, et al: Predictors of surgical site infection after
Kelly KJ, et al: Risk stratification for distal pancreatectomy utilizing liver resection: a multicentre analysis using National Surgical Quality
ACS-NSQIP: preoperative factors predict morbidity and mortality, Improvement Program data, HPB (Oxford) 14(2):136–141, 2012.
J Gastrointest Surg 15(2):250–259, discussion 259–261, 2011. Nakahira S, et al: Proposal for a sub-classification of hepato-biliary-
Kennedy TJ, et al: Role of preoperative biliary drainage of liver remnant pancreatic operations for surgical site infection surveillance following
prior to extended liver resection for hilar cholangiocarcinoma, HPB assessment of results of prospective multicenter data, J Hepatobiliary
(Oxford) 11(5):445–451, 2009. Pancreat Sci 20(5):504–511, 2013.
Kent TS, et al: The burden of infection for elective pancreatic resec- Nakayama H, et al: Subcutaneous drainage to prevent wound infection
tions, Surgery 153(1):86–94, 2013. in liver resection: a randomized controlled trial, J Hepatobiliary
Kirby JP, Mazuski JE: Prevention of surgical site infection, Surg Clin Pancreat Sci 21(7):509–517, 2014.
North Am 89(2):365–389, viii, 2009. Nanashima A, et al: Associated factors with surgical site infections after
Kitahata Y, et al: Preoperative cholangitis during biliary drainage hepatectomy: predictions and countermeasures by a retrospective
increases the incidence of postoperative severe complications after cohort study, Int J Surg 12(4):310–314, 2014.
pancreaticoduodenectomy, Am J Surg 208(1):1–10, 2014. Nordlinger B, et al: Perioperative chemotherapy with FOLFOX4 and
Kjonniksen I, et al: Preoperative hair removal—a systematic literature surgery versus surgery alone for resectable liver metastases from
review, AORN J 75(5):928–938, 940, 2002. colorectal cancer (EORTC Intergroup trial 40983): a randomised
Kobayashi S, et al: Risk factors of surgical site infection after hepatec- controlled trial, Lancet 371(9617):1007–1016, 2008.
tomy for liver cancers, World J Surg 33(2):312–317, 2009. Okabayashi T, et al: Effect of intensive insulin therapy using a closed-
Kondo K, et al: Selection of prophylactic antibiotics according to the loop glycemic control system in hepatic resection patients: a prospec-
microorganisms isolated from surgical site infections (SSIs) in a tive randomized clinical trial, Diabetes Care 32(8):1425–1427,
previous series of surgeries reduces SSI incidence after pancreatico- 2009a.
duodenectomy, J Hepatobiliary Pancreat Sci 20(3):286–293, 2013. Okabayashi T, et al: Risk factors and predictors for surgical site infec-
Konvalinka A, et al: Impact of treating Staphylococcus aureus nasal tion after hepatic resection, J Hosp Infect 73(1):47–53, 2009b.
carriers on wound infections in cardiac surgery, J Hosp Infect 64(2): Okabayashi T, et al: Effect of perioperative intensive insulin therapy for
162–168, 2006. liver dysfunction after hepatic resection, World J Surg 35(12):2773–
Kurmann A, et al: Hepatic steatosis is associated with surgical-site 2778, 2011.
infection after hepatic and colorectal surgery, Surgery 156(1):109– Padussis JC, et al: Feeding jejunostomy during Whipple is associated
116, 2014. with increased morbidity, J Surg Res 187(2):361–366, 2014.
Kurz A, et al: Perioperative normothermia to reduce the incidence of Paulus EM, et al: Routine peritoneal drainage of the surgical bed after
surgical-wound infection and shorten hospitalization. Study of elective distal pancreatectomy: is it necessary?, Am J Surg 204(4):
Wound Infection and Temperature Group, N Engl J Med 334(19): 422–427, 2012.
1209–1215, 1996. Perl TM, et al: Intranasal mupirocin to prevent postoperative
La Torre M, et al: Malnutrition and pancreatic surgery: prevalence and Staphylococcus aureus infections, N Engl J Med 346(24):1871–1877,
outcomes, J Surg Oncol 107(7):702–708, 2013. 2002.
Laupland KB, Conly JM: Treatment of Staphylococcus aureus coloniza- Pessaux P, et al: Randomized clinical trial evaluating the need for
tion and prophylaxis for infection with topical intranasal mupirocin: routine nasogastric decompression after elective hepatic resection, Br
an evidence-based review, Clin Infect Dis 37(7):933–938, 2003. J Surg 94(3):297–303, 2007.
Chapter 12 Infections in hepatic, biliary, and pancreatic surgery 222.e3
Pessaux P, et al: Identification and validation of risk factors for postop- Sugawara G, et al: Perioperative synbiotic treatment to prevent postop-
erative infectious complications following hepatectomy, J Gastrointest erative infectious complications in biliary cancer surgery: a random-
Surg 17(11):1907–1916, 2013. ized controlled trial, Ann Surg 244(5):706–714, 2006.
Petrowsky H, et al: Evidence-based value of prophylactic drainage in Sugiura T, et al: Risk factor of surgical site infection after pancreatico-
gastrointestinal surgery: a systematic review and meta-analyses, Ann duodenectomy, World J Surg 36(12):2888–2894, 2012.
Surg 240(6):1074–1084, discussion 1084–1085, 2004. Sumiyoshi K, et al: Biosynthesis and secretion of MHC class III gene
Pisters PW, et al: Effect of preoperative biliary decompression on products (complement C4 and factor B) in the exocrine pancreas,
pancreaticoduodenectomy-associated morbidity in 300 consecutive J Gastroenterol 32(3):367–373, 1997.
patients, Ann Surg 234(1):47–55, 2001. Sung JY, et al: Antibacterial activity of bile salts against common biliary
Povoski SP, et al: Preoperative biliary drainage: impact on intraopera- pathogens. Effects of hydrophobicity of the molecule and in the
tive bile cultures and infectious morbidity and mortality after pan- presence of phospholipids, Dig Dis Sci 38(11):2104–2112, 1993.
creaticoduodenectomy, J Gastrointest Surg 3(5):496–505, 1999a. Suzuki Y, et al: Randomized clinical trial of preoperative intranasal
Povoski SP, et al: Association of preoperative biliary drainage with mupirocin to reduce surgical-site infection after digestive surgery,
postoperative outcome following pancreaticoduodenectomy, Ann Br J Surg 90(9):1072–1075, 2003.
Surg 230(2):131–142, 1999b. Swoboda SM, et al: Does intraoperative blood loss affect antibiotic
Procter LD, et al: General surgical operative duration is associated with serum and tissue concentrations?, Arch Surg 131(11):1165–1171,
increased risk-adjusted infectious complication rates and length of discussion 1171–1172, 1996.
hospital stay, J Am Coll Surg 210(1):60–65, e61–e62, 2010. Togo S, et al: Perioperative infection control and its effectiveness in
Rao N, et al: A preoperative decolonization protocol for staphylococcus hepatectomy patients, J Gastroenterol Hepatol 22(11):1942–1948,
aureus prevents orthopaedic infections, Clin Orthop Relat Res 466(6): 2007.
1343–1348, 2008. Togo S, et al: Usefulness of absorbable sutures in preventing surgical
Rayes N, et al: Effect of enteral nutrition and synbiotics on bacterial site infection in hepatectomy, J Gastrointest Surg 12(6):1041–1046,
infection rates after pylorus-preserving pancreatoduodenectomy: a 2008.
randomized, double-blind trial, Ann Surg 246(1):36–41, 2007. Tzeng CW, et al: Morbidity and mortality after pancreaticoduodenec-
Richter B, et al: Nutritional support after open liver resection: a sys- tomy in patients with borderline resectable type C clinical classifica-
tematic review, Dig Surg 23(3):139–145, 2006. tion, J Gastrointest Surg 18(1):146–155, discussion 155–156, 2014.
Sadamori H, et al: Risk factors for organ/space surgical site infection Uchiyama K, et al: Risk factors for postoperative infectious complica-
after hepatectomy for hepatocellular carcinoma in 359 recent cases, tions after hepatectomy, J Hepatobiliary Pancreat Sci 18(1):67–73,
J Hepatobiliary Pancreat Sci 20(2):186–196, 2013. 2011.
Scher KS: Studies on the duration of antibiotic administration for surgi- Usami M, et al: Effects of perioperative synbiotic treatment on infec-
cal prophylaxis, Am Surg 63(1):59–62, 1997. tious complications, intestinal integrity, and fecal flora and organic
Schindl MJ, et al: The value of residual liver volume as a predictor of acids in hepatic surgery with or without cirrhosis, JPEN J Parenter
hepatic dysfunction and infection after major liver resection, Gut Enteral Nutr 35(3):317–328, 2011.
54(2):289–296, 2005. Van Buren G 2nd, et al: A randomized prospective multicenter trial of
Schmidt CM, et al: Pancreatic fistula following pancreaticoduodenec- pancreaticoduodenectomy with and without routine intraperitoneal
tomy: clinical predictors and patient outcomes, HPB Surg 2009: drainage, Ann Surg 259(4):605–612, 2014.
404520, 2009. van den Berghe G, et al: Intensive insulin therapy in critically ill
Schoellhammer HF, et al: Techniques for prevention of pancreatic leak patients, N Engl J Med 345(19):1359–1367, 2001.
after pancreatectomy, Hepatobiliary Surg Nutr 3(5):276–287, 2014. van der Gaag NA, et al: Preoperative biliary drainage for cancer of the
Scilletta R, et al: Comparative analysis of the incidence of surgical site head of the pancreas, N Engl J Med 362(2):129–137, 2010.
infections in patients with liver resection for colorectal hepatic van der Wilt AA, et al: To drain or not to drain: a cumulative meta-
metastases after neoadjuvant chemotherapy, J Surg Res 188(1):183– analysis of the use of routine abdominal drains after pancreatic
189, 2014. resection, HPB (Oxford) 15(5):337–344, 2013.
Scott-Conner CE, Grogan JB: The pathophysiology of biliary obstruc- Walz JM, et al: Surgical site infection following bowel surgery: a retro-
tion and its effect on phagocytic and immune function, J Surg Res spective analysis of 1446 patients, Arch Surg 141(10):1014–1018,
57(2):316–336, 1994. discussion 1018, 2006.
Sessler DI, Akca O: Nonpharmacological prevention of surgical wound Wang A, et al: Analysis of surgical site infection rate in pancreas opera-
infections, Clin Infect Dis 35(11):1397–1404, 2002. tion and its related risk factors, Zhongguo Yi Xue Ke Xue Yuan Xue
Shinkawa H, et al: Nutritional risk index as an independent predictive Bao 29(4):562–565, 2007.
factor for the development of surgical site infection after pancreati- Watanabe F, et al: Risk factors for intra-abdominal infection after
coduodenectomy, Surg Today 43(3):276–283, 2013. pancreaticoduodenectomy—a retrospective analysis to evaluate the
Sivaraj SM, et al: Is bactibilia a predictor of poor outcome of pancre- significance of preoperative biliary drainage and postoperative pan-
aticoduodenectomy?, Hepatobiliary Pancreat Dis Int 9(1):65–68, creatic fistula, Hepatogastroenterology 59(116):1270–1273, 2012.
2010. Webster J, Osborne S: Preoperative bathing or showering with skin
Sourrouille I, et al: Five days of postoperative antimicrobial therapy antiseptics to prevent surgical site infection, Cochrane Database Syst
decreases infectious complications following pancreaticoduodenec- Rev (2):CD004985, 2007.
tomy in patients at risk for bile contamination, HPB (Oxford) Webster J, Osborne S: Preoperative bathing or showering with skin
15(6):473–480, 2013. antiseptics to prevent surgical site infection, Cochrane Database Syst
Springer R: The Surgical Care Improvement Project—focusing on Rev (9):CD004985, 2012.
infection control, Plast Surg Nurs 27(3):163–167, 2007. Weimann A, et al: ESPEN Guidelines on Enteral Nutrition: Surgery
Stewart L, et al: Antibacterial activity of bile acids against common including organ transplantation, Clin Nutr 25(2):224–244, 2006.
biliary tract organisms, Surg Forum 37:157–159, 1986. Wilton PB, et al: Complement in local biliary tract defense: dissociation
Strasberg SM, et al: Jaundice: an important, poorly recognized risk between bile complement and acute phase reactants in cholecystitis,
factor for diminished survival in patients with adenocarcinoma of the J Surg Res 42(4):434–439, 1987.
head of the pancreas, HPB (Oxford) 16(2):150–156, 2014. Yanaga K, et al: Intraperitoneal septic complications after hepatectomy,
Su Z, et al: Factors influencing infectious complications after pancre- Ann Surg 203(2):148–152, 1986.
atoduodenectomy, J Hepatobiliary Pancreat Sci 17(2):174–179, Yang T, et al: Risk factors of surgical site infection after hepatic resec-
2010. tion, Infect Control Hosp Epidemiol 35(3):317–320, 2014.
Sudo T, et al: Perioperative antibiotics covering bile contamination Yeh DC, et al: Bacterial translocation after cirrhotic liver resection: a
prevent abdominal infectious complications after pancreatoduode- clinical investigation of 181 patients, J Surg Res 111(2):209–214, 2003.
nectomy in patients with preoperative biliary drainage, World J Surg Zerr KJ, et al: Glucose control lowers the risk of wound infection in
38(11):2952–2959, 2014. diabetics after open heart operations, Ann Thorac Surg 63(2):356–
Sudo T, et al: Specific antibiotic prophylaxis based on bile cultures is 361, 1997.
required to prevent postoperative infectious complications in pancre- Zimmitti G, et al: Systematic use of an intraoperative air leak test at
atoduodenectomy patients who have undergone preoperative biliary the time of major liver resection reduces the rate of postoperative
drainage, World J Surg 31(11):2230–2235, 2007. biliary complications, J Am Coll Surg 217(6):1028–1037, 2013.
CHAPTER 13
Clinical investigation of
hepatopancreatobiliary disease
Ali W. Majeed and Ahmed Al-Mukhtar
Patients with hepatopancreatobiliary (HPB) disease are seen in hypertension and may require urgent investigation with endos-
a variety of ways, depending on the underlying condition and copy (see Chapter 81). Shortness of breath and ankle swelling
rapidity of onset. Acute conditions are often inflammatory, and may indicate hypoalbuminemia, which can occur in many acute
pain is a predominant symptom. Jaundice is a common feature and chronic HPB disorders. A family history may point to
in many benign and malignant HPB diseases. A thorough and familial HPB disorders, and a drug history is important, espe-
complete history and clinical examination, followed by basic cially when the liver function is deranged. Social circumstances
blood tests, is a crucial starting point in the patient’s journey and employment are important to establish, especially with
from presentation to diagnosis, management, and discharge. occupational associations with HPB disorders (e.g., chronic
With the ease and availability of sophisticated high-resolution alcoholic liver and/or pancreatic disease in pub landlords).
scanning, there might develop a tendency for the physician to Accurate recording of alcohol intake must be made, because
“scan first, clinic later,” and this must be avoided. A thorough ongoing heavy alcohol use impacts treatment for chronic liver
clinical assessment will guide the appropriateness and use of or pancreatic diseases. Occasionally, patients attend the clinic
further tests and must include a careful balancing of potential smelling of alcohol in their breath, yet deny ongoing alcohol
risk versus benefit of any investigation or intervention in any use, and this should be noted. A request for a blood alcohol
particular individual, depending on their fitness, comorbidity, level could be made, provided the patient consents to it. A past
and personal or family wishes. This chapter describes the medical history should be obtained to include any major ill-
common symptoms and signs of HPB disease, the value of basic nesses and any abdominal surgery. A record of comorbidities
investigations, and how this initial assessment guides further and exercise tolerance should be made, as this will guide the
management. Clinical presentations and investigation of spe- surgeon in assessing fitness for future intervention if required
cific HPB diseases are also detailed. (see later).
CLINICAL HISTORY EXAMINATION OF THE ABDOMEN

The physician should begin the history with the presenting Any examination of the abdomen must commence with a
complaint and then proceed with a systematic inquiry into HPB general inspection of the patient (look for gynecomastia) (Fig.
symptoms, followed by a generic systematic inquiry of the other 13.1), with attention, then move to the hands. Clubbing of
bodily systems. This should be followed by a family history, the fingernails (Fig. 13.2), palmar erythema (Fig 13.3), koil-
drug history, social circumstances, employment, and travel. onychia (Fig. 13.4), and a flapping tremor of the hands are all
Abdominal pain is often a presenting symptom, and a features of chronic liver disease (see Chapters 76 and 81). The
detailed enquiry of the site, severity, radiation, and rapidity of eyes are examined next, looking for pallor and jaundice (Fig.
onset will allow some clues to a differential diagnosis. Sudden- 13.5), and the mouth and tongue inspected for dehydration.
onset severe upper abdominal pain radiating to the back is “Foetor hepaticus” is a characteristic fruity odor in patients
characteristic of acute pancreatitis (see Chapters 55 and 56), with decompensated cirrhosis and occurs due to exhaled
whereas right-side upper quadrant pain may indicate acute thiols. Attention then moves to the abdomen. Full exposure of
inflammation of the gallbladder (see Chapter 33). Any associ- the abdomen in good light is important for an adequate
ated nausea and vomiting should be recorded along with a inspection. Particular attention should be paid to any scars
history of jaundice (sclerae), color of urine and stools, and any from previous abdominal surgery, abdominal distension, and
associated itching. A history of weight loss may point to a areas of discoloration. The presence of spider nevi (Fig 13.6)
malignant process. Nausea and vomiting are common features may indicate chronic liver disease. Veins radiating from the
in acute presentations of HPB disease and may be a feature of umbilicus (caput medusa) may indicate portal venous obstruc-
gastric outlet obstruction in a malignant process. A history of tion (Figs. 13.7 and 13.8). Palpation of the abdomen should
abdominal distension may point to an ileus in acute presenta- begin with a general light palpation, looking for obvious
tions and to ascites or bowel obstruction in chronic liver disease masses or areas of tenderness. It is wise to examine the area
or a malignant process. Steatorrhea is characterized by the of pain indicated by the patient last. Deeper palpation is con-
passage of foul-smelling stools, which float on water and may ducted to seek deeper areas of tenderness and masses. This is
indicate pancreatic exocrine insufficiency (see Chapters 57, 58, followed by palpation of the liver (and gallbladder) and the
and 62). Any history of hematemesis or melena points to portal spleen.
224
Chapter 13 Clinical investigation of hepatopancreatobiliary disease 225
FIGURE 13.3 Palmar erythema.
FIGURE 13.1 Bilateral gynecomastia.
FIGURE 13.4 Koilonychia.
FIGURE 13.2 Clubbing of fingernails. FIGURE 13.5 Jaundice.

226 PART 2 DIAGNOSTIC TECHNIQUES
FIGURE 13.6 Spider nevus in a patient with cirrhosis.
FIGURE 13.9 Coronal magnetic resonance image of a Riedel lobe of

liver. This normal variant, a tongue like extension from the right lobe, may
be mistaken for a mass or an enlarged liver on physical examination.
Note the “wandering spleen.”
The healthy liver is usually impalpable; however, in very thin

individuals, the anterior edge may be palpable. Enlargement of
the liver occurs in many pathologic states, although a tongue
like extension from the right lobe—a Riedel lobe, which is of no
FIGURE 13.7 Caput medusa. pathologic significance—may be mistaken for a tumor (Fig.
13.9). Causes of hepatomegaly are listed in Box 13.1. A lobe
may undergo hypertrophy and become palpable, and this may
occur in the presence of hemiliver atrophy or after liver resec-
tion. Reduction in liver size also is important because this may
occur in cirrhosis and certain types of hepatitis. Consistency also
is relevant; a hard, knobbly liver often represents the presence
of metastases, whereas smooth enlargement may be due to cir-
rhosis (Fig. 13.10). Ascites (Fig 13.11) causes abdominal dis-
tension and should be sought by the elucidation of a fluid thrill
in tense ascites and shifting dullness in moderate ascites. Finally,
the ankles should be examined, looking for pitting edema.
Examination of the spleen should begin in the right iliac
fossa and proceed toward the left subcostal region, because this
is the direction in which splenomegaly occurs (Fig 13.12).
Rotation of the patient 45 degrees to the right may aid palpa-
tion, because the spleen then falls onto the examining right
hand. During this maneuver, the left hand should support the
rib cage and relax the skin and abdominal musculature by
drawing these down to the right. Percussion may be useful,
and if ascites is present, the spleen may be ballotable. If the
spleen is sufficiently enlarged, the notch on its anterior border
FIGURE 13.8 Computed tomography scan showing caput medusa.
may become palpable. Causes of splenomegaly are listed in
Box 13.2.
The liver is examined using a combination of palpation and
percussion from above and below to delineate its borders. Dull- Portal Hypertension
ness to percussion of the upper border extends as far as the fifth Portal hypertension is due to either intrahepatic or extrahepatic
intercostal space. Auscultation is also important because a portal venous obstruction (see Chapters 76 and 79). Intrahe-
venous hum may be heard with portal hypertension, and a bruit patic portal obstruction may be accompanied by splenomegaly
may be heard in association with hepatocellular carcinomas. and ascites. Dilated abdominal wall veins also may be found
BOX 13.1 Causes of Hepatomegaly

Variant Anatomy
Riedel lobe
Low-lying diaphragm
Inflammatory
Hepatitis
Abscesses, amebic and pyogenic
Schistosomiasis
Cirrhosis, early
Sarcoid
Biliary obstruction, especially extrahepatic
Metabolic
Amyloid
Steatohepatitis
Glycogen storage disease
Hematologic
Leukemias
Lymphomas
Myeloproliferative disorders
Sickle cell disease
Porphyrias
Tumors
Primary, benign and malignant
Secondary
Cardiovascular
Cardiac failure
Hepatic vein obstruction
BOX 13.2 Causes of Splenomegaly

Infection FIGURE 13.10 Massive hepatomegaly.
Acute: viral, bacterial
Chronic: tuberculosis, brucellosis
Parasitic: malaria, schistosomiasis
Hematologic
Leukemias
Hemolytic anemias
Hemoglobinopathies
Portal hypertension, especially extrahepatic
Neoplastic
Lymphomas
Myeloproliferative disorders
Secondary deposits
Inflammatory
Rheumatoid FIGURE 13.11 Ascites with an everted umbilicus and venous disten-
Systemic lupus sion in a patient with cirrhosis.
Amyloidosis
Extrahepatic portal hypertension is usually due to portal
vein thrombosis, and as such it is important to identify any
secondary to portosystemic anastomosis, giving rise to caput history of neonatal infection around the umbilicus, major
medusa (see Figs. 13.7 and 13.8). The more common clinical intraabdominal sepsis or pancreatitis, pancreatic cancer, or a
site of portosystemic anastomosis is at the gastroesophageal blood disorder that might lead to hypercoagulability. These
junction, leading to esophageal varices; any evidence of upper patients almost invariably have splenomegaly, often associated
gastrointestinal blood loss, whether hematemesis or melena, with pancytopenia. In such cases, the liver is normal, but ascites
should be investigated with urgent endoscopy. Rarely, hemor- may be present. A palpable spleen in the upper left abdomen
rhoidal or lower rectal varices may develop in these patients associated with portal hypertension may be found in the event
(Fig. 13.13) as a result of portosystemic anastomosis. of splenic occlusion caused by tumor or chronic pancreatitis.
TABLE 13.1 Child-Pugh Score

A B C
Serum bilirubin <2.0 mg/dL 2.0-3.0 mg/dL >3.0 mg/dL

Serum albumin >3.5 g/dL 3.0-3.5 g/dL <3.0 g/dL
Prothrombin time 10-12 sec 13-15 sec >15 sec
Ascites None Mild Severe
Encephalopathy None Mild Severe
patients. Acute alcoholic hepatitis usually occurs after a bout

of drinking and may be associated with liver tenderness, jaun-
dice, pyrexia, and leukocytosis. These patients often are seen
recurrently with epigastric pain, which can be severe. They will
usually admit to a bout of heavy drinking, which precipitates
their symptoms. Acute pancreatitis should be ruled out by
serum amylase measurement. Acute alcoholic gastritis may also
contribute to their symptoms. Advanced alcoholic liver disease
may be associated with other manifestations of hepatic impair-
ment (e.g., portal hypertension, ascites, hepatomegaly, and
bilateral gynecomastia in males (see Fig 13.1). Decompensated
liver failure results in the development of encephalopathy, jaun-
dice, ascites, fatigue, polydipsia, easy bruising, mood fluctua-
tion and often is marked by nutritional depletion.
Primary Biliary Cirrhosis

Primary biliary cirrhosis, or chronic nonsuppurative destructive
cholangitis, usually affects middle-aged women. Asymptomatic
FIGURE 13.12 Spenomegaly in a patient with acholuric jaundice. individuals may be diagnosed during routine examination,
where they are found to have hepatomegaly, elevated autoanti-
bodies, or elevated plasma alkaline phosphatase. The earliest
symptom is usually unrelenting pruritus, even before clinical
jaundice is apparent, but later the patient may experience overt
jaundice; hepatosplenomegaly; xanthelasma, especially in the
palms and around the eyes; vitiligo; and arthritis. Primary
biliary cirrhosis is often found in association with connective
tissue disease (see Chapter 112).
Primary Sclerosing Cholangitis

Ulcerative colitis and Crohn disease are associated with hepa-
tobiliary disorders, especially primary sclerosing cholangitis
(see Chapter 41). It is important to include a detailed gastro-
intestinal systems review and investigations in any patient with
unexplained liver dysfunction. Symptoms are usually nonspe-
cific, with itching, right upper quadrant discomfort, fatigue,
FIGURE 13.13 Rectal varices (an anal speculum is in place). and weight loss. Occasionally, fever, rigors, and chills may
occur. Clinical examination may reveal hepatomegaly and
Parenchymal liver disease is graded in severity with the splenomegaly.
Child-Pugh score (Table 13.1), which is a valuable tool to
assess disease progression and tolerance of intervention if Budd-Chiari Syndrome
chronic liver disease is complicated by the development of Budd-Chiari syndrome is characterized by venous outflow
hepatocellular carcinoma. obstruction to the liver, which could be intrahepatic or extra-
hepatic (see Chapter 88). Acute venous obstruction is usually
Alcoholic Liver Disease secondary to another pathologic process and presents with
The symptoms and signs of alcoholic liver disease may be abdominal pain, vomiting, hepatomegaly, ascites, and jaundice,
related directly to alcoholism or to secondary hepatocellular which may be mild. Hepatocellular failure leading to death is
dysfunction. Patients commonly experience repeated falls while usually rapid. Chronic venous obstruction of the liver usually
inebriated and may have signs of trauma and bruising. Multiple presents with gross ascites and hepatomegaly. If the inferior
“old” rib fractures are common findings on chest radiographs vena cava is blocked with tumor or thrombus, gross edema of
and are related to such behavior. Alcoholic neuropathy also the legs and superficial venous distension of the abdominal
may be present, especially given the malnourished state of many veins is apparent. Diagnosis is made on imaging.
BOX 13.3 Model for End-Stage Liver Disease (MELD)

MELD = 3.78 × ln [serum bilirubin (mg/dL)] + 11.2 × ln [INR] + 9.57
× ln [serum creatinine (mg/dL)] + 6.43 × etiology (0: cholestatic or
alcoholic; 1: otherwise)
INR, International normalized ratio; ln, natural logarithm.
Acute Liver Failure

Acute liver failure (ALF) is characterized by massive cellular
injury to a previously healthy liver, and it is associated with the
development of hepatic encephalopathy. This can occur as a
result of paracetamol (acetaminophen) toxicity, usually caused
by overdosage, from a massive viral infection or, rarely, by
FIGURE 13.14 Right upper quadrant mass caused by a large liver
cyst.
fulminant autoimmune hepatitis. Rarely, idiosyncratic reactions
to drugs can cause the illness. ALF is associated with a high
mortality and requires careful monitoring. Symptoms include
Hemochromatosis itching, jaundice, right upper quadrant pain or discomfort
Hematochromatosis is an iron overload state in which the (which may mimic acute cholecystitis), and nonspecific flulike
liver is usually affected. Excess iron deposition leads to fibrosis symptoms. Diagnosis is established by the history, massively
and cirrhosis with an increased risk of developing hepatocel- elevated serum aminotransfererases, and rapidly rising pro-
lular carcinoma. Clinically, hemochromatosis usually presents thrombin time. Orthotopic liver transplantation is the only
between the ages of 40 and 60 years and predominantly in men. effective treatment for this condition, although some cases of
Symptoms are usually nonspecific, with lethargy, increased pig- ALF can recover spontaneously. Donor organ shortage has
mentation, loss of body hair, loss of libido, arthralgia, and dictated the development of criteria that can help to predict
diabetes, usually due to the involvement of the pancreas. Clini- whether a patient with ALF has a potential to recover or
cal examination may show hepatomegaly and hyperpigmenta- whether a liver transplant is required. Various sets of criteria
tion in the axillae, groins, and genitalia as well as testicular have been proposed, of which the Model for End-Stage Liver
atrophy. Diagnosis is made by hematologic tests, including Disease (Box 13-3) and the King’s College Hospital criteria are
serum iron, serum ferritin, and serum transferrin; liver damage the best known. Most criteria utilize prothrombin time/
can be assessed with a liver biopsy. international normalized ratio, serum bilirubin, serum creati-
nine, arterial blood pH, and hepatic encephalopathy in catego-
Polycystic Disease and Large Liver Cysts rizing patients’ disease severity. Acute-on-chronic liver failure
Symptoms of polycystic disease are usually caused by the mass occurs frequently and is characterized by acute decompensation
effect of the enlarged liver or the large cyst within it (Figure of liver cirrhosis, multiorgan failure and a high mortality. This
13.14). Abdominal distension and breathlessness are common. is now considered a clinical entity, and scoring systems for its
Early satiety and vomiting can occur because of pressure on the prognostication have recently been developed (Karvellas et al,
stomach. Acute abdominal pain is usually due to infection or 2014; Theocharidou et al, 2014) (see Chapters 79, 80, and
bleeding into a cyst (see Chapters 75 and 90B). 114).
Autoimmune Hepatitis Liver Masses

Patients with autoimmune hepatitis (AIH) may be seen with Patients with large liver masses may be seen initially with right
nonspecific symptoms, including fatigue, anorexia, jaundice, upper quadrant discomfort or a palpable mass in the upper
and arthralgia, although 10% may be seen with acute liver abdomen. More often, liver masses are discovered during
failure. Two types of AIH are now recognized: Type 1 occurs imaging as part of the investigation of jaundice, nonspecific
in older patients (predominantly female) and is characterized abdominal symptoms, or follow-up of malignancy. If the liver
by antinuclear or anti–smooth muscle antibodies (ANA and mass is discovered incidentally, a full clinical history should be
SMA, respectively); type 2 AIH is seen in patients at a younger obtained, with particular attention to gastrointestinal and respi-
age and is characterized by anti–liver-kidney-microsomal type ratory symptoms. Details and duration of oral contraceptive
1 (anti–LKM-1) antibodies and can be more aggressive than and anabolic steroid use should be recorded, and the possibility
type 1. Associated autoimmune diseases can occur in as many of viral hepatitis should be considered. Relevant past history,
as 30% of patients and must be excluded. These include especially of other malignancies, should be obtained. Some
Sjögren syndrome, autoimmune thyroiditis, hemolysis, rheuma- malignancies (e.g., uveal melanoma) might stay dormant for
toid arthritis, ulcerative colitis, and idiopathic thrombocytope- many years before presenting as metastatic disease in the liver
nic purpura. (see Chapter 70). (Fig. 13.15). A thorough abdominal examination should be
performed, especially for abnormal masses and ascites, and a
Liver Disease in Pregnancy digital rectal examination is mandatory at the initial assessment.
Acute fatty liver of pregnancy may present to the surgeon with Signs of jaundice, liver insufficiency, and development of col-
liver subcapsular hematoma or rupture with massive intraab- lateral circulation should be sought. A comprehensive hemato-
dominal bleeding. Laparotomy for clot evacuation and hemo- logic and biochemical screening (Box 13-4) should include an
stasis may be required. The management is supportive, and assessment of coagulation factors and of common tumor
mortality is high. markers (α-fetoprotein, carcinoembryonic antigen, cancer
antigen (CA) 19-9, CA 125). Biopsy of liver masses should not

be performed until potentially curable conditions have been
ruled out due to the risk of extrahepatic seeding. The two most
important of these are metastatic colorectal cancer and primary
hepatocellular carcinoma (see Chapters 89 to 95).
GALLBLADDER AND BILIARY TRACT DISEASE

Gallbladder
Tenderness and guarding in the right hypochondrium exacer-
bated by inspiration (Murphy sign) suggests acute cholecystitis
(see Chapter 33). If the gallbladder is palpable in the presence
of obstructive jaundice, this suggests malignant obstruction of
the biliary tree (Couvoisier law), which is commonly due to
carcinoma in the head of the pancreas (see Chapter 62). Failure
FIGURE 13.15 Jaundice caused by liver replacement in a patient with to palpate the gallbladder does not exclude malignant disease,
metastatic ocular melanoma. however, and a nonpalpable gallbladder is the rule in malignant
obstruction at the hilus of the liver. A gallbladder that is inter-
mittently palpable may suggest the presence of a periampullary
carcinoma (Kennedy & Blumgart, 1971). Gallbladder disten-
sion and signs of sepsis in the presence of gallstones may
BOX 13.4 Laboratory Investigations in Hepatopancreatobiliary
indicate empyema of the gallbladder. In such instances, initial
Disease treatment consists of percutaneous aspiration and drainage,
Hematologic Tests with a cholecystectomy delayed for some time. Alternatively, an
Complete blood count: erythrocyte sedimentation rate, reticulocyte urgent cholecystectomy may be performed.
count, haptoglobulin levels, Coombs test
Coagulation profile (prothrombin time) Acute Cholecystitis
Acute cholecystitis may begin with a constant dull ache that
Liver Function Tests
slowly increases in severity during time, usually a couple of
Conjugated and unconjugated bilirubin levels, aspartate amino
transferase, alanine aminotransferase, γ-glutamyltransferase, alkaline
days, and may be associated with anorexia, nausea and vomit-
phosphatase ing, and pyrexia (see Chapter 33). It arises as the result of acute
Proteins (albumin, globulin) inflammation of a distended, obstructed, and usually infected
gallbladder. Patients may be more comfortable lying still
Immunologic and Serologic Tests because of the associated localized peritonitis, and coughing or
Antimitochondrial antibodies sneezing may exacerbate the pain. Acute cholecystitis is often
Smooth muscle and antinuclear antibody associated with tenderness (peritoneal irritation) in the right
Immunoglobulins upper quadrant and signs of systemic infection (pyrexia, raised
Hepatitis B surface antigens leukocyte count, raised C-reactive protein [CRP]), and it
Immunoglobulin M antibody to hepatitis A
requires antibiotic therapy to settle the infection. An infected,
Hepatitis C antibodies
Hepatitis E antibodies
obstructed gallbladder sometimes ruptures and causes general-
Cytomegalovirus antibodies ized peritonitis or a liver abscess. It is important to examine
Epstein-Barr virus antibody patients with acute cholecystitis regularly, and if the pain and
Leptospiral agglutinins tenderness do not settle rapidly with antibiotic therapy, percu-
Fasciola complement fixation test taneous drainage of the gallbladder (cholecystostomy) (see
Amebic complement fixation test Chapter 30,) or an emergency cholecystectomy (see Chapter
Hydatid complement fixation test 35) may be required. The Tokyo Guidelines for the diagnosis
Wasserman reaction and other serologic tests for syphilis of acute cholangitis and acute cholecystitis were updated in
2013 (Boxes 13.5 and 13.6) and provide a detailed classifica-
Tumor Markers
tion of various forms and grades of acute and chronic gallblad-
Carcinoembryonic antigen (CEA)
α-Fetoprotein der inflammation and infection (Kimura et al, 2013).
Cancer antigen (CA) 19-9
α1-Antitrypsin levels
Biliary Colic
Biliary colic, which is a clinical entity distinct from acute cho-
Other Tests lecystitis, usually has a crescendo-decrescendo pattern, with the
Serum amylase pain slowly building up to a peak during a few hours; this peak
Serum paracetamol (acetaminophen) level is often associated with vomiting, and the pain then gradually
Serum ammonia subsides during the ensuing few hours. Often the pain radiates
Plasma ceruloplasmin levels
to the back (Berhane et al, 2006), and this pattern of gallblad-
Iron and iron-binding capacity
der pain is thought to arise from gallbladder nociceptors, in
Spot blood alcohol
Urine (urobilinogen, hemosiderin)
response to a rise in intracholecystic pressure caused by strong
Stools (ova and parasites) contraction of the gallbladder against an obstructed neck.
Biliary colic usually settles spontaneously, leaving some residual
BOX 13.5 Tokyo Guidelines for Diagnosis of Cholangitis Cholelithiasis

A. Systemic inflammation Symptomatic cholelithiasis is the commonest indication for
A-1. Fever and/or shaking chills cholecystectomy, and asymptomatic gallstones do not need
A-2. Laboratory data: evidence of inflammatory response
intervention (Schmidt et al, 2011) (see Chapter 32). Unfortu-
B. Cholestasis
nately, the symptomatology of gallbladder stone disease has
B-1. Jaundice
B-2. Laboratory data: abnormal liver function tests been hard to define despite research spanning decades (Johnson
C. Imaging & Jenkins, 1975). A large epidemiologic investigation in
C-1. Biliary dilatation Denmark into the relationship between abdominal symptoms
C-2. Evidence of the etiology on imaging (stricture, stone, stent, etc.) and gallstones concluded that the predictive values of various
Suspected diagnosis: one item in A plus one item in either abdominal symptoms for gallstones were very low. In patients
B or C with gallstones, the prevalence of right upper quadrant abdomi-
Definite diagnosis: one item in A, one item in B and one item nal pain was similar to patients without gallstones but was
in C higher in patients who had previously undergone cholecystec-
Note:
tomy (Jorgensen, 1989; Jorgensen et al, 1991). The larger Mul-
A-2: Abnormal white blood cell counts, increase of serum C-reactive
ticenter Italian Study on Cholelithiasis showed that, in an
protein levels, and other changes indicating inflammation
B-2: Increased serum ALP, GGT, AST, and ALT levels Italian population, the presence of epigastric or right upper
Other factors that are helpful in diagnosis of acute cholangitis include quadrant pain radiating to the right shoulder, forcing the patient
abdominal pain (right upper quadrant or upper abdominal) and to rest, and intolerance to fried or fatty food were good predic-
A history of biliary disease, such as gallstones, previous biliary pro- tors of gallstones (Corazziari et al, 2008). A recent Swedish
cedures, and placement of a biliary stent. study that followed up 503 patients without gallstones for 5
In acute hepatitis, marked systematic inflammatory response is years reported that the incidence of gallstone formation was
observed infrequently. Virologic and serologic tests are required 1.39 per 100 person-years, and that no change was seen in the
when differential diagnosis is difficult. symptoms of these patients before and after the appearance of
Thresholds:
gallstones (Halldestam et al, 2009). A retrospective analysis of
A-1. Fever > 38° C
patients referred for ultrasonography based on their symptoms
A-2. Evidence of inflammatory response
WBC (×1000 microliters) < 4 or > 10 showed that approximately half of these patients had gallstones
CRP (mg/dL) ≥ 1 or gallbladder pathology (Warwick et al, 2014). Studies examin-
B-1. Jaundice T-Bil ≥ 2 (mg/dL) ing the relief of symptoms after cholecystectomy suggest that a
B-2. Abnormal liver function tests significant number of patients (as many as 40%) undergoing
ALP (IU) > 1.5 × STD cholecystectomy will not experience relief of symptoms, and
GGT (IU) > 1.5 × STD this symptomatic outcome was maintained at 10-year follow-up
AST (IU) > 1.5 × STD (Lamberts et al, 2014). Symptom questionnaire assessment
ALT (IU) > 1.5 × STD does show an improvement in 90% of patients, and it is possible
(STD is upper limit of normal value.)
that a perception of global symptoms might affect these out-
ALA, Alanine aminotransferase; ALP, alkaline phosphatase; AST, aspartate aminotransferase; comes, as indicated by patients who were American Society of
CRP, C-reactive protein; GGT, gamma glutamyl transpeptidase; T-Bil, total bilirubin; WBC, white Anesthesiologists (ASA) Grade II (they had mild systemic
blood cells. disease). As indicated by patients who were American Society
of Anesthesiologists (ASA) II reported less relief of symptoms
(Lamberts et al, 2013, 2014). A study comparing the symptom-
BOX 13.6 Acute and Chronic Cholecystitis
atic outcomes after cholecystectomy for functioning and non-
Acute edematous cholecystitis functioning gallbladders showed no difference in outcome
Acute necrotising cholecystitis (Larsen et al, 2007). A questionnaire-based study of symptom
Acute suppurative cholecystitis relief in Finnish patients suggests that patients with severe
Gallbladder perforation
preoperative symptoms are more likely to obtain relief of symp-
Biliary peritonitis
toms than those with mild symptoms (Lill et al, 2014). Various
Pericholecystic abscess
Biliary fistula questionnaire-based studies of symptomatic outcomes show a
Acute emphysematous cholecystitis “global’ improvement in symptoms after surgery; the placebo
Acute acalculous cholecystitis effect of this remains unknown (Lien et al, 2010). An important
Torsion of the gallbladder study outlined the variation in perceptions of what was consid-
Chronic calculous cholecystitis ered a valid indication for cholecystectomy (Scott & Black,
Chronic xanthogranulomatous cholecystitis 1991). These authors showed the case histories of 252 patients
to two panels, one comprising surgeons and the other a mixed
panel of doctors. The mixed panel considered 41% of the oper-
soreness in the right upper quadrant usually due to an element ations appropriate for the indications and 30% inappropriate,
of “chemical” cholecystitis. These attacks of pain can be widely but the surgeons considered 52% appropriate and 2% inap-
variable in their occurrence, and some patients may have an propriate but could not agree on the other 46%. A Korean
interval of many years between attacks; others may have almost study suggests that laparoscopic cholecystectomy led to an
constant discomfort. Some patients report that the pain is trig- improvement in colonic and dyspeptic symptoms (Kim et al,
gered by certain foods, usually fatty foods, and some patients 2014) but does not alter patterns of gastric emptying in patients
are afraid of eating for fear of triggering an attack of pain (Bagaria et al, 2013). It is important that an accurate documen-
(sitophobia). tation of dyspeptic or atypical symptoms be made, and patients
should be counseled that cholecystectomy may not relieve any pancreatitis and tumors of the pancreas can manifest in this
or all of their symptoms (Kirk et al, 2011; Zinsmeister et al, manner. Patients with chronic pancreatitis occasionally are seen
2011). Gallstone disease increases with increasing age, with jaundice, especially during an acute-on-chronic attack,
and older patients have complex disease with poorer outcome and the jaundice may be accompanied by symptoms of endo-
(Kuy et al, 2011). crine and exocrine deficiency (diabetes, steatorrhea, malab-
sorption) (see Chapters 57 and 58).
Gallstones in the Bariatric Population Malignant disease affecting the hepatic ducts at the biliary
Epidemiology studies have historically shown a higher preva- confluence in the hilus of the liver often presents with jaundice,
lence of gallstones in patients who are obese, with an eightfold but in such instances, the gallbladder is not palpable (see
increased risk of gallstones in those with a body mass index of Chapter 51B). Itching can be an initial feature of biliary
40 kg/m2 (Banim et al, 2011) or more; rapid weight loss obstruction, but in prolonged cholestasis, itching may occur in
increases the likelihood of gallbladder stones (Grover & Kothari, association with partial biliary obstruction and significantly in
2014; Moon et al, 2014) (see Chapter 32). It is interesting that primary biliary cirrhosis. Alkaline phosphatase often is mark-
the formation of gallstones occurs in the early period after these edly increased in partial biliary obstruction, even though the
procedures and is associated with impaired gallbladder motility bilirubin may be only marginally elevated. A markedly increased
(Al-Jiffry et al, 2003). With the common performance of bar- or increasing alkaline phosphatase level may be the only sign of
iatric surgical procedures, especially those that restrict endo- an early malignant or benign biliary stricture.
scopic access to the biliary tract (e.g., gastric transection and The distinction between obstructive (surgical) jaundice and
bypass), it is relevant to consider whether prophylactic removal hepatocellular (medical) jaundice is seldom the problem it was
of the gallbladder is justifiable. The incidence of gallbladder in the past. Biochemical tests of liver function (Table 13.2)
stones in these patients varies from 5% to 30%, with a high usually indicate whether the jaundice is obstructive, but this is
proportion being symptomatic and requiring cholecystectomy not always the case. Identification of the cause and level of
(Nagem et al, 2012; Tsirline et al, 2014). A meta-analysis of biliary obstruction is essential. Ultrasound is the preferred
13 studies concluded that prophylactic cholecystectomy is not initial imaging test for diagnosis of biliary obstruction and to
justified (Warschkow et al, 2013). Laparoscopic gastric bypass determine the etiology and level of obstruction. Generally, if
procedures are now routine, and HPB units are seeing an ultrasound shows gallbladder stones, magnetic resonance chol-
increasing number of patients with symptomatic or complicated angiopancreatography (MRCP) is required to confirm the pres-
biliary lithiasis. Laparoscopically assisted transgastric endo- ence of bile duct stones (see Chapter 19). If no gallstones can
scopic retrograde cholangiopancreatography (ERCP) is feasible be seen, computed tomography (CT) scanning is more useful
in this group (Grover & Kothari, 2014). for evaluation of biliary obstruction (see Chapter 18).
Biliary Obstruction Asymptomatic Bile Duct Dilatation

Classically, pain is a discriminating feature in jaundiced patients. Occasionally, patients undergoing abdominal ultrasonography
Patients with biliary tract obstruction resulting from tumor for unrelated reasons are incidentally found to have a dilated
usually have painless jaundice, whereas patients with an acute bile duct. These patients are often referred to hepatobiliary
attack of pain or a long history of intermittent episodes of surgeons for further evaluation. If there are no biliary symptoms
jaundice accompanied by pain frequently have gallstone disease. and if the liver function tests (LFTs) are completely normal,
In the event of cholangitis, rigors and fever also are present (see no further action needs to be taken. The common bile duct
Chapter 43). Jaundice unaccompanied by pain, and especially (CBD) can dilate with age and possibly previous cholecystec-
if the gallbladder is palpable, almost always indicates malig- tomy (Benjaminov et al, 2013; Hunt & Scott, 1989; Majeed
nancy, especially when associated with weight loss (see Chapter et al, 1999; McArthur et al, 2013; Wu et al, 1984). If the patient
62). Back pain usually indicates a pancreatic lesion; chronic is symptomatic and/or the liver function is deranged, an MRCP
TABLE 13.2 Liver Function Tests and Their Clinical Utility

Serum Test Normal Range Clinical Application
Bilirubin
Total 0.1-1.0 mg/dL Raised in jaundice/biliary obstruction/cholestasis
Conjugated 0.1-0.4 mg/dL Raised in biliary obstruction or cholestasis
Unconjugated 0.2-0.7 mg/dL Raised in hemolysis or Gilbert’s disease
Alkaline phosphatase 30-120 IU/L Biliary obstruction (partial or complete)
Aspartate transaminase or serum 6-40 IU/L Prognosticates liver disease. Raised in alcoholic liver disease (ratio of
glutamic oxaloacetic transferase 2 : 1 with ALT)
(AST or SGOT)
Alanine transaminase or 7-40 IU/L Raised in hepatocyte damage (viral, alcohol, hemochromatosis).
serum glutamic pyruvic transaminase Usually lower than AST in alcoholic liver disease. Very high levels in
(ALT or SGPT) fulminant liver disease (acetaminophen, viral)
Gamma glutamyl transpeptidase (GGT) 0-42 IU/L Raised in alcoholic liver disease
Serum albumin 3.5-5.3 g/dL Low in chronic liver insufficiency, pancreatic disease with poor nutrition
Prothrombin time 10-12 sec Raised in prolonged biliary obstruction. Marker of chronic liver
insufficiency
should be performed to exclude the presence of bile duct tion should be less than 40% by using a continuous intravenous
stones. Endoscopic ultrasonography (EUS) is very accurate in cholecystokinin (CCK) octapeptide infusion during a 30-minute
this setting, albeit more invasive than MRCP (Holm & Gerke, period, and there should be a positive therapeutic response
2010) but is a good option in claustrophobic patients. with absence of the recurrent pain for longer than 12 months
after cholecystectomy. Investigation of functional biliary tract
Bile Duct Stones disorders should include upper gastrointestinal endoscopy,
The human bile duct is sensitive to acute distension, and transcutaneous and endoscopic ultrasonography (to rule out
obstruction of the lower end of the bile duct by a gallstone may microlithiasis), and possibly bile sampling for microlithiasis and
cause epigastric pain that may radiate to the chest and back. cholesterol crystals. Unfortunately, no consensus exists regard-
Cholesterol bile duct stones are lighter than bile, and any dis- ing the infusion rate of CCK that in turn can directly influence
tension of the bile duct may cause these stones to float proxi- gallbladder emptying. Delgado-Aros and colleagues (2003) and
mally, releasing the obstruction. The symptoms of bile duct DiBaise and colleagues (2003) reported meta-analyses of
stones are typically intermittent jaundice associated with upper studies examining the symptomatic benefit from cholecystec-
abdominal pain. Ultrasound alone is not sufficiently sensitive tomy based on abnormal CCK-stimulated gallbladder empty-
or specific for the diagnosis of bile duct stones (Boys et al, ing. They found no good evidence to support the use of
2014), and MRCP is recommended if the liver tests are abnor- CCK-stimulated gallbladder ejection fraction to select patients
mal because abnormal LFTs are the most sensitive predictor for cholecystectomy. Similar findings were reported by Rastogi
of bile duct stones (Videhult et al, 2011) (see Chapters 36 and colleagues (2005) and Ozden and colleagues (2003). Criti-
and 37). cal evaluation of CCK-provoked gallbladder pain concludes
that this test was not a predictor of symptomatic relief after
Cholangitis cholecystectomy (Edwards, 2014; Smythe et al, 1998). It is
Acute cholangitis is an infection within a partially or completely interesting to note that in a prospective randomized study
obstructed biliary system (see Chapters 8 and 43). Charcot examining the development of symptoms in patients with
triad is the classic picture of acute cholangitis and comprises uncomplicated gallbladder stones, a CCK-stimulated ejection
jaundice, abdominal pain, rigors, and pyrexia. Hypotension and fraction greater than 60% was independently associated with
confusion are often a feature of acute suppurative cholangitis the development of symptoms.
in which the bile duct is filled with purulent bile under pressure The etiology and pathogenesis of calculous and acalculous
(Reynold pentad). The Tokyo Guidelines for the diagnosis and gallbladder pain are poorly understood despite some insights
management of acute cholangitis were originally published in by mathematical modeling (Ahmed et al, 2000; Bird et al,
2007 and have recently been updated (Takada, 2013). These 2006; Li et al, 2011, 2013; Ooi et al, 2004; Wegstapel et al,
are summarized in Box 13.5. 1999). The clinical outcomes from cholecystectomy for acalcu-
lous biliary pain are similar to those for calculous biliary pain
Acalculous (“Functional”) Biliary Pain with approximately two thirds of patients experiencing symptom
Acalculous biliary pain is considered to be the result of a “func- relief (Ahmed, 2011; Wybourn, 2013). We would therefore cur-
tional” motility disorder of the gallbladder and/or sphincter of rently recommend that patients experiencing recurrent and
Oddi (SO) (Behar et al, 2006). Gallbladder motor dysfunction severe symptoms as defined by the Rome III criteria should be
giving rise to abdominal pain has been defined by the Rome III offered cholecystectomy but should be carefully counseled
Committee on Functional Biliary and Pancreatic Disorders. regarding the uncertainty of its outcome, and this counseling
These criteria are listed in Box 13.7. Biliary and pancreatic pain should be clearly documented.
should be defined by site, severity, modality of onset, and dura-
tion, and by the absence of typical symptoms of gastrointestinal Sphincter of Oddi Dysfunction
reflux disorder, functional dyspepsia, and irritable bowel syn- SO dysfunction may be present in patients with an intact biliary
drome. There should also be the absence of gallstones, biliary tree, but it usually comes to light as a potential diagnosis when
sludge, or microlithiasis. An abnormal gallbladder ejection frac- symptoms continue after a cholecystectomy. It could be argued
that the diagnosis of SO disorder should not be made in patients
with a gallbladder in situ. The potential risks of cholecystec-
tomy are outweighed by the potential risk of endoscopic biliary
intervention, and even SO manometry and the symptoms
BOX 13.7 Rome III Criteria for Functional Gallbladder Disorder between the two conditions can be indistinguishable. SO dys-
Diagnostic criteria. Must include all of the following:
function has been subclassified into a biliary or pancreatic type
1. Criteria for functional gallbladder and sphincter of Oddi disorder (Box 13.8). Biliary type SO dysfunction in postcholecystectomy
2. Gallbladder is present patients has been arbitrarily subclassified into types I, II, and
3. Normal liver enzymes, conjugated bilirubin, and amylase/lipase III. Biliary type I SO dysfunction is defined as pain, elevated
LFTs documented on two or more occasions, delayed contrast
Rome III Criteria for Functional Biliary Sphincter of Oddi drainage, and a dilated CBD with a corrected diameter of
Disorder 12 mm or more at ERCP. Biliary type II SO dysfunction is char-
Diagnostic criteria must include both of the following: acterized as pain only with one or two of the criteria in type I.
1. Criteria for functional gallbladder and sphincter of Oddi disorder Biliary type III SO dysfunction is defined as recurrent biliary-type
2. Normal amylase/lipase
pain with none of the criteria from type I. Toouli (2009) con-
Supportive criterion
Elevated serum transaminases, alkaline phosphatase, or conjugated
cluded that accurate sphincter manometric data is crucial to
bilirubin temporarily related to at least two pain episodes establish a diagnosis of SO dysfunction. Invasive investigation
of sphincter dysfunction and/or sphincterotomy is associated
with a risk of pancreatitis in as many as 30% of patients. Inser- pancreatic enzymes are elevated soon after an attack of pain
tion of prophylactic pancreatic stents is advocated as a means and whether the pancreatic duct is dilated greater than 5 mm.
of reducing this risk, and some evidence exists to support this In patients seen with pancreatitis, it is important to exclude
practice. Injection of botulinum toxin into the sphincter also common causes of acute pancreatitis (see Acute Pancreatitis
has been advocated as a “therapeutic test” before a sphincter- later). Noninvasive investigations include secretin-stimulated
otomy is performed. In a literature review, Sgouros and Pereira ultrasonography of the pancreatic duct and secretin-induced
(2006) concluded that biliary sphincterotomy reliably treats MRCP. Invasive investigations include pancreatic SO manom-
symptoms in 85%, 69%, and 37% of patients with types I, II, etry, botulinum toxin, or stent drainage. Total endoscopic pan-
and III SO dysfunction, respectively, and pancreatic sphincter- creatic SO sphincterotomy is the currently recommended
otomy may benefit as many as 75% of patients. They state that treatment (Behar, 2006).
sphincter manometry is the gold standard in the investigation.
Sphincter manometry may itself cause pancreatitis, and PANCREAS
noninvasive investigation with MRCP, ultrasonography, and
hepatoiminodiacetic acid choledochoscintigraphy should be Acute Pancreatitis
conducted first. The Milwaukee criteria are a useful summary Acute pancreatitis is usually accompanied with severe upper
of expected outcomes after sphincterotomy for biliary SO dys- abdominal pain that develops rapidly, becomes constant, and
function (Table 13.3). Pancreatic type SO dysfunction is also persists for several hours (see Chapters 55 and 56). The pain
subclassified into types I, II, and III according to whether typically starts in the epigastrium but may soon spread through-
out the abdomen and radiate into the back. Sometimes the
patient may find relief by sitting forward, but usually he or she
lies still and curled up in bed. The pain can be so sudden in
BOX 13.8 Rome III Criteria for Functional Gallbladder and
onset that, at times, it can simulate a perforated peptic ulcer,
Sphincter of Oddi Disorders
and indeed, acute pancreatitis can mimic almost every cause of
Clinical Features acute abdominal pain and may even produce chest pain that
Episodes of right upper quadrant pain or epigastric pain that must simulates myocardial infarction, pneumonia, or pleurisy. Fre-
have ALL the following features: quent vomiting and sequestered fluid in the small bowel may
Episodes lasting 30 minutes or longer lead to rapid and severe dehydration. Even keeping the stomach
Recurrent symptoms occurring at different intervals (not daily)
empty with a nasogastric tube may not prevent the patient from
The pain builds up to a steady level
The pain is moderate to severe enough to interrupt the patient’s daily
retching, and marked nausea is almost a constant feature. Hic-
activities or lead to an emergency department visit coughs can also occur due to diaphragmatic irritation second-
The pain is not relieved by bowel movements ary to extension of the inflammatory process tracking up via the
The pain is not relieved by postural change retroperitoneum.
The pain is not relieved by antacids Patients with acute pancreatitis may exhibit signs of pro-
Exclusion of other structural disease that would explain the found shock with tachycardia, tachypnea, hypotension, and
symptoms confusion. At the other extreme, patients may appear well, with
little in the way of physical signs. The patient may be afebrile
Supportive Criteria on admission, but progression of the inflammatory process
The pain may present with one or more of the following:
leads to fever, facial flushing, and mild jaundice. Abdominal
Associated nausea and vomiting
Radiates to the back and/or right infra-subscapular region
examination usually reveals marked tenderness in the epigas-
Awakens from sleep in the middle of the night trium with guarding, but marked rigidity as seen in a perforated
ulcer is often lacking. The presence of a paralytic ileus may
TABLE 13.3 Milwaukee Criteria Based on the Frequency of Abnormal SO Manometry and Pain Relief by Sphincterotomy
Probably of Pain Relief by
Frequency of Sphincterotomy if Manometry
Biliary SOD Type Abnormal Manometry Abnormal Normal Manometry before Sphincter Ablation
Type 1 75%-95% 90%-95% 90%-95% Unnecessary

Biliary type pain
Abnormal AST or ALT > 2× normal
Delayed drainage of ERCP contrast
from biliary tree > 45 min
Dilated CBD > 12 mm
Type 2 55%-65% 85% 35% Highly recommended
Biliary type pain
Only 1 or 2 of above criteria
Type 3 25%-60% 55%-65% <10 Mandatory
Only biliary type pain
ALT, Alanine aminotransferase; AST, aspartate aminotransferase; CBD, common bile duct; ERCP, endoscopic retrograde cholangiopancreatography; SO, sphincter of Oddi; SOD, sphincter of Oddi
dysfunction.
From Hogan (1988) and Geenen (1989).
TABLE 13.4 Grading of Pancreatic Injury

Type of
Grade* Injury Description of Injury
I Hematoma Minor contusion without duct injury

Laceration Superficial laceration without duct injury
II Hematoma Major contusion without duct injury or

tissue loss
Laceration Major laceration without duct injury or
tissue loss
III Laceration Distal transection or parenchymal injury

with duct injury
FIGURE 13.16 Cullen sign (periumbilical bruising) in a patient with IV Laceration Proximal transection or parenchymal
acute pancreatitis. injury involving ampulla
V Laceration Massive disruption of pancreatic head
*Advance one stage for multiple injuries up to grade III.
“idiopathic” pancreatitis. Acute pancreatitis can also follow

ERCP (see Chapter 29); abdominal surgery, especially to the
stomach, duodenum, and biliary tree; coronary artery bypass
surgery; trauma; and certain viral illnesses, especially mumps.
The association with alcohol abuse varies according to country
and urbanization, being as high as 65% in parts of urban
America. Biliary disease is responsible for greater than 30% of
cases in a general population, and sieving the feces of patients
with gallstone-associated pancreatitis shows gallstones present
in a high proportion of them (Kelly, 1976), prompting many
to speculate that the pancreatitis resulted from the temporary
lodging of a stone in the ampulla of Vater. Postoperative pan-
FIGURE 13.17 Grey-Turner sign (flank bruising) in a patient with hem- creatitis accounts for 5% to 10% of cases (Thompson et al,
orrhagic pancreatitis. 1988), and hyperamylasemia can occur in approximately 10%
of patients undergoing ERCP, although only 1% will experience
severe acute pancreatitis. The cause of traumatic pancreatitis is
result in abdominal distension with a tympanic percussion note again subject to geographic variation (Jones, 1985); penetrating
and diminished bowel sounds. Later features may include a trauma is more common in the United States, and blunt trauma
discoloration around the umbilicus (Cullen sign; Fig. 13.16) or and injury as a result of traffic accidents are more common in
in the flanks (Grey-Turner sign; Fig. 13.17) as a result of bruis- the United Kingdom. Pancreatic injuries were graded by Moore
ing within the fascial planes, pancreatic ascites, and a pleural and colleagues in 1990, and this classification is recommended
effusion. A mass is rarely present, unless the inflammatory by the American Association for Trauma Surgery (Table 13.4).
process has resulted in a pancreatic abscess or sizeable pseudo- Other rare causes include hyperparathyroidism, hyperlipid-
cyst. General features of acute pancreatitis may include confu- emia, and certain drugs, including steroids, azothiaprine,
sion, disorientation, hypoxemia, and even adult respiratory diuretics, fenofibrate, mesalazine, and angiotensin-converting
distress syndrome. The associated hypocalcemia rarely results enzyme inhibitors (Douros et al, 2013; Ksiądzyna, 2011;
in clinical signs of neuromuscular irritability. Nakashima & Howard, 1977), although caution is advised
Rarely, acute pancreatitis can occur without severe pain, and against attributing pancreatitis to drugs (Tenner, 2014).
in these cases, the diagnosis can be missed, and such failure However, once all these causes have been excluded, for approx-
may be associated with a poor prognosis (Lankisch et al, 1991). imately 25% of patients, no clear etiology is identified. The sex
It is important to be aware of the condition and consider it in ratio for acute pancreatitis is virtually equal but tends to have
any patient who experiences sudden shock or anuria, especially a higher peak age incidence in women, reflecting the higher
if it follows biliary surgery, ERCP, coronary artery bypass, or incidence of gallstones; in men the lower peak age reflects their
trauma. A detailed history is therefore vital, particularly includ- propensity to alcohol abuse.
ing the past medical history and family history, especially in
children. Any history of alcohol abuse, biliary colic, jaundice, Diagnosis and Severity Scoring
recent surgery, or ERCP should be actively sought. Approxi- A raised serum or urinary amylase level is essential for the
mately half of the cases of acute pancreatitis are caused by diagnosis of acute pancreatitis, and the value of measurement
gallstones and a further 25% caused by alcohol. The etiology of serum amylase is for diagnosis only. The amylase level does
is not clear in 20% to 25% of patients, and these are labeled not reflect severity of pancreatitis and does not indicate the
likely clinical course. Occasionally, the diagnosis is made by CT

scan in patients who are seen late with abdominal pain after the
hyperamylasemia has subsided. Raised urinary amylase can be
helpful in such cases. Once pancreatitis is diagnosed, its severity
needs to be graded. The main factors to consider in severity
grading at presentation are the systemic effects of the initial
inflammatory insult. These are respiratory (pO2, tachypnea),
renal (oliguria, urea, creatinine), and cardiovascular (tachycar-
dia, hypotension). Numerous clinical scoring systems have
been proposed. Of these, the Revised Atlanta Classification is
comprehensive (Banks, 2013; Talukdar, 2014). Other scoring
systems (e.g., Bedside Index of Severity in Acute Pancreatitis)
have also been proposed (Senapati, 2014; Zhang, 2014). These
systems are designed to predict the course of the acute pancre-
atitis from a mild self-limiting attack to severe sepsis, necrosis,
and multiorgan failure. Although age is not considered in the
severity scoring, older patients with significant comorbidity tol-
erate moderate to severe pancreatitis poorly and have a high
mortality.
Pancreatic pseudocysts may develop as a sequel of acute
pancreatitis. These may cause symptoms due to their size
(gastric or duodenal compression) or sepsis, or they may erode
into vessels causing pseudoaneurysms with the risk of cata-
strophic bleeding.
Chronic Pancreatitis
There are few pathognomonic or diagnostic signs for chronic
pancreatitis, and diagnosis usually depends on the history and
special investigations (see Chapters 57 and 58). The patient
may appear malnourished, with signs of weight loss, they may
be icteric as a result of compression of the CBD by the fibrotic
process in the head of the pancreas, and erythema ab igne may
be present on the back (Fig. 13.18) or abdomen. Uncommonly,
a palpable mass, pseudocyst, or splenomegaly may be evident,
and occasionally pancreatic ascites, if there has been rupture of
a pseudocyst or of the duct system. FIGURE 13.18 This patient has long-standing chronic pancreatitis that
affects the entire length of the pancreas with severe flank and back pain
Chronic pancreatitis is a difficult condition both to diagnose
on the left side. Prolonged application of local heat has resulted in
and to treat. It is crucial to exclude other associated conditions, marked rash of erythema ab igne.
such as pancreatic cancer. Once the diagnosis is established, it
is important to (1) identify any predisposing cause, (2) stage
the severity of the disease, and (3) document any morphologic discomfort, but others may simply complain of intermittent
and functional changes against which any progression can be attacks. In some, severe bouts may be brought on by ingesting
assessed. One of the most common predisposing factors is alcohol, although often no clear association is made between
alcohol abuse, although chronic pancreatitis is also associated attacks and oral intake. In many the pain is so debilitating that
with biliary disease, congenital abnormalities such as an annular it interferes with the patient’s domestic, social, and occupa-
pancreas (Dowsett et al, 1989), ductal abnormalities, and tional activities and can have a devastating impact on the
rarely, hyperparathyroidism. An autosomal dominant heredi- quality of life. Indeed, in many it results in a change in person-
tary form of the condition may also be a factor, so a family ality with a manipulative tendency often associated with an
history is crucial (Teich, 2008). Polymorphisms at PRSS1- addiction to opiates. These patients can be particularly difficult
PRSS2 and CLDN2-MORC4 gene loci were found to be associ- to assess, especially if they are already addicted to alcohol.
ated with alcoholic and nonalcoholic chronic pancreatitis Other presenting symptoms include marked steatorrhea
(Derikx et al, 2015; Rai et al, 2014). Rarely is any specific drug resulting from diminished lipase production and inadequate
or dietary factor implicated. fat absorption (Di Magno, 1982). The stool frequency may be
Pain is the most common presenting symptom of chronic increased to as often as six times a day, and the stools them-
pancreatitis and is often the most difficult to control. The pain selves become pale and bulky with an offensive odor, and they
is usually situated in the epigastrium and subcostal regions and float, making them difficult to flush away. However, this is not
commonly radiates through to the back. Pain can be severe and a constant feature of chronic pancreatitis; some patients may
penetrating in nature, and often the patient will lean forward even suffer from constipation as a result of the increased use of
or curl up to relieve the discomfort. In severe cases, the patient opiate analgesia.
may even kneel on all fours to gain relief. Occasionally, the pain The incidence of diabetes mellitus is a variable and often
may radiate to the left shoulder tip. Variation in the pattern of late feature. As chronic pancreatitis progresses, there may be
pain in these patients is enormous, and some are never free of loss of endocrine function, initially requiring dietary control
and subsequently oral hypoglycemic agents or even insulin. and allows the patient to proceed to endoscopy and endoscopic
Other more general features of chronic pancreatitis may include sphincterotomy when appropriate (see Chapter 29). In patients
weight loss, nausea and vomiting, jaundice, and even gastroin- with acute pancreatitis, the duodenum and small bowel often
testinal bleeding as a result of splenic vein thrombosis, although are filled with gas, and detection of CBD stones is less
this is rare. reliable.
Nutritional depletion is a common feature in patients with Serum amylase is of no value in predicting the severity of an
chronic pancreatitis and is reflected in cachexia and low serum attack, and scoring systems have been developed, such as the
albumin levels. This should be managed in conjunction with Ranson and Glasgow scores. Scoring systems are accurate in
specialized HPB (hepato-pancreato-biliary) nutritionists and predicting outcome only after 48 hours. Inflammatory markers
may include a period of fine-bore nasogastric feeding before such as CRP are also are valuable, but similarly to the Ranson
any intervention is contemplated. and Glasgow scores, such markers may not assess severity accu-
rately until 48 hours after onset. The Acute Physiology, Age,
Pancreatic Cancer and Chronic Health Evaluation score is a more accurate predic-
Regrettably, physical signs of pancreatic cancer are often absent tor of severity at admission, but its complexity limits its routine
or infrequent, until the lesion is unresectable (see Chapters 61 clinical use.
and 62). The patient may be symptomatic with weight loss, but In patients who remain unwell 1 week after onset, or in
unless the lesion is situated in the head of the pancreas such those who deteriorate after initial improvement, pancreatic
that it obstructs the CBD early, causing obstructive jaundice, ischemia and necrosis should be suspected, and a contrast-
few other signs may be evident. With distal bile duct obstruc- enhanced CT scan should be performed. Patients judged to
tion, the gallbladder may become distended (Courvoisier gall- have a severe attack of pancreatitis at onset are at high risk of
bladder). An abdominal mass and ascites are late signs and developing a systemic inflammatory response syndrome and
usually herald inoperability. Systemic signs such as flitting require very careful and repeated observation with attention to
thrombophlebitis migrans may be apparent, but this is uncom- fluid balance, nutrition, and sepsis. Prophylactic antibiotic
mon and nonspecific. All too often cachexia is the only physical therapy in this group of patients is useful and may prevent the
finding. development of infected pancreatic necrosis. Patients with
Later clinical features of pancreatic cancer are pain, weight clinically and biochemically severe pancreatitis should be
loss, and jaundice. The pain is often severe and gnawing and managed in conjunction with critical care physicians in a suit-
radiates to the back. It is often worse at night, when the patient able environment. Patients may go on to experience pancreatic
is lying flat, which usually indicates tumor extension to the pseudocysts that may cause pain or other effects as a result of
retroperitoneum, often signifying unresectability. Weight loss pressure on neighboring organs, especially the stomach.
may be marked, and in many cases of tumors in the body of Nausea, vomiting, or early satiety should alert the physician,
the pancreas, weight loss may be the only symptom, until pain as these pseudocysts are easily detectable on CT scanning.
indicates invasion of the peripancreatic tissue. The more rapid Continuing evidence of systemic sepsis usually indicates the
the weight loss, the less likely the lesion is to be resectable presence of infected pancreatic necrosis, which can be con-
(Bakkevold et al, 1992). Jaundice can be associated with severe firmed by fine needle aspiration.
pruritus, but this classic presentation only occurs when the In patients with chronic pancreatitis, serologic blood tests
lesion is in the head of the pancreas; when jaundice occurs with such as amylase, lipase, and elastase are not helpful, but CRP
tumors of the body or tail of the pancreas, it usually signifies may indicate active inflammation. Fecal elastase may indicate
hilar liver or nodal metastases. exocrine deficiency, whereas routine hematologic and biochem-
Approximately 5% of patients with pancreatic cancer will ical tests, including glycosylated hemoglobin, are important to
have diabetes mellitus within the 2 years before diagnosis provide a baseline for endocrine assessment.
(Pannala et al, 2009). Other signs such as ascites, an epigastric It may be difficult to differentiate between pancreatic carci-
mass, or enlarged supraclavicular nodes (Virchow sign) are noma and some cases of chronic pancreatitis as there are no
indications of inoperability. Because of the lack of specific signs clear-cut markers for this, although lymphoplasmacytic scleros-
and symptoms, very commonly a diagnostic dilemma arises as ing pancreatitis is usually accompanied by an elevated serum
to whether a lesion is chronic pancreatitis or pancreatic cancer. immunoglobulin G (see Chapter 57). The tumor marker CA
Frequently, diagnostic imaging cannot provide the answer, and 19-9 has a sensitivity of 80% and a specificity of 75% for pan-
even at operation, it may not be apparent which condition the creatic cancer (Kim et al, 1999), but as the level tends to
surgeon is facing; this occurred in as many as 15% of cases in correspond with tumor volume, potentially curative tumors
some series (Mao et al, 1995). may be associated with normal levels (Eloubeidi et al, 2004;
Schlieman et al, 2003), which limits its use as a screening tool.
INVESTIGATION OF PATIENTS WITH Very high levels of CA 19-9 are associated with a higher inci-
PANCREATIC DISEASE dence of metastatic disease (Maithel et al, 2008), and laparos-
copy is advocated to determine the presence of subradiologic
In patients with acute pancreatitis, diagnosis usually is made metastatic disease in patients with a CA 19-9 level greater than
based on the history and the serum amylase and lipase levels. 150 kU/L (nonjaundiced patients) and greater than 300 kU/L
An elevated serum amylase can fall quite rapidly, and it may in jaundiced patients (Halloran et al, 2008).
be normal if measured 2 to 3 days after the onset of the attack. A contrast-enhanced CT scan provides information as to the
Urinary amylase may be elevated in these patients; however, a extent and nature of any mass and in particular helps differenti-
contrast-enhanced CT scan often confirms the diagnosis but is ate solid from cystic lesions, vascular involvement or encase-
limited for identifying gallstones. An ultrasound scan is usually ment, and lymph node metastases (see Chapter 18). No
the first investigation of choice, because it detects gallstones significant advantage of gadolinium-enhanced magnetic
resonance imaging over multidetector CT has been seen in the in such morbidity that the patient never returns home—a poor
staging of pancreatic cancer (Park et al, 2009), but multidetec- postoperative outcome. The same applies to major hepatic
tor CT has a better diagnostic yield (Satoi et al, 2009). MRCP resections. Calculation of risk for cancer resections must include
can be useful in the differential diagnosis of pancreatic cystic all the potential oncologic outcomes: recurrent malignancy after
masses and duct delineation, and EUS may contribute substan- resection, anticipated survival without resection, survival with
tially to the assessment of ampullary tumors (Kahaleh et al, alternative treatment modalities. As well as clinical judgment,
2004), for image-guided aspiration of pancreatic cysts, and to there are a number of scoring systems that have been devised
obtain biopsies of intrapancreatic tumors (see Chapter 19). over time to assist in quantifying the risk associated with surgery:
ASA grading, Lee’s Revised Cardiac Risk Index (RCRI), and
ASSESSMENT OF FITNESS FOR MAJOR Physiologic and Operative Severity Score for the enUmeration
HEPATOPANCREATOBILIARY SURGERY of Mortality and Morbidity (POSSUM), to name a few. More
recently, objective fitness assessments have been undertaken to
Major hepatic and pancreatic resections (the majority of which assist in the prediction of the poorer outcome: 6-minute walk
will be undertaken for cancer) inflict a large physiologic and test (6MWT), intermittent shuttle walk (ISW), and cardiopul-
metabolic insult on patients with a significant risk of both peri- monary exercise testing (CPET). Of these, CPET has shown
operative mortality and morbidity. It is therefore useful to quan- the most potential to be a reliable predictor of mortality and
tify such risk so that the patient and their families can make major morbidity in patients undergoing major hepatic and pan-
informed choices (see Chapters 24 and 25). For example, creatic resections (Snowden, 2013).
surgery in an elderly patient with a borderline resectable pan-
creatic cancer requiring vascular resection and reconstruction References are available at expertconsult.com
may not lead to death within 30 days of surgery but may result
Chapter 13 Clinical investigation of hepatopancreatobiliary disease 238.e1
REFERENCES Halloran CM, et al: Carbohydrate antigen 19-9 accurately selects

patients for laparoscopic assessment to determine resectability of
Ahmed M, Diggory R: Acalculous gallbladder disease: the outcomes pancreatic malignancy, Br J Surg 95(4):453–459, 2008.
of treatment by laparoscopic cholecystectomy, Ann R Coll Surg Engl Hogan WJ, Geenen JE: Biliary dyskinesia, Endoscopy 20(Suppl 1):179–
93(3):209–212, 2011. 183, 1988.
Ahmed R, et al: In vitro responses of gallbladder muscle from patients Holm AN, Gerke H: What should be done with a dilated bile duct?
with acalculous biliary pain, Digestion 61(2):140–144, 2000. Curr Gastroenterol Rep 12(2):150–156, 2010.
Al-Jiffry BO, et al: Changes in gallbladder motility and gallstone forma- Hunt DR, Scott AJ: Changes in bile duct diameter after cholecystec-
tion following laparoscopic gastric banding for morbid obestity, Can tomy: a 5-year prospective study, Gastroenterology 97(6):1485–1488,
J Gastroenterol 17(3):169–174, 2003. 1989.
Bagaria D, et al: Effect of laparoscopic cholecystectomy on gastric Johnson AG, Jenkins A: Flatulent dyspepsia: a tiered symptom complex?
emptying and dyspepsia, Surg Endosc 27(9):3116–3120, 2013. Med Hypotheses 1(6):204–206, 1975.
Bakkevold KE, et al: Carcinoma of the pancreas and papilla of Vater: Jones RC: Management of pancreatic trauma, Am J Surg 150(6):698–
presenting symptoms, signs, and diagnosis related to stage and tumour 704, 1985.
site: a prospective multicentre trial in 472 patients. Norwegian Pan- Jorgensen T: Abdominal symptoms and gallstone disease—an epide-
creatic Cancer Trial, Scand J Gastroenterol 27(4):317–325, 1992. miological investigation, Hepatology 9(6):856–860, 1989.
Banim PJR, et al: The aetiology of symptomatic gallstones quantifica- Jorgensen T, et al: Persisting pain after cholecystectomy—a prospective
tion of the effects of obesity, alcohol and serum lipids on risk. Epi- investigation, Scand J Gastroenterol 26(1):124–128, 1991.
demiological and biomarker data from a UK prospective cohort study Kahaleh M, et al: Factors predictive of malignancy and endoscopic
(EPIC-Norfolk), Eur J Gastroenterol Hepatol 23(8):733–740, 2011. resectability in ampullary neoplasia, Am J Gastroenterol 99(12):2335–
Banks PA, et al Group APCW: Classification of acute pancreatitis–2012: 2339, 2004.
revision of the Atlanta classification and definitions by international Karvellas CJ, Bagshaw SM: Advances in management and prognostica-
consensus, Gut 62(1):102–111, 2013. tion in critically ill cirrhotic patients, Curr Opin Crit Care 20(2):210–
Behar JC, et al: Functional gallbladder and sphincter of Oddi disorders, 217, 2014.
Gastroenterology 130(5):1498–1509, 2006. Kelly TR: Gallstone pancreatitis: pathophysiology, Surgery 80(4):488–
Benjaminov F, et al: Effects of age and cholecystectomy on common 492, 1976.
bile duct diameter as measured by endoscopic ultrasonography, Surg Kennedy A, Blumgart LH: Carcinoma of the ampulla of Vater and
Endosc 27(1):303–307, 2013. duodenum, J Clin Pathol 24:765, 1971.
Berhane TV, et al: Pain attacks in non-complicated and complicated Kim GH, et al: Fate of dyspeptic or colonic symptoms after laparo-
gallstone disease have a characteristic pattern and are accompanied scopic cholecystectomy, J Neurogastroenterol Motil 20(2):253–260,
by dyspepsia in most patients: the results of a prospective study, 2014.
Scand J Gastroenterol 41(1):93–101, 2006. Kim H, et al: A new strategy for the application of CA19-9 in the dif-
Bird NC, et al: Investigation of the functional three-dimensional ferentiation of pancreaticobiliary cancer: analysis using a receiver
anatomy of the human cystic duct: a single helix? Clin Anat 19(6): operating characteristic curve, J Gastroenterol Hepatol 94:1941–1946,
528–534, 2006. 1999.
Boys JA, et al: Can ultrasound common bile duct diameter predict Kimura Y, et al: TG13 current terminology, etiology, and epidemiology
common bile duct stones in the setting of acute cholecystitis? Am J of acute cholangitis and cholecystitis, J Hepatobiliary Pancreat Sci
Surg 207(3):432–435, discussion 435, 2014. 20:8–23, 2013.
Corazziari E, et al: Gallstones, cholecystectomy and irritable bowel Kirk G, et al: Preoperative symptoms of irritable bowel syndrome
syndrome (IBS) MICOL population-based study, Dig Liver Dis predict poor outcome after laparoscopic cholecystectomy, Surg
40(12):944–950, 2008. Endosc 25(10):3379–3384, 2011.
Delgado-Aros S, et al: Systematic review and meta-analysis: does gall- Ksiądzyna D: Drug-induced acute pancreatitis related to medications
bladder ejection fraction on cholecystokinin cholescintigraphy commonly used in gastroenterology, Eur J Intern Med 22(1):20–25,
predict outcome after cholecystectomy in suspected functional 2011.
biliary pain? Aliment Pharmacol Ther 18(2):167–174, 2003. Kuy S, et al: Age matters: a study of clinical and economic outcomes
Derikx MH, et al: Polymorphisms at PRSS1-PRSS2 and CLDN2- following cholecystectomy in elderly Americans, Am J Surg 201(6):
MORC4 loci associate with alcoholic and non-alcoholic chronic 789–796, 2011.
pancreatitis in a European replication study, Gut 64(9):1426–1433, Lamberts MP, et al: Persistent and de novo symptoms after cholecys-
2015. tectomy: a systematic review of cholecystectomy effectiveness, Surg
DiBaise JK, et al: Does gallbladder ejection fraction predict outcome Endosc 27(3):709–718, 2013.
after cholecystectomy for suspected chronic acalculous gallbladder Lamberts MP, et al: Patient-reported outcomes of symptomatic chole-
dysfunction? A systematic review, Am J Gastroenterol 98(12):2605– lithiasis patients following cholecystectomy after at least 5 years of
2611, 2003. follow-up, Surg Endosc 28(12):3443–3450, 2014.
Di Magno EP: Controversies in the treatment of exocrine pancreatic Lankisch PG, et al: Undetected fatal acute pancreatitis: why is the
insufficiency, Dig Dis Sci 27:737–742, 1982. disease so frequently overlooked? Am J Gastroenterol 86:322–326,
Douros A, et al: Drug-induced acute pancreatitis: results from the 1991.
hospital-based Berlin case-control surveillance study of 102 cases, Larsen TK, et al: The influence of gallbladder function on the symp-
Aliment Pharmacol Ther 38(7):825–834, 2013. tomatology in gallstone patients, and the outcome after cholecystec-
Dowsett JF, et al: Annular pancreas: a clinical, endoscopic and immu- tomy or expectancy, Dig Dis Sci 52(3):760–763, 2007.
nohistochemical study, Gut 30(1):130–135, 1989. Li WG, et al: A mechanical model for CCK-induced acalculous gall-
Edwards MA, et al: Pain provocation and low gallbladder ejection frac- bladder pain, Ann Biomed Eng 39(2):786–800, 2011.
tion with CCK cholescintigraphy are not predictive of chronic acal- Li WG, et al: Anisotropic behaviour of human gallbladder walls, J Mech
culous gallbladder disease symptom relief after cholecystectomy, Dig Behav Biomed Mater 20:363–375, 2013.
Dis Sci 59(11):2773–2778, 2014. Lien H-H, et al: Changes in quality-of-life following laparoscopic cho-
Eloubeidi MA, et al: Endoscopic ultrasound-guided fine needle aspira- lecystectomy in adult patients with cholelithiasis, J Gastrointest Surg
tion biopsy of suspected cholangiocarcinoma, Clin Gastroenterol 14(1):126–130, 2010.
Hepatol 2(3):209–213, 2004. Lill S, et al: Elective laparoscopic cholecystectomy for symptomatic
Geenen JE, et al: The efficacy of endoscopic sphincterotomy after uncomplicated gallstone disease: do the symptoms disappear? Surg
cholecystectomy in patients with sphincter-of-Oddi dysfunction, Endosc 28(6):1816–1820, 2014.
N Engl J Med 320:82–87, 1989. Maithel SK, et al: Preoperative CA 19-9 and the yield of staging
Grover BT, Kothari SN: Biliary issues in the bariatric population, Surg laparoscopy in patients with radiographically resectable pancreatic
Clin North Am 94(2):413–425, 2014. adenocarcinoma, Ann Surg Oncol 15(12):3512–3520, 2008.
Halldestam I, et al: Incidence of and potential risk factors for gallstone Majeed AW, et al: The preoperatively normal bile duct does not dilate
disease in a general population sample, Br J Surg 96(11):1315–1322, after cholecystectomy: results of a five year study, Gut 45(5):741–
2009. 743, 1999.
238.e2 PART 2 DIAGNOSTIC TECHNIQUES
Mao C, et al: Observations on the development patterns and the Smythe A, et al: A requiem for the cholecystokinin provocation test?
consequences of pancreatic exocrine adenocarcinoma, Arch Surg Gut 43(4):571–574, 1998.
130:125–134, 1995. Snowden CP, et al: Cardiorespiratory fitness predicts mortality and
McArthur TA, et al: The common duct dilates after cholecystectomy hospital length of stay after major elective surgery in older people,
and with advancing age reality or myth? J Ultrasound Med 32(8):1385– Ann Surg 257(6):999–1004, 2013.
1391, 2013. Takada T, et al: TG13: Updated Tokyo Guidelines for the manage-
Moon RC, et al: Comparison of cholecystectomy cases after Roux-en- ment of acute cholangitis and cholecystitis, J Hepatobiliary Pancreat
Y gastric bypass, sleeve gastrectomy, and gastric banding, Surg Obes Sci 20(1):1–7, 2013.
Relat Dis 10(1):64–68, 2014. Talukdar R, et al: Clinical utility of the Revised Atlanta Classification
Moore EE, et al: Organ injury scaling, II: pancreas, duodenum, small of acute pancreatitis in a prospective cohort: have all loose ends been
bowel, colon, and rectum, J Trauma 30(11):1427–1429, 1990. tied? Pancreatology 14(4):257–262, 2014.
Nagem R, Lazaro-da-Silva A: Cholecystolithiasis after gastric bypass: Teich NM: Hereditary chronic pancreatitis, Best Pract Res Clin Gastro-
a clinical, biochemical, and ultrasonographic 3-year follow-up study, enterol 22:115–130, 2008.
Obes Surg 22(10):1594–1599, 2012. Tenner S: Drug induced acute pancreatitis: does it exist? World J Gas-
Nakashima Y, Howard JM: Drug-induced acute pancreatitis, Surg troenterol 20(44):16529–16534, 2014.
Gynecol Obstet 145(1):105–109, 1977. Theocharidou E, et al: The Royal Free Hospital score: a calibrated
Ooi RC, et al: The flow of bile in the human cystic duct, J Biomech prognostic model for patients with cirrhosis admitted to intensive
37(12):1913–1922, 2004. care unit. Comparison with current models and CLIF-SOFA score,
Ozden N, DiBaise JK: Gallbladder ejection fraction and symptom Am J Gastroenterol 109(4):554–562, 2014.
outcome in patients with acalculous biliary-like pain, Dig Dis Sci Thompson JS, et al: Postoperative pancreatitis, Surg Gynecol Obstet
48(5):890–897, 2003. 167(5):377–380, 1988.
Pannala RB, et al: New-onset diabetes: a potential clue to the early Toouli J: Sphincter of Oddi: function, dysfunction, and its manage-
diagnosis of pancreatic cancer, Lancet Oncol 10:88–95, 2009. ment, J Gastroenterol Hepatol 24(Suppl 3):S57–S62, 2009.
Park HS, et al: Preoperative evaluation of pancreatic cancer: compari- Tsirline VB, et al: How frequently and when do patients undergo cho-
son of gadolinium-enhanced dynamic MRI with MR cholangiopan- lecystectomy after bariatric surgery? Surg Obes Relat Dis 10(2):313–
creatography versus MDCT, J Magn Reson Imaging 30(3):586–595, 321, 2014.
2009. Videhult P, et al: Are liver function tests, pancreatitis and cholecystitis
Rai P, et al: Frequency of SPINK1 N34S mutation in acute and recur- predictors of common bile duct stones? Results of a prospective,
rent acute pancreatitis, J Hepatobiliary Pancreat Sci 21(9):663–668, population-based, cohort study of 1171 patients undergoing chole-
2014. cystectomy, HPB (Oxford) 13(8):519–527, 2011.
Rastogi A, et al: Controversies concerning pathophysiology and man- Warschkow R, et al: Concomitant cholecystectomy during laparoscopic
agement of acalculous biliary-type abdominal pain, Dig Dis Sci Roux-en-Y gastric bypass in obese patients is not justified: a meta-
50(8):1391–1401, 2005. analysis, Obes Surg 23(3):397–407, 2013.
Satoi S, et al: Pre-operative patient selection of pancreatic cancer Warwick AM, et al: Natural history of acalculous biliary symptoms,
patients by multi-detector row CT, Hepatogastroenterology 56(90): Ann R Coll Surg Engl 95(7):511–514, 2014.
529–534, 2009. Wegstapel HB, et al: The relationship between in vivo emptying of the
Schlieman MG, et al: Utility of tumor markers in determining resect- gallbladder, biliary pain, and in vitro contractility of the gallbladder
ability of pancreatic cancer, Arch Surg 138:951–956, 2003. in patients with gallstones: is biliary colic muscular in origin? Scand
Schmidt M, et al: A 24 year controlled follow-up of patients with silent J Gastroenterol 34(4):421–425, 1999.
gallstones showed no long-term adverse events leading to cholecys- Wu CC, et al: Effect of aging on common bile duct diameter: a real-
tectomy, Scand J Gastro 46(7–8):949–954, 2011. time ultrasonographic study, J Clin Ultrasound 12(8):473–478, 1984.
Scott EA, Black N: Appropriateness of cholecystectomy in the Wybourn CA, et al: Laparoscopic cholecystectomy for biliary dyskine-
United Kingdom: a consensus panel approach, Gut 32(9):1066– sia: which patients have long term benefit? Surgery 154(4):761–767,
1070, 1991. 2013.
Senapati D, et al: A prospective study of the Bedside Index for Severity Zhang J, et al: Comparison of the BISAP scores for predicting the
in Acute Pancreatitis (BISAP) score in acute pancreatitis: an Indian severity of acute pancreatitis in Chinese patients according to the
perspective, Pancreatology 14(5):335–339, 2014. latest Atlanta classification, J Hepatobiliary Pancreat Sci 21(9):689–
Sgouros SP, Pereira SP: Systematic review: sphincter of Oddi dysfunc- 694, 2014.
tion: non-invasive diagnostic methods and long-term outcome after Zinsmeister AR: Factors that predict relief from upper abdominal pain
endoscopic sphincterotomy, Aliment Pharmacol Ther 24(2):237–246, after cholecystectomy, Clin Gastroenterol Hepatol 9(10):891–896,
2006. 2011.
CHAPTER 14
Emerging techniques in diagnostic imaging
Richard Kinh Gian Do
The field of radiology has undergone tremendous growth in the With single energy (or traditional) CT, the x-rays are gener-
new millennium, with continuous advances in diagnostic ated from accelerating electrons in a x-ray tube subject to a
imaging that include innovations in medical devices, such as peak voltage, kVp, that determines a predictable energy spec-
the creation of dual energy–source multidetector computed trum of the emitted x-rays. By adding a second x-ray beam with
tomography (CT); innovations in imaging agents, such as new a different kVp, one can not only create CT images based on
hepatocyte-specific gadolinium-binding contrast agents for electron density but also calculate the concentration of distinct
magnetic resonance imaging (MRI); standardization of quanti- elements present in the body. Dual-energy CT (DECT) thus
tative functional imaging, such as for MR diffusion-weighted takes advantage of the “signature” x-ray interaction profile of
imaging, and innovations in image analysis, such as texture distinct elements, such as calcium (Ca20) and iodine (I53).
analysis, as a radiomics approach to tumor heterogeneity. In The potential applications of DECT are numerous but are
this chapter, we will review a few emerging techniques in diag- primarily separated into two functions. The first is in changing
nostic imaging with relevance to hepatopancreaticobiliary the image contrast between iodine and other elements, for
tumors. example, by increasing the conspicuity of tumors that enhance
using iodinated contrast. At lower x-ray energy, the attenuation
DUAL-ENERGY COMPUTED TOMOGRAPHY of iodinated contrast is magnified compared with soft tissue,
which can alter the conspicuity of subtle enhancing lesions.
Medical imaging with CT has benefited from the development Potential applications include improving hepatocellular carci-
of multidetector technology, leading to faster scanning of the noma (HCC) detection (Altenbernd et al, 2011), liver metas-
abdomen with higher-resolution images, resulting in submilli- tasis detection (Robinson et al, 2010), and pancreatic cancer
meter isotropic image voxels that allow multiplanar reformat- detection (Macari et al, 2010) (Fig. 14.1). A second application
ting and three-dimensional volume rendering (see Chapter 18). of DECT relies on the quantification of specific elements, such
New dose modulation techniques have also reduced the radia- as calcium or iodine. Quantifying iodine can improve our ability
tion dose exposure for patients. However, one of the more to measure treatment response, for example, in therapies in
exciting techniques to have emerged in CT imaging relevant to which tumor vascularization (and enhancement) is affected and
hepatopancreaticobiliary tumors is dual-energy CT (DECT) changes in tumor attenuation are informative (Uhrig et al,
(Coursey et al, 2010; Heye et al, 2012). To understand this 2013). Fat quantification, in patients at risk for hepatic steato-
technique, it is worth discussing a few physical principles sis, for example, is also another potential application of this
underlying CT imaging. technique (Hur et al, 2014). Although DECT is gaining in
CT is based on the application of x-rays, first recognized by popularity as the applications multiply, a number of obstacles
Wilhelm C. Roentgen in 1895. X-rays represent electromag- remain, including standardization and differences in hardware
netic waves (photons) of very high energy and very short wave- and software between vendors (Morgan, 2014). Nevertheless,
lengths that can pass through most objects, allowing us to “see” the potential quantification of certain elements by DECT,
through the body. The degree of x-ray attenuation (or interac- including iodine uptake, is attractive to radiologists who are
tion) by different elements in our body is related to the number increasingly exploring quantitative tools in medical imaging.
of electrons present in each element. The higher the atomic
number of an element, the greater the number of electrons FUNCTIONAL IMAGING WITH MAGNETIC
present that can potentially interact with x-rays. In simplistic RESONANCE IMAGING
terms, x-rays will be more frequently scattered or absorbed by
the photoelectric effect when they travel through bones, which The similarities between MRI and CT are evident: cross-
are high in calcium (Ca20), than when traveling through other sectional high-resolution imaging of abdominal organs, using
soft tissues made of hydrogen (H1), carbon (C6), nitrogen (N7), intravenous contrast to highlight differences in enhancement.
and oxygen (O8), which are lower in atomic number. The Analogous to advances in CT, progress in MRI technology has
density of atoms is another factor contributing to x-ray attenu- benefited from hardware upgrades generating more rapid and
ation; lungs are radiolucent (dark) on radiographs and CT higher-resolution images. However, the added value for MRI
because of the lower density of atoms (and electrons) present versus CT arises from the ability to image patients frequently
in air. On the other hand, the utility of iodinated intravenous without the risk of ionizing radiation and from new soft tissue
contrast used in CT is partly based on the high attenuation of contrast mechanisms (see Chapter 19).
iodine (I53), which can more frequently scatter and absorb MRI is based on the principles of nuclear magnetic reso-
x-rays, thereby providing information on the vasculature and nance (NMR), which harnesses the inherent magnetic proper-
the enhancement of organs. ties of protons (H1) and their electromagnetic interactions. For
239
A B
C D
FIGURE 14.1 Computed tomography (CT) images of a patient with infiltrative hepatocellular carcinoma obtained by dual-source dual-energy CT.
Arterial phase (A and B) and portal venous phase images (C and D) on liver windows demonstrate an ill-defined right posterior hepatic mass with
arterial enhancement, better appreciated on the arterial phase of lower (60) KEV images (A) than traditional higher (77) KEV images (B). At a lower
KEV (A and C), image noise is increased compared with higher KEV images (C and D).
example, tissue contrast can be generated from differences in tiple cancers (Hylton, 2006). Tumor response to chemotherapy
chemical environment that affect tissue relaxivity (e.g., T1- vs. has traditionally been assessed by measurements of tumor size,
T2-weighted imaging), differences in microscopic and macro- such as through guidelines from the Response Assessment Cri-
scopic motion (e.g., diffusion-weighted imaging [DWI] and teria in Solid Tumors rules. With cytotoxic therapies, reduction
vascular flow-sensitive imaging), electromagnetic environment in size of tumors is expected with a good response to therapy.
(e.g., susceptibility imaging), and many others. The soft tissue However, as new targeted therapies may inhibit vascularization
contrast mechanisms generated by MRI are numerous and or act as cytostatic agents, traditional response criteria based
growing, with new sequences exploiting different biologic states on tumor shrinkage may underestimate therapeutic effective-
(e.g., oxygenation of hemoglobin) being developed continu- ness. With DCE-MRI, quantification of tumor vascularity is
ously. Two applications of MRI that emphasize function (and measured to better predict tumor response to antiangiogenic or
not just form) that have received increasing attention are targeted therapy. DCE-MRI is based on the repeated imaging
dynamic contrast-enhanced MRI (DCE-MRI) and diffusion- of an organ or body part during several minutes, during which
weighted MR imaging (DWI or DW-MRI). the initial uptake and subsequent distribution of intrave-
DCE-MRI is a technique that noninvasively characterizes nously administered contrast are recorded (Fig. 14.2). Although
the vasculature of organs and tumors and has the potential to dynamic contrast-enhanced CT is also technically feasible, the
provide predictive or prognostic response biomarkers in mul- increased ionizing radiation associated with obtaining multiple
Chapter 14 Emerging techniques in diagnostic imaging 241
FIGURE 14.2 Consecutive oblique coronal magnetic resonance (MR) images were obtained of the abdomen every 5 seconds, before (first panel,
top left) and immediately after intravenous contrast administration. Contrast can be seen arriving in the abdominal aorta (middle panel, top row),
before enhancing the periphery of the right hepatic lobe peripheral cholangiocarcinoma, on subsequent frames. The entire tumor was imaged during
this dynamic contrast-enhanced MRI sequence, with only a single representative slice through the middle of the tumor shown at various time points.
CT images has limited its development. With MRI, the arrival blood plasma or vascular space (VS) and extracellular extravas-
and distribution of gadolinium (Gd) contrast throughout the cular space (EES); kep, the rate constant between EES and VS;
vasculature and interstitial space is continuously imaged, and and ve, the fractional vascular volume. For example, parame-
through pharmacokinetic modeling, the changes in signal inten- ters can be measured at baseline and compared on follow-up
sity of an organ (or tumor) of interest are converted into perfu- posttreatment scans. Because perfusion parameters are derived
sion parameters (Do et al, 2009). from DCE-MRI, this technique was initially applied for evalu-
Quantitative perfusion parameters obtained from DCE-MRI ation of antiangiogenic therapy, but its use has expanded to
reflect the rate of exchange of Gd from vessels to the interstitial other targeted chemotherapy agents. Although there has been
space and include Ktrans, a volume transfer constant between an increasing number of studies evaluating the utility of
DCE-MRI in hepatobiliary tumors (Konstantinidis et al, 2014; imaging techniques may be standardized across different
Taouli et al, 2013), the imaging protocols for DCE-MRI have medical centers, the variability in imaging interpretation
differed among medical centers, limiting comparison of studies between different radiologists can remain substantial. Thus an
and dissemination of this functional imaging technique. Efforts emerging body of research has focused on the diagnostic per-
are ongoing to reduce the variability and improve standardiza- formance of radiologists. One of the criticisms of published
tion of imaging parameters by the Quantitative Imaging Bio- retrospective radiology studies that report on the accuracy of
marker Alliance, formed by the Radiological Society of North imaging modalities for different clinical questions has been the
America. use of consensus interpretation among expert readers (Bankier
An extension of DCE-MRI that is unique to liver imaging et al, 2010). Individual interpretation, rather than consensus
is the use of hepatobiliary contrast agents to simultaneously interpretation, is the norm in routine clinical practice. The
assess vascular supply (through both the hepatic artery and referring clinician is thus dependent on the training and
portal vein) as well as hepatic function (Sourbron et al, 2012). expertise of the diagnostic radiologist who is assigned to inter-
Although the majority of MR contrast agents containing Gd pret their patient’s imaging study at a particular point in time.
are excreted through the kidneys, gadoxetate (Gd-EOB-DTPA) A method to reduce the potential variability among diagnos-
undergoes both renal and biliary excretion approximately tic radiologists is the establishment of diagnostic imaging crite-
evenly. Thus it is often referred to as a dual-function contrast ria. For example, numerous criteria for the diagnosis of HCC
agent, allowing the evaluation of both tumor and liver paren- have been published by various societies (Cruite et al, 2013)
chymal enhancement in the same study (Sirlin et al, 2014). The (see Chapter 91). The role of imaging diagnosis for HCC is
role of gadoxetate has expanded and is further discussed in unique in its potential impact in patient care, guiding the choice
Chapter 19. of treatment and even affecting the allocation of liver trans-
DWI is an alternative and complementary functional plants. The American College of Radiology (ACR) first released
imaging technique to DCE-MRI that can help predict and a set of guidelines for the diagnosis of HCC termed the Liver
monitor tumor response to therapy (Hamstra et al, 2007; Imaging Reporting and Data System (LI-RADS) in 2013 (Mitch-
Padhani et al, 2009; Thoeny & Ross, 2010). With DWI, one ell et al, 2014). The ACR will regularly update LI-RADS
can measure the apparent diffusion coefficients (ADCs), guidelines to help standardize the reporting and data collection
which provide an estimate of the magnitude of water diffusion of CT and MR imaging of HCC (http://www.acr.org/quality-
in biologic tissues. When ADC is measured in tumors, it pro- safety/resources/LIRADS). Soon after its initial release, it
vides an indirect measure of tumor cellularity at baseline became clear that there were slight discrepancies among
and necrosis following treatment. There is an inverse relation- LI-RADS, the criteria published by the American Association
ship between ADC and tumor cellularity, which is explained for the Study of Liver Disease (AASLD) (Bruix et al, 2011)
by the presence of cell membranes in proliferating tumors, and the Organ Procurement and Transplantation Network
which act as barriers to water motion. Thus rapidly growing (OPTN) (Wald et al, 2013), which were later addressed in the
tumors with higher concentrations of lipid barriers will be 2014 revision. Nevertheless, for all sets of three criteria above,
lower in ADC. Most treatments will lead to disruptions of cell the imaging diagnosis of HCC relies primarily on the assess-
membrane integrity with resulting greater water diffusion and ment of tumor arterial phase hypervascularity and the detection
subsequent increases in ADC. In HCC (Chapiro et al, 2014) of “washout appearance” on portal venous phase or delayed
and colorectal liver metastases (Cui et al, 2008; Koh et al, imaging.
2007), ADC can be used to assess response to chemoemboli- The concept of washout on CT and MR imaging is straight-
zation and chemotherapy, respectively. A study using DWI forward, described as decreased enhancement in a lesion when
to monitor HCC response to sorafenib is also promising compared with the liver parenchyma after initial enhancement
(Schraml et al, 2009). As an indirect measure of tumor cel- on the arterial phase. With LI-RADS, the suggestion is to rec-
lularity and necrosis, DWI is ideally suited as a biomarker to ognize this finding only when it is unequivocally present. The
assess therapy response. Further discussion on DWI can be possibility that different radiologists may interpret “washout
found in Chapter 19. appearance” with some variability was well known (Liu et al,
2013). In fact, a large interobserver variability study since the
DIAGNOSTIC CRITERIA first release of LI-RADS demonstrated that there was only
moderate agreement between radiologists for the detection of
Although functional imaging techniques such as DCE-MRI “washout appearance” (Davenport et al, 2014). The agree-
and DWI are quantitative methods to measure tumor proper- ment for the definitive diagnosis of HCC based on imaging
ties, the majority of diagnostic radiology still relies on the sub- criteria for LI-RADS, AASLD, and OPTN guidelines was also
jective interpretations of images. For example, CT and MRI moderate to substantial at best. Thus despite efforts at stan-
have a critical role in the initial assessment of hepatobiliary dardization with the development of diagnostic guidelines,
lesions, such as in the characterization of a solitary liver mass variability in the performance of diagnostic radiologists can
that is incidentally discovered. A hypervascular solitary tumor remain substantial in specific clinical scenarios and a source of
may represent a benign neoplasm (e.g., focal nodular hyper- uncertainty for clinical management. Further efforts to reduce
plasia), a tumor with malignant potential (e.g., hepatocellular this variability may come in the form of accreditation pro-
adenoma), or a primary malignancy, such as HCC. The tem- grams, similar to that required for breast imagers through the
poral pattern of tumor enhancement, the heterogenous ACR Mammography Accreditation Program and the federal
appearance of the tumor, the conspicuity of the tumor with Mammography Quality Standards Act. These studies are
respect to the adjacent liver parenchyma on different MRI important acknowledgements of the variability inherent in
sequences are a few of the imaging features used by interpret- diagnostic imaging that extend beyond the imaging techniques
ing radiologists as diagnostic clues. Although CT and MR themselves.
Chapter 14 Emerging techniques in diagnostic imaging 243
A B
FIGURE 14.3 Example of textural differences in liver parenchyma on computed tomography images from two patients (A and B). Sample with
magnified rectangular regions of interest demonstrate greater heterogeneity in the second patient (B) than the first patient (A), which can be quanti-
fied by texture parameters.
RADIOMICS emergence of genetic variations within tumors. Quantifying

tumor heterogeneity on imaging, to uncover differences in
Variability in the interpretation of medical imaging partly stems genetic background of individual tumors, can thus form the
from the inherent subjectivity of visual assessments performed basis for patient-specific therapies in cancer treatment.
by diagnostic radiologists. One method to reduce this variability Multiple studies have begun to uncover the potential
will come from the use of computer software to aid in the inter- of texture analysis for outcome analyses in multiple tumors.
pretation of medical images and the quantification of imaging Texture analysis of an image can be defined as the measurement
features. Radiomics is a growing field of study, with a focus on of variations in pixel intensity levels (i.e., gray-scale level) as a
improving image analysis through the extraction of large amounts function of space. Thus, rather than saying that one tumor looks
of advanced quantitative features of medical images, through more heterogeneous than another simply by subjective visual
automatic or semiautomatic software that can provide more and assessment, one can provide quantitative measures of heteroge-
better information than a physician (Lambin et al, 2012). The neity based on accepted texture parameters, such as contrast,
quantification of imaging features can include the description of entropy, and other higher-order statistics (Fig. 14.3). The poten-
tumor enhancement at the individual pixel level on CT or MRI, tial of radiomics to capture tumor heterogeneity in lung and head
the measurement of tumor heterogeneity by texture analysis, the and neck cancers, and their association with gene-expression
description of tumor border with respect to the surrounding patterns has been described (Aerts et al, 2014). To date, limited
organ, and many others. The features that are reproducible and radiogenomics studies have been performed in hepatopancreati-
most informative are then analyzed for their relationship with cobiliary tumors (Rao et al, 2014; Segal et al, 2007).
treatment outcomes or gene expression (i.e., radiogenomics). Texture analysis is not limited to the imaging of tumors.
An underlying hypothesis for radiomics is that genomic and During the development of liver cirrhosis, the liver acquires a
proteomics patterns of malignancies can be expressed in macro- characteristic nodular contour, and the liver parenchyma is also
scopic image-based features. For tumors, one of the features altered in appearance with an increasingly reticular enhance-
frequently encountered in CT and MRI that is the subject of ment pattern. A number of studies have investigated the poten-
radiomics research is their imaging heterogeneity. tial of texture analysis to quantify liver fibrosis by MRI (Bahl
Until recently, the heterogeneity of tumors on imaging was et al, 2012; House et al, 2015). Texture analysis of liver paren-
commonly observed but seldom assessed qualitatively by diag- chyma also has the potential to predict liver failure after major
nostic radiologists or quantitatively by computer software. hepatic resection (Simpson et al, 2014). Thus quantitative
Tumor heterogeneity observed on imaging has the potential to image analyses can be extended beyond radiomics for tumor
reflect molecular and cellular dynamics that may be specific to prognostication, to the potential prediction of hepatic and pan-
individual patients and may be predictive of response to tar- creatic function.
geted therapies that are increasingly in use (Gatenby et al,
2013). Gatenby and colleagues propose that heterogeneity in SUMMARY
tumor enhancement is based on perfusion deficits, which can
generate significant microenvironmental selection pressures, The field of radiology has undergone tremendous progress
and that adaptive response to heterogeneity can lead to the in recent decades, with continual investments in imaging
technologies to produce more rapid and higher-resolution scan- This acknowledgement is leading to the emergence of standard-
ning. Novel imaging devices, contrast agents, and functional ization in radiologic reports and image acquisition parameters,
techniques continue to emerge and provide improvements in as well as greater interest in the use of computer-based analyses
biologic insight for medical decision making. However, the and the development of radiomics.
variability inherent to imaging technique and diagnostic radi-
ologists themselves can affect the diagnostic performance of an
imaging study, and is increasingly the subject of investigation. References are available at expertconsult.com
Chapter 14 Emerging techniques in diagnostic imaging 244.e1
REFERENCES from two prospective clinical trials, Ann Surg Oncol 21(8):2675–
2683, 2014.
Aerts HJ, et al: Decoding tumour phenotype by noninvasive imaging Lambin P, et al: Radiomics: extracting more information from medical
using a quantitative radiomics approach, Nat Commun 5:4006, 2014. images using advanced feature analysis, Eur J Cancer 48(4):441–446,
Altenbernd J, et al: Dual-energy-CT of hypervascular liver lesions in 2012.
patients with HCC: investigation of image quality and sensitivity, Liu YI, et al: Quantitatively defining washout in hepatocellular carci-
Eur Radiol 21(4):738–743, 2011. noma, AJR Am J Roentgenol 200(1):84–89, 2013.
Bahl G, et al: Noninvasive classification of hepatic fibrosis based on Macari M, et al: Dual-source dual-energy MDCT of pancreatic adeno-
texture parameters from double contrast-enhanced magnetic reso- carcinoma: initial observations with data generated at 80 kVp and at
nance images, J Magn Reson Imaging 36(5):1154–1161, 2012. simulated weighted-average 120 kVp, AJR Am J Roentgenol
Bankier AA, et al: Consensus interpretation in imaging research: is 194(1):W27–W32, 2010.
there a better way? Radiology 257(1):14–17, 2010. Mitchell DG, et al: LI-RADS (Liver Imaging Reporting and Data
Bruix J, Sherman M, D. American Association for the Study of Liver: System): summary, discussion, and consensus of the LI-RADS
Management of hepatocellular carcinoma: an update, Hepatology Management Working Group and future directions, Hepatology
53(3):1020–1022, 2011. 61(3):1056–1065, 2014.
Chapiro J, et al: Radiologic-pathologic analysis of contrast-enhanced Morgan DE: Dual-energy CT of the abdomen, Abdom Imaging
and diffusion-weighted MR imaging in patients with HCC after 39(1):108–134, 2014.
TACE: diagnostic accuracy of 3D quantitative image analysis, Radi- Padhani AR, et al: Diffusion-weighted magnetic resonance imaging as
ology 273(3):746–758, 2014. a cancer biomarker: consensus and recommendations, Neoplasia
Coursey CA, et al: Dual-energy multidetector CT: how does it work, 11(2):102–125, 2009.
what can it tell us, and when can we use it in abdominopelvic Rao SX, et al: Whole-liver CT texture analysis in colorectal cancer:
imaging? Radiographics 30(4):1037–1055, 2010. does the presence of liver metastases affect the texture of the remain-
Cruite I, et al: Imaging-based diagnostic systems for hepatocellular ing liver? United European Gastroenterol J 2(6):530–538, 2014.
carcinoma, AJR Am J Roentgenol 201(1):41–55, 2013. Robinson E, et al: Dual source dual energy MDCT: comparison
Cui Y, et al: Apparent diffusion coefficient: potential imaging bio- of 80 kVp and weighted average 120 kVp data for conspicuity of
marker for prediction and early detection of response to chemo- hypo-vascular liver metastases, Invest Radiol 45(7):413–418, 2010.
therapy in hepatic metastases, Radiology 248(3):894–900, 2008. Schraml C, et al: Diffusion-weighted MRI of advanced hepatocellular
Davenport MS, et al: Repeatability of diagnostic features and scoring carcinoma during sorafenib treatment: initial results, AJR Am J
systems for hepatocellular carcinoma by using MR imaging, Radiol- Roentgenol 193(4):W301–W307, 2009.
ogy 272(1):132–142, 2014. Segal E, et al: Decoding global gene expression programs in liver cancer
Do RK, et al: Dynamic contrast-enhanced MR imaging of the liver: by noninvasive imaging, Nat Biotechnol 25(6):675–680, 2007.
current status and future directions, Magn Reson Imaging Clin N Am Simpson AL, et al: Texture feature analysis for prediction of postoperative
17(2):339–349, 2009. liver failure prior to surgery, San Diego, CA, 2014, SPIE 2014 Medical
Gatenby RA, et al: Quantitative imaging in cancer evolution and Imaging: Image Processing Conference.
ecology, Radiology 269(1):8–15, 2013. Sirlin CB, et al: Consensus report from the 6th International forum for
Hamstra DA, et al: Diffusion magnetic resonance imaging: a biomarker liver MRI using gadoxetic acid, J Magn Reson Imaging 40(3):516–
for treatment response in oncology, J Clin Oncol 25(26):4104–4109, 529, 2014.
2007. Sourbron S, et al: Combined quantification of liver perfusion and func-
Heye T, et al: Dual-energy CT applications in the abdomen, AJR Am tion with dynamic gadoxetic acid-enhanced MR imaging, Radiology
J Roentgenol 199(Suppl 5):S64–S70, 2012. 263(3):874–883, 2012.
House MJ, et al: Texture-based classification of liver fibrosis using Taouli B, et al: Hepatocellular carcinoma: perfusion quantification
MRI, J Magn Reson Imaging 41(2):322–328, 2015. with dynamic contrast-enhanced MRI, AJR Am J Roentgenol
Hur BY, et al: Quantification of the fat fraction in the liver using dual- 201(4):795–800, 2013.
energy computed tomography and multimaterial decomposition, Thoeny HC, Ross BD: Predicting and monitoring cancer treatment
J Comput Assist Tomogr 38(6):845–852, 2014. response with diffusion-weighted MRI, J Magn Reson Imaging
Hylton N: Dynamic contrast-enhanced magnetic resonance imaging as 32(1):2–16, 2010.
an imaging biomarker, J Clin Oncol 24(20):3293–3298, 2006. Uhrig M, et al: Monitoring targeted therapy using dual-energy CT:
Koh DM, et al: Predicting response of colorectal hepatic metastasis: semi-automatic RECIST plus supplementary functional information
value of pretreatment apparent diffusion coefficients, AJR Am J by quantifying iodine uptake of melanoma metastases, Cancer
Roentgenol 188(4):1001–1008, 2007. Imaging 13(3):306–313, 2013.
Konstantinidis IT, et al: Regional chemotherapy for unresectable intra- Wald C, et al: New OPTN/UNOS policy for liver transplant allocation:
hepatic cholangiocarcinoma: a potential role for dynamic magnetic standardization of liver imaging, diagnosis, classification, and report-
resonance imaging as an imaging biomarker and a survival update ing of hepatocellular carcinoma, Radiology 266(2):376–382, 2013.
CHAPTER 15
Ultrasound of the liver, biliary tract, and pancreas
Jill S. Gluskin
ULTRASOUND BASICS visibility in patients with a higher body mass index (Choudhry
et al, 2000).
Ultrasound is a favored imaging modality because of its tre- Acoustic artifacts have been described, many of which are
mendous versatility, low cost, real-time capability, and portabil- clinically useful. Acoustic enhancement is an artifact that occurs
ity. Ultrasound is considered safe at clinical, diagnostic levels, when there is increased through-transmission of sound waves
with no confirmed harmful biologic effects on the operator or in fluid-containing structures, making tissue behind the fluid
patient. Use of Doppler ultrasound also allows assessment of appear artificially bright. This occurs because ultrasound units
blood flow dynamics. Despite these advantages, certain limita- adjust for a certain attenuation of sound with depth of travel
tions influence the applicability of ultrasound. Ultrasound such that there is erroneous amplification of echoes deep to
waves are unable to penetrate bone or air, which can obscure fluid-containing structures. Acoustic enhancement is useful to
lesions and limit the field of view. The quality of ultrasound characterize structures as cystic (Fig. 15.2). Acoustic shadow-
imaging and its interpretation are also operator dependent; ing occurs when a structure attenuates sound more rapidly than
ultrasound imaging and diagnosis is greatly influenced by skill surrounding tissues and casts a dark acoustic shadow beyond
and experience. the object. This occurs with strong reflectors, such as calcifica-
tions, or strong attenuators, such as dense tumors. Acoustic
Principles of Ultrasound Interpretation shadowing is a feature of gallstones and, in conjunction with
In diagnostic ultrasound applications, brief pulses of acoustic mobility, aids in distinction between gallstones and gallbladder
energy are emitted by the transducer, transmitted into the body, polyps (Fig. 15.3).
reflected from acoustic interfaces, and received by the trans- Some artifacts suggest the presence of structures that are
ducer. The average velocity of sound in soft tissues is 1540 m/s, not present. Reverberation artifact occurs when the ultra-
with reduced velocities in fat and increased velocities in bone. sound beam is repeatedly reflected back and forth between
The ultrasound unit uses precise timing to interpret and process two parallel highly reflective surfaces, such as walls of a fluid-
information included in the returning echoes, including fre- filled structure, producing artifactual echoes due to propaga-
quency, amplitude, and phase information. Ultrasound units tion assumption by the ultrasound processor. Reverberations
assume a uniform propagation velocity of sound to compute may produce the false appearance of echoes within fluid or
the depth of the interface reflecting the pulse and then to gener- make a cystic structure appear solid. Comet tail artifact is a
ate an imaging display. This pulse-echo principle allows infer- type of reverberation that occurs when two reflective inter-
ence of the position, nature, and motion of the interface from faces are closely spaced, such as within a punctate crystal,
which the received echo originated. The intensity of the producing posterior echoes that are parallel, evenly spaced
reflected echo is interpreted as varying shades of gray. echogenic bands with a triangular tapered shape. Comet tail
Different ultrasound transducers are optimized for specific artifact allows identification of surgical clips and is a feature of
frequencies. Lower-frequency transducers have poorer resolu- gallbladder adenomyomatosis (Fig. 15.4). Twinkle artifact is a
tion with greater depth of penetration, and thus are used color Doppler artifact that helps to detect and verify crystals
to image deeper structures such as abdominopelvic tissues. and calcifications, particularly if a calculus does not demon-
Higher-frequency transducers have better spatial resolution, strate acoustic shadowing. Twinkle artifact occurs posterior to
but higher-frequency sound attenuates rapidly and has poorer strong reflectors and appears as turbulent color Doppler flow
tissue penetration. Thus it is best used for superficial soft with a mix of red and blue pixels; however, spectral Doppler
tissues. The depth of the target structure is a major determinant tracing demonstrates noise. Additional artifacts have been
of the transducer frequency. described and are outside the scope of this chapter; the reader
is referred to specialized texts (Campbell et al, 2004; Feldman
Gray Scale Ultrasound Terminology and Artifacts et al, 2009; Rubens et al, 2006).
Echogenicity of tissue refers to the reflection or transmission of
ultrasound waves relative to surrounding tissues. Based on gray Doppler Ultrasound
scale imaging, a structure on the image display can be charac- Doppler ultrasound is used to identify and evaluate blood flow
terized as anechoic (uniformly black), hypoechoic (dark gray), in vessels based on the backscatter of blood cells. Echoes
or hyperechoic (light gray) (Fig. 15.1). returning from moving targets change in frequency compared
Tissue harmonic imaging is a gray scale ultrasound setting with the transmitted pulse. This change in frequency is propor-
that is useful in transabdominal ultrasound for depicting cystic tional to the velocity of the reflector relative to the transducer
lesions and lesions containing echogenic material such as air, and is defined by the Doppler effect. Doppler imaging allows
calcification, or fat. Harmonic imaging also improves lesion assessment of vessel patency, direction of blood flow, flow
245
hv
FIGURE 15.1. Transverse gray scale ultrasound image of the liver

shows the right hepatic vein (hv) as an anechoic fluid-containing struc-
ture. Note that the walls of the veins are hyperechoic. The hyperechoic
curved line posteriorly represents the diaphragm.
FIGURE 15.3. Gallstones (curved arrow) layering in the gallbladder

produce an acoustic shadow (straight arrows).
FIGURE 15.2. Benign hepatic cyst with a thin, smooth wall. Note the
posterior acoustic enhancement (arrows) that is a characteristic feature
of a cystic lesion.
FIGURE 15.4. Gallbladder adenomyomatosis with bright reflectors in

velocity, and spectral waveforms. Three different Doppler
the gallbladder wall (arrows) producing comet tail artifact secondary to
displays are available: color, power, and spectral Doppler sound reverberation.
(Fig. 15.5).
Color Doppler provides information about the direction of
motion and differences in flow velocity. Echoes from red blood of the Doppler setting and causes the color display to wrap
cells are displayed in colors designated for flow toward or away around the displayed scale and appear as if there is reversal in
from the transducer. Color assignment and range of displayed flow direction. Color Doppler findings of vascular stenoses
velocities can be modified by the user. Hue and saturation of include color aliasing, disorganized color intraluminally, and
the color display progressively change as velocity increases with color bruit in the surrounding soft tissues, representing distur-
color desaturation toward white at the highest values. Color bances associated with high-velocity jets and poststenotic tur-
aliasing occurs at sites where velocity exceeds the highest value bulence. These findings of stenoses require confirmation by
Chapter 15 Ultrasound of the liver, biliary tract, and pancreas 247
0 19.6 .046
5
19.6
cm/s
.046 GB
PV
10
IVC
A B
15
.13
MHV
.13
PW:3MHz θ = 58°
.40
m/s
C.60
FIGURE 15.5. Doppler ultrasound of the liver. A, Color Doppler mode shows mean flow velocity and direction of flow toward or away from the
transducer, as indicated in the bar display at left of the image. The portal vein (PV) indicates hepatopetal flow. The inferior vena cava (IVC) shows
flow away from the transducer. The hepatic artery (arrowhead) has mixed signals due to high-velocity flow and aliasing. B, Power Doppler shows
amplitude of flow but not direction. It is useful for small vessels and low flow, as shown in this image of vessels in the gallbladder (GB) wall (arrow-
heads). C, Spectral Doppler shows flow pattern over time; when angle corrected, it can measure true velocity. A sample cursor is placed within the
vessel, as shown in this transverse image of the middle hepatic vein (MHV). Flow variations are due to changes in right atrial pressure.
gray scale and pulsed-wave Doppler analysis. Limitations color Doppler image. Arterial waveforms are characterized as
of color Doppler imaging include dependence on angle of high resistance by limited flow during diastole or low resistance
insonation, inability to display the entire Doppler spectrum in by continuous flow during diastole. At sites of stenosis, flow is
the image, and artifacts caused by aliasing and noise. not laminar; instead, flow becomes turbulent, and the spectral
Power (or amplitude) Doppler is a complementary tech- Doppler waveform reflects the red blood cells moving at varying
nique that displays total amplitude of the echo signal but not velocities. Instead of a narrow and well-defined tracing, spectral
flow direction. Power Doppler signal is more sensitive for flow broadening is seen at sites of stenosis with filling-in of the area
detection than color Doppler and is less dependent on the under the spectral waveform.
angle of insonation. It is not subject to aliasing and is less sen- The resistive index (RI) is a useful metric obtained with
sitive to noise. Power Doppler is most useful in showing spectral Doppler that reflects the strength of the inflow pressure
areas of low flow, depicting slow flow in an area of subtotal and the resistance of the downstream microvascular bed. This
occlusion, and demonstrating intralesional vascular patterns. calculation is performed on arterial spectral Doppler waveforms
Modern scanners have programmed settings that automatically and measures the difference between the peak systolic and end-
optimize the velocity scale and pulse repetition frequency to diastolic velocities. A lower-than-expected resistive index may
increase sensitivity for low flow, whether in color or power result from (1) decreased systolic velocity, indicating impaired
Doppler modes. inflow, or (2) increased diastolic flow due to disorganized or
In spectral Doppler, a sample volume cursor is placed within increased overall vascularity downstream. An elevated resistive
the target vessel and displays a waveform of the entire range of index near 1 implies nearly absent diastolic flow and high-
velocities during time, rather than the mean velocity as in a resistance downstream.
LONG LEFT LONG
GB
4A 4B
2 3
PV
IVC
LH
A B
LONG RT
5
8
6
RHV
RK
7
C
FIGURE 15.6. Normal liver longitudinal ultrasound images. A, Left hepatic lobe. Segments are numbered. LH, Left hepatic vein. B, Left lobe
segment IV. Hepatic segments are numbered. GB, Gallbladder; IVC, inferior vena cava; PV, portal vein. C, Right lobe. Hepatic segments are num-
bered. RHV, Right hepatic vein; RK, right kidney.
LIVER ULTRASOUND The Couinaud segmental anatomy of the liver is shown in

Figs. 15.6 and 15.7 (see Chapter 2). The vascular landmarks
Normal Liver Anatomy that define the Couinaud segmental anatomy are well seen by
A normal liver is smooth in contour and uniform in echo- ultrasound (Lafortune et al, 1991; Soyer et al, 1994). Ultra-
genicity. Hepatic parenchyma is hypoechoic to the spleen and sound images are obtained through available acoustic windows
either isoechoic or minimally hyperechoic to renal parenchyma. avoiding bone and air, which will vary the appearance of the
Liver size is most commonly determined sonographically by a liver compared with the standard transverse planes of com-
longitudinal image of the right lobe. Gosink and Leymaster puted tomography (CT) and magnetic resonance (MR).
(1981) found that when the liver measures 15.5 cm or greater Images are best obtained in the subcostal or intercostal
in the midhepatic line, hepatomegaly is present in 75% of approach, with the patient in a supine or left lateral decubitus
patients. In the right midclavicular line, the normal mean position. The proximal portion of the left portal vein is extra-
length is 10 cm, with a standard deviation of 1.5 cm (Niederau hepatic and courses anteriorly and leftward, and thus it is
et al, 1983). In most patients the measurement of liver length often best visualized in the subxiphoid region, whereas the
suffices, but hepatic shape can be variable, and thus three- right portal vein and its bifurcation are best visualized with a
dimensional ultrasound volumetric analysis can aid evaluation subcostal approach. When the left portal vein arises early,
(Treece et al, 2001; Wilson et al, 2009). there is a common pitfall for the subcostal view of the right
TRANS TRANS
4
4 2 3
L
L
M
8 1
R
M
R
IVC
A B
TRANS
TRANS
4B
L
1 5 RA
R
IVC
A RP IVC
RK 6
C D
FIGURE 15.7. Transverse images of the liver. A, Transverse sonogram at the level of the hepatic veins. Hepatic segments are numbered. IVC,
inferior vena cava; L, left hepatic vein; M, middle hepatic vein; R, right hepatic vein. B, Transverse section of the left lobe at the level of the portal
vein bifurcation. L, Left portal vein; M, uppermost main portal vein; R, right portal vein;. C, Transverse image of the left portal vein (LPV) and right
portal vein (RPV) at the bifurcation. Segments IVB and I are indicated. A, Aorta; IVC,inferior vena cava; L, left hepatic vein; R, right hepatic vein; RK,
right kidney;. D, Transverse image of the inferior right lobe segments V and VI. IVC, inferior vena cava; K, Right kidney; RA, anterior right portal vein;
RP, posterior right portal vein.
anterior and posterior sectoral branches to be mistaken as the masses are benign (see Chapters 46, 75, and 90B). Hypoecho-
left and right portal veins. genicity is a feature of many malignancies. Hyperechoic masses
include both benign etiologies, such as hemangioma or focal
Hepatic Masses fat, as well as malignant hepatic neoplasms, such as hepatocel-
Liver Mass Detection and Characterization lular carcinoma (HCC) and metastases.
Echogenicity is the dominant factor determining lesion conspi- Coregistration of ultrasound, CT, and MRI images can
cuity on ultrasound. Markedly hypoechoic or hyperechoic facilitate lesion localization and intervention. Electromagnetic
lesions may be readily detected and characterized by ultra- sensors placed near the scanning area and sensors attached
sound, even when small (Eberhardt et al, 2003). Larger masses to the ultrasound probe allow tracking of transducer position
that have echogenicity similar to adjacent liver may be more and orientation. Anatomic landmarks are identified on initial
difficult to appreciate. Lesion size and body habitus are signifi- scanning, and with navigation software the landmarks are
cant factors that influence lesion detection (Eberhardt et al, coregistered with CT or MR images that have been down-
2003). On gray scale ultrasound, liver masses are differentiated loaded into the ultrasound unit (Ewertsen et al, 2008). While
by internal architecture and are described as cystic, hypoechoic, the ultrasound scan is performed in real time, CT or MR
or hyperechoic relative to the liver. The majority of cystic images are reconstructed into a similar orientation plane
ULTRASOUND MRI
FIGURE 15.8. Coregistration technology shows a transverse ultrasound image of the liver and a magnetic resonance image that have been coreg-
istered to display identical orientation and plane of section. (Courtesy GE Healthcare, Little Chalfont, United Kingdom.)
(Fig. 15.8). Images may be displayed side by side or in a tained enhancement and washout may occur in any phase.
blended, overlapped format. In a study of contrast-enhanced Healthy liver parenchyma enhances brightly during the PV
ultrasound fused in real time with CT, Jung and colleagues phase. Lesions enhance differently than background paren-
(2009) found that more lesions were identified with image chyma and thus will show increased conspicuity within the
fusion than with CT alone. Similar instrumentation allows enhanced parenchyma. Sustained enhancement, meaning that
global positioning system navigation for needle tracking and the lesion remains isoechoic or hyperechoic to the liver, is a
volume acquisition. favorable sign and associated with benignity. In the PV phase,
benign hemangiomas and focal nodular hyperplasia (FNH)
Role of Contrast-Enhanced Ultrasound in Evaluation of Focal have sustained enhancement equal to or greater than the liver
Liver Lesions (see Chapter 90). Washout during the PV phase, meaning the
Contrast-enhanced ultrasound (CEUS) has evolved as a lesion becomes hypoechoic compared with adjacent liver, is an
problem-solving tool for the characterization of focal liver unfavorable sign and correlated with malignant histology such
lesions due to its noninvasive real-time imaging capability. as HCC and metastases (Brannigan et al, 2004; Quaia et al,
Ultrasound contrast agents are intravenously injected micro- 2004; von Herbay et al, 2010; Wilson et al, 2009). These pat-
bubbles that are highly echogenic and oscillate in an ultrasound terns of enhancement can be used to differentiate benign from
field to enhance ultrasound signal in gray scale, color, and malignant focal hepatic lesions and will be described in the
spectral Doppler. Bubbles are capable of transpulmonary lesion subsections (Brannigan et al, 2004; Ding et al, 2005;
passage for left-side blood pool enhancement. Ultrasound con- Wilson & Burns, 2006).
trast agents are purely intravascular and do not diffuse into the CEUS improves both characterization of focal liver masses
interstitium. Multiple injections can be performed to assess and liver lesion detection, with improved ability to see a greater
different features of a lesion (i.e., kinetic curves) or assess dif- number and smaller lesions than on nonenhanced ultrasound.
ferent lesions. Ultrasound is sensitive to extremely low concen- Contrast agents also improve diagnostic accuracy and diagnos-
trations of contrast due to amplification of echoes from tic confidence (Albrecht et al, 2003; Albrecht et al, 2004;
resonating microbubbles in the blood pool. To obtain images Blomley et al, 1999; Bryant et al, 2004; Harvey et al, 2000;
of best quality, ultrasound scanning techniques should optimize Quaia et al, 2004; Strobel et al, 2009). Diagnostic performance
imaging settings, which can vary depending on the contrast of CEUS has been compared with that of CT and MRI, with
agent. Low mechanical index (LMI) settings are usually utilized similar accuracy in recognition of malignancy when performed
to minimize bubble destruction. with expert operators (Trillaud et al, 2009). There is high con-
Microbubbles stay in the circulation, allowing continuous cordance for lesion appearance in arterial phase imaging,
assessment as the contrast agent traverses the imaging field. and less so in the portal phase; this is usually attributed to
Vascular phases are divided into the arterial (10 to 40 ultrasound contrast agents remaining intravascular, compared
seconds), portal venous (PV) (40 to 120 seconds), and late with CT and MR contrast agents that may diffuse into the
parenchymal (>120 seconds) phases (Wilson & Burns, 2006). interstitium (Wilson & Burns, 2006). Ultrasound contrast
Observations are made of the liver and focal liver lesions, agents are not nephrotoxic, and impaired renal function is not
including wash-in, peak enhancement, and washout. Sus- a contraindication (Jang & Yu, 2009).
In summary, CEUS provides real-time dynamic assessment

capability with high spatial and temporal resolution, making it
a valuable component of multimodality imaging. Microbubble
contrast agents are widely used in Europe and Asia. In the
United States, the Food and Drug Administration (FDA) has
not approved microbubble contrast agents for noncardiac use;
thus utilization for abdominal imaging requires off-label uses of
FDA-approved products. The clinical role of CEUS using
existing contrast agents is still being vetted in the United States,
even as new agents are being developed and modified. Ultra- *
sound contrast research is a rapidly evolving and expanding
field, with many potential applications, including imaging,
potentiation therapy, and drug delivery.
Benign Liver Lesions

Cystic Masses
Benign hepatic cysts demonstrate smooth, barely perceptible
walls, posterior acoustic enhancement, and no internal echoes
on ultrasound (see Fig. 15.2) (see Chapters 75 and 90B). A
partial septation, thin septation, or puckered wall may be FIGURE 15.9. Hepatic abscess in a patient after embolization for
present. A confident diagnosis of simple hepatic cyst can be hepatic neuroendocrine metastases (arrowheads). The abscess (aster-
isk) contains complex fluid and has posterior acoustic enhancement
made on ultrasound, and no further evaluation is required.
(arrows).
Blood vessels and arteriovenous fistulas or aneurysms may
appear cystic on gray scale ultrasound, yet can easily be dif-
ferentiated from simple cysts by demonstrating internal blood hemangiomas are vascular, flow is extremely slow, and usually
flow on Doppler imaging. Lymphomatous masses are markedly no Doppler signal is evident. Atypical hemangiomas, seen in
hypoechoic and may mimic simple cysts on gray scale ultra- 20% of cases, may have a thin echogenic rim and a lacy or
sound, yet will demonstrate internal vascularity on color granular central appearance from foci of collagen (Fig. 15.10B).
Doppler. When cysts are innumerable and also present in the Larger hemangiomas often lack characteristic features because
kidneys or pancreas, polycystic disease should be considered. of central fibrosis, necrosis, and myxomatous degeneration.
Cystic masses are deemed complex when there are internal When the background liver becomes hyperechoic as a result of
echoes within the cyst fluid, irregular or thickened walls, mural steatosis, hemangiomas may appear hypoechoic to liver paren-
nodularity, solid areas, or calcification. Differential diagnosis of chyma (Bartolotta et al, 2007; Liu et al, 2009).
complex cystic masses includes hemorrhage, infection, and With CEUS, hemangiomas classically show arterial phase
neoplasm. Cysts with internal hemorrhage may have internal peripheral nodular enhancement. During PV and late phases,
solid components, septations, and layering debris, but with there is centripetal progression of enhancement with complete
no detectable vascularity. Hepatic hematomas from trauma or partial fill-in and no washout (Fig. 15.11) (Brannigan et al,
initially appear hyperechoic and over time will evolve to 2004; Quaia et al, 2004; Wilson & Burns, 2006; Wilson &
become hypoechoic and cystic (Korner et al, 2008; McGahan Burns, 2010).
et al, 2006). If an ultrasound shows a classic appearance of hemangioma
Abscesses may appear solid initially, then cystic with debris; and the patient has no risk factors, history of underlying liver
the wall is usually vascular (Fig. 15.9) (see Chapter 72). Echi- disease (hepatitis, alcohol abuse, fatty liver, etc.), or malig-
nococcal cysts have a variable appearance and may appear as nancy, then no follow-up imaging needs to be performed (Leifer
simple fluid-filled cysts, may contain wavy membranes from a et al, 2000). A study of 213 patients with ultrasound examina-
rupture and detached endocyst, may contain daughter cysts tions that showed a typical hemangioma appearance and no risk
and/or echogenic material, or they may show calcification for hepatic malignancy found only one patient with malignancy
(Lewall & McCorkell, 1985). When hydatid cysts are infected, on long-term follow-up and concluded that typical hemangio-
they lose their characteristic sonographic appearance and mas in low-risk patients do not require follow-up. This sugges-
become diffusely hyperechoic (see Chapter 74). tion does not apply to patients with cirrhosis, hepatitis, or
Cystic neoplasms, such as biliary cystadenoma and cystad- chronic liver disease who are at increased risk for HCC, nor
enocarcinoma, may be multilocular, with cystic locules demon- does it apply to patients with a history of cancer. In a study of
strating different echogenicities depending on the cystic fluid 1982 patients with cirrhosis, ultrasound depicted hemangioma-
content (see Chapter 90B). They may have mural nodularity, like lesions in 44 patients; on follow-up, half of these proved to
nodular thickened septations, and mural or septal calcifications be HCCs, and half were hemangiomas (Caturelli et al, 2001).
(Buetow et al, 1995; Levy et al, 2002). Some metastases may Thus in patients at risk for HCC, any echogenic lesion merits
also be cystic, as discussed later. further evaluation and follow-up.
Hepatic Hemangioma Focal Nodular Hyperplasia

Hemangiomas are characteristically homogeneously hypere- Focal nodular hyperplasia (FNH) has a smooth lobulated
choic circumscribed masses with subtle acoustic enhancement contour and variable echogenicity (see Chapter 90A). When
(Fig. 15.10A) (see Chapter 90A). Increased echogenicity is due an FNH is echoic, it is difficult to visualize, yet subtle vascu-
to multiple vascular interfaces within hemangiomas. Although lar displacement or contour abnormalities may draw attention
A B
FIGURE 15.10. Hemangiomas. A, Longitudinal sonogram of the right lobe shows a brightly echogenic hemangioma (arrow) with a circumscribed
border. B, Atypical hemangioma (arrow) with areas of internal heterogeneity, a result of fibrosis or myxomatous degeneration, and a thin echogenic
rim (arrowhead).
to the lesion (Buetow et al, 1996; Hussain et al, 2004). A key Malignant Liver Neoplasms
to ultrasound diagnosis of FNH is the characteristic Doppler
Hepatocellular Carcinoma
appearance of a central feeding artery with tortuous spoke-
wheel vascularity (Fig. 15.12). A central scar may be evident, Ultrasound is used to screen for hepatocellular carcinoma
yet is usually better depicted on CT, MRI, or CEUS. Classic (HCC) in high-risk patients (see Chapter 91) (Bruix & Sherman,
contrast enhancement of FNH in the arterial phase includes a 2011; European Association for the Study of Liver, 2012).
central stellate-enhancing artery. The lesion becomes homo- When used as a surveillance test, ultrasound has a sensitivity
geneous during late arterial phase and on PV phase there is ranging from 58% to 89% (Bolondi, 2003). A meta-analysis of
sustained enhancement, often with a hypoechoic central scar surveillance ultrasound for detecting early-stage HCC demon-
(Brannigan et al, 2004; Dietrich et al, 2005; Kamaya et al, strated an average sensitivity of 63%, with higher sensitivity for
2009; Quaia et al, 2004; Ungermann et al, 2007; Wilson ultrasound performed every 6 months than with annual surveil-
& Burns, 2006; Wilson & Burns, 2010). Persistent portal lance (Singal et al, 2009). Although ultrasound can detect 1- to
phase enhancement helps differentiate FNH from malignant 2-cm nodules in patients with chronic liver disease, there is a
liver lesions (Kim et al, 2009; von Herbay et al, 2010) high false-positive diagnosis, mostly due to high-grade dysplas-
(Fig. 15.13). tic nodules (Serste et al, 2012). A study comparing ultrasound
screened versus unscreened patients with cirrhosis showed
Hepatic Adenoma improved median survival (17 vs. 12 months) and improved
Hepatic adenomas (see Chapter 90A) have variable echo- survival rates at 1, 3, and 5 years for HCC arising in the
genicity but are often hyperechoic due to high lipid content. screened Child-Pugh class B patients but not in the class C
Internal hemorrhage within adenomas may produce cystic patients (Trevisani et al, 2007).
areas, and calcification may also be present (Grazioli et al, HCC may be solitary, multiple, or diffuse, and may also
2001). Intratumoral veins are often identified on Doppler extend into the bile ducts and cause biliary obstruction or
imaging (Bartolozzi et al, 1997; Golli et al, 1994). On CEUS hemobilia (Kojiro et al, 1982). Small HCC tumors (< 5 cm)
arterial phase, adenomas characteristically show diffuse or cen- are often hypoechoic but may be hyperechoic from fatty meta-
tripetal hypervascular enhancement from dysmorphic arteries. morphosis (Caturelli et al, 2001; Choi et al, 1993). Larger
PV phase appearance is variable, either showing equal enhance- HCC are often heterogeneous because of liquefaction necrosis
ment or soft washout relative to the liver (Bartolotta et al, 2007; and fibrosis (Tanaka et al, 1983). Sonographic depiction of
Kim et al, 2006; Kim et al, 2008; Ricci et al, 2008; Wilson & HCC may be limited in the setting of severe cirrhosis (Bennett
Burns, 2010). et al, 2002).
RT RT
A B
RT
C
FIGURE 15.11. Giant atypical hemangioma. A, Baseline gray scale image demonstrates a mainly isoechoic left lobe mass (arrows) found in a
patient with abdominal pain. B, Contrast-enhanced ultrasound in early arterial phase (19 seconds) demonstrates that the mass has globular peripheral
enhancement but no central enhancement. C, In the portal venous phase (56 seconds), the mass demonstrates more marked peripheral globular
enhancement consistent with hemangioma. (Courtesy Paul S. Sidhu, King’s College Hospital, London.)
A B
FIGURE 15.12. Focal nodular hyperplasia (FNH). A, Gray scale image of a hypoechoic FNH. FNH lesions are often subtle. The echogenic hepatic
parenchyma consistent with steatosis increases conspicuity of the FNH in this patient. B, Color Doppler image demonstrating a tortous vessel with
spoke-wheel pattern that demonstrated arterial waveform on spectral Doppler.
HCC has a propensity for venous involvement with throm- et al, 1999; Smith et al, 2008). Regional lymphadenopathy is
bosis of the portal or hepatic veins (Fig. 15.14). Both acute common.
bland thrombus and tumor thrombus can expand the portal
vein, so vein expansion is not a distinguishing feature. Spectral Liver Metastases
Doppler interrogation of bland thrombus will only show noise. Sonographic appearance of metastases may vary according to
Spectral Doppler of tumor thrombus may reveal hepatofugal the primary malignancy (see Chapters 92, 93, and 94). The
arterial flow, a pathognomonic finding (Giorgio et al, 2004; most common appearance of metastases is hypoechoic lesions
Lencioni et al, 1995; Shah et al, 2007; Tanaka et al, 1993). or lesions with peripheral hypoechoic halos (Fig. 15.16). The
CEUS offers diagnostic advantage in the detection of malignant hypoechoic halo corresponds to fibrosis, compressed sinusoids,
portal vein and hepatic vein thrombosis compared with and tumor neovascularity surrounding the metastatic deposit
conventional gray scale, color Doppler, and spectral Doppler (Kruskal et al, 2000; Wernecke et al, 1992). Hypoechoic
ultrasound (Rossi et al, 2006). metastases are commonly due to metastatic breast, lung, pan-
CEUS improves HCC characterization and is highly con- creatic, gastric, and esophageal carcinoma. In addition to
cordant with CT findings, but contrast-enhanced ultrasound is metastases, lymphoma and HCC should be considered in the
best at identifying lesions greater than 2 cm (Giorgio et al, differential diagnosis of hypoechoic lesions or lesions with
2004; Inoue et al, 2009; Jang et al, 2009; Quaia et al, 2007). hypoechoic halos. Lymphoma is hypoechoic because of its
On arterial phase, HCC typically shows dysmorphic arteries uniform cellularity and may appear as diffusely infiltrative or as
and is hypervascular, sometimes isovascular. HCC classically multiple well-defined hypoechoic lesions; this latter pattern is
demonstrates a washout appearance on PV phase, yet some- more common in non-Hodgkin lymphoma or acquired immu-
times demonstrates delayed washout or no washout. Observa- nodeficiency syndrome (AIDS)-related lymphoma (Townsend
tion during the portal phase should be extended as long as et al, 1989).
300 s, because as many as 22% of HCCs have delayed washout Echogenic metastases tend to arise from HCC or gastroin-
(Jang et al, 2007) (Fig. 15.15). testinal tract malignancies. Vascular metastases are also typi-
Fibrolamellar HCC, a variant of HCC with better prognosis, cally hyperechoic, such as those from renal cell carcinoma, islet
characteristically presents as a large solitary mass in an adoles- cell tumors, carcinoids, and choriocarcinoma (Marchal et al,
cent or young adult without underlying liver disease. On 1985; Rubaltelli et al, 1980; Tanaka et al, 1990). Calcified
ultrasound, a fibrolamellar HCC is well defined, lobulated, and metastases may be found with mucinous adenocarcinoma of
with variable echogenicity. Calcifications may be present the colon as well as other primaries. When metastases are dif-
(Chung et al, 2009). A central scar may be visible on gray scale fusely infiltrative, such as in breast or lung carcinoma, it can
ultrasound in one third to two thirds of patients, but scars are be difficult to detect individual lesions. Diffuse hepatic metas-
better seen with contrast-enhanced ultrasound (McLarney tases are especially prone to pseudocirrhosis; treated or
A B
C D
FIGURE 15.13. Focal nodular hyperplasia (FNH) on contrast-enhanced ultrasound. A, An isoechoic mass (arrows) is seen in the left hepatic lobe
on gray scale imaging. B, Twelve seconds after contrast injection, the FNH (arrows) shows enhancement in a central scar and a characteristic spoke-
wheel vascular pattern. C, The FNH (arrows) enhances brightly at 18 seconds. D, Five minutes after contrast injection, the FNH (arrows) is still
enhanced and is echogenic relative to the liver. (Courtesy GE Healthcare, Little Chalfont, United Kingdom.)
untreated metastases can elicit a fibrotic response mimicking that have diminished in size or changed in echogenicity as a
cirrhosis (Fennessy et al, 2004; Young et al, 1994). result of chemotherapy.
Cystic hepatic metastases are seen in ovarian and pancreatic
cystadenocarcinomas, as well as mucinous colon carcinomas. Sonography of Diffuse Liver Disease
Ovarian cystic metastases are characteristically located periph- Fatty Liver Disease
erally due to tumor implants on the hepatic surface. Metastases Fatty liver is commonly encountered incidentally on sonograms
undergoing central necrosis, such as sarcomas, also may appear of the liver and in patients referred for imaging evaluation of
cystic, yet thick complex walls and internal contents are usually abnormal liver function tests (see Chapters 71 and 100). Fat
evident and distinguish the metastasis from a simple cyst. deposition within hepatocytes results in increased echogenicity
CEUS improves detection of individual metastases com- of the liver, decreased acoustic penetration, and loss of the
pared with gray scale ultrasound and in some cases may reveal echogenic borders of portal vessels (Fig. 15.18A and B). Sono-
more lesions than are evident on CT (Fig. 15.17) (Albrecht graphic diagnosis of fatty liver is challenging because echo-
et al, 2003). Liver metastases are frequently detected sono- genicity is not a quantitative measure. Appearance will vary
graphically, yet CT is the preferred modality to determine depending on whether fat deposition is focal or generalized and
extent of disease, number of hepatic lesions, and extrahepatic on the presence of associated liver disease. Although unproven,
disease. Response to therapy is also better evaluated by CT and some investigators advocate comparing the relative echogenicity
MRI because treatment-related fatty liver may obscure lesions between the liver and right kidney with that between the spleen
TR LIVER
Long PV
RT
A B
Long PV .20
.20
PV
C D
FIGURE 15.14. Hepatocellular carcinoma (HCC) with portal vein thrombosis. A, Transverse image shows a hypoechoic HCC in the right hepatic
lobe (arrows). B, Transverse image of the main portal vein reveals only a trickle of blood flow (arrow). The lumen is nearly filled with echogenic
thrombus. C, Longitudinal image of the main portal vein shows the extent of thrombus (arrow). D, Color Doppler image of the main portal vein
confirms arterial flow (arrow) within the hypoechoic thrombus; this is pathognomonic for tumor thrombus. PV, Portal vein.
and left kidney. Because the normal spleen is more echogenic also be lobar, rounded and masslike, or perivascular in distribu-
than liver, if the difference in echogenicity between the liver and tion. In the latter two cases, MRI is useful to confirm that the
right kidney is greater than the difference between the spleen focal abnormality truly represents fat deposition and not a mass
and left kidney, then the liver is abnormally echogenic (Tchel- lesion. Focal fat deposition should not have mass effect, and
epi et al., 2002). therefore intrahepatic vessels should traverse these areas of
In fatty liver disease, the liver contour remains smooth. increased echogenicity without displacement. However, this
Hepatomegaly may be present. Steatosis degrades the quality finding is not entirely specific, because occasionally some mass
of sonographic evaluation of the hepatic parenchyma because lesions may likewise fail to displace vessels.
of decreased acoustic penetration and focal lesions may escape Certain regions may be spared within a diffusely fatty liver
detection. In addition, lesions that typically appear echogenic (Fig. 15.18C). These are typically areas of nonportal venous
against a background of normal liver, such as hemangiomas, inflow, such as around the gallbladder (cystic veins) and in the
may appear hypoechoic in the fatty liver. region of the porta hepatis. Focal fatty sparing appears as a
Fat deposition in the liver may be diffuse and uniform, hypoechoic region with geographic margins and segmental dis-
patchy, or focal. Areas of focal fat deposition are commonly tribution, usually located in segment IV anterior to the portal
wedge shaped and echogenic and occur in characteristic loca- bifurcation or adjacent to the gallbladder, and leaves vessels
tions, such as along the ventral surface of segment IV and undisturbed. These distinctive features allow a more confident
adjacent to the falciform ligament. Focal fat deposition may diagnosis.
ivc
A B
C D
FIGURE 15.15. Hepatocellular carcinoma. A, Transverse gray scale image reveals a large hepatocellular carcinoma (HCC); the mass (m) splays
the hepatic veins. ivc, Inferior vena cava. B, Contrast-enhanced ultrasound, early arterial portal venous phase, shows dysmorphic vessels in the
hypervascular hepatocellular carcinoma (HCC) (arrows). C, Peak arterial portal phase image shows brightly enhancing HCC. D, The HCC (arrows)
has very slow and weak washout, not shown until more than 4 minutes from the end of the saline flush. (Courtesy Stephanie R. Wilson.)
Viral Hepatitis appearance known as the “starry night” liver for the striking
Most imaging modalities, including sonography, are neither prominence of the periportal echoes against the abnormally
sensitive nor specific for the diagnosis of acute hepatitis. Defini- dark background liver (Mortele & Ros, 2001; Tchelepi et al,
tive diagnosis of hepatitis (see Chapter 70) is usually based on 2002). Hepatitis may also produce secondary changes in the
clinical and laboratory findings, with imaging useful for exclud- gallbladder, with irregular lamellated thickening of the gallblad-
ing other diagnoses (Mortele & Ros, 2001). The most common der wall, especially in hepatitis A (Tchelepi et al, 2002).
sonographic abnormality in hepatitis is hepatomegaly. Peripor-
tal edema may be appreciated sonographically, but it is better Cirrhosis
seen on CT or MRI. Occasionally, sonography may demon- Sonography is neither sensitive nor specific for the diagnosis of
strate very low hepatic parenchymal echogenicity, producing an cirrhosis, yet certain findings may suggest its presence. Surface
Shear wave elastography relies on the displacement of tissues

induced by a force, either external pressure or the radiation
force from a focused ultrasound beam. The displacement of
tissues induces elastic shear waves, which propagate and are
detected by the ultrasound transducer. The velocity of shear
waves correlates with tissue elasticity; the stiffer the tissue, the
faster the shear wave propagates. Shear wave velocity is
expressed in meters per second or in kilopascals using the
Young modulus. (Sarvazyan et al, 2011).
Two primary methods for performing elastography with
ultrasound have been devised. Transient elastography (TE,
FibroScan [Echosens, Paris]) uses an ultrasound transducer
mounted on the axis of a vibrator. Transient vibrations induce
elastic shear waves in the underlying tissues. The device uses a
pulse-echo technique to provide an estimated velocity of the
shear waves. The TE technique is rapid, with high intraobserver
and interobserver reproducibility but is limited in patients with
ascites (Fraquelli et al, 2007; Goyal et al, 2009).
Real-time elastography utilizes radiation force from a focused
ultrasound beam to generate shear waves. The ultrasound
transducer uses a Doppler-like technique to measure shear
FIGURE 15.16. Liver metastases with hypoechoic halo. Metastatic
wave propagation. Unlike TE, real-time elastography can be
neuroendocrine tumors (arrows) in the liver have a hypoechoic halo
peripherally and a target appearance. A hypoechoic halo usually indi- integrated into conventional sonographic equipment and incor-
cates malignant lesion or, less commonly, an infectious etiology. porated into a routine liver ultrasound examination. The opera-
tor searches for an acoustic window in real-time and liver
stiffness measurements are guided by the gray scale image. The
nodularity is the most reliable sign of cirrhosis (see Chapter region of interest is placed at least 2 cm beneath the anterior
76), and nodularity along the deep surfaces of the liver is a more capsule to prevent reverberation artifacts, avoiding focal liver
sensitive sign of cirrhosis than superficial nodularity (Filly et al, lesions and large vessels. Real-time elastography is not limited
2002) (Fig. 15.18D). Volume redistribution includes right side by ascites. This technique is commercially available by several
of the lobe volume loss with relative hypertrophy of the left side vendors with promising results for detection of cirrhosis and
of the lobe and caudate. Other sonographic features of cirrhosis different stages of liver fibrosis (Ferraioli et al, 2014; Fraquelli
include coarsened echotexture and heterogeneous parenchymal et al, 2007; Friedrich-Rust et al, 2007).
echogenicity (Caturelli et al, 2003). Sonoelastography remains an investigational tool for evalu-
Morphologic features of portal hypertension include spleno- ation of focal liver lesions. Both benign and malignant lesions
megaly, ascites, reversal of PV flow direction, and varices. The can be compressible or stiff compared with healthy liver. In
portal vein diameter may exceed 13 mm, although this is not addition, the stiffness of the background parenchyma varies
always the case. Enlarged, tortuous hepatic arteries are often significantly with fibrosis. In some series, the difference in elas-
present in cirrhotic livers and should not be mistaken for dilated tographic measurements is statistically significant; however, the
bile ducts or venous collaterals; color and spectral Doppler are overlap in values between benign and malignant focal liver
useful for making this distinction. lesions makes the technique unreliable for diagnosis (Park et al,
2013; Yu & Wilson, 2012).
Role of Sonoelastography
The prognosis and management of patients with chronic liver Vascular Evaluation
disease largely depend on the extent and progression of liver Portal vein blood flow should be toward the liver (antegrade or
fibrosis. In chronic hepatitis C, fibrosis influences the indica- hepatopetal) and should create a waveform that always remains
tion for antiviral treatment (European Association for the Study above the baseline. Cardiac variability creates hepatic venous
of Liver, 2011). Liver biopsy is the gold standard for liver pulsatility that is transmitted through the liver parenchyma.
fibrosis assessment but is invasive, samples a small portion of This creates a phasic pattern in the normal portal vein wave-
the liver, and is not optimal for repeat assessments or assess- form that is decreased in conditions such as cirrhosis and fatty
ment of uneven distribution of fibrosis. Sonoelastography is a liver (Colli et al, 1994). The portal vein and hepatic artery
noninvasive tool allowing assessment of liver fibrosis by mea- should flow in the same direction. Retrograde flow in the portal
suring tissue elasticity. It allows assessment of a larger paren- vein is confirmed on color Doppler by showing opposite direc-
chymal volume than a liver biopsy sample. Of note, liver tion of flow of the portal vein and hepatic artery. The velocity
elasticity can be affected by other factors than fibrosis, includ- of the portal vein ranges from 16 to 40 cm/s (McNaughton &
ing acute viral hepatitis, edema, vascular congestion, and extra- Abu-Yousef, 2011).
hepatic cholestasis (Arena et al, 2008; Coco et al, 2007; Hepatic venous waveform is phasic due to cardiac pulsation,
Millonig et al, 2008; Millonig et al, 2010). Patients are exam- including alternating antegrade and retrograde flow variations
ined for sonoelastographic measurements in the supine posi- with predominantly antegrade flow (see Fig. 15.5C). When the
tion, with the right arm elevated above the head for optimal hepatic veins or inferior vena cava (IVC) are narrowed, the
intercostal access. There may limited assessment in obese waveform becomes flattened; this is a sensitive sign for venous
patients or those with narrow intercostal spaces. compromise.
A B
C
FIGURE 15.17. Hepatic metastases evaluated with contrast-enhanced ultrasound. A, Gray scale image reveals a hypoechoic mass (arrow) repre-
senting a metastasis. B, In the arterial portal phase after contrast injection, the liver metastases (arrows) have peripheral rim enhancement, and an
additional lesion is now evident. C, The metastases (arrows) show complete washout of contrast in less than a minute; image taken at 54 seconds.
(Courtesy Stephanie R. Wilson.)
Hepatic arterial waveform is a low-resistance waveform with hypoechoic, isoechoic, or hyperechoic. The portal vein diam-
a brisk systolic upstroke and continuous antegrade diastolic eter may be normal or expanded. Doppler imaging may show
flow, with expected RI of 0.55 to 0.7. absent color flow or areas in the vessel that do not fill with color
completely. Tumor thrombus may be distinguished from bland
Portal Vein Thrombosis and Cavernous Transformation thrombus by the presence of arterial waveforms within the
Portal vein thrombosis may be caused by hypercoagulable thrombus, a finding that is specific, but not very sensitive (Rossi
states, inflammatory diseases, sepsis, myeloproliferative disor- et al, 2006) (see Fig. 15.14). CEUS improves both detection
ders, neoplasms, and most commonly portal hypertension and characterization of thrombi in the portal system (Rossi
(Rossi et al, 2006) (Fig. 15.19). Gray scale findings include et al, 2006). Very slow flow in the PV system may result in a
solid material within the portal vein lumen that may be anechoic, false-positive sonographic diagnosis of portal vein thrombus.
gb
A B
pv
ivc
C D
FIGURE 15.18. Diffuse liver disease. A, Hepatic steatosis, longitudinal right lobe. Hepatic parenchyma of fatty liver is very echogenic relative to the
right kidney. B, Severe hepatic steatosis produces marked attenuation of sound, such that the posterior liver is poorly visualized. gb, Gallbladder.
C, Focal sparing in an otherwise fatty liver. Transverse image shows a hypoechoic region (arrow) in segment IV anterior to the porta. ivc, Inferior vena
cava; pv, portal vein. D, Cirrhosis. The liver is small with coarsened echogenicity and a nodular surface (arrows). a, Aorta.
CT or MRI may be used for confirmation of portal vein throm- segments I and IV and atrophy of the left lateral segment and/
bus, especially in the evaluation of patients prior to liver or periphery of the liver (Tublin et al, 2008; Vilgrain, 2001;
transplantation. Vilgrain et al, 2006).
In the setting of long-standing portal vein thrombosis, a
collateral hepatopetal venous circulation develops in the hepatic Budd-Chiari Syndrome and Venoocclusive Disease
hilum through recanalization of the thrombus and enlargement Budd-Chiari is a group of disorders characterized by hepatic
of collateral veins (de Groen et al, 1999). This phenomenon is venous outflow obstruction, with obstruction at the level of the
known as cavernous transformation of the portal vein, and it IVC, hepatic veins, or hepatic venules (see Chapter 87).
manifests on sonography as numerous serpiginous vessels in the Hepatic venoocclusive disease is rare and involves diffuse nar-
periportal region and nonvisualization of the main portal trunk rowing at the venule level. Etiologies of Budd-Chiari are varied,
(Fig. 15.20A). Overall flow in these collateral vessels is toward including hematologic abnormalities causing venous thrombus,
the liver (hepatopetal). In patients with cirrhosis, spectral venous obstruction by intrahepatic tumor, venous webs, or
Doppler should be used to confirm the venous character of the injury to hepatic venules from irradiation and chemotherapy
varicosities because enlarged and tortuous hepatic arteries may (Cura et al, 2009).
have a similar appearance on color Doppler. In chronic portal Venous obstruction and ascites are key features of Budd-
vein thrombosis and cavernous transformation, the morphology Chiari syndrome in the acute phase. The liver becomes enlarged
of the liver may become distorted, with enlargement of and heterogeneous in echogenicity. Imaging of the hepatic vein
A B
FIGURE 15.19. Left portal vein thrombus due to portal hypertension in this patient with cirrhosis. A, Nonocclusive echogenic material in the left
portal vein on gray scale ultrasound image. B, Correlative finding on contrast-enhanced computed tomography.
confluence may show narrowed hepatic veins, tumor encase- liver, such arterial collateral vessels are initially absent, requir-
ment, intraluminal thrombus, or absence of normal hepatic ing at least 2 weeks to develop. Therefore, in the early post-
venous structures. On color Doppler, hepatic vein flow altera- transplant period, bile ducts are supplied solely by the hepatic
tions and venous collaterals are readily evident. Spectral artery.
Doppler analysis may show flattened waveforms from more The frequency of early hepatic arterial thrombosis is higher
central obstruction or areas of high-velocity flow due to stenosis in children than adults and has been decreasing with improve-
(Grant et al, 1989; Ralls et al, 1992) (Fig. 15.20B-D). In the ments in perioperative care. A review of 71 clinical studies of
more chronic phase, intrahepatic and extrahepatic venous col- early hepatic arterial thrombosis published between 1990 and
laterals develop, and regenerating nodules are also present 2007 found an overall incidence of 4.4%, with a rate of 8.3%
(Brancatelli et al, 2002; Brancatelli et al, 2007). The caudate in children and 2.9% in adults (Bekker et al, 2009). Causes of
lobe is typically spared because of its unique hepatic venous hepatic arterial thrombosis include technical anatomic causes
drainage into the IVC; intrahepatic venous collaterals will form (kinking, stenotic anastomosis) and nonanatomic causes, such
to drain through the caudate lobe, resulting in caudate hyper- as immunologic factors, clotting abnormalities, and infections,
trophy (Bargallo et al, 2003). Approximately 25% of patients particularly cytomegalovirus (Pastacaldi et al, 2001). The
with Budd-Chiari syndrome also have portal vein thrombosis median time to detection of hepatic arterial thrombosis is
(Mahmoud et al, 1997). approximately 7 days (Bekker et al, 2009). Most cases of
hepatic artery thrombosis are identified through routine sono-
Liver Transplant Evaluation graphic surveillance, when patients are either asymptomatic or
Ultrasound is the modality of choice for evaluation of liver have elevation of liver function tests. Bile duct ischemia as a
transplants. Early detection of transplant complications is result of hepatic artery thrombosis results in recurrent cholan-
essential for preservation of graft function; therefore sono- gitis, intrahepatic abscess, bile duct stenosis, or bile leak. Ful-
graphic expertise must be readily available in any transplant minant hepatic necrosis is uncommon (Garcia-Criado et al,
setting. The frequency of sonographic surveillance varies by 2002).
institution, although initial assessment of vascular patency is On sonographic evaluation, the normal hepatic artery wave-
generally performed within 1 day following transplant (see form shows a brisk systolic upstroke and continuous diastolic
Chapters 113 and 120). flow with a resistive index between 0.5 and 0.8. The normal
The overall incidence of vascular complications following systolic acceleration time, measured from end diastole to the
liver transplantation is approximately 9% (Caiado et al, 2007; first systolic peak, is usually less than 0.08 s. Hepatic artery
Quiroga et al, 2001), and Doppler ultrasound is the mainstay thrombosis is diagnosed on ultrasound when flow is absent in
of evaluation. Hepatic arterial thrombosis is a serious complica- the proper hepatic artery and intrahepatic arteries on color and
tion because it impairs the viability of the bile ducts within the spectral Doppler.
transplanted liver. In the native liver, arterial collaterals, primar- Stenosis of the hepatic artery occurs in an estimated 5% to
ily from the phrenic arteries, can sustain bile ducts following 11% of transplant recipients and usually develops at the anas-
hepatic artery ligation (Bekker et al, 2009). In the transplanted tomotic site within 3 months of transplantation (Caiado et al,
gb
A B
TRANS
ivc
PW:3MHz θ = 55°
.60
m/s
C D .60
FIGURE 15.20. Vascular disorders of the liver. A, Cavernous transformation of the portal vein. The portal supply has a serpiginous appearance
because of numerous collateral vessels. gb, Gallbladder. B, Outflow obstruction in a patient with venoocclusive disease. On transverse view, the
intrahepatic inferior vena cava (arrow) is markedly narrowed, and portions of the right hepatic vein (arrowhead) are obliterated. C, The middle hepatic
vein (arrows) is narrowed. ivc, Inferior vena cava. D, The middle hepatic vein shows flattened Doppler flow consistent with partial venous
obstruction.
2007). If the anastomotic site can be directly visualized, color than 0.5 and a prolonged acceleration time in excess of 0.08 s
Doppler may show focal aliasing and spectral Doppler will (Caiado et al, 2007; Tamsel et al, 2007). Occasionally, a tardus-
demonstrate elevated velocities. More commonly, the hepatic parvus waveform may be present in the intrahepatic arteries in
artery can only be visualized in the porta hepatis downstream the setting of chronic hepatic artery thrombosis. This implies
of the anastomosis. With anastomotic stenosis, there will be a that a collateral arterial supply has developed, and its damp-
tardus-parvus waveform in the downstream proper hepatic and ened flow resembles the waveform seen in hepatic artery ste-
intrahepatic arteries (Fig. 15.21). This waveform is character- nosis (Caiado et al, 2007). Therefore both the main hepatic
ized by a blunted, delayed peak systolic velocity with RI less artery in the porta hepatis and the intrahepatic right and left
.20
LIVER
.20 artifact
V1 = 0.233m/s
V2 = 0.138m/s
Rl = 0.41 V = –0.27m/s
S/D = 1.69 ±Mean = –0.13m/s
PW:2MHz 1Hz θ = 54° RHV

θ = 9° LHA
.30
m/s
A.30 B
.51
IVC HV ANAST
.51
PW:2.5MHz
.50
m/s
1.0
C
FIGURE 15.21. Vascular complications after liver transplantation. A, Left hepatic artery (LHA) tardus-parvus waveform indicates hepatic artery
stenosis or thrombosis. The waveform has blunted arterial upstroke and high diastolic flow with resistive index less than 0.5. B, Duplex spectral
Doppler image of the right hepatic vein (RHV) in a patient after orthotopic liver transplant with piggyback anastomosis (ANAST) shows monophasic
flattened flow, indicating anastomotic stricture. C, Duplex Doppler with the sample placed over the area of flow aliasing shows high-velocity turbulent
flow (arrows) at the stenotic venous anastomosis, localizing the area of focal narrowing. IVC, Inferior vena cava.
hepatic arteries must be interrogated on all Doppler studies of The common bile duct is best seen from the right intercos-
transplanted livers. In the early postoperative period, the tal approach with the patient in the left lateral decubitus posi-
hepatic artery RI may be elevated, a finding sometimes associ- tion, which allows use of the liver or gallbladder as an
ated with an older donor age or prolonged cold ischemia time acoustic window. However, the left lateral decubitus position
(Caiado et al, 2007; Shaw et al, 2003). brings bowel gas to the right side and may obscure the distal
On the venous side, portal and hepatic vein transplant duct. Alternatively, upright or right lateral decubitus positions
complications are relatively rare (see Fig. 15.21). Portal vein fill the stomach with fluid, improving visualization of the
stenosis and thrombosis each have an incidence of approxi- distal common bile duct. The right and left hepatic ducts lie
mately 1% to 2% (Caiado et al, 2007; Tamsel et al, 2007). anterior to the portal vein confluence and are best visualized
Portal vein thrombosis may result from technical problems at on transverse subcostal ultrasound images. Second-order
surgery, prior portal vein thrombosis, increased downstream branches may also be seen adjacent to the segmental portal
resistance (e.g., hepatic venous outflow impairment), sluggish vein branches.
portal vein inflow, or hypercoagulable states. Diagnosis of Several techniques help enhance visualization of the biliary
portal vein thrombosis is made when flow is absent in the ducts. Tissue harmonic imaging improves contrast and reduces
various Doppler modes or an echogenic thrombus is visual- reverberation and side lobe artifacts when evaluating biliary
ized on gray scale ultrasound. Suspected thrombosis may ducts (Ortega et al, 2001). Use of compression and optimal
be confirmed with another imaging modality, such as CT technical parameters of frequency, gain, filter, and scale will
or MRI. produce the most favorable ultrasound images.
Transjugular Intrahepatic Portosystemic Shunt Evaluation Gallbladder

Transjugular intrahepatic portosystemic shunt (TIPS) (see Gallstones and Biliary Sludge
Chapter 87) placement is a therapeutic option that can be used Gallstones are mobile, echogenic, and demonstrate posterior
to decompress the PV system in patients with complications of acoustic shadowing (Fig. 15.23; see Fig. 15.3). It is important
portal hypertension. The shunt conventionally connects the to scan patients in different positions to differentiate gallstones
right portal vein to the right hepatic vein, diverting blood flow from polyps, because gallstones are mobile, whereas polyps are
from the portal to the systemic circulation. Flow in the portal fixed. Size of the stone is a principle determinant of shadowing,
veins should be directed toward the TIPS. Color Doppler and it is frequently difficult to obtain shadowing from stones
should therefore show reversed flow direction in the branch smaller than 3 mm. Technical factors that will optimize dem-
portal veins and antegrade flow in the main portal vein. onstration of acoustic shadowing from small gallstones include
Doppler evaluation is obtained to assess shunt patency and increasing transducer frequency, setting the focal zone at depth
TIPS stenosis. TIPS evaluation with ultrasound is usually per- of the gallstones, positioning the patient such that multiple
formed the day after the TIPS procedure and then at regular stones will cluster together, and demonstrating twinkle artifact
intervals for routine surveillance. Symptoms of increased portal on color Doppler. Stones in the gallbladder neck may be dif-
hypertension may indicate TIPS stenosis or thrombosis, usually ficult to visualize but should roll into the gallbladder body or
prompting sonographic evaluation. The flow velocity within the fundus when the patient is examined in the left lateral decubitus
TIPS ranges from 90 to 200 cm/s (Abbitt, 2002; Tchelepi et al, position. Fixed stones in the gallbladder neck are frequently
2002). Absent flow on color and spectral Doppler is 100% associated with cholecystitis. If large stones or multiple stones
sensitive and 96% specific for shunt thrombosis (Abbitt, 2002). fill the entire gallbladder lumen, there may be little surrounding
Findings of TIPS stenosis include angle-corrected shunt veloc- bile. When this occurs, gallstones are diagnosed by the “wall
ity less than 50 cm/s, greater than 250 cm/s, or reversal of flow echo shadow” sign produced by echoes from the anterior gall-
in the proximal portion of the hepatic vein (Abbitt, 2002; Dodd bladder wall, echogenic anterior surface of the stone, and pos-
et al, 1995; Fidelman et al, 2012; Zizka et al, 2000). Other terior acoustic shadowing produced by the stone. A thin
authors have found that an increase or decrease in peak stent crescent of hypoechoic bile separates the gallbladder wall from
velocity of greater than 50 cm/s compared with baseline is 93% the echogenic stones (see Fig. 15.23D). In contradistinction, a
sensitive and 77% specific for shunt stenosis (Dodd et al, porcelain gallbladder has calcification in the gallbladder wall
1995). itself, so there is no hypoechoic crescent of bile separating the
wall from the echo.
Gallbladder sludge is viscous echogenic bile that is nonshad-
ULTRASOUND OF THE GALLBLADDER AND owing and sometimes takes on a rounded shape called tumefac-
BILIARY TREE tive sludge. Galllstones develop in approximately 15% of patients
with sludge. Sludge can obscure the interfaces of small stones.
Anatomy and Technique Sludge will change with positional variation and moves slowly,
The gallbladder lies on the undersurface of the liver at the whereas gallstones roll quickly on real-time ultrasound.
junction of the right and left hepatic lobes, which is defined
superiorly by the middle hepatic vein and inferiorly by the Cholecystitis
interlobar fissure. Sonographically, it is best visualized from the Acute cholecystitis (see Chapter 33) is due to obstruction of
right intercostal approach with the patient in the supine or left the cystic duct or gallbladder neck, usually by a gallstone. The
lateral decubitus position (Fig. 15.22). The spiral valves of gallbladder becomes distended with a chemical inflammation
Heister are often visualized in the gallbladder neck. The gall- of the mucosa, producing a tense and tender gallbladder with
bladder may fold upon itself or have a fundal phrygian cap. The wall edema and inflammation. For the diagnosis of cholecysti-
gallbladder is best evaluated when distended, and therefore tis, ultrasound has a sensitivity of 80% to 100% sensitivity,
patients should fast before the scan. specificity of 60% to 100%, and positive predictive value of
A B
C
FIGURE 15.22. Gallbladder. A, Normal longitudinal gallbladder image with visualization of the gallbladder neck. B, Thickened gallbladder wall
(arrows) resulting from inflammation. Sludge (arrowheads) layers posteriorly. C, Gallbladder polyps (arrows). Note the lack of acoustic shadowing.
90% to 94% for the diagnosis (Harvey & Miller, 1999; Ralls greater than 15,000 cells/mL (Fagan et al, 2003). Ultrasound
et al, 1985). Ultrasound findings of acute calculous cholecysti- features of gangrenous cholecystitis include floating intralumi-
tis include gallstones, gallbladder wall thickening greater than nal membranes from sloughed mucosa, shadowing foci from
3 mm, pericholecystic fluid, and a positive sonographic Murphy air in the gallbladder wall, disrupted gallbladder wall, and peri-
sign. Gallstones, wall thickening, and Murphy sign together cholecystic abscess formation (Jeffrey et al, 1983).
have a positive predictive value of 92% to 95% (Ralls et al, Gallbladder wall thickening is nonspecific and may be seen
1985; Smith et al, 2009; Teefey et al, 1991) (Fig. 15.24). in a wide range of gallbladder disease and extracholecystic
In emphysematous cholecystitis, echogenic air bubbles in pathologic conditions. Diffuse gallbladder wall thickening in
the gallbladder wall produce reverberation artifact. In this patients without primary gallbladder wall disease occurs in a
entity, there is possibility of gallbladder necrosis, gangrene, and variety of systemic processes, such as hypoalbuminemia, con-
perforation. Gangrenous cholecystitis occurs more often in gestive heart failure, hepatitis, and pancreatitis (van Breda
patients with diabetes mellitus or a white blood cell count Vriesman et al, 2007).
liver
3 4
2
1
RK
A B
W
E
m
C D
FIGURE 15.23. Gallstones and gallbladder carcinoma. A, Longitudinal view of the gallbladder shows layering stones (numbered) with acoustic
shadow. A small amount of sludge (arrow) is also seen in the dependent portion of the gallbladder. B, Views of the fundus show focal mural thicken-
ing (arrows) suspicious for tumor. RK, Right kidney. C, Flow is seen within the mass (m) on color Doppler imaging. D, Gallstones fill the gallbladder
lumen, producing a Wall-Echo-Shadow (WES) sign from the anterior gallbladder wall, the echogenic anterior surface of gallstones, and posterior
acoustic shadowing by the gallstones.
Hyperplastic Cholecystoses and Gallbladder Polyps thickening is most common in the fundus; when it occurs in
Hyperplastic cholecystoses (see Chapter 49), such as choles- the gallbladder body, there is annular constriction producing
terolosis and adenomyomatosis, can cause focal or polypoid an hourglass-shaped gallbladder. Masslike areas of adenomyo-
gallbladder wall thickening. Cholesterolosis results from abnor- matosis are called adenomyomas.
mal cholesterol deposits in the gallbladder wall creating wall The majority of incidentally detected polypoid gallbladder
irregularities or polypoid intraluminal masses. Cholesterolosis lesions are nonneoplastic and represent cholesterol polyps or
usually presents as multiple small (1 to 10 mm) nonshadowing inflammatory polyps (Corwin et al, 2011). Rarely, these may
polyps arising from the nondependent wall with echogenic be neoplastic, such as adenomas, and malignant transformation
speckles and lobular contour (see Fig. 15.22C). Cholesterol to adenocarcinoma is a concern. Adenomas tend to be solitary
polyps are benign with no malignant potential (Ito et al, 2009; and uniformly hyperechoic yet become more heterogeneous as
Levy et al, 2002; Terzi et al, 2000). Adenomyomatosis is a they increase in size and may either be pedunculated or sessile.
benign hyperplastic cholecystosis with no malignant potential Thickening of the gallbladder wall adjacent to an adenoma may
that results from hyperplasia of both the mucosa and muscula- suggest malignancy.
ris propria of the gallbladder wall. Intramural diverticula are In patients with polypoid gallbladder lesions, risk factors for
called Rokitansky-Aschoff sinuses, which trap bile that accumu- malignancy include age of patient (> 60 years of age), coexis-
lates cholesterol crystals appearing as cystic spaces in a thick- tence of gallstones, and size of the polypoid lesion (>10 mm in
ened gallbladder wall with a characteristic comet tail artifact diameter) (Terzi et al, 2000). The prevalence of malignancy in
(Levy et al, 2002) (see Fig. 15.4). The gallbladder wall thicken- polyps greater than 10 mm ranges from 37% to 88% (Ishikawa
ing in adenomyomatosis may be focal or diffuse. Focal wall et al, 1989; Koga et al, 1988). Surgical consultation for
A B
FIGURE 15.24. Acute cholecystitis. A, Longitudinal sonogram shows a distended gallbladder with thickened irregular wall (arrows) and layering
sludge (arrowheads). B, Computed tomographic scan shows gallbladder wall thickening (arrows).
A B
FIGURE 15.25. Gallbladder carcinoma. A, Gray scale longitudinal image of the gallbladder shows a solid irregular mass in the fundus. B, Transverse
color Doppler image of this lesion demonstrates internal vascularity, suspicious for tumor. Pathology demonstrated gallbladder adenocarcinoma.
asymptomatic polyps greater than 10 mm and for symptomatic into the adjacent liver (Wibbenmeyer et al, 1995) (Fig. 15.25).
gallbladder polyps irrespective of size has been proposed (Wiles When the mass occupies the gallbladder lumen, it can result in
et al, 2014). Often, asymptomatic gallbladder polyps smaller a displaced or “trapped” stone that is fixed in position due to
than 10 mm are followed sonographically, and those smaller intraluminal tumor. Secondary signs of gallbladder cancer
than 6 mm may not be followed or followed at extended inter- include discontinuity of the echogenic mucosal lining, absence
vals (Pedersen et al, 2012). of echogenic specks seen in cholesterol crystals, and high-
velocity arterial flow greater than 60 cm/s. Selective mucosal
Gallbladder Carcinoma calcification of the gallbladder wall is significantly associated
Carcinoma of the gallbladder (see Chapter 49) may appear as with gallbladder cancer, whereas diffuse intramural calcifica-
a sessile or polypoid mass, a thickened gallbladder wall, or tion is not (Stephen & Berger, 2001). Gallbladder carcinoma
infiltrative mass that fills the gallbladder lumen and extends often contiguously extends into hepatic segments IVB and V,
or into the hepatic hilum, possibly directly involving the main be no change or only a slight increase in diameter after chole-
bile duct to cause secondary biliary obstruction. Adjacent ade- cystectomy (Abbitt, 2002; Feng & Song, 1995; Mueller et al,
nopathy may also be present. Ultrasound performs well for 1981 ). The effect of age is controversial, yet even in the elderly
detection of the gallbladder primary and local tumor spread, the common duct diameter usually does not exceed 7 mm
but CT is necessary to more accurately determine resectability, (Bachar et al, 2003; Horrow et al, 2001; Perret et al, 2000).
distant disease spread, and peritoneal metastases (Bach et al, Variable threshold values in literature are due to differences in
1998). sample size, patient population, and control of confounding
In the era of laparoscopic surgery for gallstone disease, it is variables.
important to assess the gallbladder carefully on preoperative Intrahepatic biliary dilatation is less well defined. Intrahe-
ultrasound to exclude occult gallbladder cancer and to plan an patic ducts should measure greater than 2 mm or greater than
appropriate surgical approach. Approximately 47% of gallblad- 40% of the adjacent portal vein to be considered dilated. The
der carcinomas are detected incidentally at laparoscopic chole- “double-track” sign is caused by dilated bile ducts parallel to
cystectomy (Duffy et al, 2008). For these patients, reexploration the portal vein branches (Fig. 15.26). Intrahepatic bile duct
with definitive resection and reexcision of laparoscopic port dilatation is best appreciated in transverse images of the left
sites is recommended (D’Angelica et al, 2009; Duffy et al, hepatic lobe. The pattern of bile duct dilatation should be
2008; Winston et al, 1999). assessed to determine whether it is symmetric in both lobes or
localized to one portion of the liver. To determine the level of
Biliary Ducts obstruction, the left and right main ducts are followed trans-
Choledochal Cysts versely to their junction, and transverse views of the common
Choledochal cysts (see Chapter 46) are classified according to hepatic and common bile ducts are obtained. With this approach
the location of the dilated ductal segment: type I is fusiform it is possible to differentiate intrahepatic from extrahepatic
extrahepatic duct dilatation, type II is diverticulum of the extra- obstruction, and usually the etiology of the obstruction can be
hepatic duct, type III is a choledochocele from a dilated termi- determined (Fig. 15.27).
nal portion of the common bile duct, type IV is multifocal Bile duct dilatation is most often caused by an obstructive
dilatation, and type V is cystic dilatation of the intrahepatic bile process. Bile duct calculi appear as intraluminal filling defects
ducts that is synonymous with Caroli disease (Todani et al, (see Fig. 15.26B), which may be intrahepatic or extrahepatic.
1977). For Caroli disease, the cysts may be large, and the con- Calculi may form or reflux into the intrahepatic ducts, and
nection with the bile duct is not always evident sonographically. small calculi can be mistaken for air (see Chapters 36 and 37).
On imaging, it is important to demonstrate the connection with Because there is little bile surrounding the ductal calculi, and
the bile ducts to differentiate this condition from polycystic because the stones may be small, acoustic shadowing may not
disease. Arterial signals from the fibrovascular bundle at the always be elicited. Transabdominal ultrasound remains opera-
margin of the saccules may also aid in diagnosis (Miller et al, tor dependent, but, when performed by experienced examiners,
1995). has high diagnostic accuracy for choledocholithiasis (Rickes
et al, 2006). Debris or thick bile within the ducts may cause
Biliary Obstruction internal echoes within ducts and fluid levels, but debris does
Measurement of the extrahepatic duct is usually performed not shadow and will shift with positional variation. Hemobilia
near the crossing of the hepatic artery, with measurement of may layer or appear echogenic and masslike if the clot is orga-
the lumen from inner wall to inner wall. The common bile duct nized (see Chapter 125). Intraductal tumor from colorectal
lumen normally measures 6 mm in diameter or less. There may metastases or HCC may also appear as an intraductal mass.
gb
cbd
A B
FIGURE 15.26. Biliary obstruction from choledocholithiasis. A, Transverse sonogram of the liver reveals the “double-track” sign (circled areas),
consistent with intrahepatic biliary dilatation. B, Longitudinal view of the common bile duct (cbd) shows an echogenic stone (arrow) that produces
acoustic shadowing (arrowheads). gb, Gallbladder; v, portal vein.
m
m
v
v
v
ivc
A B
C
FIGURE 15.27. Biliary obstruction and vascular encasement from adenopathy at the porta hepatis. A, Mass (m) obstructs a mildly
dilated common bile duct (arrows) and involves the main portal vein (v). B, Nodal masses (m) encase the portal vein (v) that is markedly narrowed
(arrows). ivc, Inferior vena cava. C, Color Doppler image shows narrowed hepatic artery (a; arrow) with dilatation proximal to the encased segment.
v, Portal vein.
Tumor often expands the duct and does not produce acoustic In patients with biliary obstruction who are being considered
shadowing (Ghittoni et al, 2010; Jhaveri et al, 2009). Obstruc- for surgical resection or palliative biliary drainage, the distribu-
tion as a result of biliary ascariasis is associated with tubular tion of biliary ductal dilatation should be carefully evaluated to
structures within the bile duct, and movement of the worms is determine management. Any isolated biliary ductal segments
pathognomonic (Lim, 1990). Pancreatic carcinoma, gallblad- that do not communicate with the main ducts should be noted,
der carcinoma, and cholangiocarcinoma can all cause biliary because isolated segments may alter surgical approach, and
dilatation. biliary drainage may require placement of multiple catheters.
The pattern of biliary obstruction and the appearance of the (Jarnagin et al, 2001).
duct wall are also useful for diagnosis. Recurrent pyogenic
cholangitis is evidenced by dilated ducts with intraductal calculi Cholangiocarcinoma
and segmental dilatation, frequently in the left lobe (see Chapter Cholangiocarcinoma is a ductal tumor of the intrahepatic or
39). Lobar atrophy may also be present. Primary sclerosing extrahepatic bile ducts (see Chapter 47). Tumors of the periph-
cholangitis causes a beaded appearance of the ducts with wall eral intrahepatic ducts tend to present at a larger size, whereas
thickening, strictures, and discontinuous areas of dilatation; tumors of the extrahepatic duct tend to be smaller and often
intraluminal echoes in the dilated ducts represents debris such manifest as obstructive jaundice. Cholangiocarcinomas may be
as pus, sludge, or sloughed epithelium (see Chapter 41). Mural hyperechoic or hypoechoic.
thickening of the ducts also is seen with AIDS-associated chol- The intrahepatic peripheral form of cholangiocarcinoma
angitis. For tumors adjacent to the umbilical portion of the left presents as a focal hepatic mass, which may be solitary or with
portal vein, subtle bile duct dilatation or wall thickening may satellite lesions. Intrahepatic cholangiocarcinoma has associ-
be a clue to biliary ductal involvement. ated peripheral biliary ductal dilatation in 25% of cases, may
have capsular retraction, lack a hypoechoic halo, and lack Because of its midline and posterior location, the pancreas
venous thrombosis; these features help differentiate peripheral is often partially obscured by overlying gas in the stomach when
cholangiocarcinoma from HCC (Chung et al, 2009; Soyer the patient is in the supine position. Optimally, patients should
et al, 1995) (see Chapter 50). be fasting for several hours before the scan to reduce bowel gas.
Hilar cholangiocarcinomas occur at the biliary confluence Applying pressure with the transducer will displace bowel gas.
and classically produce segmental upstream dilatation with It is best to begin abdominal ultrasound with pancreatic images,
abrupt cutoff and nonunion of the right and left dilated ducts because the deep inspiratory scans used to image the liver lead
at the porta hepatis. The majority of hilar cholangiocarcino- to aerophagia and diminish the likelihood of pancreatic visual-
mas are isoechoic, and some larger tumors may have ization. Positional maneuvers, such as sitting the patient
hypoechoic rim, especially if liver is involved (Bloom et al, upright, may gather fluid into the gastric antrum and displace
1999). The extent of tumor involving bile ducts is demon- air into the fundus, affording better visualization. Patients may
strated sonographically by the location of tumor and the dis- be asked to drink 500 mL of water and examined 10 to 15
tribution of bile duct obstruction (Hann et al, 1997). Hilar minutes afterward, to use the fluid-filled stomach as an acoustic
tumors are often associated with hepatic lobe atrophy (see window.
Chapter 51). The following should be assessed in pancreatic ultrasound
On ultrasound, extrahepatic cholangiocarcinoma may appear examination: parenchymal abnormalities, the distal common
as infiltrative spread along the duct walls, nodular mural thicken- bile duct in the region of the pancreatic head, the pancreatic
ing, or papillary (Hann et al, 1997). The papillary form, which duct for dilatation or other abnormalities, and the peripancre-
is a polypoid expansile intraductal mass, has a better prognosis atic region for adenopathy or fluid. The pancreas is rarely seen
and surgical outcome (Jarnagin et al, 2005). With experienced sonographically in its entirety, and therefore when primary
operators, ultrasound can correctly identify 96% of extrahepatic pancreatic pathology is suspected, CT or MRI are preferred
cholangiocarcinomas and can demonstrate tumor infiltration modalities.
along the duct wall (Lim, 2003; Robledo et al, 1996).
Tumor involvement of the biliary ducts is best evaluated by Pancreatitis
ultrasound before placement of biliary drainage catheters, Acute pancreatitis is a common cause of abdominal pain and
because pneumobilia or stent artifact may obscure tumor is usually diagnosed with clinical and laboratory findings (see
margins, and the pattern of biliary obstruction may not be Chapter 55). Sonographic findings of pancreatitis are usually
evident after decompression. The portal vein is frequently subtle and include heterogeneous parenchymal echogenicity or
directly involved and may be encased or occluded. Both portal focal alterations in echogenicity; focal masslike abnormalities
vein status and ductal spread help determine surgical approach with contour convexity; indistinct anterior border; and peripan-
(de Groen et al, 1999; Jarnagin et al, 2001) (Fig. 15.28). creatic fluid (Finstad et al, 2005). Acute pancreatitis may be
Ultrasound contrast agents aid diagnosis and staging of focal or diffuse, and differentiation from a focal mass may not
cholangiocarcinoma (Lim et al, 2006; Xu et al, 2010). Khalili be possible. Additional findings evaluated in the setting of pan-
and colleagues (2003) reported that correct prediction of creatitis are biliary findings such as stones or ductal dilatation,
resectability improved from 69% on noncontrast ultrasound to ascites, and pleural effusions. The main purpose of sonography
96% after contrast. Staging workup generally involves other in evaluation of suspected pancreatitis is to detect biliary and
imaging modalities. gallbladder abnormalities. Although CT is the preferred tech-
nique for gauging the severity and extent of acute pancreatitis,
ULTRASOUND OF THE PANCREAS sonography is useful to determine whether the disease was trig-
gered by gallstones, because the sensitivity of CT for detection
Anatomy and Technique of gallstones has been reported to be as low as 25% (Chan et al,
The pancreas is oriented transversely in the epigastric anterior 2006).
pararenal space and lies deep to the stomach and left lobe of Complications of acute pancreatitis may be seen with sonog-
liver (see Chapter 2). The splenic vein marks the dorsal/ raphy, such as peripancreatic fluid collections and vascular
posterior border of the pancreatic body and tail (Fig. 15.29). complications (e.g., splenic or portal vein thrombosis, splenic
The pancreatic neck is marked by the confluence of splenic and artery pseudoaneurysm). Pancreatic parenchymal and extra-
superior mesenteric veins. Pancreatic head and uncinate process pancreatic fluid collections develop and evolve during time, and
may wrap around the portosplenic confluence such that pan- may be sterile, infected, or hemorrhagic. The classic pseudocyst
creatic tissue lies both anterior and posterior to the vein (Sirli is a circumscribed, smooth-walled spheric anechoic mass with
& Sporea, 2010). Celiac trunk marks the cranial border of the posterior acoustic enhancement (Fig. 15.30), sometimes mul-
gland. The pancreatic tail has a slight cranial trajectory and may tilocular, and during time may develop mural calcification.
extend to the splenic hilum. Color Doppler should always be used to evaluate any epigastric
Normal pancreatic echogenicity is fine, homogeneous, and cystic mass, including those seen in the setting of pancreatitis,
isoechoic or hyperechoic to liver. The pancreatic duct may be to distinguish fluid collections from other complications of
visible, particularly in thin patients, as an anechoic tubular pancreatitis such as pseudoaneurysms. Color Doppler will also
structure oriented along the long axis of the pancreas, usually enable detection of splenic vein thrombosis, which occurs in
measuring no larger than 2 mm in diameter, which tapers or 1% to 3% of patients following acute pancreatitis (Balthazar,
becomes invisible in the tail. On transverse images, the common 2002). The presence of pancreatic necrosis, abscess, or hemor-
bile duct can usually be visualized in cross-section in the pos- rhage cannot be determined reliably with ultrasound; CT or
terior aspect of the head of the pancreas (see Fig. 15.29). The MRI provides more accurate evaluation. CEUS may be used
gastroduodenal artery can often be seen in cross-section along to evaluate acute pancreatitis to identify complications such as
the right anterior aspect of the pancreatic head. pancreatic necrosis, which appears as heterogeneous areas of
pv
ivc
A B
rpv
C
FIGURE 15.28. Klatskin tumor. A, Longitudinal view of the common hepatic duct shows a tapered segment (arrow) consistent with tumor stricture.
ivc, Inferior vena cava; pv, portal vein. B, Transverse image at the biliary confluence reveals an echogenic mass (arrow) at the bifurcation and dilated
bile ducts in both lobes. C, The left portal vein is narrowed and encased (arrows), as shown on color Doppler transverse image. rpv, Right portal
vein.
nonenhancement (Ardelean et al, 2014). A benefit of contrast- may occur in chronic pancreatitis and can be difficult to dif-
enhanced ultrasound is that it can be performed at the patient’s ferentiate from adenocarcinoma. The finding of calcifications
bedside. within a hypoechoic focal mass is helpful for differentiating the
Chronic inflammation of the pancreas eventually causes two entities and suggests chronic pancreatitis, rather than car-
scarring and destruction of pancreatic parenchyma, producing cinoma, as the etiology (see Chapter 57).
a small atrophic gland with poorly defined margins and patchy Autoimmune pancreatitis classically appears as focal or
mixed echogenicity. The main pancreatic duct may become diffuse gland enlargement with indistinct anterior margin and
dilated and irregular (Bolondi et al, 1989). Calcifications may nondilated pancreatic duct (Araki et al, 2006; Sahani et al,
be intraductal or intraparenchymal and may show twinkle arti- 2004) (Fig. 15.31). This type of pancreatitis is typically res
fact on color Doppler. Focal hypoechoic inflammatory masses ponsive to steroids, and recent studies of contrast-enhanced
v v
a
A
ivc
A
ivc
A B
FIGURE 15.29. Normal pancreas. A, Transverse image of the normal pancreas and adjacent vascular anatomy. Pancreatic body (arrows) is shown.
a, Superior mesenteric artery; A, aorta; ivc, inferior vena cava; v, splenic vein. B, Common bile duct (cursors; arrow) in the head of the pancreas.
A, Aorta; ivc, inferior vena cava; v, splenic vein.
ivc
*
a
A B
FIGURE 15.30. Pancreatic pseudocyst. A, Transverse sonogram of the pancreas reveals a cyst (asterisk) in the tail and extension into the body
(arrows). a, Aorta; ivc, inferior vena cava. B, Computed tomography scan confirms extensive pseudocyst (arrows).
sonography suggest that the degree of vascularity evident fol-

lowing contrast injection correlates directly with the pathologic
grade of inflammation and inversely with the grade of fibrosis
(Numata et al, 2004).
Pancreatic Neoplasms
Solid Neoplasms of the Pancreas
Although best staged by CT, adenocarcinoma of the pancreatic
head may be detected on sonography, because ultrasound can
be the initial imaging study for evaluation of jaundice, right
v
upper quadrant pain, or epigastric pain. Pancreatic adenocar-
cinoma classically appears as an ill-defined hypoechoic mass
associated with abrupt dilation of the upstream pancreatic duct
and atrophy of the more distal pancreas (Fig. 15.32) (see ivc
Chapters 61 and 62). The “double-duct sign,” which is dilation a
of the pancreatic duct and common bile duct, is a suspicious
imaging finding for pancreatic head adenocarcinoma. Tumors
in the body or tail of the pancreas are inconsistently visualized
on sonography because of variables such as patient body
habitus and overlying bowel gas. Color Doppler sonography FIGURE 15.31. Autoimmune pancreatitis. Transverse image of the
typically does not show increased vascularity within pancreatic pancreas shows diffuse pancreatic enlargement, well-defined margins,
and lobulated appearance (arrows). a, Aorta; ivc, inferior vena cava; v,
splenic vein.
A B
C D
FIGURE 15.32. Examples of pancreatic adenocarcinoma. Hypoechoic solid pancreatic head mass causing upstream main pancreatic ductal dilatation
(A), common bile duct dilatation (B), and intrahepatic biliary dilatation (C). D, A different patient with a hypoechoic mass in the pancreatic tail.
adenocarcinoma. Nevertheless, Doppler evaluation can be tumors. In patients with multiple endocrine neoplasia type I,
useful to visualize the peripancreatic vasculature, such as the real-time intraoperative scanning can also help identify addi-
celiac axis and superior mesenteric vessels, which are often tional subcentimeter neuroendocrine tumors that may escape
encased in locally advanced disease. Pancreatic applications of CT detection (Shin et al, 2009).
ultrasound elastography are undergoing investigation, with the There are several mimics of pancreatic masses on ultraso-
endoscopic ultrasound approach more widely studied than the nography. Either focal fatty sparing or fatty replacement of the
transabdominal approach. Initial studies propose use of elas- pancreas can appear masslike, as well as focal pancreatitis, such
tography to distinguish malignant lesions, which are densely as groove pancreatitis. Additional pitfalls include congenital
fibrotic and firm, from benign lesions such as focal inflamma- contour abnormalities, complicated pseudocysts, and lipomas
tion, which tend to be less fibrous. Another proposed use for that may mimic pancreatic masses.
endoscopic ultrasound elastography is to aid in selection of a
target for fine needle aspiration. Cystic Neoplasms of the Pancreas
In contrast to adenocarcinomas, neuroendocrine tumors of Cystic pancreatic neoplasms (see Chapter 60) fall into two
the pancreas are commonly hypervascular masses, isoechoic major categories: serous and mucinous cystic neoplasms. Dif-
or hypoechoic, well defined, and smoothly marginated (see ferentiation may be made on imaging studies based on the
Chapter 65). Functional tumors typically present early and at internal architecture and size of the cysts.
a smaller size than nonfunctioning tumors. Larger tumors may Serous cystadenomas are generally benign neoplasms with
demonstrate internal necrosis and calcifications. Intraoperative well-circumscribed lobulated borders. Internal appearance
ultrasound may be used to localize pancreatic neuroendocrine depends on size of individual cystic components that range
mass
A B
FIGURE 15.33. Serous cystadenoma of the pancreas. Transverse intraoperative ultrasound images of the pancreas. A, Mass in the pancreatic
tail has a small cyst (arrow). B, On magnified higher frequency images, the mass has honeycomb architecture.
A B
C D
FIGURE 15.34. Main duct intraductal papillary mucinous neoplasm (IPMN). Gray scale (A) and color Doppler (B) images demonstrate main pan-
creatic duct dilatation and tortuosity. Correlative coronal magnetic resonance images (C and D) showing dilated pancreatic duct throughout its course
without obstructing mass.
from 1 to 20 mm in size. Numerous tiny cystic spaces may not the potential to develop carcinomas, and any solid compo-
be resolved individually and create multiple interfaces, such nent or mural nodularity is concerning for malignancy (see
that the cystic component is often echogenic or solid appearing; Chapter 60).
internal architecture is often described as having a honeycomb Endoscopic ultrasound is useful in evaluation of cystic pan-
or spongy appearance (Fig. 15.33). Serous tumors composed creatic neoplasms due to better demonstration of internal cyst
of relatively larger cysts will appear as partially solid masses with architecture and potential for fluid aspiration (see Chapter 16).
peripheral cystic areas on sonography. A central stellate scar is
a characteristic feature of many serous cystadenomas and may CONCLUSION
be visible as an echogenic or calcified focus with shadowing
(Hutchins & Draganov, 2009; Yeh et al, 2001). Ultrasound is a useful and safe imaging modality to evaluate
Mucinous cystic neoplasms usually appear as unilocular or and follow up abnormalities of the liver, gallbladder, biliary
multilocular cystic lesions, with thick septations and thick tree, and pancreas. In particular, the gallbladder and biliary tree
walls, and are usually located in the pancreatic body and tail. have inherently excellent contrast resolution due to their cystic
Mucinous cystic neoplasms have malignant potential, and a nature relative to adjacent solid organs. Doppler ultrasound is
solid mural nodule is suspicious for malignancy. Intraductal an important and valuable component of the hepatic vascular
papillary mucinous neoplasms (IPMNs) are cystic neoplasms evaluation. Abdominal sonography is often used as an initial
that arise from the epithelium of the pancreatic ductal examination to determine the site of abnormality and can
system, characterized by intraductal papillary growth and often provide a diagnosis or a short differential diagnosis.
abundant mucin production, leading to ductal dilatation. The addition of cross-sectional CT or MR imaging to abdomi-
Side-branch–type IPMNs commonly appear as small unilocu- nal ultrasonography aids in diagnosis when sonography is
lar cysts or multilocular cystic masses within the pancreatic inconclusive.
parenchyma. Main-duct–type IPMNs may solely appear as
main pancreatic duct dilatation (Fig. 15.34). IPMNs have References are available at expertconsult.com.
Chapter 15 Ultrasound of the liver, biliary tract, and pancreas 275.e1
REFERENCES Caturelli E, et al: Coarse nodular US pattern in hepatic cirrhosis: risk

for hepatocellular carcinoma, Radiology 226(3):691–697, 2003.
Abbitt PL: Ultrasonography of the liver. An update on new applica- Caturelli E, et al: Hemangioma-like lesions in chronic liver disease:
tions, Clin Liver Dis 6(1):17–28, 2002. diagnostic evaluation in patients, Radiology 220(2):337–342,
Albrecht T, et al: Improved detection of hepatic metastases with pulse- 2001.
inversion US during the liver-specific phase of SHU 508A: multi- Chan WC, et al: Gallstone detection at CT in vitro: effect of peak
center study, Radiology 227(2):361–370, 2003. voltage setting, Radiology 241(2):546–553, 2006.
Albrecht T, et al: Guidelines for the use of contrast agents in ultra- Choi BI, et al: Small hepatocellular carcinomas and associated nodular
sound, Ultraschall Med 25(4):249–256, 2004. lesions of the liver: pathology, pathogenesis, and imaging findings,
Araki J, et al: Natural course of autoimmune pancreatitis without AJR Am J Roentgenol 160(6):1177–1187, 1993.
steroid therapy showing hypoechoic masses in the uncinate process Choudhry S, et al: Comparison of tissue harmonic imaging with con-
and tail of the pancreas on ultrasonography, J Ultrasound Med ventional US in abdominal disease, Radiographics 20(4):1127–1135,
25(8):1063–1067, 2006. 2000.
Ardelean M, et al: Contrast enhanced ultrasound in the pathology of Chung YE, et al: Hepatocellular carcinoma variants: radiologic-
the pancreas—a monocentric experience, Med Ultrason 16(4):325– pathologic correlation, AJR Am J Roentgenol 193(1):W7–W13,
331, 2014. 2009.
Arena U, et al: Acute viral hepatitis increases liver stiffness values Coco B, et al: Transient elastography: a new surrogate marker of liver
measured by transient elastography, Hepatology 47(2):380–384, fibrosis influenced by major changes of transaminases, J Viral Hepat
2008. 14(5):360–369, 2007.
Bach AM, et al: Gallbladder cancer: can ultrasonography evaluate Colli A, et al: Abnormalities of Doppler waveform of the hepatic veins
extent of disease?, J Ultrasound Med 17(5):303–309, 1998. in patients with chronic liver disease: correlation with histologic
Bachar GN, et al: Effect of aging on the adult extrahepatic bile duct: findings, AJR Am J Roentgenol 162(4):833–837, 1994.
a sonographic study, J Ultrasound Med 22(9):879–882, quiz 883-875, Corwin MT, et al: Incidentally detected gallbladder polyps: is follow-up
2003. necessary? Long-term clinical and US analysis of 346 patients, Radi-
Balthazar EJ: Complications of acute pancreatitis: clinical and CT ology 258(1):277–282, 2011.
evaluation, Radiol Clin North Am 40(6):1211–1227, 2002. Cura M, et al: Diagnostic and interventional radiology for Budd-Chiari
Bargallo X, et al: Sonography of the caudate vein: value in diagnosing syndrome, Radiographics 29(3):669–681, 2009.
Budd-Chiari syndrome, AJR Am J Roentgenol 181(6):1641–1645, D’Angelica M, et al: Analysis of the extent of resection for adenocar-
2003. cinoma of the gallbladder, Ann Surg Oncol 16(4):806–816, 2009.
Bartolotta TV, et al: Characterization of benign hepatic tumors arising de Groen PC, et al: Biliary tract cancers, N Engl J Med 341(18):1368–
in fatty liver with SonoVue and pulse inversion US, Abdom Imaging 1378, 1999.
32(1):84–91, 2007. Dietrich CF, et al: Differentiation of focal nodular hyperplasia
Bartolozzi C, et al: Differentiation of hepatocellular adenoma and focal and hepatocellular adenoma by contrast-enhanced ultrasound, Br J
nodular hyperplasia of the liver: comparison of power Doppler Radiol 78(932):704–707, 2005.
imaging and conventional color Doppler sonography, Eur Radiol Ding H, et al: Imaging of focal liver lesions: low-mechanical-index
7(9):1410–1415, 1997. real-time ultrasonography with SonoVue, J Ultrasound Med 24(3):
Bekker J, et al: Early hepatic artery thrombosis after liver transplanta- 285–297, 2005.
tion: a systematic review of the incidence, outcome and risk factors, Dodd GD 3rd, et al: Detection of transjugular intrahepatic portosys-
Am J Transplant 9(4):746–757, 2009. temic shunt dysfunction: value of duplex Doppler sonography, AJR
Bennett GL, et al: Sonographic detection of hepatocellular carcinoma Am J Roentgenol 164(5):1119–1124, 1995.
and dysplastic nodules in cirrhosis: correlation of pretransplantation Duffy A, et al: Gallbladder cancer (GBC): 10-year experience at
sonography and liver explant pathology in 200 patients, AJR Am J Memorial Sloan-Kettering Cancer Centre (MSKCC), J Surg Oncol
Roentgenol 179(1):75–80, 2002. 98(7):485–489, 2008.
Blomley MJ, et al: Improved imaging of liver metastases with stimu- Eberhardt SC, et al: Utility of sonography for small hepatic lesions
lated acoustic emission in the late phase of enhancement with the found on computed tomography in patients with cancer, J Ultrasound
US contrast agent SH U 508A: early experience, Radiology 210(2): Med 22(4):335–343, quiz 345-336, 2003.
409–416, 1999. European Association for the Study of Liver: EASL Clinical Practice
Bloom CM, et al: Role of US in the detection, characterization, Guidelines: management of hepatitis C virus infection, J Hepatol
and staging of cholangiocarcinoma, Radiographics 19(5):1199–1218, 55(2):245–264, 2011.
1999. European Association for the Study of Liver, European Organisation
Bolondi L, et al: Sonography of chronic pancreatitis, Radiol Clin North for Research and Treatment of Cancer: EASL-EORTC clinical prac-
Am 27(4):815–833, 1989. tice guidelines: management of hepatocellular carcinoma, Eur J
Bolondi L: Screening for hepatocellular carcinoma in cirrhosis, J Hepat Cancer 48(5):599–641, 2012.
39(6):1076–1084, 2003. Ewertsen C, et al: Biopsy guided by real-time sonography fused with
Brancatelli G, et al: Benign regenerative nodules in Budd-Chiari syn- MRI: a phantom study, AJR Am J Roentgenol 190(6):1671–1674,
drome and other vascular disorders of the liver: radiologic-pathologic 2008.
and clinical correlation, Radiographics 22(4):847–862, 2002. Fagan SP, et al: Prognostic factors for the development of gangrenous
Brancatelli G, et al: Budd-Chiari syndrome: spectrum of imaging find- cholecystitis, Am J Surg 186(5):481–485, 2003.
ings, AJR Am J Roentgenol 188(2):W168–W176, 2007. Feldman MK, et al: US artifacts, Radiographics 29(4):1179–1189,
Brannigan M, et al: Blood flow patterns in focal liver lesions at 2009.
microbubble-enhanced US, Radiographics 24(4):921–935, 2004. Feng B, Song Q: Does the common bile duct dilate after cholecystec-
Bruix J, Sherman M: Management of hepatocellular carcinoma: an tomy? Sonographic evaluation in 234 patients, AJR Am J Roentgenol
update, Hepatology 53(3):1020–1022, 2011. 165(4):859–861, 1995.
Bryant TH, et al: Improved characterization of liver lesions with liver- Fennessy FM, et al: Hepatic capsular retraction in metastatic carci-
phase uptake of liver-specific microbubbles: prospective multicenter noma of the breast occurring with increase or decrease in size of
study, Radiology 232(3):799–809, 2004. subjacent metastasis, AJR Am J Roentgenol 182(3):651–655, 2004.
Buetow PC, et al: Biliary cystadenoma and cystadenocarcinoma: Ferraioli G, et al: Shear wave elastography for evaluation of liver fibro-
clinical-imaging-pathologic correlations with emphasis on the impor- sis, J Ultrasound Med 33(2):197–203, 2014.
tance of ovarian stroma, Radiology 196(3):805–810, 1995. Fidelman N, et al: The transjugular intrahepatic portosystemic shunt:
Buetow PC, et al: Focal nodular hyperplasia of the liver: radiologic- an update, AJR Am J Roentgenol 199(4):746–755, 2012.
pathologic correlation, Radiographics 16(2):369–388, 1996. Filly RA, et al: Sonographic evaluation of liver nodularity: inspection
Caiado AH, et al: Complications of liver transplantation: multimodal- of deep versus superficial surfaces of the liver, J Clin Ultrasound
ity imaging approach, Radiographics 27(5):1401–1417, 2007. 30(7):399–407, 2002.
Campbell SC, et al: Slow flow or no flow? Color and power Doppler Finstad TA, et al: Sonography of acute pancreatitis: prevalence of find-
US pitfalls in the abdomen and pelvis, Radiographics 24(2):497–506, ings and pictorial essay, Ultrasound Q 21(2):95–104, quiz 150, 153-
2004. 104, 2005.
Fraquelli M, et al: Reproducibility of transient elastography in the Kim T, et al: Liver masses with central or eccentric scar, Semin Ultra-
evaluation of liver fibrosis in patients with chronic liver disease, Gut sound CT MR 30(5):418–425, 2009.
56(7):968–973, 2007. Kim TK, et al: Benign liver masses: imaging with microbubble contrast
Friedrich-Rust M, et al: Real-time elastography for noninvasive assess- agents, Ultrasound Q 22(1):31–39, 2006.
ment of liver fibrosis in chronic viral hepatitis, AJR Am J Roentgenol Kim TK, et al: Focal nodular hyperplasia and hepatic adenoma: dif-
188(3):758–764, 2007. ferentiation with low-mechanical-index contrast-enhanced sonogra-
Garcia-Criado A, et al: Radiology in liver transplantation, Semin Ultra- phy, AJR Am J Roentgenol 190(1):58–66, 2008.
sound CT MR 23(1):114–129, 2002. Koga A, et al: Diagnosis and operative indications for polypoid lesions
Ghittoni G, et al: Intrabile duct metastasis from colonic adenocarci- of the gallbladder, Arch Surg 123(1):26–29, 1988.
noma without liver parenchyma involvement: contrast enhanced Kojiro M, et al: Hepatocellular carcinoma presenting as intrabile duct
ultrasonography detection, Abdom Imaging 35(3):346–348, 2010. tumor growth: a clinicopathologic study of 24 cases, Cancer 49(10):
Giorgio A, et al: Contrast-enhanced sonographic appearance of hepa- 2144–2147, 1982.
tocellular carcinoma in patients with cirrhosis: comparison with Korner M, et al: Current role of emergency US in patients with major
contrast-enhanced helical CT appearance, AJR Am J Roentgenol trauma, Radiographics 28(1):225–242, 2008.
183(5):1319–1326, 2004. Kruskal JB, et al: Hepatic colon cancer metastases in mice: dynamic in
Golli M, et al: Hepatocellular adenoma: color Doppler US and patho- vivo correlation with hypoechoic rims visible at US, Radiology
logic correlations, Radiology 190(3):741–744, 1994. 215(3):852–857, 2000.
Gosink BB, Leymaster CE: Ultrasonic determination of hepatomegaly, Lafortune M, et al: Segmental anatomy of the liver: a sonographic
J Clin Ultrasound 9(1):37–44, 1981. approach to the Couinaud nomenclature, Radiology 181(2):443–
Goyal N, et al: Non-invasive evaluation of liver cirrhosis using ultra- 448, 1991.
sound, Clin Radiol 64(11):1056–1066, 2009. Leifer DM, et al: Follow-up of patients at low risk for hepatic malig-
Grant EG, et al: Budd-Chiari syndrome: the results of duplex and color nancy with a characteristic hemangioma at US, Radiology 214(1):167–
Doppler imaging, AJR Am J Roentgenol 152(2):377–381, 1989. 172, 2000.
Grazioli L, et al: Hepatic adenomas: imaging and pathologic findings, Lencioni R, et al: Portal vein thrombosis after percutaneous ethanol
Radiographics 21(4):877–892, discussion 892-874, 2001. injection for hepatocellular carcinoma: value of color Doppler sonog-
Hann LE, et al: Cholangiocarcinoma at the hepatic hilus: sonographic raphy in distinguishing chemical and tumor thrombi, AJR Am J
findings, AJR Am J Roentgenol 168(4):985–989, 1997. Roentgenol 164(5):1125–1130, 1995.
Harvey CJ, et al: Hepatic malignancies: improved detection with pulse- Levy AD, et al: From the archives of the AFIP. Benign tumors and
inversion US in late phase of enhancement with SH U 508A-early tumorlike lesions of the gallbladder and extrahepatic bile ducts:
experience, Radiology 216(3):903–908, 2000. radiologic-pathologic correlation. Armed Forces Institute of Pathol-
Harvey RT, Miller WT Jr: Acute biliary disease: initial CT and ogy, Radiographics 22(2):387–413, 2002.
follow-up US versus initial US and follow-up CT, Radiology Lewall DB, McCorkell SJ: Hepatic echinococcal cysts: sonographic
213(3):831–836, 1999. appearance and classification, Radiology 155(3):773–775, 1985.
Horrow MM, et al: Is age associated with size of adult extrahepatic bile Lim JH: Radiologic findings of clonorchiasis, AJR Am J Roentgenol
duct: sonographic study, Radiology 221(2):411–414, 2001. 155(5):1001–1008, 1990.
Hussain SM, et al: Focal nodular hyperplasia: findings at state-of-the- Lim JH: Cholangiocarcinoma: morphologic classification according to
art MR imaging, US, CT, and pathologic analysis, Radiographics growth pattern and imaging findings, AJR Am J Roentgenol 181(3):
24(1):3–17, discussion 18-19, 2004. 819–827, 2003.
Hutchins GF, Draganov PV: Cystic neoplasms of the pancreas: a diag- Lim JH, et al: Early bile duct carcinoma: comparison of imaging fea-
nostic challenge, World J Gastroenterol 15(1):48–54, 2009. tures with pathologic findings, Radiology 238(2):542–548, 2006.
Inoue T, et al: Value of liver parenchymal phase contrast-enhanced Liu LP, et al: Focal hypoechoic tumors of fatty liver: characterization
sonography to diagnose premalignant and borderline lesions and of conventional and contrast-enhanced ultrasonography, J Ultra-
overt hepatocellular carcinoma, AJR Am J Roentgenol 192(3):698– sound Med 28(9):1133–1142, 2009.
705, 2009. Mahmoud AE, et al: Poor prognosis and limited therapeutic options
Ishikawa O, et al: The difference in malignancy between pedunculated in patients with Budd-Chiari syndrome and portal venous system
and sessile polypoid lesions of the gallbladder, Am J Gastroenterol thrombosis, Eur J Gastroenterol Hepatol 9(5):485–489, 1997.
84(11):1386–1390, 1989. Marchal G, et al: Correlation of sonographic patterns in liver metasta-
Ito H, et al: Polypoid lesions of the gallbladder: diagnosis and followup, ses with histology and microangiography, Invest Radiol 20(1):79–84,
J Am Coll Surg 208(4):570–575, 2009. 1985.
Jang HJ, et al: Enhancement patterns of hepatocellular carcinoma at McGahan JP, et al: Appearance of solid organ injury with contrast-
contrast-enhanced US: comparison with histologic differentiation, enhanced sonography in blunt abdominal trauma: preliminary expe-
Radiology 244(3):898–906, 2007. rience, AJR Am J Roentgenol 187(3):658–666, 2006.
Jang HJ, et al: Small nodules (1-2 cm) in liver cirrhosis: characteriza- McLarney JK, et al: Fibrolamellar carcinoma of the liver: radiologic-
tion with contrast-enhanced ultrasound, Eur J Radiol 72(3):418– pathologic correlation, Radiographics 19(2):453–471, 1999.
424, 2009. McNaughton DA, Abu-Yousef MM: Doppler US of the liver made
Jarnagin WR, et al: Staging, resectability, and outcome in 225 patients simple, Radiographics 31(1):161–188, 2011.
with hilar cholangiocarcinoma, Ann Surg 234(4):507–517, discus- Miller WJ, et al: Imaging findings in Caroli’s disease, AJR Am J Roent-
sion 517-509, 2001. genol 165(2):333–337, 1995.
Jarnagin WR, et al: Papillary phenotype confers improved survival after Millonig G, et al: Extrahepatic cholestasis increases liver stiffness
resection of hilar cholangiocarcinoma, Ann Surg 241(5):703–712, (FibroScan) irrespective of fibrosis, Hepatology 48(5):1718–1723,
discussion 712-704, 2005. 2008.
Jeffrey RB, et al: Gangrenous cholecystitis: diagnosis by ultrasound, Millonig G, et al: Liver stiffness is directly influenced by central venous
Radiology 148(1):219–221, 1983. pressure, J Hepatol 52(2):206–210, 2010.
Jhaveri KS, et al: Intrahepatic bile duct dilatation due to liver metas- Mortele KJ, Ros PR: Imaging of diffuse liver disease, Semin Liver Dis
tases from colorectal carcinoma, AJR Am J Roentgenol 193(3):752– 21(2):195–212, 2001.
756, 2009. Mueller PR, et al: Postcholecystectomy bile duct dilatation: myth or
Jung EM, et al: New real-time image fusion technique for characteriza- reality?, AJR Am J Roentgenol 136(2):355–358, 1981.
tion of tumor vascularisation and tumor perfusion of liver tumors Niederau C, et al: Sonographic measurements of the normal liver,
with contrast-enhanced ultrasound, spiral CT or MRI: first results, spleen, pancreas, and portal vein, Radiology 149(2):537–540,
Clin Hemorheol Microcirc 43(1-2):57–69, 2009. 1983.
Kamaya A, et al: Hypervascular liver lesions, Semin Ultrasound CT MR Numata K, et al: Contrast-enhanced sonography of autoimmune pan-
30(5):387–407, 2009. creatitis: comparison with pathologic findings, J Ultrasound Med
Khalili K, et al: Hilar biliary obstruction: preliminary results with 23(2):199–206, 2004.
Levovist-enhanced sonography, AJR Am J Roentgenol 180(3):687– Ortega D, et al: Tissue harmonic imaging: is it a benefit for bile duct
693, 2003. sonography?, AJR Am J Roentgenol 176(3):653–659, 2001.
Chapter 15 Ultrasound of the liver, biliary tract, and pancreas 275.e3
Park H, et al: Characterization of focal liver masses using acoustic Soyer P, et al: Imaging of intrahepatic cholangiocarcinoma: 1. Periph-
radiation force impulse elastography, World J Gastroenterol 19(2):219– eral cholangiocarcinoma, AJR Am J Roentgenol 165(6):1427–1431,
226, 2013. 1995.
Pastacaldi S, et al: Hepatic artery thrombosis after orthotopic liver Stephen AE, Berger DL: Carcinoma in the porcelain gallbladder: a
transplantation: a review of nonsurgical causes, Liver Transpl 7(2): relationship revisited, Surgery 129(6):699–703, 2001.
75–81, 2001. Strobel D, et al: Tumor-specific vascularization pattern of liver metas-
Pedersen MR, et al: Ultrasound follow-up for gallbladder polyps less tasis, hepatocellular carcinoma, hemangioma and focal nodular
than 6 mm may not be necessary, Dan Med J 59(10):A4503, hyperplasia in the differential diagnosis of 1,349 liver lesions in
2012. contrast-enhanced ultrasound (CEUS), Ultraschall Med 30(4):376–
Perret RS, et al: Common bile duct measurements in an elderly popu- 382, 2009.
lation, J Ultrasound Med 19(11):727–730, quiz 731, 2000. Tamsel S, et al: Vascular complications after liver transplantation:
Quaia E, et al: Characterization of focal liver lesions with contrast- evaluation with Doppler US, Abdom Imaging 32(3):339–347, 2007.
specific US modes and a sulfur hexafluoride-filled microbubble con- Tanaka S, et al: Hepatocellular carcinoma: sonographic and histologic
trast agent: diagnostic performance and confidence, Radiology correlation, AJR Am J Roentgenol 140(4):701–707, 1983.
232(2):420–430, 2004. Tanaka S, et al: Color Doppler flow imaging of liver tumors, AJR Am
Quaia E, et al: Diagnostic value of hepatocellular nodule vascularity J Roentgenol 154(3):509–514, 1990.
after microbubble injection for characterizing malignancy in patients Tanaka K, et al: Diagnosis of portal vein thrombosis in patients with
with cirrhosis, AJR Am J Roentgenol 189(6):1474–1483, 2007. hepatocellular carcinoma: efficacy of color Doppler sonography
Quiroga S, et al: Complications of orthotopic liver transplantation: compared with angiography, AJR Am J Roentgenol 160(6):1279–
spectrum of findings with helical CT, Radiographics 21(5):1085– 1283, 1993.
1102, 2001. Tchelepi H, et al: Sonography of diffuse liver disease, J Ultrasound Med
Ralls PW, et al: Real-time sonography in suspected acute cholecystitis. 21(9):1023–1032, quiz 1033-1024, 2002.
Prospective evaluation of primary and secondary signs, Radiology Teefey SA, et al: Sonography of the gallbladder: significance of striated
155(3):767–771, 1985. (layered) thickening of the gallbladder wall, AJR Am J Roentgenol
Ralls PW, et al: Budd-Chiari syndrome: detection with color Doppler 156(5):945–947, 1991.
sonography, AJR Am J Roentgenol 159(1):113–116, 1992. Terzi C, et al: Polypoid lesions of the gallbladder: report of 100 cases
Ricci P, et al: Behavior of hepatocellular adenoma on real-time low- with special reference to operative indications, Surgery 127(6):622–
mechanical index contrast-enhanced ultrasonography with a second- 627, 2000.
generation contrast agent, J Ultrasound Med 27(12):1719–1726, Todani T, et al: Congenital bile duct cysts: classification, operative
2008. procedures, and review of thirty-seven cases including cancer arising
Rickes S, et al: Impact of the operator’s experience on value of high- from choledochal cyst, Am J Surg 134(2):263–269, 1977.
resolution transabdominal ultrasound in the diagnosis of choledo- Townsend RR, et al: Abdominal lymphoma in AIDS: evaluation with
cholithiasis: a prospective comparison using endoscopic retrograde US, Radiology 171(3):719–724, 1989.
cholangiography as the gold standard, Scand J Gastroenterol 41(7): Treece G, et al: 3D ultrasound measurement of large organ volume,
838–843, 2006. Med Image Anal 5(1):41–54, 2001.
Robledo R, et al: Extrahepatic bile duct carcinoma: US characteristics Trevisani F, et al: Surveillance for early diagnosis of hepatocellular
and accuracy in demonstration of tumors, Radiology 198(3):869– carcinoma: is it effective in intermediate/advanced cirrhosis?, Am J
873, 1996. Gastroenterol 102(11):2448–2457, quiz 2458, 2007.
Rossi S, et al: Contrast-enhanced versus conventional and color Trillaud H, et al: Characterization of focal liver lesions with SonoVue-
Doppler sonography for the detection of thrombosis of the portal enhanced sonography: international multicenter-study in compari-
and hepatic venous systems, AJR Am J Roentgenol 186(3):763–773, son to CT and MRI, World J Gastroenterol 15(30):3748–3756, 2009.
2006. Tublin ME, et al: Altered liver morphology after portal vein thrombo-
Rubaltelli L, et al: The role of vascularization in the formation of sis: not always cirrhosis, Dig Dis Sci 53(10):2784–2788, 2008.
echographic patterns of hepatic metastases: microangiographic and Ungermann L, et al: Focal nodular hyperplasia: spoke-wheel arterial
echographic study, Br J Radiol 53(636):1166–1168, 1980. pattern and other signs on dynamic contrast-enhanced ultrasonog-
Rubens DJ, et al: Doppler artifacts and pitfalls, Radiol Clin North Am raphy, Eur J Radiol 63(2):290–294, 2007.
44(6):805–835, 2006. van Breda Vriesman AC, et al: Diffuse gallbladder wall thickening:
Sahani DV, et al: Autoimmune pancreatitis: imaging features, Radiol- differential diagnosis, AJR Am J Roentgenol 188(2):495–501, 2007.
ogy 233(2):345–352, 2004. Vilgrain V: Ultrasound of diffuse liver disease and portal hypertension,
Sarvazyan A, et al: An overview of elastography—an emerging branch Eur Radiol 11(9):1563–1577, 2001.
of medical imaging, Curr Med Imaging Rev 7(4):255–282, 2011. Vilgrain V, et al: Atrophy-hypertrophy complex in patients with cavern-
Serste T, et al: Accuracy and disagreement of computed tomography ous transformation of the portal vein: CT evaluation, Radiology
and magnetic resonance imaging for the diagnosis of small hepato- 241(1):149–155, 2006.
cellular carcinoma and dysplastic nodules: role of biopsy, Hepatology von Herbay A, et al: Contrast-enhanced ultrasound with SonoVue:
55(3):800–806, 2012. differentiation between benign and malignant focal liver lesions in
Shah ZK, et al: Enhancing and expansile portal vein thrombosis: value 317 patients, J Clin Ultrasound 38(1):1–9, 2010.
in the diagnosis of hepatocellular carcinoma in patients with multiple Wernecke K, et al: Pathologic explanation for hypoechoic halo seen on
hepatic lesions, AJR Am J Roentgenol 188(5):1320–1323, 2007. sonograms of malignant liver tumors: an in vitro correlative study,
Shaw AS, et al: Ultrasound of non-vascular complications in the post AJR Am J Roentgenol 159(5):1011–1016, 1992.
liver transplant patient, Clin Radiol 58(9):672–680, 2003. Wibbenmeyer LA, et al: Sonographic diagnosis of unsuspected gall-
Shin LK, et al: Intraoperative ultrasound of the pancreas, Ultrasound bladder cancer: imaging findings in comparison with benign gall-
Q 25(1):39–48, quiz 48, 2009. bladder conditions, AJR Am J Roentgenol 165(5):1169–1174, 1995.
Singal A, et al: Meta-analysis: surveillance with ultrasound for early- Wiles R, et al: Growth rate and malignant potential of small gallbladder
stage hepatocellular carcinoma in patients with cirrhosis, Aliment polyps—systematic review of evidence, Surgeon 12(4):221–226,
Pharmacol Ther 30(1):37–47, 2009. 2014.
Sirli R, Sporea I: Ultrasound examination of the normal pancreas, Med Wilson SR, Burns PN: An algorithm for the diagnosis of focal liver
Ultrason 12(1):62–65, 2010. masses using microbubble contrast-enhanced pulse-inversion sonog-
Smith MT, et al: Best cases from the AFIP: fibrolamellar hepatocellular raphy, AJR Am J Roentgenol 186(5):1401–1412, 2006.
carcinoma, Radiographics 28(2):609–613, 2008. Wilson SR, Burns PN: Microbubble-enhanced US in body imaging:
Smith EA, et al: Cross-sectional imaging of acute and chronic gallblad- what role?, Radiology 257(1):24–39, 2010.
der inflammatory disease, AJR Am J Roentgenol 192(1):188–196, Wilson SR, et al: Volume imaging in the abdomen with ultrasound:
2009. how we do it, AJR Am J Roentgenol 193(1):79–85, 2009.
Soyer P, et al: Surgical segmental anatomy of the liver: demonstration Winston CB, et al: Recurrent gallbladder carcinoma along laparoscopic
with spiral CT during arterial portography and multiplanar recon- cholecystectomy port tracks: CT demonstration, Radiology 212(2):
struction, AJR Am J Roentgenol 163(1):99–103, 1994. 439–444, 1999.
Xu HX, et al: Enhancement pattern of hilar cholangiocarcinoma: Yu H, Wilson SR: New noninvasive ultrasound techniques: can they
contrast-enhanced ultrasound versus contrast-enhanced computed predict liver cirrhosis?, Ultrasound Q 28(1):5–11, 2012.
tomography, Eur J Radiol 75(2):197–202, 2010. Zizka J, et al: Value of Doppler sonography in revealing transjugular
Yeh HC, et al: Microcystic features at US: a nonspecific sign for micro- intrahepatic portosystemic shunt malfunction: a 5-year experience in
cystic adenomas of the pancreas, Radiographics 21(6):1455–1461, 216 patients, AJR Am J Roentgenol 175(1):141–148, 2000.
2001.
Young ST, et al: CT of the liver in patients with metastatic breast
carcinoma treated by chemotherapy: findings simulating cirrhosis,
AJR Am J Roentgenol 163(6):1385–1388, 1994.
CHAPTER 16
Endoscopic ultrasound of the biliary tract
and pancreas
Suhail Bakr Salem, Mark Andrew Schattner, and Hans Gerdes
IMAGING AND DIAGNOSIS

The normal pancreatic parenchyma has a homogeneous
The diagnosis of benign and malignant diseases of the pancreas echogenic appearance (Fig. 16.1), and tumors usually appear
and biliary tree classically have relied on a detailed history and hypoechoic, often with irregular borders, in sharp distinction
complete physical examination, with correlation of the results from healthy tissue (Fig. 16.2). Large tumors of the pancreas
of clinical chemistries. Imaging of the hepatic and pancreatic may be difficult to evaluate completely because of limited pen-
parenchyma and ductal anatomy has, however, evolved as criti- etration of the tumor by US. Conversely, small tumors of the
cal for accurate diagnosis and for guiding therapy. Routine pancreas that are often missed by CT or MRI are readily
radiography does not provide the soft-tissue resolution required, imaged by EUS. For example, islet cell tumors, which are often
but ultrasonography (US; see Chapter 15), computed tomog- encapsulated and small, are frequently detected on EUS and
raphy (CT; see Chapter 18), and magnetic resonance imaging appear as well-demarcated hypoechoic lesions (see Chapter
(MRI; see Chapter 19) have become important noninvasive 65). Other neuroendocrine tumors, such as gastrinomas, may
modalities in the routine investigation of almost any symptom, be isoechoic within the pancreatic parenchyma and difficult to
physical sign, or laboratory abnormality potentially relating to identify without careful, tedious, real-time imaging (Fig. 16.3).
pathologic conditions of the biliary tree or pancreas. Ampullary tumors are also often seen and staged on EUS
Invasive procedures for imaging the biliary and pancreatic because of their relation to the duodenal wall, CBD, and pan-
ductal systems, primarily percutaneous cholangiography (PTC) creatic duct (Fig. 16.4) (see Chapter 59). Extrahepatic bile duct
(see Chapter 20) or endoscopic retrograde cholangiopancrea- tumors can also be detected and described in detail with EUS
tography (ERCP) (see Chapter 29) remain important therapeu- imaging (Fig. 16.5).
tically, but diagnostically, these have been almost entirely Cysts of the pancreas are generally anechoic and well demar-
replaced by less invasive modalities. Endoscopic ultrasonogra- cated and thus easily identified even when small (see Chapter
phy (EUS) has become an essential tool for diagnosis and 60). Some cysts may have internal echoes or solid areas, which
treatment. Its high-resolution images complement the more raise concern for a mucinous lesion or associated tumor (Figs.
general findings of cross-sectional imaging and result in a 16.6 and 16.7). Cysts can easily be distinguished from vascular
higher sensitivity for diagnosis of early-stage disease and detec- structures using the Doppler flow. Serous cystadenomas may
tion of smaller lesions. Linear array echoendoscopes permit also appear isoechoic with the pancreas and require careful
guided passage of needles and devices through the endoscope, imaging for proper identification (Fig. 16.8). Pseudocysts can
allowing biopsies to be obtained and permitting therapeutic vary in size and sonographic features but usually lack a discrete
interventions. This chapter discusses the techniques of radial wall and septa, particularly if acute. However, internal echoes
and linear endosonography in the diagnosis, staging, and treat- often are seen as a result of necrotic debris (Fig. 16.9).
ment of benign and malignant disease of the pancreas and
biliary tree.
Endoscopic Ultrasound-Guided Fine-Needle Aspiration
and Biopsy
ENDOSCOPIC ULTRASOUND TECHNIQUE The development of linear array echoendoscopes, which scan
an area orthogonally in line with the endoscope (thus in line
The pancreas is located posterior to the stomach and is readily with the biopsy channel), has allowed the development of endo-
seen by EUS imaging through the wall of the stomach and scopic ultrasound-guided fine-needle aspiration (EUS-FNA)
duodenum. With the transducer in the duodenum, the pancre- and needle biopsy. The indications for EUS-guided aspiration
atic head and uncinate process, ampulla of Vater, pancreatic or biopsy include pathologic confirmation of a suspected
ducts, common bile duct (CBD), and the surrounding vascular pancreatic or periampullary cancer, evaluation of pancreatic
and nodal structures can be visualized. With the transducer in masses or lymph nodes of unclear etiology, and aspiration of
the stomach, the pancreatic body and tail, gallbladder, and left pancreatic cysts to help distinguish mucinous from inflamma-
lobe of the liver are seen. Additionally, the celiac, splenic, tory or serous lesions. EUS-guided needle puncture has also
hepatic, and superior mesenteric arteries, as well as the splenic, provided the platform for therapeutic EUS-guided techniques
superior mesenteric, and portal veins are all seen in detail (see (see sections Endoscopic Ultrasound-Guided Therapy and Novel
Chapter 1). Therapeutics).
276
Chapter 16 Endoscopic ultrasound of the biliary tract and pancreas 277
Endoscopic Ultrasound Fine Needle Aspiration Technique

Within the duodenum or stomach, the EUS probe is positioned
in close proximity to the target lesion, typically less than 3 cm
away. The area is then interrogated with Doppler flow to ensure
the absence of significant vascular structures in the needle path.
A 25-, 22-, or 19-gauge needle can then be directed into the
target lesion. The tip and shaft of the needles used for FNA
produce a bright, hyperechoic image. This allows the needle to
be followed in real time to ensure precise positioning within the
target lesion (Fig. 16.10). Ideally, a cytopathologist or cytotech-
nologist should be present at the time of the FNA to determine
the cellular adequacy of the specimen. Alternatively, multiple
punctures (up to seven) should be performed to ensure an
adequate cytologic specimen (LeBlanc et al, 2004). Cyst fluid
can also be aspirated and sent for cytology, tumor markers, and
chemical analysis (Fig. 16.11). Fine needle biopsies can also be
obtained, as a result of the development of new needles
FIGURE 16.1. Normal endoscopic ultrasound image of the body of (Iglesias-Garcia et al, 2011). These needles allow tissue with
the pancreas with thin main pancreatic duct (arrow). preserved architecture to be acquired, as opposed to purely
cytologic specimens.
DIAGNOSIS OF PANCREATIC CANCER

CBD
Endoscopic Ultrasound Fine-Needle Aspiration of Solid
Pancreatic Lesions
M Solid masses of the pancreas may represent a primary pan
creatic cancer, neuroendocrine tumor, metastatic lesion, or
focal pancreatitis. These masses may be difficult to visualize
PV SMV
on noninvasive imaging, when small. Endoscopic ultrasound
allows high-resolution imaging of the pancreas as well as facili-
tating guidance of FNA. Resectable pancreatic lesions for which
surgery is indicated generally do not require preoperative EUS
with FNA. However, patients with locally advanced pancreatic
cancer require histologic confirmation of the diagnosis before
FIGURE 16.2. Solid, irregular, hypoechoic mass (M) in the pancreatic
the initiation of chemotherapy or radiation. Tissue acquisition
head seen with abrupt termination of a dilated common bile duct (CBD)
for molecular testing is also playing an increasing role in guiding
and contact with the portal confluence (portal vein [PV] and superior
mesenteric vein [SMV]). Cytology obtained by endoscopic ultrasound– therapy. In addition, in cases in which the diagnosis is uncer-
guided fine needle aspiration proved to be adenocarcinoma. tain, such as autoimmune pancreatitis, EUS with FNA can
provide useful information to help stratify those patients who
would benefit from surgery.
EUS-FNA is highly accurate in diagnosing patients with
suspected pancreatic cancer, with a sensitivity of 77% to 95%,
specificity of 94% to 99%, positive predictive value (PPV) of
98% to 100%, and a negative predictive value (NPV) of 86%
to 92% (Brugge et al, 2010; Chen et al, 2012; Puli et al, 2013;
Turner et al, 2010; Wiersema et al, 1997; Yusuf et al, 2009)
(Table 16.1). In comparison, percutaneous CT-guided biopsy
SV has a pooled sensitivity of 87% and NPV of only 58% (Hartwig
et al, 2009) (see Chapter 22). Similarly, the sensitivity of cytol-
ogy obtained from brushings during cholangiography is poor
(approximately 40%) and cannot adequately exclude cancer
(Wakatsuki et al, 2005). EUS-FNA should be considered as the
initial approach to lesions in or adjacent to the pancreas. In a
prospective study of 102 patients with suspected pancreatic
cancer who had a negative CT-guided biopsy, EUS-FNA had
a sensitivity of 95%, specificity of 100%, PPV of 100%, and a
NPV of 92% (Gress et al, 2001). With EUS-FNA, the risk of
FIGURE 16.3. Small, well-circumscribed, hypoechoic appearance of needle-tract tumor seeding and peritoneal carcinomatosis is
an insulinoma (arrow) in the body of the pancreas with splenic vein below minimized, as the FNA is generally performed through a
(SV). segment of duodenal or stomach wall that will be removed as
A B
FIGURE 16.4. Endoscopic (A) and endoscopic ultrasound (B) imaging of an ampullary tumor (arrow).
FIGURE 16.5. Hypoechoic mass appearance of cholangiocarcinoma

FIGURE 16.6. Mucinous cystic lesion in the head of the pancreas with
of the common bile duct (solid arrow) with biliary stent visible (open
mural nodule (arrow).
arrow).
FIGURE 16.7. Multiseptated mucinous cystic lesion in the head of the FIGURE 16.8. Typical microcystic appearance of a serous cystade-
pancreas (arrow). noma in the head of the pancreas (arrow).
TABLE 16.1 Performance Characteristics for EUS in the

Diagnosis and Staging of Pancreatic, Ampullary, and Extrahepatic
Biliary Tumors
Sensitivity Specificity Accuracy
Pancreatic Tumors
Diagnosis*
Tumor >3 cm 86%-95% 94%-99% 86%-94%
Tumor <3 cm 85%-95% 87%-100% 93%-100%
Tumor staging — — 72%-98%
Nodal staging — — 44%-66%
Ampullary Tumors — —
Diagnosis — — 93%-100%
Resectability — — 61%-88%
Bile Duct Tumors — —
Diagnosis* — — 43%-86%
FIGURE 16.9. Large debris-filled pseudocyst adjacent to the tail of the TN staging — — 60%-80%
pancreas.
*Includes the use of fine-needle aspiration to obtain cytology. EUS, Endoscopic ultrasound;
TN, tumor node.
part of a resection, should the patient be an appropriate surgical

candidate (Fornarni et al, 1989). Although the accuracy of
EUS-FNA is well established, fine-needle biopsy has demon-
strated comparable sensitivity and specificity of 90% and 100%,
respectively, with added benefit of preserved tissue architecture
in cases where diagnosis is more challenging (Iglesias-Garcia
et al, 2011). Although not part of routine clinical practice
yet, molecular genetic analysis promises to enhance the diag-
nostic capacity of EUS-FNA (see Chapter 9B). In one study,
87% of patients with pancreatic ductal adenocarcinoma
were found to have KRAS mutations (Ogura et al, 2012).
When combined with standard cytohistopathologic assessment,
KRAS mutational analysis increased the sensitivity and specific-
ity of EUS-FNA by 5% and 6%, respectively. Despite these
enhancements to diagnostic accuracy, false-positive cytologic
evaluations leading to unnecessary operations may still occur,
especially in cases of autoimmune pancreatitis (Learn et al,
FIGURE 16.10. Bright appearance of fine needle aspiration of a solid 2011). Specimens obtained by EUS must be interpreted in
mass in the head of the pancreas. context, including clinical history, laboratory and serologic
data, as well as radiographic studies to minimize risk of
misdiagnosis.
Endoscopic Ultrasound Fine Needle Aspiration of
Pancreatic Cystic Lesions
Cystic lesions of the pancreas remain a diagnostic and thera-
peutic challenge (see Chapter 60). The differential diagnosis
includes retention/simple cysts, pseudocysts, cystic neoplasms,
and cystic degeneration of solid masses (Table 16.2). EUS
provides detailed imaging of the entire pancreas, including
location and number of cysts, size, ductal dilatation or com-
munication, signs of chronic pancreatitis, cyst wall thickness,
mural nodules, papillary projections, and intracystic structures
such as septations, debris, and mass components. Retention
cysts are generally small, thin-walled, and unilocular; they carry
no malignant potential and can be left untreated. Pseudocysts
develop as a result of acute inflammation (see Chapters 55 and
56). On EUS, they are thick walled as they become chronic
and may have debris within the cyst cavity (see Fig. 16.9).
FIGURE 16.11. Fine needle aspiration of cyst fluid. They can be seen in or adjacent to the pancreas and often
TABLE 16.2 Imaging Characteristics and Fluid Analysis of Pancreatic Cystic Lesions
Communication with
EUS Morphology Pancreatic Duct Fluid Amylase Fluid CEA Fluid Cytology
Retention cysts Unilocular, thin walled No Variable Low Normal duct or centroacinar cells
Pseudocyst Unilocular, thick-walled with Yes Very high Low Inflammatory cells, no epithelial
debris cells
Serous Microcystic with central No Low Low Small cuboidal cells, positive
cystadenoma calcification glycogen stain
Mucinous Macrocystic with thick Occasionally Low High Atypical ductal cells, positive
cystadenoma septations mucin stain
IPMN Dilated, tortuous pancreatic Yes High High Atypical ductal cells, positive
duct and/or side branches mucin stain
CEA, Carcinoembryonic antigen; EUS, endoscopic ultrasound; IPMN, intraductal papillary mucinous neoplasm.
communicate with the pancreatic duct. FNA will yield thick optimal cutoff is not universally agreed upon, a cyst-fluid CEA
fluid with inflammatory cells, but no epithelial cells and tumor concentration of more than 192 ng/mL predicts a mucinous
marker levels (carcinoembryonic antigen [CEA]) in the cyst lesion with a sensitivity of 73%, specificity of 84%, and an
fluid are low or undetectable; by contrast, fluid amylase is accuracy of 79% (Brugge et al, 2004; Cizginer et al, 2011).
usually markedly elevated. Although helpful for diagnosing mucinous lesions, cyst-fluid
Regarding cystic neoplasms, it is important to distinguish CEA levels do not correlate with the presence of malignancy.
mucinous and papillary tumors from serous lesions (see Another limitation of cyst-fluid CEA analysis is that it requires
Chapter 60). Serous tumors have no malignant potential and aspiration of at least 1 mL of fluid, which can be difficult if the
do not require resection unless symptoms occur or they cavity is small or the fluid is very viscous, making it difficult to
encroach on vascular structures. Typically, serous cystadeno- pull through a fine needle. More recently, cyst-fluid DNA anal-
mas are composed of a honeycomb of microcysts with no com- ysis has shown promise in differentiating mucinous from non-
munication with the pancreatic duct (see Fig. 16.8). They can mucinous cysts. This analysis can be done on as little as 200 µL
be large and exhibit central stellate calcification on imaging of fluid. The presence of a KRAS mutation, high DNA content,
studies. Papillary lesions and solid pseudopapillary tumors, and loss of polymorphic alleles are indicators of mucinous cysts.
carry the risk of malignant transformation and should be In a large validation study, early KRAS mutation followed by
referred for resection. Mucinous lesions—mucinous cystic allelic loss was 96% specific and 37% sensitive for a malignant
neoplasm (MCN) and intraductal papillary mucinous neo- cyst (Khalid et al, 2009). Interleukin-1β (IL-1β) has recently
plasm (IPMN)—also harbor malignant potential. Mucinous been identified as a potential biomarker for high-grade dyspla-
cysts are generally composed of one or more large cystic spaces sia or malignancy (Maker et al, 2011). Cyst fluid aspirated
with thickened walls or septa (see Fig. 16.7). Features that ahead of surgical resection was found to contain higher con-
may indicate malignant potential of MCNs include size greater centrations of IL-1β in specimens with evidence of high-grade
than 4 cm, mural nodules, mass-forming lesions, and periph- dysplasia or carcinoma. Nonetheless, a reliable and sufficiently
eral egg-shell calicifcations (Reddy et al, 2004). In the case of sensitive biomarker for malignancy is still lacking.
IPMNs, distinguishing main-duct from side-branch types is
important clinically because the former are associated with
Complications of Endoscopic Ultrasound
malignancy in as many as 62% of patients (Tanaka et al, Fine-Needle Aspiration
2012). Although cross-sectional imaging may help in this EUS-FNA of the pancreas is a safe procedure. The reported
regard by demonstrating a dilated pancreatic duct characteris- rate of pancreatitis is 0% to 2% (Eloubeidi et al, 2004b; Gress
tic of main-duct IPMN, communication with the main pan- et al, 2002a; O’Toole et al, 2001; Williams et al, 1999). In a
creatic duct can also be demonstrated on EUS. multicenter analysis of almost 5000 EUS-FNAs of solid pan-
Although these anatomic characteristics are helpful, EUS creatic masses, the incidence of pancreatitis was 0.3%, and the
cyst morphology alone is insufficient to characterize the lesion majority of cases were clinically mild (Eloubeidi et al, 2004b).
(Ahmad et al, 2003; Brugge et al, 2004; Tanaka et al, 2012). EUS-guided aspiration of pancreatic cysts is similarly safe, with
Aspiration of cyst fluid via EUS-FNA can further aid in the a pancreatitis rate of less than 1% and an overall complication
diagnosis. Cyst fluid can be analyzed for tumor markers, rate of about 2% (Lee et al, 2005). Bacteremia is uncommon
amylase, and molecular markers, and it provides material for following EUS-FNA of solid lesions, with a rate similar to that
cytopathologic assessment. Cyst fluid tends to be paucicellular; for diagnostic upper endoscopy; thus prophylactic antibiotic
thus the sensitivity and negative predictive value of cytology is administration is unwarranted (Janssen et al, 2004; Levy et al,
low (Centeno et al, 1997; Sedlack et al, 2002). The presence 2003). However, early data evaluating EUS-guided cyst aspira-
of thick, viscous fluid and a positive mucin stain is suggestive tion showed increased infectious complications, leading to the
of a mucinous lesion. The concentration of several tumor recommendation of routine antibiotic prophylaxis for all
markers in pancreatic cyst fluid has been examined, and CEA patients undergoing this procedure (Adler et al, 2005). Major
has been found to be the most useful. Higher concentrations extraluminal hemorrhage is a rare complication and occurs at
of CEA most accurately predict mucinous lesions, although the a rate of less than 1% (Affi et al, 2001).
staging modality. The same can be said for determining vascular

TABLE 16.3 AJCC Staging of Pancreatic Cancer
invasion and resectability of large pancreatic tumors (>3 cm).
TNM Definitions The performance characteristics of EUS for the diagnosis
Primary Tumor (T)
and staging of pancreatic tumors is summarized in Table 16.1.
For example, using the TNM system of cancer staging, Muller
TX Cannot be assessed
and colleagues (1994) showed that EUS has an accuracy of
T0 No evidence of primary tumor
75% for staging pancreatic cancer, compared with 56% for CT
Tis Carcinoma in situ and 57% for MRI. A prospective study comparing EUS with
T1 Tumor confined to pancreas, ≤2 cm in diameter multidetector CT for diagnosis and staging of pancreatic cancer
T2 Tumor confined to pancreas, >2 cm in diameter also found EUS to be more accurate for diagnosis and tumor
T3 Extrapancreatic extension, no celiac axis or SMA staging (DeWitt et al, 2004). Of the 80 evaluated patients, the
involvement sensitivity for detecting a pancreatic mass was 98% for EUS
T4 Involvement of celiac axis or SMA and 86% for CT (p = 0.012). Both modalities were equivalent
Regional Lymph Nodes (N) for nodal staging (44% vs. 47%) and for predicting resectability
NX Cannot be assessed (88% vs. 92%). In contrast, another recent study favored CT,
N0 Regional lymph node metastases absent with an accuracy of 74% for tumor staging, 62% for node
N1 Regional lymph node metastases present staging, and 83% for resectability, compared with 62%, 65%,
Distant Metastases (M) and 72%, respectively for EUS (Soriano et al, 2004).
Perhaps the most useful application of EUS is the diagnosis
MX Cannot be assessed
of small pancreatic tumors (<3 cm) in patients with equivocal
M0 Distant metastases absent
radiologic imaging (see Table 16.1). Several studies have dem-
M1 Distant metastases present
onstrated that EUS is superior to CT scanning for evaluation
AJCC TNM of pancreatic lesions smaller than 3 cm. For example, Muller
Stage 0 Tis, N0, M0 and colleagues (1994) reported sensitivities of 93% for EUS,
Stage I 67% for MRI, and 53% for CT using surgical pathology as the
IA T1, N0, M0 gold standard. In addition, EUS far exceeds CT and MRI in
IB T2, N0, M0 the diagnosis and staging of tumors of the ampulla of Vater to
Stage II determine whether local or radical resection may be appropriate
IIA T3, N0, M0 (see Table 16.1). Other reports, as well as our own experience,
IIB T1-T3, N1, M0 have shown the usefulness of EUS in evaluating patients with
Stage T4, N0-N1, M0
dilated pancreatic and/or bile ducts but no demonstrable cause.
III EUS in this setting has been useful in the identification of small
Stage T1-T4, N0-N1, M1 tumors of the ampulla, pancreas, or bile duct and nonneoplastic
IV causes, such as stone disease, chronic pancreatitis, or abnormal
anatomy, as may be found in patients with chronic narcotic use
AJCC, American Joint Committee on Cancer; SMA, superior mesenteric artery;
TNM, tumor-node-metastasis. (Zylberberg et al, 2000). Therefore, when a pancreatic mass is
From Edge SB, Byrd DR, Compton CC, eds. AJCC Cancer Staging Manual. 7th ed. New York,
suspected but cannot be identified by standard imaging modali-
NY:Springer, 2010. ties, EUS should be considered if further clarification is clini-
cally warranted. A large retrospective series reported a high
specificity for EUS in the diagnosis of pancreatic cancer, with
a negative predictive value of 100%, when the results of EUS
STAGING OF PANCREATIC CANCER showed a normal pancreas (Klapman et al, 2005).
The method of staging generally practiced in the United States

Endoscopic Ultrasound Determination of
is that published by the American Joint Committee on Cancer Vascular Involvement
(see Chapter 62). It follows the tumor-node-metastasis (TNM) Some of the earlier descriptions of EUS in assessing pancreatic
staging system and is outlined in Table 16.3. cancer focused on its accuracy in identifying vascular involve-
ment. In this regard, Yasuda and colleagues (1993) reported an
Endoscopic Ultrasound Tumor-Node-Metastasis Staging accuracy of 79% for EUS compared with 41% for CT and 72%
Some of the first reports of EUS imaging of the pancreas for angiography. Rosch and colleagues (1992) showed that
described its ability to detect and diagnose pancreatic and peri- EUS correctly identified portal vein invasion in 95% of 40
ampullary tumors (Kaufman & Sivak, 1989; Yasuda et al, patients undergoing operation for ampullary and pancreatic
1988). However, subsequent studies evaluated the ability of cancer, compared with 85% for angiography, 73% for CT, and
EUS to stage tumors and assess resectability (Rosch et al, 1992; 55% for transcutaneous US. These authors also showed that
Tio & Tytgat, 1986). Although the general approach to the arterial encasement was less reliably detected by EUS.
preoperative evaluation of pancreatic cancer has focused on the Early studies were criticized because EUS findings were
TNM stage, the usefulness of such staging is questionable, compared with manual assessment by the surgeon at the
particularly because T stage does not necessarily correlate with time of the operation, rather than by pathologic confirmation
resectability. Initial studies favored EUS versus CT in terms of of vascular invasion. Aslanian and others (2005) carefully
staging accuracy. However, with the development of high- studied this issue by retrospectively comparing preoperative
resolution, multidetector technology, CT has equaled, if not EUS findings with the surgical pathology of specimens, includ-
surpassed, the accuracy of EUS and is once again the preferred ing tumors adherent to the vasculature. In this series, the
reported sensitivity, specificity, PPV, and NPV of EUS for extrahepatic cholangiocarcinoma (Mohamadnejad et al, 2011).
vascular invasion were 50%, 58%, 28%, and 82%, respectively, Tumor detection was superior with EUS compared with tripha-
but the results were better when only the portal, superior mes- sic CT scan and MRI. EUS-FNA added significant diagnostic
enteric, and splenic veins were considered. A meta-analysis of yield, particularly with distal bile duct tumors, with an overall
29 studies, including 1308 patients, evaluating the assessment sensitivity of 73%. In addition, EUS determined resectability
of vascular invasion by EUS, showed a pooled sensitivity of with a sensitivity and specificity of 53% and 97%, respectively.
73% and pooled specificity of 90.2% (Puli et al, 2007). Another Intraductal US can also provide staging information with
recent report compared EUS and MRI staging of resectability respect to depth of invasion, the maximal longitudinal extent
based on arterial invasion or encasement. The sensitivity, speci- of the lesion, and the presence of invasion into other organs
ficity, NPV, PPV, and accuracy of MRI were 78%, 96%, 92%, and major blood vessels, particularly portal vein invasion, with
88%, and 91%, respectively, compared with 33%, 100%, 81%, a reported accuracy of 60% to 80% (Inui & Miyoshi, 2005;
100%, and 82% for EUS, but these differences were not sta- Tamada et al, 1995, 1998). Standard echoendoscopes are less
tistically significant (p = 0.13) (Borbath et al, 2005). accurate for tumor staging but can detect regional lymphade-
nopathy and allow tissue acquisition via FNA.
Preoperative Reassessment after Neoadjuvant
Chemoradiotherapy ENDOSCOPIC ULTRASOUND-GUIDED THERAPY
Studies on the treatment of pancreatic cancer have recently
focused on neoadjuvant treatment with chemotherapy and Celiac Plexus Neurolysis
radiation therapy, followed by attempted radical resection (see In patients with pain as a result of pancreatic cancer, EUS-
Chapter 68). Several studies have investigated the ability of guided celiac plexus neurolysis has been shown to be a safe and
EUS to reassess patients who have completed neoadjuvant effective alternative to surgical or percutaneous approaches (see
therapy, with most suggesting that EUS is no better than CT Chapter 62). The celiac axis is easily identified using a linear-
for determining resectability. In one report, EUS correctly array echoendoscope positioned in the stomach (Fig. 16.12),
assessed residual tumor stage in 40% and node status in 90%, and it provides a quick method of palliating cancer-related
but it incorrectly suggested venous invasion in 43% (Bettini pain in patients undergoing staging and diagnosis of pancreas
et al, 2005). These results suggest that inflammatory changes cancer. Two recent meta-analyses demonstrated improved sus-
in the tumor bed, pancreas, and lymph nodes alter the anatomy tained pain relief in 72% to 80% of patients at a follow-up
and blur the distinction between normal tissue planes and range of 1 to 6 months (Kaufman et al, 2010; Puli et al,
between tumor and normal tissue. 2009). In addition, decreased analgesic requirements and
fewer opioid-induced side effects can be expected after EUS-
DIAGNOSIS AND STAGING OF guided neurolysis. Recent studies have suggested that bilateral
CHOLANGIOCARCINOMA ethanol injection to the right and left of the celiac axis is
more effective than unilateral or central injection (Puli et al,
Primary bile duct cancers (see Chapters 51 and 59) usually 2009; Sahai et al, 2009), and no reports of neurologic compli-
present with painless jaundice. Although transcutaneous US cations or pancreatitis have been reported with this technique
and CT scanning reliably show biliary dilatation, they are less (O’Toole & Schmulewitz, 2009; Puli et al 2009). A random-
accurate for delineating tumors, especially if tumors are smaller ized, double-blinded, controlled trial was performed of early
than 2 cm. ERCP is the dominant invasive modality used to EUS-guided neurolysis to prevent pain progression in patients
evaluate extrahepatic bile duct strictures. Diagnostic specimens
can be obtained by rubbing a cytology brush against the stric-
ture, but the sensitivity is low, with a yield of only 40% to 50%
(Ponchon et al, 1995; Victor et al, 2012; Wakatsuki et al,
2005). EUS has now emerged as a useful diagnostic and staging
technique in this regard through the use of linear and radial
echoendoscopes, as well as high-frequency radial miniprobes C
(passed into the bile duct over a guidewire through the working
channel of an ERCP scope), to perform intraductal US.
For detecting benign causes of biliary obstruction, such as SMA
AO
CBD stones, EUS, with a sensitivity of 89% to 94% and specific-
ity of 94% to 100%, is more accurate than transcutaneous US,
CT, magnetic resonance cholangiopancreatography (MRCP),
ERCP, or intraoperative cholangiography (Garrow et al, 2007;
Tse et al, 2008) (see Chapters 15, 18, 19, and 23). If a biliary
stricture is visualized, malignancy is suggested by the presence
of an irregular, thickened (>3 mm) bile duct wall (see Fig. 16.5).
Hypoechoic infiltration invading through the biliary wall layers
or an adjacent pancreatic mass can also be seen. The accuracy
of EUS-FNA in the diagnosis of bile duct strictures has a
reported sensitivity ranging from 43% to 86% (DeWitt et al,
2006; Eloubeidi et al, 2004a; Rosch et al, 2004) (see Table FIGURE 16.12. Endoscopic ultrasound appearance of the longitudinal
16.1). More recently, results of a large single-center experience course of the aorta (AO) with celiac (C) and superior mesenteric artery
suggest that EUS can be very useful in the assessment of (SMA) origins.
with inoperable pancreatic cancer (Wyse et al, 2011). Patients et al, 2009; Gress et al, 2002b). This precludes the need for
who received EUS-guided neurolysis had significantly less pain intraoperative US and may allow the surgeon to perform a more
at 3 months as well as a tendency toward less narcotic use. limited pancreatic resection. A role for EUS-guided intratu-
moral placement of gold fiducials in localizing pancreatic
Drainage of Pseudocysts and Peripancreatic Collections tumors for targeted radiation therapy has also been demon-
The role for EUS as a minimally invasive and effective tech- strated (Pishvaian et al, 2006; Sanders et al, 2010; Yan & Van
nique for guiding drainage of peripancreatic fluid collections Dam, 2008). This modality of radiotherapy minimizes toxicity
after acute pancreatitis (Lopes et al, 2007; Varadarajulu et al, to the surrounding tissue by allowing delivery of highly confor-
2007) (see Chapters 27, 30, and 56) or after pancreatic resec- mal radiation treatments.
tion has been established. A fine needle is used to puncture the
fluid collection, and a guidewire is used for transluminal stent- NOVEL THERAPEUTICS
ing. Using real-time imaging and Doppler flow, intervening
organs and vascular structures can be avoided. Thus there are EUS-guided ethanol ablation of cysts has been previously
fewer complications, such as bleeding and perforation, com- reported (DeWitt et al, 2009; Gan et al, 2005). Patients with
pared with the percutaneous approach. Pseudocysts often asymptomatic, unilocular pancreatic cysts were treated by injec-
contain a large amount of necrotic debris that typically requires tion and lavage of the cyst with ethanol through an FNA needle,
endoscopic debridement and is not effectively drained by stent with complete cyst resolution documented in 30% to 35% of
placement alone. This technique has also been applied to the patients. More recently, ethanol has been combined with pacli-
drainage of peripancreatic fluid collections after distal pancre- taxel (Oh et al, 2011), with complete resolution achieved in
atectomy (Tilara et al, 2014; Varadarajulu et al, 2009) (see 62% to 78% of patients. Radiofrequency ablation of pancreatic
Chapter 66). EUS-guided drainage of pancreatic fluid collec- cysts has also been described recently in a small series of patients
tions compares favorably with percutaneous drainage tech- (Pai et al, 2013). Despite these advances, concerns remain
niques, with similarly high success rates (Kwon et al, 2013) about residual epithelium, which could remain after ablation.
(see Chapter 27). EUS-guided drainage eliminates the need In addition, evidence that ablation reduces the risk of malig-
for an external drainage apparatus as well as the risk of cysto- nancy, the need for resection, or continued surveillance is
cutaneous fistula compared with percutaneous drainage tech- lacking.
niques. Recent improvements in stent design have made In patients with advanced abdominal malignancies, simul-
drainage of pancreatic fluid collections more simple and effec- taneous duodenal and biliary obstruction can occur. Tradition-
tive (Shah et al, 2015). Ultimately, a collaborative multidisci- ally, these patients have required percutaneous or surgical
plinary approach taking into consideration the size and location biliary drainage because transpapillary decompression via
of the fluid collection will determine the best approach for ERCP is often not possible. EUS-guided transduodenal or
drainage. transgastric biliary drainage is a novel approach that allows
internal drainage (Will et al, 2007; Yamao et al, 2008; Park
Tumor Localization et al, 2009). Similar to pseudocyst drainage, the dilated CBD
Preoperative localization of small pancreatic tumors by fine or left hepatic duct is localized with EUS and punctured with
needle tattooing can be performed safely using EUS (Farrell an FNA needle. A cholangiogram is performed (Fig. 16.13A),
A B
FIGURE 16.13. A, Cholangiogram obtained by endoscopic ultrasound (EUS)-guided puncture of the common bile duct by fine-needle aspiration
in preparation for EUS-guided biliary drainage. B, EUS image after deployment of self-expanding metal stent.
followed by guidewire insertion, dilation of the tract, and biliary tree. It provides detailed images of the pancreas and bile
deployment of a self-expandable metal stent across the choledo- ducts and complements the findings of noninvasive radio-
choduodenostomy or hepaticogastrostomy tract (Fig. 16.13B). graphic imaging. For patients seen initially with suspected pan-
A recent study directly compared ERCP with EUS-guided creatic or bile duct cancer, multidetector CT scanning assists
biliary drainage (Dhir et al, 2015). Short-term outcomes have in identifying those with obvious masses, metastatic disease,
been shown to be comparable to ERCP, with a similar rate of and vascular involvement. The usefulness of EUS in this setting
adverse events and lower risk of pancreatitis. EUS-guided is limited, except to provide tissue diagnosis through the use of
biliary drainage was also recently compared with percutaneous EUS-guided fine-needle aspiration. For patients with small
drainage in patients who failed ERCP (Khashab et al, 2015) tumors, ampullary lesions, or equivocal findings detected by
(see Chapter 29). EUS-guided biliary drainage was equally CT or MRI, the high sensitivity of EUS helps provide a diag-
effective, with fewer adverse events, reduced need for reinter- nosis, with a high accuracy for tumor staging and determining
ventions, and reduced cost. resectability. Because EUS has poor sensitivity for identifying
EUS-guided intratumoral therapy of solid masses is being distant metastatic disease, CT scanning and laparoscopy will
actively investigated. Several therapeutic agents have been eval- remain important tools in the evaluation of patients with pan-
uated, including activated lymphocyte cultures, viral vectors, creatic and bile duct cancer who are being considered for
oncolytic viruses, and radioactive seeds (Chang et al, 2000; surgery. EUS-FNA is a safe and valuable tool in the evaluation
Hecht et al, 2003; Hecht et al, 2012; Jin et al, 2008; Senzer of patients with a suspected solid or cystic mass of the pancreas
et al, 2004). Early studies have established safety and demon- or extrahepatic bile ducts. It has a higher diagnostic sensitivity
strated partial response or disease stabilization in most patients. and specificity than noninvasive imaging and allows tissue sam-
The feasibility and safety of performing radiofrequency ablation pling as part of the same procedure. As molecular testing
and cryotherapy of the pancreas by using a EUS platform has improves, the utility of EUS-FNA as a diagnostic modality is
been established in animal models (Carrara et al, 2008; Gold- likely to grow. Similarly, EUS-FNA has evolved from a diag-
berg et al, 1999). nostic to a therapeutic procedure in the management of biliary
and pancreatic disease.
SUMMARY
EUS is an essential tool in the preoperative evaluation of
patients with abnormalities that involve the pancreas and the
Chapter 16 Endoscopic ultrasound of the biliary tract and pancreas 284.e1
REFERENCES Gress F, et al: Endoscopic ultrasonography-guided fine-needle aspira-

tion biopsy of suspected pancreatic cancer, Ann Intern Med
Adler DG, et al: ASGE guideline: complications of EUS, Gastrointest 134(6):459–464, 2001.
Endosc 61(1):8–12, 2005. Gress F, et al: EUS-guided fine-needle aspiration of the pancreas:
Affi A, et al: Acute extraluminal hemorrhage associated with EUS- evaluation of pancreatitis as a complication, Gastrointest Endosc
guided fine needle aspiration: frequency and clinical significance, 56(6):864–867, 2002a.
Gastrointest Endosc 53(2):221–225, 2001. Gress FG, et al: Preoperative localization of a neuroendocrine tumor
Ahmad NA, et al: Interobserver agreement among endosonographers of the pancreas with EUS-guided fine needle tattooing, Gastrointest
for the diagnosis of neoplastic versus non-neoplastic pancreatic cystic Endosc 55(4):594–597, 2002b.
lesions, Gastrointest Endosc 58(1):59–64, 2003. Hartwig W, et al: Preoperative tissue diagnosis for tumours of the
Aslanian H, et al: EUS diagnosis of vascular invasion in pancreatic pancreas, Br J Surg 96(1):5–20, 2009.
cancer: surgical and histologic correlates, Am J Gastroenterol 100(6): Hecht JR, et al: A phase I/II trial of intratumoral endoscopic ultrasound
1381–1385, 2005. injection of ONYX-015 with intravenous gemcitabine in unresect-
Bettini N, et al: Preoperative locoregional re-evaluation by endoscopic able pancreatic carcinoma, Clin Cancer Res 9(2):555–561, 2003.
ultrasound in pancreatic ductal adenocarcinoma after neoadjuvant Hecht JR, et al: EUS or percutaneously guided intratumoral TNFerade
chemoradiation, Gastroenterol Clin Biol 29(6–7):659–663, 2005. biologic with 5-fluorouracil and radiotherapy for first-line treatment
Borbath I, et al: Preoperative assessment of pancreatic tumors using of locally advanced pancreatic cancer: a phase I/II study, Gastrointest
magnetic resonance imaging, endoscopic ultrasonography, positron Endosc 75(2):332–338, 2012.
emission tomography and laparoscopy, Pancreatology 5(6):553–561, Iglesias-Garcia J, et al: Feasibility and yield of a new EUS histology
2005. needle: results from a multicenter, pooled, cohort study, Gastrointest
Brugge WR, et al: Diagnosis of pancreatic cystic neoplasms: a report Endosc 73(6):1189–1196, 2011.
of the cooperative pancreatic cyst study, Gastroenterology 126(5):1330– Inui K, Miyoshi H: Cholangiocarcinoma and intraductal sonography,
1336, 2004. Gastrointest Endosc Clin N Am 15(1):143–155, 2005.
Brugge WR, et al: Diagnosis of pancreatic neoplasia with EUS and Janssen J, et al: Frequency of bacteremia after linear EUS of the upper
FNA: a report of accuracy, Gastrointest Endosc 71(1):91–98, 2010. GI tract with and without FNA, Gastrointest Endosc 59(3):339–344,
Carrara S, et al: Endoscopic ultrasound-guided application of a new 2004.
hybrid cryotherm probe in porcine pancreas: a preliminary study, Jin Z, et al: Endoscopic ultrasonography-guided interstitial implanta-
Endoscopy 40(4):321–326, 2008. tion of iodine-125 seeds combined with chemotherapy in the treat-
Centeno BA, et al: Cytologic diagnosis of pancreatic cystic lesions: a ment of unresectable pancreatic carcinoma: a prospective pilot study,
prospective study of 28 percutaneous aspirates, Acta Cytol 41(4):972– Endoscopy 40(4):314–320, 2008.
980, 1997. Kaufman AR, Sivak MV Jr: Endoscopic ultrasonography in the dif-
Chang KJ, et al: Phase I clinical trial of allogeneic mixed lymphocyte ferential diagnosis of pancreatic disease, Gastrointest Endosc 35(3):
culture (cytoimplant) delivered by endoscopic ultrasound-guided 214–219, 1989.
fine-needle injection in patients with advanced pancreatic carci- Kaufman M, et al: Efficacy of endoscopic ultrasound-guided celiac
noma, Cancer 88(6):1325–1335, 2000. plexus block and celiac plexus neurolysis for managing abdominal
Chen J, et al: Diagnostic accuracy of endoscopic ultrasound-guided pain associated with chronic pancreatitis and pancreatic cancer,
fine-needle aspiration for solid pancreatic lesion: a systematic review, J Clin Gastroenterol 44(2):127–134, 2010.
J Cancer Res Clin Oncol 138(9):1433–1441, 2012. Khalid A, et al: Pancreatic cyst fluid DNA analysis in evaluating pan-
Cizginer S, et al: Cyst fluid carcinoembryonic antigen is an accurate creatic cysts: a report of the PANDA study, Gastrointest Endosc
diagnostic marker of pancreatic mucinous cysts, Pancreas 40(7):1024– 69(6):1095–1102, 2009.
1028, 2011. Khashab MA, et al: A comparative evaluation of eus-guided biliary
DeWitt J, et al: Comparison of endoscopic ultrasonography and mul- drainage and percutaneous drainage in patients with distal malignant
tidetector computed tomography for detecting and staging pancre- biliary obstruction and failed ERCP, Dig Dis Sci 60(2):557–565,
atic cancer, Ann Intern Med 141(10):753–763, 2004. 2015.
DeWitt J, et al: EUS-guided FNA of proximal biliary strictures after Klapman JB, et al: Negative predictive value of endoscopic ultrasound
negative ERCP brush cytology results, Gastrointest Endosc 64(3):325– in a large series of patients with a clinical suspicion of pancreatic
333, 2006. cancer, Am J Gastroenterol 100(12):2658–2661, 2005.
DeWitt J, et al: EUS-guided ethanol versus saline solution lavage for Kwon YM, et al: Management of peripancreatic fluid collections fol-
pancreatic cysts: a randomized, double-blind study, Gastrointest lowing partial pancreatectomy: a comparison of percutaneous versus
Endosc 70(4):710–723, 2009. EUS-guided drainage, Surg Endosc 27(7):2422–2427, 2013.
Dhir V, et al: Multicenter comparative evaluation of endoscopic place- Learn PA, et al: Pitfalls in avoiding operation for autoimmune pancre-
ment of expandable metal stents for malignant distal common bile atitis, Surgery 150(5):968–974, 2011.
duct obstruction by ERCP or EUS-guided approach, Gastrointest LeBlanc JK, et al: Optimal number of EUS-guided fine needle passes
Endosc 81(4):913–923, 2015. needed to obtain a correct diagnosis, Gastrointest Endosc 59(4):475–
Eloubeidi MA, et al: Endoscopic ultrasound-guided fine needle aspira- 481, 2004.
tion biopsy of suspected cholangiocarcinoma, Clin Gastroenterol Lee LS, et al: EUS-guided fine needle aspiration of pancreatic cysts: a
Hepatol 2(3):209–213, 2004a. retrospective analysis of complications and their predictors, Clin
Eloubeidi MA, et al: Acute pancreatitis after EUS-guided FNA of solid Gastroenterol Hepatol 3(3):231–236, 2005.
pancreatic masses: a pooled analysis from EUS centers in the United Levy MJ, et al: Prospective risk assessment of bacteremia and other
States, Gastrointest Endosc 60(3):385–389, 2004b. infectious complications in patients undergoing EUS-guided FNA,
Farrell JJ, et al: EUS-guided fine-needle tattooing for preoperative Gastrointest Endosc 57(6):672–678, 2003.
localization of early pancreatic adenocarcinoma, Gastrointest Endosc Lopes CV, et al: Endoscopic-ultrasound-guided endoscopic transmural
69(1):176–177, 2009. drainage of pancreatic pseudocysts and abscesses, Scand J Gastroen-
Fornarni F, et al: Complications of ultrasonically guided fine-needle terol 42(4):524–529, 2007.
abdominal biopsy. Results of a multicenter Italian study and review Maker AV, et al: Cyst fluid interleukin-1beta (IL1beta) levels predict
of the literature. The Cooperative Italian Study Group, Scand J the risk of carcinoma in intraductal papillary mucinous neoplasms
Gastroenterol 24(8):949–955, 1989. of the pancreas, Clin Cancer Res 17(6):1502–1508, 2011.
Gan SI, et al: Ethanol lavage of pancreatic cystic lesions: initial pilot Mohamadnejad M, et al: Role of EUS for preoperative evaluation of
study, Gastrointest Endosc 61(6):746–752, 2005. cholangiocarcinoma: a large single-center experience, Gastrointest
Garrow D, et al: Endoscopic ultrasound: a meta-analysis of test per- Endosc 73(1):71–78, 2011.
formance in suspected biliary obstruction, Clin Gastroenterol Hepatol Muller MF, et al: Pancreatic tumors: evaluation with endoscopic US,
5(5):616–623, 2007. CT, and MR imaging, Radiology 190(3):745–751, 1994.
Goldberg SN, et al: EUS-guided radiofrequency ablation in the pan- Ogura T, et al: Clinical impact of K-ras mutation analysis in EUS-
creas: results in a porcine model, Gastrointest Endosc 50(3):392–401, guided FNA specimens from pancreatic masses, Gastrointest Endosc
1999. 75(4):769–774, 2012.
Oh HC, et al: Endoscopic ultrasonography guided ethanol lavage with Tamada K, et al: Preoperative staging of extrahepatic bile duct cancer
paclitaxel injection treats patients with pancreatic cysts, Gastroenter- with intraductal ultrasonography, Am J Gastroenterol 90(2):239–246,
ology 140(1):172–179, 2011. 1995.
O’Toole TM, Schmulewitz N: Complication rates of EUS-guided Tamada K, et al: Characterization of biliary strictures using intraductal
celiac plexus blockade and neurolysis: results of a large case series, ultrasonography: comparison with percutaneous cholangioscopic
Endoscopy 41(7):593–597, 2009. biopsy, Gastrointest Endosc 47(5):341–349, 1998.
O’Toole DL, et al: Assessment of complications of EUS-guided fine- Tanaka M, et al: International consensus guidelines 2012 for the man-
needle aspiration, Gastrointest Endosc 53(4):470–474, 2001. agement of IPMN and MCN of the pancreas, Pancreatology 12(3):
Pai M, et al: Endoscopic ultrasound guided radiofrequency ablation, 183–197, 2012.
for pancreatic cystic neoplasms and neuroendocrine tumors, World Tilara A, et al: Endoscopic ultrasound-guided transmural drainage of
J Gastrointest Surg 7(4):52–59, 2015. postoperative pancreatic collections, J Am Coll Surg 218(1):33–40,
Park do H, et al: EUS-guided biliary drainage with one-step placement 2014.
of a fully covered metal stent for malignant biliary obstruction: a Tio TL, Tytgat GN: Endoscopic ultrasonography in staging local
prospective feasibility study, Am J Gastroenterol 104(9):2168–2174, resectability of pancreatic and periampullary malignancy, Scand J
2009. Gastroenterol Suppl 123:135–142, 1986.
Pishvaian AC, et al: EUS-guided fiducial placement for CyberKnife Tse F, et al: EUS: a meta-analysis of test performance in suspected
radiotherapy of mediastinal and abdominal malignancies, Gastroin- choledocholithiasis, Gastrointest Endosc 67(2):235–244, 2008.
test Endosc 64(3):412–417, 2006. Turner BG, et al: Diagnosis of pancreatic neoplasia with EUS and
Ponchon T, et al: Value of endobiliary brush cytology and biopsies for FNA: a report of accuracy, Gastrointest Endosc 71(1):91–98, 2010.
the diagnosis of malignant bile duct stenosis: results of a prospective Varadarajulu S, et al: Role of EUS in drainage of peripancreatic fluid
study, Gastrointest Endosc 42(6):565–572, 1995. collections not amenable for endoscopic transmural drainage, Gas-
Puli SR, et al: Diagnostic accuracy of EUS for vascular invasion in trointest Endosc 66(6):1107–1119, 2007.
pancreatic and periampullary cancers: a meta-analysis and system- Varadarajulu S, et al: EUS for the management of peripancreatic fluid
atic review, Gastrointest Endosc 65(6):788–797, 2007. collections after distal pancreatectomy, Gastrointest Endosc 70(6):
Puli SR, et al: EUS-guided celiac plexus neurolysis for pain due to 1260–1265, 2009.
chronic pancreatitis or pancreatic cancer pain: a meta-analysis and Victor DW, et al: Current endoscopic approach to indeterminate biliary
systematic review, Dig Dis Sci 54(11):2330–2337, 2009. strictures, World J Gastroenterol 18(43):6197–6205, 2012.
Puli SR, et al: How good is endoscopic ultrasound-guided fine-needle Wakatsuki T, et al: Comparative study of diagnostic value of cytologic
aspiration in diagnosing the correct etiology for a solid pancreatic sampling by endoscopic ultrasonography-guided fine-needle aspira-
mass? A meta-analysis and systematic review, Pancreas 42(1):20–26, tion and that by endoscopic retrograde pancreatography for the
2013. management of pancreatic mass without biliary stricture, J Gastro-
Reddy RP, et al: Pancreatic mucinous cystic neoplasm defined by enterol Hepatol 20(11):1707–1711, 2005.
ovarian stroma: demographics, clinical features, and prevalence of Wiersema MJ, et al: Endosonography-guided fine-needle aspiration
cancer, Clin Gastroenterol Hepatol 2(11):1026–1031, 2004. biopsy: diagnostic accuracy and complication assessment, Gastroen-
Rosch T, et al: Staging of pancreatic and ampullary carcinoma by terology 112(4):1087–1095, 1997.
endoscopic ultrasonography. Comparison with conventional sonog- Will U, et al: Treatment of biliary obstruction in selected patients by
raphy, computed tomography, and angiography, Gastroenterology endoscopic ultrasonography (EUS)-guided transluminal biliary
102(1):188–199, 1992. drainage, Endoscopy 39(4):292–295, 2007.
Rosch T, et al: ERCP or EUS for tissue diagnosis of biliary strictures? Williams DB, et al: Endoscopic ultrasound guided fine needle aspira-
A prospective comparative study, Gastrointest Endosc 60(3):390–396, tion biopsy: a large single centre experience, Gut 44(5):720–726,
2004. 1999.
Sahai AV, et al: Central vs. bilateral endoscopic ultrasound-guided Wyse JM, et al: Randomized, double-blind, controlled trial of early
celiac plexus block or neurolysis: a comparative study of short-term endoscopic ultrasound-guided celiac plexus neurolysis to prevent
effectiveness, Am J Gastroenterol 104(2):326–329, 2009. pain progression in patients with newly diagnosed, painful, inoper-
Sanders MK, et al: EUS-guided fiducial placement for stereotactic able pancreatic cancer, J Clin Oncol 29(26):3541–3546, 2011.
body radiotherapy in locally advanced and recurrent pancreatic Yamao K, et al: EUS-guided choledochoduodenostomy for palliative
cancer, Gastrointest Endosc 71(7):1178–1184, 2010. biliary drainage in patients with malignant biliary obstruction: results
Sedlack R, et al: Utility of EUS in the evaluation of cystic pancreatic of long-term follow-up, Endoscopy 40(4):340–342, 2008.
lesions, Gastrointest Endosc 56(4):543–547, 2002. Yan BM, Van Dam J: Endoscopic ultrasound-guided intratumoural
Senzer N, et al: TNFerade biologic, an adenovector with a radiation- therapy for pancreatic cancer, Can J Gastroenterol 22(4):405–410,
inducible promoter, carrying the human tumor necrosis factor alpha 2008.
gene: a phase I study in patients with solid tumors, J Clin Oncol Yasuda K, et al: The diagnosis of pancreatic cancer by endoscopic
22(4):592–601, 2004. ultrasonography, Gastrointest Endosc 34(1):1–8, 1988.
Shah RJ, et al: Safety and efficacy of endoscopic ultrasound-guided Yasuda K, et al: Staging of pancreatic carcinoma by endoscopic ultra-
drainage of pancreatic fluid collections with lumen-apposing covered sonography, Endoscopy 25(2):151–155, 1993.
self-expanding metal stents, Clin Gastroenterol Hepatol 13(4):747– Yusuf TE, et al: Retrospective analysis of the utility of endoscopic
752, 2015. ultrasound-guided fine-needle aspiration (EUS-FNA) in pancreatic
Soriano A, et al: Preoperative staging and tumor resectability masses, using a 22-gauge or 25-gauge needle system: a multicenter
assessment of pancreatic cancer: prospective study comparing endo- experience, Endoscopy 41(5):445–448, 2009.
scopic ultrasonography, helical computed tomography, magnetic Zylberberg H, et al: Dilated bile duct in patients receiving narcotic
resonance imaging, and angiography, Am J Gastroenterol 99(3):492– substitution: an early report, J Clin Gastroenterol 31(2):159–161,
501, 2004. 2000.
CHAPTER 17
Role of nuclear medicine in diagnosis and
management of hepatopancreatobiliary disease
Simone Krebs and Mark Dunphy
Nuclear medicine specialists prescribe radioactive pharmaceu- other guidelines, especially when these make judgments on the
ticals, radiopharmaceuticals, for diagnostic imaging and internal diagnostic accuracy or clinical impact of nuclear medicine;
radiotherapy of a variety of diseases. This chapter discusses sometimes guidelines cite outdated nuclear medicine research
clinical applications of diagnostic nuclear medicine imaging for from decades past. Such guidelines might be designed in recog-
the care of patients with hepatic, pancreatic, and biliary (hepa- nition that nuclear medicine clinical practice varies worldwide,
topancreatobiliary, HPB) diseases. Radioembolization of liver as reflected in the often widely varying diagnostic sensitivities
tumors with radiolabeled microspheres is discussed separately and specificities reported by different medical centers perform-
(see Chapter 96B). Other standard forms of nuclear medicine ing a particular NMI procedure. Although the same term is used
therapy, involving radiopharmaceuticals administered in thera- (e.g., fluorodeoxyglucose [FDG] PET/computed tomography
peutic amounts of radioactivity, are not specific to HPB diseases, [CT]), when extremely different diagnostic performance out-
but a few are mentioned when relevant (e.g., somatostatin comes are reported between centers, the differences in diagnostic
receptor–targeted radionuclide therapy), with citations for performance are likely attributable in significant part to differ-
further information. ences in hardware (e.g., scanners) or methodology (e.g.,
In general, the role of diagnostic nuclear medicine imaging administered tracer doses, software-based data-processing
(NMI), or scintigraphy, including positron emission tomography algorithms, adjunctive medications) that may or may not be
(PET), is to provide HPB clinicians with a noninvasive method considered “state of the art.” This chapter discusses the essential
to aid in detecting and localizing certain types of HPB disease criteria defining state-of-the-art for the major HPB-related
and evaluating HPB organ function, as well as evaluating the diagnostic and therapeutic nuclear medicine procedures, as well
effects of treatment. In general, NMI can be considered a clini- as the diagnostic accuracy and clinical relevance of state-of-the-
cal assay of cellular biology in the tissues of patients; the in vivo art nuclear medicine in the care of the patient with HPB disease.
tissue accumulation, or uptake, of most radiopharmaceuticals State-of-the-art clinical nuclear medicine is based on a rela-
depends on the biomolecular composition of living cells, in tively expensive infrastructure. A notable example is PET
body tissues, as well as tissue perfusion. imaging with the standard radiopharmaceutical fluorine 18
The diagnostic accuracy of scintigraphy varies according to (F-18, 18F) FDG used for imaging HPB tumors (among other
the specific scintigraphic study (including the specific radio- applications). To perform FDG PET, a highly expensive cyclo-
pharmaceutical used and how it is assayed) and the specific tron facility, costing multiple millions of US dollars (2015) for
disease or condition being studied. The HPB specialist must construction, besides annual costs for staff, supplies, and
integrate diagnostic data from any scintigraphic study of a maintenance, is required for production of the radioactive
particular patient with signs, symptoms, and data from other isotope 18F, used in synthesis of FDG. After FDG is adminis-
relevant assays for optimal diagnostic accuracy and therapeutic tered to a patient, the patient is then, in state-of-the-art practice,
decision making. NMI has a major positive impact on patient scanned for disease using a PET camera combined with an
care, improving therapeutic strategy. x-ray computed tomography (CT) camera, a hybrid PET-CT
This chapter discusses the published clinical evidence camera system that is also highly expensive to purchase and
regarding the impact of nuclear medicine in HPB diseases and maintain (again, millions of US dollars). Even in the wealthiest
focuses on state-of-the-art nuclear medicine and, as such, nations with comparable well-developed nuclear medicine
concentrates predominantly on published medical literature infrastructures, the availability of certain state-of-the-art nuclear
from the past 10 years. In our experience, most clinical nuclear medicine diagnostic and therapeutic procedures often varies
medicine research publications before 10 years ago often significantly for the “average” citizen, depending on socioeco-
employ methodology and technology that is no longer reflective nomic factors (an important discussion but beyond the scope
of current state-of-the-art clinical practice in nuclear medicine. of this chapter). Other diagnostic and therapeutic nuclear
The state-of-the-art in nuclear medicine, in its diagnostic and medicine procedures important to management of HPB dis-
therapeutic procedures, has improved rapidly in the past 10 eases, however, are relatively low cost and are available to
years and continues to evolve and innovate, including major patients in most nations. If unsure of the availability of any
improvements in commercially available nuclear imaging particular nuclear medicine procedure discussed here, the
camera systems (particularly the advent of hybrid “fusion reader can contact his or her local radiology department or
imaging” camera systems), image data processing, new types of search the Internet to find the nearest medical center offering
instrumentation, and clinical introduction of new radiopharma- the procedure of interest.
ceuticals, both for diagnostic imaging and nuclear therapy. After a concise introduction to the pharmacology and tech-
Therefore, we strongly advise the reader to note the dates of nology of diagnostic imaging and therapy in radiopharmaceu-
nuclear medicine references cited in HPB bibliographies and ticals and the general role of nuclear medicine in HPB diseases,
285
we discuss current state-of-the-art nuclear medicine procedures elemental radioisotope atom typically is incorporated within a
for specific HPB clinical indications. For diagnostic imaging molecule by chemical bonding. The molecule is said to be
procedures, discussion focuses on how well a particular clinical radiolabeled.
NMI study performs for a specific HPB indication, in terms of As with any pharmaceutical, each type of radiopharmaceuti-
its diagnostic accuracy (sensitivity, specificity) and potential cal has in vivo pharmacokinetic (PK) properties specific to and
pitfalls, including necessary patient preparation, when appli- determined by its molecular structure and associated physico-
cable. For the few nuclear medicine therapeutic procedures chemical properties. PK properties include the radiopharma-
applicable specifically to HPB diseases, this chapter focuses on ceutical’s distribution in tissues throughout the body
summarizing medical literature regarding clinical therapeutic (biodistribution), metabolism, and bodily elimination (by hepa-
efficacy. Each section discusses specific standard nuclear medi- tobiliary and urinary excretion for all relevant radiopharmaceu-
cine radiopharmaceuticals and procedures that have clinical ticals). The in vivo PK properties are also determined, to some
applications for particular HPB diseases, as used in state-of- degree, by the physicochemical properties of excipients (vehi-
the-art clinical practice. The chapter also includes a concise cles) in the radiopharmaceutical formulation (e.g., formulation
look at select new, currently investigational radiopharmaceuti- of an orally administered radiopharmaceutical compound may
cals relevant to HPB disease. affect its bioavailability and biodistribution), as well as by the
Radiopharmaceuticals can be placed into three major catego- route of administration (e.g., peripheral intravenous injection,
ries of applications in HPB disease: detection and evaluation of hepatic arterial catheter infusion).
cancerous HPB tumors, treatment of HPB cancers, and evalu- The mass-amount of radioactive molecules (per se) in any
ation of HPB organ function (and indirectly for detection of prescribed radiopharmaceutical formulation is only a trace
disease entities causing HPB organ dysfunction). The complete amount, typically in the picogram (pg) range. This tiny mass-
pharmacopoeia used in current clinical practice, approved dose of radioactive molecule is incapable of exerting detectable
by the US Food and Drug Administration (FDA), includes pharmacologic effects on body tissues in vivo, but the typical
numerous different radiopharmaceuticals, each with diagnostic picogram amounts of radioactive molecules emit radioactivity
and/or therapeutic applications in one or more of the major sufficient for diagnostic imaging and therapeutic applications.
medical specialties, most not specifically relevant to HPB dis- With exceptions, the nonradioactive constituents of radiophar-
eases. It is beyond the scope of this chapter to discuss all general maceutical compounds typically used only for clinical diagnostic
nuclear medicine tests that an HPB specialist or their patients imaging are present in somewhat higher mass-amounts but are
may encounter clinically. Only those radiopharmaceuticals of still scant, typically less than 100 micrograms (µg), and allergic
major importance to the diagnosis and treatment of HPB dis- reactions, other side effects, or pharmacodynamic effects are
eases are discussed here. Similarly, we include no comprehensive rarely reported. Nuclear medicine specialists may prescribe the
review of the novel tracers described in medical literature; radiopharmaceutical compound to be administered in conjunc-
rather, we focus on select radiopharmaceutical “front-runners” tion with a relatively high and biologically effective mass-amount
that HPB specialists may see at their own medical center in the (e.g., milligrams or grams) of a nonradioactive, or unlabeled,
next 10 years. version of the same compound or a related compound, with
therapeutic intent (relevant compounds are discussed later).
RADIOPHARMACEUTICALS The term “radiotracer,” “tracer dose,” or “radiotracer dose”
commonly refers to the use of a radiolabeled molecule in tiny
Nuclear medicine specialists prescribe radiopharmaceuticals for (trace) amounts to study molecular biology. The trace amount
diagnostic imaging and internal radiotherapy of a variety of of radioactivity and the trace mass of the administered radio-
diseases, with this chapter exclusively focused on HPB applica- tracer are unable to affect (and therefore unable to interfere
tions. A radiopharmaceutical is a radioactive compound con- with measurements of) the biomolecular system or target being
taining a radionuclide, also (loosely) referred to as a “radioisotope” assayed. Following this common convention, in this chapter we
(radioactive isotope). A radioisotope is an energetically unstable use radiotracer to refer to radiopharmaceutical administered
atom that will achieve a stable or more stable, lower-energy state specifically as a trace, biologically ineffective amount of radio-
(transitioning from a parent to a daughter state) by releasing activity useful solely for diagnostic imaging. We use therapeutic
(radiating) energy (radiation), in some form (e.g., emitting a radiopharmaceutical to refer to administration of a relatively high
gamma ray, positron particle, or beta particle, as discussed amount of radioactivity with the intent of inducing therapeutic
later). The release of energy by the (parent) radioisotope atom radiobiologic effects in vivo, as discussed later. The radioactivity
may be called a physical decay, disintegration, or transition. The emitted by a therapeutic radiopharmaceutical may be useful for
energy decay makes the elemental atom either become a differ- diagnostic imaging as well as radiotherapy. The clinical diag-
ent isotope of the same element (e.g., the radioisotope techne- nostic imaging of therapeutic radiopharmaceuticals’ biodistri-
tium 99m decays to the stable isotope technetium 99) or become bution in vivo in patients (e.g., to detect or quantify tumor
a different element, by transmutation (e.g., the radioisotope 18F targeting by the radiopharmaceutical) is known as theranostics.
decays to become a stable form of the element oxygen, Most radiopharmaceuticals in current clinical practice rel-
i.e., oxygen 18). Other forms of nuclear decay are possible evant to HPB diseases are administered intravenously. One
(e.g., transitions from a higher-energy unstable radioisotope to notable exception is hepatic arterial infusion of radiotracer.
a lower-energy, but still unstable, daughter radioisotope). A
radiopharmaceutical is administered in a trace amount (with no DIAGNOSTIC IMAGING IN NUCLEAR MEDICINE
detectable radiobiologic effects) or therapeutic amount, for use
as a diagnostic imaging agent or therapeutic agent. A radiophar- The diagnostic images of the patient analyzed by the nuclear
maceutical also contains other active and inactive ingredients in medicine specialist are the result of technology, pharmacology,
the compound formulation. In the radiopharmaceutical, the and biology. This section focuses on what the clinical imaging
Chapter 17 Role of nuclear medicine in diagnosis and management of hepatopancreatobiliary disease 287
specialist and referring clinician need to know to interpret decay with its characteristic radioactive emissions. The emitted
nuclear scan images properly and to avoid misinterpretations energy detected by a modern nuclear scanner (e.g., PET
that might negatively impact patient care, as well as diagnostic scanner, gamma camera) is always in the form of photons. The
pitfalls by properly preparing patients. detection of photons by the camera system generates signals
In general, diagnostic nuclear imaging is a noninvasive pro- processed and analyzed by computer software to determine the
cedure that uses scanning hardware to examine the distribution direction/location from which the photons originated (whether
of a radiopharmaceutical within the internal environment of the in a two-dimensional plane or a three-dimensional space). The
body. As discussed, imaging the in vivo distribution of a signal data analysis is processed to create imagery that visualizes
radiopharmaceutical can be considered an in vivo assay of the in vivo biodistribution of the radiopharmaceutical in the
radiopharmaceutical pharmacokinetics, not just in blood (as for patient studied.
conventional plasma pharmacokinetics) but also in tissues/ Does the biodistribution of the radioisotope atoms visual-
organs throughout the body. In diagnostic imaging, it is not the ized by the nuclear scan represent the biodistribution of the
radiopharmaceutical itself that is of interest (although measuring administered radiopharmaceutical (molecules)? If the radio-
tumor concentrations of a radiolabeled therapeutic can be of pharmaceutical undergoes no in vivo chemical transformation
major interest, for research in tumor dose-response correla- to another form (e.g., catabolite or metabolite) before imaging
tions). Rather, the in vivo radiopharmaceutical distribution is of the patient, the answer is yes. Otherwise, the radioisotope
most often used as a diagnostic imaging biomarker; that is, the biodistribution imagery may represent a composite of biodis-
tissue concentration of tracer detected by the scanner is used as tributions, including those of the (unmodified) administered
noninvasive biologic marker, a radiologic surrogate for obtaining radiopharmaceutical and the radioactive products of in vivo
tissue (e.g., by biopsy) to assay an in vivo biophysiologic process chemical reactions (i.e., reaction products that still incorporate
or tissue biomolecular target of interest. In general, as with any the radioisotopic atom). Usually, in vivo metabolism of the
pharmaceutical in the modern age of “molecular medicine,” a radiopharmaceutical causes in vivo production of metabolites,
particular radiopharmaceutical compound is typically designed one or more of which include the radioisotope; these are radio-
with the intent of specifically targeting one or more in vivo metabolites. Such metabolism may be the diagnostic imaging
biophysiologic processes and tissue biomolecular targets specific target of the nuclear scan (e.g., PET imaging with F-18 FDG
to a disease or physiologic condition of interest. to detect tumor concentrations of the FDG metabolite). Some
No radiopharmaceutical compound yet designed has been radiometabolites may be radiolabeled molecules, or in vivo
found to bind exclusively to one particular biologic molecule. metabolism may yield radioisotope in free, unattached elemen-
Some compounds, however, such as radiolabeled antibodies, tal form. These radiometabolites often have different in vivo PK
radiolabeled “small molecules,” and other types of agents, do properties from the intact parent radiopharmaceutical. Thus,
bind with very high selectivity and affinity to relatively few the radiotracer biodistribution visualized by nuclear imagery
biologic molecules and not at all to other types of molecules will represent a combination of biodistributions: that of the
and are called “targeted agents.” Still, the biophysiologic pro- intact radiopharmaceutical and that of one or more radiome-
cesses and biologic molecules targeted by such agents for tabolites. In vivo metabolism occurs but typically does not
diagnostic imaging (or “targeted therapy”) of a particular interfere with diagnostic interpretation. On the contrary,
condition of interest can almost invariably be found in other metabolism may yield a radiometabolite “trapped” in a tissue
physiologic or pathologic conditions, again precluding 100% of interest, such as enzymatic trapping of the PET imaging
specificity. For example, the biologic molecule prostate-specific FDG in tumor cells; the cytoplasmic enzyme hexokinase yields
membrane antigen (PSMA, now more properly referred to as the radiometabolite fluorodeoxyglucose-6-phosphate, which is
glutamate carboxypeptidase II), once thought to be uniquely trapped intracellularly.
expressed by prostate tissues and thus a good biomarker for Because the radioisotope is the same for the parent radio-
prostate cancer (e.g., for imaging by PSMA-targeted radiola- pharmaceutical and each of its radiometabolites, image data
beled antibody), was later found to be expressed by certain analysis alone cannot separate the composite biodistributions
other tissues in the body and in the neovasculature of most in nuclear imagery. In clinical practice, however, this is rarely
tumors. However, PSMA is highly expressed in only a few types relevant, because radiopharmaceuticals that have achieved
of nonprostatic tissues and therefore still possesses high selec- widespread clinical use did so by having favorable in vivo
tivity for prostatic tissues. Therefore, “perfect” specificity should pharmacokinetics, typically including in vivo metabolism that
not be expected for diagnostic imaging agents, even radiola- yields a biodistribution that predominantly represents that of
beled antibodies, considering the underlying imperfect pharma- one parent compound or relevant radiometabolite, with one
cologic and biologic specificity, as well as potentially misleading predominant biologic “meaning.” This chapter discusses the
imaging artifacts. meaning of each radiopharmaceutical scintigraphic biomarker
Nevertheless, scintigraphy can be highly accurate, noninva- scan relevant in HPB diseases. Once administered to a patient,
sive in vivo virtual histochemistry, as long as scintigraphic evi- the radioisotope used for diagnostic imaging emits radiation
dence suggesting a particular type of disease is consistent with that can be detected by a nuclear scanner.
the clinical picture (i.e., pretest likelihood of that disease being Diagnostic imaging with radiopharmaceuticals, in standard
present, based on signs and symptoms from clinical and labora- clinical practice, may be referred to in various ways, including
tory examinations). If not, further confirmation of scintigraphic (1) general terms such as nuclear imaging or scintigraphy, (2)
findings may be warranted to ensure accurate therapeutic deci- reference to one of two general types of scintigraphic camera
sion making. technology (PET, SPECT/SPET), and (3) procedure involved
The imaging signal detected by nuclear scanners is produced (e.g., theranostic imaging).
by the radiopharmaceutical administered to the patient. The The term scintigraphy (Latin scintilla, “spark”) in medicine
radioisotope in the radiopharmaceutical undergoes its physical refers to the light produced by crystalline detectors in clinical
scintigraphic cameras when those crystals are struck by gamma energy produced by positron-electron annihilation is always the
rays emitted from radiopharmaceuticals (e.g., as emitted from same, 1022 kiloelectron volts (keV), regardless of the positron-
within a patient scanned after receiving a radiopharmaceutical emitting isotope involved (e.g., F-18, carbon 11, gallium 68).
injection). These scintillations produced in the crystalline The 1022-keV annihilation energy splits instantly into two
detectors are recognized and processed by the camera system photons: gamma rays of 511 keV each. This pair of gamma rays
to yield nuclear imagery. travels in opposite directions. The modern PET system is basi-
Of the basic types of scans found in a radiology department cally composed of a ring of crystalline detectors surrounding a
(e.g., plain radiography, CT, MRI, US), diagnostic nuclear gantry bed. Exiting the body at the speed of light, a pair of these
imaging scans are typically of the longest duration, in terms of annihilation photons may strike two of the detectors on the
both the time the patient must physically spend with the scanner PET ring, each strike-incident occurring practically at the same
and the time required for the entire study (start to finish), often moment. This coincidence of two strikes, of about 511 keV each,
with necessary delays before scanning or between scanning (i.e., is recognized by the PET circuitry as a true signal, whereas
if the patient is scanned more than once after a single radio- signals without a coincidence pattern or with incorrect energies
pharmaceutical administration) to allow the radiopharmaceuti- are dismissed as noise. A true signal informs the system that an
cal time to undergo desired in vivo physiologic processes. The annihilation occurred along the spatial line between the two
total duration of a diagnostic nuclear imaging study thus detectors. By analyzing millions of such signals gathered during
depends on a variety of technical, biologic, as well as typical PET imaging, along all the ring of different “lines of response”
clinical logistical variables. Most frequently, a radiopharmaceu- (imaginary lines between pairings of detectors,) sophisticated
tical is administered intravenously, by bolus injection. After the computer algorithms can deduce the distribution and concen-
injection, a standard time-delay may be necessary before the tration of radioactivity within the patient (i.e., biodistribution
patient undergoes scanning, to allow the radiopharmaceutical of isotope). Importantly, PET scanners cannot distinguish
to spread throughout the body and achieve a biodistribution between different positron-emitting isotopes (e.g., 18F, 11C,
68
considered optimal for imaging. To acquire data for a single Ga). The energy signal generated by different positron-
image, the time that a patient spends “in front of the camera” emitting isotopes and detected by PET scanners is always the
must be of sufficient duration for the scanner to collect a sta- same: coincident 511 keV gamma rays. Thus, if one patient
tistically robust number of radioactive signals, or counts, to received two different radiopharmaceuticals simultaneously
ensure that the derived imagery will be satisfactory for visual (e.g., one radiolabeled with 18F and the other with 68Ga), and
analysis. Low-count images are visually “noisy.” How long it the patient was then scanned, the PET scanner could not sepa-
takes for the camera to collect enough photons for a sufficient- rate the radioactive signal emitted by the 18F compound from
quality diagnostic image depends primarily on intrinsic proper- the radioactive signal emitted by the 68Ga compound. There-
ties of the radioisotope involved, how much radiopharmaceutical fore, patients do not receive two different PET radiopharma-
is administered, how well the radiopharmaceutical concentrates ceuticals simultaneously, because the result is one PET scan of
in tissues of interest (e.g., tumors) compared with surrounding mixed “meaning.” In this scenario, one cannot know how much
tissues in vivo, how well the camera system detects photons, of a radioactive signal in one tissue site represents one PET
and how the photon data are constructed into imagery. Depend- tracer versus the other.
ing on the type of nuclear imaging study, before imaging even The design of the standard (non-PET) gamma camera is
starts, there may be a standardized delay after the radiophar- tailored to detect isotopes that decay by emitting (unpaired)
maceutical administration to allow the radiopharmaceutical gamma rays of relatively lower energy, typically energy lower than
time to achieve an in vivo biodistribution considered optimal the 511 keV energy (photopeaks) of PET imaging. This is often
(i.e., 1 hour after FDG injection before PET scan acquisition). referred to as single-photon imaging, to distinguish such imaging
Lastly, the imaging specialist decides whether to have the from PET imaging that involves detecting a pair of two photons
patient undergo scanning at additional time points or using simultaneously. In contrast, single-photon imaging refers to
special techniques, if thought necessary to increase the diagnos- camera detection of (unpaired) photons from nonpositron-
tic accuracy of the study. The referring clinician’s staff can help emitting radioisotopes. Some non-PET radioisotopes have
prepare patients mentally by advising them of the prolonged radioactive decay characterized by emitting photons of more
duration typical of diagnostic nuclear imaging. than one energy level; for example, the common single-photon
Nuclear scanners may be categorized into two general types: imaging isotope indium 111 simultaneously emits two types of
PET scanners and standard gamma cameras. Their designs are photons relevant for imaging—photons differing in energy—with
tailored to image two fundamentally different types of radio- two energy levels (photopeaks) of 173 and 247 keV.
pharmaceuticals (radioisotopes): those that emit positrons (for After the scanner system processes the photon data to
PET cameras) and those that emit gamma rays (for standard determine the spatial origin of the signals in vivo, a nuclear
gamma cameras). medicine technologist at a computer workstation uses software
In PET imaging, the radiation detected by the crystals of to reconstruct the data for visual display and analysis by the
PET cameras comes from gamma rays that are produced by physician imaging specialist. How the in vivo radiopharmaceu-
positron-electron annihilations that occur, after a decaying radio- tical biodistribution is displayed depends on how the imaging
isotope emits a positron. The positron emitted by the PET data was acquired. This biodistribution imagery can have a flat,
radiopharmaceutical encounters its antimatter: any electron in two-dimensional (2D), or planar, appearance; or a (virtual)
its immediate environment. The two particles “annihilate” three-dimensional (3D) appearance (e.g., allowing display of
during the encounter, converting all particulate mass (m) to sections of data in conventional transaxial, coronal, and sagittal
pure energy (E). This annihilation reaction is an everyday views, similar to CT). The imagery may represent biodistribu-
example of Einstein’s theory of relativity, E = mc2. Because the tion at one or a few time points, or imagery may display time-
mass of positrons and electrons are universally constant, the dependent changes in biodistribution in cinematic fashion.
Planar images of radiotracer biodistribution in the anterior- receives an 18F FDG injection for FDG PET, based on the
posterior plane will result in an image in which in vivo tracer known half-life of 18F (~1.9 hours) and the amount of tracer we
accumulations in two or more organs or other tissues may administer to a patient (maximal tracer dose of 12 mCi). Thus,
overlap (in the 2D plane) and thus potentially obscure detec- one must wait until the PET isotope has disappeared (decayed)
tion or evaluation of the radiotracer uptake of interest (e.g., sufficiently before attempting to image another radiopharma-
tumor detection). Tomographic (SPECT and PET) nuclear ceutical. A common rule of thumb is waiting 10 half-lives, a
imaging can help avoid this potential issue by permitting tracer half-life referring to the physical half-life of the isotope.*
biodistribution to be evaluated in three dimensions. However, In general, however, single-photon radiopharmaceutical
the limited spatial resolution of scintigraphic imaging may make imaging studies can be performed immediately before PET
it difficult to localize a particular tracer accumulation in a small scans, without needing to wait for the single-photon radiophar-
tissue structure (e.g., tracer uptake in a small tumor may be maceutical to disappear; this is because the photons of non-PET
hidden if the tumor is located within or immediately adjacent radioisotopes typically are relatively low energy and thus cannot
to a normal organ that also accumulates tracer). Additionally, interfere with detection of the relatively high-energy 511 keV
for single-photon imaging agents, SPECT often requires a photons of PET radiopharmaceuticals. A few non-PET radio-
significantly longer duration scan than a 2D planar scan using isotopes do emit high-energy photons at or above the 511 keV
standard SPECT camera systems, and often 2D imaging may range, but even the presence of these high-energy photons from
be sufficient for the clinical data desired. The necessity for a prior non-PET radiopharmaceutical injection do not neces-
SPECT imaging is guided by the reason for a particular exami- sarily preclude immediately performing a PET study. Because
nation, available clinical research, and the particular patient the high-energy photons of the non-PET radioisotope are not
context. For PET, by definition, tomography (3D imaging) is produced as pairs, these photons will strike PET detectors in
always used, involving ring-type dedicated PET camera systems random directions. The PET scanner, however, will ignore
with sophisticated signal analysis algorithms (discussed later). photons that are not detected in the coincidence pattern typical
For a brief time, planar (2D) imaging and even 3D imaging of annihilation photon pairs. As such, the PET scanner often
with single-photon emission with PET radiopharmaceuticals can detect the desired PET radiopharmaceutical signal without
(e.g., FDG) were explored by using gamma cameras, which are significant interference from any residual non-PET radiophar-
still standard for nuclear imaging with gamma-emitting (non- maceutical in the patient’s system. The opposite, however, is
PET) radiopharmaceuticals. However, gamma camera imaging not the case: residual PET radiopharmaceutical will interfere
of PET radiopharmaceuticals is clearly inferior diagnostically with detection of another PET or a non-PET radiopharmaceu-
and not state-of-the-art practice. tical until a certain number of half-lives elapse. The referring
clinician ordering two or more diagnostic imaging nuclear
“Can I Order Two Nuclear Medicine Scans medicine scans for a single patient in a short period must be
on the Same Day?” aware of this potential problem of “conflicting” radioisotopes
The distinction between positron-emitting (PET) and solely and should consult with an imaging specialist or others expe-
gamma photon–emitting (non-PET) isotopes is relevant for rienced on how best to schedule and prioritize the different
referring specialists primarily because of this question, fre- scans, if uncertain.
quently posed to nuclear medicine specialists or their schedul-
ing staff. Unlike positron-emitting radioisotopes, the gamma
Advantages of Positron Emission Versus
ray–emitting isotopes of radiopharmaceuticals used for single- Single-Photon Imaging
photon imaging come in a variety of energy levels. Gamma The advantage of PET imaging versus single-photon gamma
cameras can distinguish and potentially use these energy levels imaging is that the PET coincidence-detection method permits
to separate the biodistributions of one radioisotope from more precise determination of where the radiation originated.
another, if a patient were to receive two radiopharmaceuticals Thus, the scan imagery reconstructed from PET data has a
with different radioisotopes simultaneously; that is, by having much better spatial resolution (typically 4-5 mm, vs. <1 mm on
the camera only accept detected photons of the energy level CT) than that reconstructed from single-photon gamma data
characteristic of the particular radioisotope of interest, then (typically 12-15 mm). According to the Nyquist principle, this
doing the same for the other radioisotopes. However, the energy superior resolution results in PET-acquired data providing
emissions of different single-photon imaging radioisotopes can superior quantification of radioactivity concentrations in imaging
overlap, particularly if one of the radioisotopes emits relatively data analyses compared with single-photon imaging. As one
high-energy emissions, because some emitted rays will lose potential advantage versus PET imaging, single-photon imaging
energy and fall into the energy (keV) range of the other radio- can simultaneously detect and distinguish two or more different
isotope. The 511 keV coincidence gamma rays of PET radio- gamma-emitting radiopharmaceuticals in a single patient in
pharmaceuticals are relatively high energy and will interfere
with imaging of single-photon radiopharmaceuticals, whether
to a significant degree depends on other technical factors. Thus,
after a PET scan, some period of delay is necessary, to allow *Physical half-life is an intrinsic property of the radioisotope: the rate at which it
undergoes its nuclear decay. For example, for an amount of radioisotope in a
the PET radiopharmaceutical to undergo physical decay and bottle, after one physical half-life, there will be half as much radioisotope (half as
biologic clearance from the patient, before performing another much radioactivity) remaining in the bottle. Biologic half-life is a PK parameter:
the rate at which the patient’s body clears a (radio)pharmaceutical (e.g., by
imaging study with a single-photon radiopharmaceutical or urination/defecation). Thus, after a patient receives a dose of radiopharmaceuti-
another PET radiopharmaceutical. cal, the rate at which it disappears from the body is determined by a combination
For example, for the common radiopharmaceutical, 18F of the physical half-life and the biologic half-life, called the effective half-life. After
10 physical half-lives of 18F (~19 hours), more 18F FDG will have disappeared
FDG, our institution allows a subsequent different nuclear from a patient’s body than predicted by physical half-life alone, because 18F FDG
imaging study typically 20 hours or more after the patient is also eliminated by urinary excretion.
vivo, whereas PET imaging cannot distinguish between differ- can vary considerably; it can be of standard diagnostic quality
ent PET isotopes. Gamma rays emitted by non-PET isotopes (i.e., exactly the same technical-quality CT scan as obtained
for single-photon imaging can have a variety of signature energy from a separate, stand-alone, state-of-the-art CT scanner), or it
levels, which can distinguished and separated by signal process- can be of inferior diagnostic quality, sometimes to a great degree.
ing. Single-photon imaging has the added practical advantage The CT scan might be acquired with a lower current (mA),
of being more disseminated and available to clinicians and yielding relatively noisier images, with less detail and greater
patients worldwide. susceptibility to certain artifacts (e.g., beam hardening). Chest
In general, PET imaging is less available because of the imaging may be acquired at pulmonary end expiration rather
higher costs of imaging hardware and the need for access to an than the standard maximal pulmonary inspiration. The CT scan
institutional or commercial source of PET isotopes, which must might be acquired with the patient’s upper extremities posi-
be local due to the short half-life of isotopes such as 18F (~2 tioned along the torso if the patient cannot tolerate having the
hours). However, the PET infrastructure is rapidly growing arms raised (the standard chest CT position) for the 15 to 25
worldwide. Currently, an estimated 80% of the US population minutes of a standard torso FDG PET/CT. In this position, the
is within the service area of a PET imaging facility. arms are “in the way” of the CT current and can reduce the
Policies of regulatory agencies also can pose obstacles to quality of the torso CT images obtained. Also, the CT scan
patient access to nuclear medicine procedures, both general and might not use the oral or intravenous iodinated contrast material
specific. For example, cancer patients in certain parts of Europe routinely used in standard CT protocol. Acquiring this relatively
overall have access to a wider range of PET tumor-imaging inferior-quality CT scan is in fact common clinical practice,
agents than US cancer patients. The U.S. pharmacopoeia of often acquired intentionally, even possibly using a hybrid CT
these agents is essentially one agent, the FDA-approved and scanner fully capable of acquiring standard diagnostic-quality
widely available FDG. Many other PET tumor-imaging agents CT scans. It is usually done, for example, if the patient already
are used clinically in the United States, but essentially only at had a recent diagnostic-quality CT scan. In that case, a com-
select few medical centers, under the auspices of FDA-approved panion lower-dose, noncontrast companion CT may be suffi-
Investigational New Drug (IND) applications that detail the cient to the basic needs of the PET or SPECT scan. In synergic
specific conditions under which the PET agents will be used fusion imaging particularly, a “nondiagnostic” companion CT
and for a prespecified number of patients. Moreover, synthesis usually (but not always) provides sufficient anatomic detail to
of these IND-sponsored PET radiopharmaceuticals must meet identify what tissues are involved in radiotracer uptake within
stringent FDA regulatory standards that only certain commer- an organ and provides attenuation correction, a modification of
cial entities and few medical centers have the capacity to meet. the scintigraphic data based on tissue densities and depth
In certain European countries, regulatory standards regarding measured by CT that improves scintigraphic image quality,
radiopharmaceutical synthesis for human use recognize the especially for quantification (e.g., for PET quantitative mea-
inherently low-risk nature of administering microgram amounts surements known as the standardized uptake value [SUV]).
of radiopharmaceutical compounds and are relatively less Decay photons emitted from within deeper or denser body
stringent than US standards. tissues will lose more energy from tissue interactions than
photons from superficial or less dense tissues. The attenuation
Fusion Imaging correction attempts to account for that artifact to provide more
Scintigraphic imaging of PET and single-photon emission accurate measurement of tissue tracer concentrations.
radiopharmaceuticals is often combined with computed tomog- In essence, the nondiagnostic companion CT scan is a
raphy (CT) imaging, for fusion imaging: PET/CT and SPECT/ blurry image; some images might have a slight blur, and the
CT, respectively. In fusion imaging, the 3D imagery of PET or reader still can identify subcentimeter tissue structures (e.g.,
SPECT is combined with CT data so that tracer biodistribution/ small pulmonary nodules), whereas others might be fairly
localization is visualized within the internal anatomy. Clinical worthless (e.g., whole subregions of the body may be affected
studies have overall demonstrated that fusion imaging can have by beam-hardening or other CT artifact). The physician who
a synergistic effect on the accuracy of scintigraphy and CT image orders a fusion PET/CT or SPECT/CT must know whether
analyses for various clinical applications. A notable general the companion CT is of diagnostic quality or must be specifi-
example is improved accuracy for detection of radiotracer-avid cally ordered to be a diagnostic-quality CT scan, or whether a
tumors; often, on fusion imagery, scintigraphy (PET or SPECT) diagnostic-quality CT scan must be obtained separately. Simi-
highlights findings poorly detected or easily overlooked on CT, larly, the imaging specialist providing a PET/CT or SPECT/
or vice versa. The CT information also serves as data for an CT report can assist the referring physician by stating whether
important technical function, called attenuation correction, which the companion CT is considered equivalent to a standard
improves the quantitative accuracy of measurements derived diagnostic-quality CT scan. The imaging specialist often will
from PET or SPECT analyses. Fusion imaging has done much describe companion CT findings, even in a poor-quality
to rescue diagnostic nuclear medicine from its former moniker nondiagnostic companion CT, if relevant CT findings are
(deserved or not) of “unclear medicine.” detectable; however, the referring physician should not con-
In state-of-the-art practice, a SPECT or PET scanner is clude that the presence of CT findings described in a fusion
physically joined to a CT scanner so that the CT scan is PET/CT or SPECT/CT report means that CT used was of
acquired immediately before the scintigraphic imaging, without diagnostic quality. In our experience, the quality of the com-
the patient changing position, laying on the scanner bed as it panion CT varies considerably between medical centers and
passes within the conjoined central tunnel of the hybrid scanner. between fusion scanner models. The referring clinician is
One important caveat remains regarding the CT scans advised to assess the quality of the companion CT when evalu-
involved in fusion PET/CT and SPECT/CT imaging: the ating the results of a fusion imaging study, particularly if the
companion CT scan. The quality of the companion CT image study was performed at an outside facility.
Beyond the companion CT scan, first-generation clinical imaging (e.g., tumor detection) or incidental findings (e.g.,
PET/magnetic resonance imaging (MRI) and SPECT/MRI pneumonia in a patient referred for tumor detection). Various
camera systems are already available at a few major medical pathologies may have a similar scintigraphic appearance with a
centers and will begin disseminating to other medical centers particular radiopharmaceutical. For example, a patient with
during the next 10 years. At this point, clinical experience is cancer referred for tumor detection by nuclear scintigraphy may
relatively scant, with published clinical research predominantly have a nodular, noncancerous lesion that avidly accumulates a
focused on technical validations and optimizations of the radiopharmaceutical, appearing as an abnormal focus of radio-
companion MRI. activity in the patient’s body, mimicking a cancerous tumor.
The diagnostic specificity of nuclear scintigraphy varies
“Are All PET and SPECT Scanners Created Equal?” according to the radiopharmaceutical used, the disease entity
How does one know if one’s medical center has a state-of-the-art or biologic process being assayed, and the pretest likelihood of
PET or SPECT scanner? Currently, fusion PET/CT and false-positive results. For example, specificity of FDG PET is
SPECT/CT scanners are the state of the art. The medical lit- lower for patients who live in areas where tuberculosis (TB) is
erature indicates that the diagnostic accuracy of fusion imaging endemic; TB lesions absorb FDG and can mimic tumor sites
for tumors that is derived from a physically integrated, hybrid on FDG PET. The specificity of conventional hepatobiliary
PET/CT is superior to the diagnostic accuracy of combined scintigraphy for detection of gallbladder cystic duct obstruction
reading of PET and CT data obtained from separate scanners, (in the context of suspected cholecystitis) is lower for patients
even if fused by software. Each year now, new models of com- who do not fast properly for the procedure. Obstruction is
mercial PET/CT scanners enter the marketplace. excluded on scintigraphy if the radiopharmaceutical, after
There is some support for the belief that these innovations hepatobiliary excretion, is detected filling the gallbladder
improve patient care. For example, clinical research shows that lumen (having entered via cystic duct). However, normal
modern iterative methods are superior to older, filtered back- digestive processes associated with recent feeding trigger physi-
projection methods for reconstruction of PET images, in terms ologic contraction of the gallbladder, creating a pressure gradi-
of the ability of imaging specialists to detect tumor sites. Oth- ent through the cystic duct that prevents excreted-radiotracer
erwise, the benefits are theoretic but tenable, robustly supported from filling the gallbladder. The absence of gallbladder
by data models and phantom studies. filling caused by physiologic processes mimics a pathologic
cystic duct obstruction (e.g., caused by cystocholelithiasis) on
PHARMACOLOGY AND BIOLOGY IN scintigraphy.
RADIOPHARMACEUTICAL IMAGING Radiation Dose in Nuclear Medicine
Radiopharmaceuticals constitute a large pharmacopoeia of When administered solely for diagnostic imaging, conventional
different radioactive agents used for imaging, radiotherapy, or radiopharmaceuticals expose the receiving patient to a low
both (theranostic agents). An administered dose of a typical radiation dose, typically one or more orders of magnitude below
diagnostic nuclear imaging agent typically contains a trace the level of radiation exposure conventionally accepted as being
(microgram) amount of molecular compound with no expected associated with an increased risk of harmful radiation effects,
biologic effect and minimal risk of toxicity; this is labeled with based on decades of dosimetric research. Certain radiopharma-
isotope that emits radiation suitable for imaging. FDA-approved ceuticals are administered at relatively high doses of radioactiv-
radiopharmaceuticals, administered in standard diagnostic ity with therapeutic intent; the administered radioactivity is
imaging doses, expose patients to levels of internal irradiation sufficiently high (and concentrates in body tissues at levels
considered low by radiologic safety standards. sufficient) to induce acute radiobiologic effects on diseased
As with other pharmaceuticals, each radiopharmaceutical tissues and other organs within a patient, with potential thera-
is distinguished by its pharmacologic profile (e.g., route peutic benefit as well as adverse side effects.
of administration/bioavailability, metabolism, biodistribution, The radiation dose absorbed by the patient from the radio-
pharmacokinetics), including potential interactions with other pharmaceutical is determined by the specific type of radioactive
drugs or substances. Sometimes, a particular diagnostic nuclear isotope involved, the administered activity (amount) of radio-
imaging study may require special patient preparation, hours pharmaceutical (quantified in becquerels, typically megabec-
or days before the radiopharmaceutical is administered; for querels [MBq], or curies, typically millicuries [mCi]), and the
example, patients may need to abstain from certain over-the- pharmacokinetics and distribution of radiopharmaceutical
counter or prescription drugs that may affect the in vivo throughout the body (biodistribution), which is again deter-
behavior of a certain radiopharmaceutical. Patient preparation, mined by the physicochemical characteristics of the radiophar-
when necessary, is specific to the particular radiopharmaceuti- maceutical molecule (or element).
cal employed in the nuclear imaging study. The oncologist
unfamiliar with a particular study should contact an imaging DIAGNOSTIC NUCLEAR IMAGING IN
specialist for information on patient preparation. HEPATOPANCREATOBILIARY ONCOLOGY
The expertise of the nuclear imaging specialist lies in his or
her in depth knowledge of the diversity of the nuclear medicine Radiopharmaceuticals localize to tumors according their physi-
pharmacopoeia (e.g., in vivo bioavailability, metabolism, biodis- cochemical molecular properties and route of administration
tribution) of each radiopharmaceutical and the variety of condi- (e.g., intravenous, oral). These factors also dictate the distribu-
tions under which these may be altered from the norm (e.g., tion of the radioactive compounds throughout the rest of the
pathologic conditions, normal variations in patient physiology, body (biodistribution) and their clearance from the body
interactions with endogenous or exogenous substances). Patho- (almost always involving urinary and hepatobiliary excretion).
logic conditions may include the indication for diagnostic Only a few radiopharmaceuticals with oncologic applications
are relevant to and frequently used in patients with HPB cancer staging, therapy evaluation) as well as clinical applications
care, where state-of-the-art nuclear medicine practice is avail- unique to nuclear imaging (e.g., tumor pharmacokinetics and
able, with significant variation in the availability of specific biologic phenotype, disease prognosis, biomolecular pharmaco-
radiopharmaceuticals between countries, predominantly for dynamics, drug biodistribution). These unique clinical applica-
economic and regulatory reasons. tions are specific to particular radiopharmaceuticals and cancer
The complete pharmacopoeia of radiopharmaceuticals types; in HPB cancers, these unique applications remain inves-
includes those that have no oncologic clinical application but tigational, although in certain non-HPB cancers, particular
that are important in the general medical care of patients with examples have become routine clinical practice, such as use of
HPB cancer (e.g., preoperative myocardial perfusion imaging, FDG PET to characterize residual soft tissue masses in lym-
“nuclear stress test”). This section focuses on oncology-specific phoma or germ cell tumors as viable tumor or nonviable treated
applications of diagnostic and therapeutic nuclear medicine for disease.
clinical management of HPB cancers. Ongoing cancer imaging research is exploring nuclear
Diagnostic radiology in general has vital clinical applications imaging for clinical applications of tracking the distribution and
in the care of patients with HPB cancer, notably tumor detec- pharmacokinetics of radiolabeled cancer drugs inside living
tion; disease staging; treatment planning; and evaluation of patients (in vivo) by scintigraphy (particularly by PET imaging),
tumor response to various forms of anticancer treatment. For allowing clinicians to follow cancer drugs in vivo to evaluate
radiologic imaging intended for these primary clinical applica- drug uptake and retention or clearance by tumors and organs.
tions, major HPB oncology guidelines recommend use of Providing unprecedented pharmacokinetic data about the
standard CT (see Chapter 18), MRI (Chapter 19), and/or ability of cancer drugs to target tumors, such novel scintigraphic
ultrasound (US; Chapter 15) for radiologic evaluation of HPB assays may expand the oncologic applications of nuclear
cancer patients (National Comprehensive Cancer Network imaging to guiding patient selection, dosing strategy, and
[NCCN], 2015). These guidelines do not provide a standard role response assessment (Dunphy & Lewis, 2009). This chapter
for diagnostic nuclear medicine imaging (e.g., PET), except for concentrates on the role and efficacy of state-of-the-art diag-
a few minor specific applications. CT, US, and MRI, as well as nostic nuclear imaging techniques and technology, with pre-
plain radiography, are superior to nuclear medicine imaging views of new nuclear imaging agents in development.
modalities for definition of anatomy. Nuclear medicine imaging
is the preeminent clinical imaging modality for evaluation in Fluorodeoxyglucose Positron Emission Tomography
vivo physiology and molecular biology in general. We sometimes In the past 15 years, FDG PET has rapidly emerged as a revo-
state that PET/CT is superior to CT, which is not the same as lutionary imaging modality in clinical oncology, demonstrating
stating PET is superior to CT (noting again, however, that the diagnostic efficacy in tumor staging and tumor-response evalu-
CT of PET/CT often is of lower quality than standard CT), ation for histologies across the spectrum of cancer. FDG,
In general, HPB cancer guidelines recognize the potential 2-deoxy-2-(18F)fluoro-D-glucose, is an analogue of glucose;
value of FDG PET/CT imaging as a diagnostic adjuvant, if fluorine-18 occupies the molecular 2′ position in which a
first-line radiologic imaging (CT, MRI, US) yields indetermi- hydroxyl group is found in glucose (Fig. 17.1). The substitution
nate results. Importantly, the guidelines do not endorse a affects the metabolism of FDG compared with glucose. In vivo,
standard role for FDG PET as a first-line imaging modality, intravenous (IV) FDG extravasates into tissues, followed by its
often because clinical research regarding the added value of uptake into tissue cells by glucose transporter proteins. Blood
FDG PET in these major applications in HPB oncology is FDG concentrations decrease to relatively low levels 45 to 90
simply insufficient, not because clinical trials have already minutes after injection; at this point, further FDG uptake in
shown that FDG PET has no such value. Moreover, PET/CT most tissues is relatively minor, and after that time, FDG
technology and methodology have undergone considerable concentrations in most tissues and tumors remain relatively
evolution, improving diagnostic accuracy significantly in recent stable. Acquiring a single FDG PET scan, beginning 45 to 90
years, and these remain in flux. Recent clinical research, as minutes after injection, has become the standard clinical
discussed here, may prompt future revisions of HPB oncology approach. Usually, as for HPB cancer imaging, the scan spans
guidelines. from “eyes to thighs,” including the entire head or extremities
only if there is a patient-specific clinical reason.
Nuclear Oncology After FDG enters the cell, intracellular hexokinase converts
Broadly, oncologic applications of nuclear imaging include FDG to FDG-6-phosphate, making it too polar to traverse the
those common to CT, MRI, and US (tumor detection, disease cell membrane. In most cells, this traps the tracer, in the form
OH OH
O O
HO HO
HO HO
18F
OH
OH OH
FIGURE 17.1. Molecular structures of glucose (left) and fluorine-18 fluorodeoxyglucose (right).
of FDG-6-phosphate, intracellularly. After phosphorylation of the same as the glucose metabolism of normal cell counter-
FDG, the product FDG-6-phosphate is not metabolized further parts). The Warburg hypothesis, however, is outdated: the shift
to other forms, to any relevant degree, and will not become a of glucose metabolism from the mitochondrial TCA cycle to
substrate for the enzymes that use glucose-6-phosphate (G6P) cytosolic glycolysis is not an inefficient use of glucose; rather,
as a substrate. One notable exception is G6P in the hepatocytes it is a “repurposing” of glucose. In multiple cell lines, abnormal
of normal liver tissues. FDG-6-phosphate is converted back to cytosolic glycolysis has become understood as advantageous to
FDG to a detectable degree, and the nonpolar FDG can then cancer cell proliferation. Cytosolic glycolysis yields fewer ATP
potentially diffuse back across the plasma membrane, leaving molecules, but it yields glucose-derived metabolites during the
the hepatocyte and, possibly, the entire liver. Clinical FDG multiple intermediate steps of glycolysis that the cell can use in
PET demonstrates this, as liver concentrations of FDG begin other anabolic pathways as components for synthesizing mac-
to decline approximately 1 hour after injection. In other healthy romolecules necessary for building cellular biomass before cell
tissues, dephosphorylation and cellular efflux of FDG is rela- division and tumor growth. To meet the bioenergetic needs of
tively minor. In most tumor types that have been studied at two these cells, instead of predominantly relying on glucose, these
or more delayed time points after injection, FDG appears cells depend on other nutrient molecules, notably glutamine,
effectively trapped. In tissues affected by infectious or inflam- to fuel the TCA cycle (Wise & Thompson, 2010).
matory disease, on the other hand, tissue concentrations of The Warburg effect explains the avidity of tumors for FDG,
tracer reportedly often decline significantly with time, similar the glucose analogue, when visualized by PET, but only in part.
to liver, beginning approximately 1 hour after injection. This PET visualizes the FDG-avidity of tumors at the macroscopic
suggests that detecting FDG efflux in suspicious FDG-avid tissue level (again, with spatial resolution of ~4 mm). The
tissues by PET imaging at two or more delayed time points may “FDG-avid tumor” visualized by PET and described in PET/
be a radiologic biomarker for distinguishing FDG-avid inflam- CT reports represents a complex composite of FDG avidities
matory tissues (e.g., infected tissues, posttreatment inflamma- of tumor cells and nonneoplastic cell constituents within the
tory tissue infiltrates) that can mimic FDG-avid tumor (i.e., tumor internal microenvironment, under the influence of
false-positive PET results) from true FDG-avid tumor. The complex biomolecular and other processes. Detected FDG
clinical reports testing this hypothesis are somewhat inconsis- avidity in a particular tumor often does primarily represent
tent, however, and clinical research remains ongoing; at present, tumor cell FDG-avidity (i.e., the sum of FDG-uptake from all
FDG PET imaging at a single time point is the norm. tumor cells within the tumor), more than the FDG-avidity of
The basic rationale for using FDG PET for tumor detection other cells in the same tumor. However, the FDG avidity of
is the observation that neoplastic cells often (but not always) other constituents of the tumor microenvironment sometimes
accumulate FDG more than the nonneoplastic cells of origin, contributes a clinically significant degree, particularly in the
and that the difference in FDG concentration between the posttreatment setting, potentially causing diagnostic confusion,
tumor and surrounding normal tissues in an organ is detectable as discussed. Within the neoplastic cell, FDG avidity can be
by PET. This avidity of tumors for FDG manifests on PET induced by oncogenic mutations changes directly or indirectly
imagery as a “hot spot,” or a focus, of FDG accumulation that affecting glucose metabolism. For example, oncogenic muta-
is of abnormally high concentration relative to other, healthy tions that drive cell proliferation, typical of aggressive tumors,
tissues. often are associated with the shift of glucose metabolism to
Why do many tumors concentrate FDG so avidly, and why cytosolic glycolysis and increased glucose consumption; tumor
are other tumors not so avid? The physician-scientist and Nobel cell FDG avidity can also be increased by epigenetic metabolic
laureate Otto Warburg long ago observed in multiple tumor cell changes induced by conditions in the tumor microenvironment
lines that he studied an abnormally high rate of glycolysis in (e.g., tumor hypoxia). In certain cases, overall tumor FDG-
cancer cells compared with their normal cellular counterparts, avidity can be caused, in relatively large part or even primarily,
even in the presence of normal levels of oxygen. In normal cells by tumor cellular constituents other than the neoplastic cells,
with adequate environmental oxygen, glucose metabolism is such as infiltrating inflammatory cells, especially when the
typically directed into the mitochondrial tricarboxylic acid tumor cells do not have intrinsically high FDG avidity and
(TCA) cycle; glucose–TCA cycle metabolism yields the when inflammatory cells are present in relatively high tissue
maximal amount of energy substrate (adenosine triphosphate, concentrations. For example, inflammatory cells can accumu-
ATP) from each glucose molecule metabolized for meeting the late around the necrotic cores of tumors before treatment and
bioenergetic needs of the cell. The TCA cycle depends on can infiltrate heavily throughout tumors after treatment. Stan-
oxygen to function; in normal cells, if environmental oxygen is dard FDG PET guidelines often advise that posttreatment PET
low, glucose metabolism instead occurs in the cytosol by an be deferred for 6 to 8 weeks after chemotherapy and 2 to 3
oxygen-independent glycolytic process that yields much less months after radiotherapy. It was empirically observed in clini-
ATP per glucose molecule. In cancer cells, however, Warburg cal PET trials that successfully treated tumors often demon-
observed that glucose metabolism occurred predominantly by strate apparently suspicious residual FDG avidity in the first
glycolysis in the cytosol, regardless of whether or not the tumor few weeks after treatment because of inflammatory cells infil-
cells were well oxygenated. This preference of tumor cells is the trating the treated tumors, presumably to clear the necrotic/
Warburg effect. According to the Warburg hypothesis, cancer cell apoptotic debris associated with successful treatment.
metabolism of glucose was inefficient because it yielded fewer Whenever using PET to characterize or localize tumors,
ATP molecules per glucose cell; and this inefficiency was caused the oncologist (and imaging specialist) should be aware
by a defect in the mitochondrial metabolism of cancer cells. of key factors that affect the apparent FDG avidity of a
The Warburg effect remains a valid observation, although not tumor and the diagnostic sensitivity of FDG PET: tumor size;
a universal phenomenon among all cancer cell lines and types cancer treatment(s); and background organ FDG avidity. As
(i.e., glucose metabolism of some cancer cell lines is essentially the Nyquist principle indicates, the apparent FDG avidity of
subcentimeter tumors will be underestimated, because such during therapy have frequently predicted favorable clinical
small lesions fall below the spatial resolution of PET technol- outcomes, whereas stable or increasing FDG avidity portend
ogy. It is still possible for PET to detect a subcentimeter tumor, worse outcomes across a variety of cancers. Evaluation of
if the tumor is so FDG avid that the FDG accumulation in tumor FDG avidity, after cytotoxic therapy without a pretreat-
tumor is detectably higher than that in background tissues of ment PET study for comparison can be performed but can
the organ involved. Frequently, however, subcentimeter tumors yield potentially confusing findings; for example, reactive
lack apparent FDG avidity and may be reported as “too small lymph nodes and partially treated metastatic adenopathy can
to characterize by FDG PET.” have similar appearances on PET (heterogeneous nonspecific
Cancer treatments, depending on action and efficacy, also FDG avidity). Tumors may have marked residual FDG avidity
affect apparent tumor FDG avidity (Liu et al, 2006). Tumor after treatment, which may provoke concerns about tumor
FDG avidity represents the sum of the FDG avidity of constitu- resistance. However, if a pretreatment FDG PET had been
ent tumor cells. Various studies indicate FDG PET is unable obtained, the residual FDG avidity might have been observed
to detect microscopic residual disease; for example, a partially as a marked decrease from baseline tumor FDG avidity, sug-
treated tumor containing FDG-avid cells may be of macro- gesting a favorable tumor response. In other words, the change
scopic size on CT or MRI, but contains a depleted cell popula- in tumor FDG avidity before versus after treatment can be
tion with a sum FDG avidity that appears minimal to nil on more predictive of tumor response than merely the posttreat-
PET imagery. For staging, FDG PET is expected to be less ment FDG avidity alone. As mentioned, certain tumor histolo-
accurate after therapy than before therapy (Liu et al, 2006). gies seem frequently to lack FDG-avidity, despite the presence
FDG PET can also be false positive, detecting FDG uptake at of viable, macroscopic neoplastic disease. A lack of tumor
a former tumor site in the absence of residual disease. With FDG avidity on a posttreatment PET scan can be potentially
systemic therapy, residual FDG uptake may indicate inflamma- misleading as an indicator of tumor response, unless a pretreat-
tory cells (extremely FDG avid, when active) infiltrating tissues ment PET scan has demonstrated the tumor being treatment
to remove the debris of treated disease. Radiotherapy and was originally FDG avid.
surgery, for tumor treatment or resection, both evoke local
tissue inflammation that can be greatly FDG avid, mimicking Colorectal Cancer Metastasis to Liver
local residual or recurrent neoplastic disease on PET imagery. The most recent guidelines from NCCN (2015), European
FDG PET is usually deferred for several weeks after surgery or Society of Medical Oncology (ESMO) (van Cutsem et al,
radiotherapy, when evaluation for local disease is desired. 2014), and European Registration of Cancer Care (EURECCA)
Certain organs and organ systems have marked FDG avidity agree on the potential roles and limitations of FDG PET/CT
consistently or variably that may exceed that of primary tumors in clinical management of colorectal cancer (CRC) (van de
and metastases, obscuring tumor detection. The brain is con- Velde et al, 2014). This discussion focuses on the CRC patient
sistently FDG avid because it normally depends glucose with potentially resectable liver metastases.
metabolism; FDG PET has limited sensitivity for detection of In cancer patients with liver metastases, but no extrahepatic
brain metastases. Myocardium, skeletal musculature, adipose metastases, complete resection of liver metastases improves
tissues, adrenal glands, and alimentary tract tissues have vari- long-term survival better than other treatments currently avail-
able FDG avidity that may or may not interfere with PET able (see Chapter 92). Before surgery, it is essential to confirm
analysis. Bone marrow can have high levels of FDG avidity, that liver metastases are likely resectable (based on CT and/or
usually associated with a hematopoietic response to ongoing or MRI imaging) and that no extrahepatic metastatic disease is
recent systemic cancer therapy or marrow stimulants (e.g., present. In the search for CRC metastases, metastatic disease
GCSF). When evaluation of osseous structures is vital, experts is detected most often in the liver, followed by the lungs;
advise deferring FDG PET for several weeks after the last dose metastases in central nervous system, bones, adrenal glands,
of systemic therapy or marrow stimulant, to allow marrow and spleen occur in less than 10% of CRC patients (Kemeny
effects and associated FDG avidity to subside. FDG is excreted et al, 2014).
through the urinary tract; the radioactive signal from excreted- For CRC staging, FDG PET/CT is currently considered a
FDG in the collecting systems typically obscures PET evalua- diagnostic adjunct to dedicated structural CT and MRI, which
tion of the kidneys and urinary bladder. The liver, lungs, and remain primary and indispensable imaging modalities for their
other tissues have lesser degrees of background FDG avidity superior spatial resolution, providing the most precise structural
that usually do not obscure tumor detection significantly. detail about a patient’s anatomy that can currently be obtained
The diagnostic specificity of FDG PET is limited by the noninvasively, information essential for surgical planning and
existence of other pathologies that demonstrate marked FDG fro evaluating tumor structural response (shrinkage or growth)
avidity, which may mimic neoplastic disease on scintigraphy. to therapy. Despite its inferior spatial resolution, however, FDG
Such etiologies broadly include various types of infectious or PET/CT is often found to be the most sensitive modality for
other inflammatory diseases and hyperplastic or dysplastic detection of colorectal metastases, which radiologically is the
conditions. Certain normal variants in FDG biodistribution or primary determinant of resectable versus nonresectable disease.
a focal accumulation of excreted FDG in the urinary tract may FDG PET/CT had the highest sensitivity for detection of
also mimic an FDG-avid neoplasm. colorectal liver metastases (mean weighted sensitivity of 90%-
Besides tumor detection, for disease (re)staging, FDG PET 92%), with equivalent overall specificity, compared with US,
may be used to evaluate tumor response to cancer therapy. CT, and MRI in a meta-analysis (Kinkel et al, 2002) cited by
Frequently, FDG PET is performed twice, before and after current EURECCA guidelines (van de Velde et al, 2014). It
therapy, for comparison, using changes in tumor FDG avidity should be noted that a meta-analysis from 2002 would include
as an index of changes in tumor cell population size (i.e., CT and MRI techniques that are no longer considered state of
tumor response). Marked decreases in tumor FDG avidity the art. Regardless, FDG PET/CT currently is not considered
part of conventional CRC staging; rather, guidelines suggest

that PET/CT should be used on an as-needed basis to clarify
CT or MR findings indeterminate for metastatic disease
(NCCN, 2015).
To be cited as “relevant” for guidelines describing state-of-
the-art initial CRC staging, referenced FDG PET/CT clinical
research should involve (1) only CRC patients, (2) only pre-
treatment FDG PET/CT, and (3) only PET/CT, especially with
current PET/CT hardware and software methodology, and
certainly not PET only (recognized as diagnostically inferior to
PET/CT).
For detection of colorectal metastases, FDG PET/CT per-
forms best in the pretherapy (initial staging) setting (Akhurst
et al, 2005; Nahas et al, 2008) and is extremely sensitive in
detection of liver and lung metastases(Nahas et al, 2008), the
two most common viscera to be involved by metastatic disease
at initial presentation (Jemal et al, 2006).
In general, FDG PET/CT has long been recognized as
superior to conventional CT for detecting hepatic metastases
(Abdel-Nabi et al, 1998; Even-Sapir et al, 2002; Kinkel et al,
2002), with decreasing sensitivity for both modalities in char-
acterizing subcentimeter hepatic lesions (Fong et al, 1999;
Ruers et al, 2002). FDG PET/CT and MRI are equivalent in
diagnostic sensitivity for liver metastases, at least regarding
whether liver metastases are present or absent, for an individual
patient. MRI appears to identify more subcentimeter liver
lesions than FDG PET/CT (i.e., with greater sensitivity).
FIGURE 17.2. Fluorine-18 fluorodeoxyglucose (FDG) PET scan of
Current NCCN guidelines for CRC do not advocate routine
colon cancer patient on chemotherapy treatment, being considered for
use of FDG PET/CT for initial staging or posttreatment local treatment of known liver metastases. This three-dimensional
follow-up surveillance of patients with no evidence of distant maximal intensity projection (MIP) PET image spans from midskull region
metastatic disease by standard IV contrast-enhanced (CE) CT to proximal thigh regions. Two known liver metastases demonstrate focal
or by plasma tumor biomarker assays, except potentially as an hypermetabolic activity (arrowheads). FDG PET also detects extrahe-
arbiter for equivocal CT findings, or if a patient cannot have patic disease, a subcentimeter right adrenal lesion with intense hyper-
iodinated IV contrast for CT (Abdel-Nabi et al, 1998). NCCN metabolic activity (arrow; standardized uptake value, 14.2), new
guidelines advise clinicians to consider FDG PET/CT for the compared to a prior FDG PET scan, missed by recent standard com-
workup of patients with potentially resectable metastases (syn- puted tomography, and consistent with a metastasis.
chronous or metachronous) to avoid futile surgery (NCCN,
2015). ESMO 2014 colorectal cancer guidelines agree with
current NCCN guidelines regarding the role of FDG PET-CT
(van Cutsem et al, 2014). refute) suspicious findings on FDG PET that lack a clear cor-
Patients planning to undergo hepatic resection based on relate on CT.
conventional imaging will be found to have extrahepatic disease The NCCN guidelines suggest that FDG PET/CT provides
by FDG PET in 18% to 32% of cases (Fig. 17.2) (Boykin et al, no clinical added value for evaluating the response of colon
1999; Lai et al, 1996; Ruers et al, 2002), changing manage- cancer to chemotherapy. The guideline authors noted the
ment in 20% to 40% of cases in early clinical PET trials (Fong potential for false-negative and false-positive results, citing a
et al, 1999; Ruers et al, 2002) and changing management in single review article that, contrary to the guidelines, suggests
curative-intent surgery in as many as 25% of patients in a later that FDG-PET “seems promising for monitoring [colorectal
trial (Joyce et al, 2006). Fernandez and colleagues (2004) carcinoma] patient response to therapy but larger studies are
found that patients with hepatic metastases who underwent necessary” (Delbeke & Martin, 2004).
FDG PET for preoperative staging had a much higher 5 year NCCN guidelines also suggest clinicians consider FDG
survival rate than historic controls. A randomized trial of PET/CT for evaluation of patients with serial elevations of
patients with resectable metachronous metastases found no serum carcinoembryonic antigen (CEA) levels without an
impact of FDG PET/CT on survival, but surgical management identifiable source after standard IV contrast CT. A systematic
was changed by PET/CT in 8% of patients (resection aborted review and meta-analysis of 11 trials (510 patients) reported
because PET/CT revealed nonresectable disease in 2.7%; PET pooled estimates of sensitivity and specificity for FDG PET/
increased the extent of liver resection, underestimated by CT for detection of occult tumor recurrence (after conventional
conventional CT, in 1.5%; additional organ surgery undertaken, CT workup) of 94% (95% confidence interval [CI], 89%-97%)
guided by PET/CT, in 3.4%). However, PET/CT had false- and 77% (95% CI, 66%-86%), respectively (Lu et al, 2013).
positive results in an additional 8.4% of patients, indicating the
need for caution, and often biopsy or additional imaging, before Unresectable Liver Cancer
allowing PET/CT findings to determine resectability (Moulton Although surgical resection often provides the best patient
et al, 2014). For example, MRI can be used to confirm (or outcomes for cancerous liver tumors, some patients are not
surgical candidates (e.g., due to extent of tumor, underlying concordant with 2015 NCCN guidelines, suggesting a limited,
liver disease, or comorbid conditions). For the patient with an as-needed role for nuclear imaging in liver cancer staging.
unresectable primary or secondary liver malignancy, nuclear One potential rationale for advising no routine role for FDG
medicine offers both diagnostic and therapeutic options. PET in HCC is that HCC tumors can be undetectable (false
Radionuclide therapy with yttrium 90 (90Y)–labeled resin negative) on FDG PET. An untreated, fully viable HCC tumor
or glass microspheres, delivered by hepatic intraarterial cath- mass, easily detectable by CT or MRI, may concentrate FDG
eter infusion as a radioembolization treatment for liver tumors, no more avidly than surrounding normal, nontumorous liver,
has FDA-approved indications for patients with liver-only or and thus the HCC can appear indistinct or invisible on PET.
liver-predominant tumor burden (see Chapter 96B). Soma- In other HCC patients, tumors can demonstrate detectable,
tostatin receptor (SSTR)–targeted radionuclide therapy is a visually distinct FDG-avidity on PET; even in such cases,
treatment for malignant carcinoid tumors, including primary however, the visual intensity of the HCC tumor uptake is only
or secondary liver tumors (and not restricted to liver-only/ mildly or moderately greater than the uptake in the surrounding
predominant disease) (see Chapter 93). Considered investiga- normal liver. Research suggests potential biologic explanations
tional in the United States but more widely available for the typically low FDG avidity of HCC, notably HCC
in Europe, SSTR-targeted radionuclide therapy is discussed overexpression of the enzyme glucose-6-phosphatase (thus
later. inhibiting metabolic trapping of FDG as FDG-6-phosphate,
FDG PET/CT has been suggested as a promising radiologic which is otherwise a poor substrate for glucose-6-phosphatase)
assay for detecting a favorable liver tumor response to 90Y (Torizuka et al, 1995). In particular, well-differentiated HCC
microsphere radioembolization (Kennedy et al, 2012), as well can demonstrate this non–FDG-avid PET phenotype, whereas
as for evaluating hepatic tumor response to other forms of higher-grade HCC tumors more frequently have detectable
treatment. However, current hepatobiliary oncology guidelines FDG avidity (Lee et al, 2004; Torizuka et al, 1995). Therefore,
cite insufficient published clinical evidence to support routine although a lack of FDG avidity may be frustrating from a
use of FDG PET/CT in hepatobiliary tumor-response evalua- diagnostic viewpoint, it also appears to be associated with a less
tion, whether for unresectable or resectable tumors. Standard aggressive tumor and a more favorable patient prognosis
CE CT and MRI are considered indispensable for tumor (Shiomi et al, 2001).
assessment, and FDG PET/CT is still deemed to be of unproven Another potential rationale for advising no routine role for
added value. We expect future guidelines will deem use of FDG FDG PET in HCC is that HCC is usually a liver-limited
PET/CT for evaluation of hepatic tumor “metabolic response” disease. The main clinical impact of FDG PET/CT in other
as appropriate for tumors that demonstrate FDG avidity. types of cancers (e.g., non–small cell lung cancer) is often in
Additionally, FDG PET/CT imaging is currently use within the detecting regional and distant metastases.
interventional radiology (IR) suite (with the scanner located Beyond FDG, separate PET clinical trials with investiga-
physically within the IR suite) of a few major medical centers tional radiopharmaceuticals carbon 11 (11C) acetate (Hwang
as an investigational procedure to assess the effectiveness of et al, 2009) and 11C choline (Yamamoto et al, 2008) provided
percutaneous tumor ablation therapy immediately after abla- preliminary data suggesting that these novel tracers are better
tion, permitting immediate repeat tumor ablation as needed than FDG PET for detection of well-differentiated HCC. At
(Ryan et al, 2013). This technique should gain acceptance as the same time, FDG PET appeared better than both novel
clinical outcome data accumulate. tracers for detection of poorly differentiated HCC, suggesting
Nuclear scintigraphy with a SSTR-targeted imaging agent, FDG PET and either tracer may be complementary options for
notably indium 111 (111In)–radiolabeled octreotide (111In pen- HCC scintigraphy. However, 11C has a physical half-life of only
tetreotide), has potential clinical application in evaluation of 20 minutes, precluding commercial supply of 11C imaging
liver tumors, usually for detection of carcinoid tumors, which agents. To perform patient studies with 11C PET tracers, a
may be undetected by FDG PET. As with FDG PET, 111In medical center must have a multimillion-dollar cyclotron facil-
octreotide scintigraphy is not recommended by consensus ity and radiochemistry expertise to synthesize 11C tracer on site.
guidelines for evaluation of hepatic tumor response, although
it is typically used to predict carcinoid tumor response to Tumor-Response Evaluation
(nonradioactive) therapy with octreotide analogues (discussed As stated, no current major guidelines advocate routine use of
later). nuclear imaging, including PET, for HCC tumor imaging,
including evaluating tumor response to therapy. FDG PET/CT
has become a tumor-response radiologic biomarker with a major
Management of Patients With
clinical impact on the management of a growing number of
Hepatocellular Carcinoma other cancers, now standard for breast cancer, lymphoma,
Initial Presentation (See Chapter 91) esophageal cancer, gastric cancer, and other malignancies. In
Hepatocellular carcinoma (HCC) is detected and characterized these types of cancer, overall robust clinical literature demon-
by US, CT, and MRI in standard practice. NCCN guidelines strates that a therapy-induced decrease in tumor FDG avidity,
(2015) do not provide a role for FDG PET/CT in HCC, as a biomarker of a decrease in tumor metabolism and tumor
without discussion of rationale. NCCN guidelines advise con- cell mass, correlates with histopathologic response and is often
ventional bone scintigraphy with technetium 99m (99mTc)– prognostically powerful, often when conventional changes in
radiolabeled bisphonates, such a methylenediphosphate (MDP), tumor volume (e.g., anatomic criteria such as RECIST) mea-
for staging of HCC patients with suspicious skeletal symptoms. sured by CT or MRI failed to predict either pathologic response
The 2012 European Association for the Study of the Liver or prognosis (Wahl et al, 2009).
(EASL) and the European Organization for Research and Prerequisites for using FDG PET/CT to assess tumor
Treatment of Cancer (EORTC) joint guidelines are generally response to treatment are that the tumor demonstrates FDG
avidity before treatment, as a biomarker of tumor metabolic PET/CT has reported sensitivity no greater than 90% (usually
activity, and that the FDG avidity is sufficiently high so that much lower) for HCC in the pretreatment setting.
PET can detect a therapy-induced decrease in tumor FDG
avidity, as a biomarker of a prognostically favorable decrease in Cholangiocarcinoma (See Chapters 50 and 51)
tumor metabolism and tumor cell mass. For example, well- Cholangiocarcinoma is the second most common primary
designed PET response criteria called the PERCIST guidelines hepatic tumor. Cholangiocarcinoma is usually an FDG-avid,
(Wahl et al, 2009) define sufficient pretreatment FDG avidity PET-positive disease. A 2014 meta-analysis of 10 studies found
as being tumor FDG avidity (quantified by the standardized that FDG PET/CT, on a per-patient basis, had an overall
uptake value [SUV]) that is greater than 1.5 times higher than sensitivity of 82% and specificity of 75% for detection of
the PET-measured average FDG avidity (SUV) of the individual cholangiocarcinoma, and diagnostic sensitivity was higher for
patient’s liver plus 2 standard deviations (SD) of that average intrahepatic tumors than for perihilar and other extrahepatic
liver avidity (the SD figures automatically calculated by PET tumors (Annunziata et al, 2014). FDG-avid inflammation
analysis software). along the biliary tree (i.e., cholangitis) can mimic or obscure
Whereas other cancer types (e.g., breast, lymphoma) are detection of FDG-avid biliary tumor (see Chapter 43).
frequently FDG avid and PET scan positive, HCC tumors FDG PET/CT appears sensitive for detection of intrahepatic
frequently demonstrate minimal to nil FDG avidity, undetect- and extrahepatic metastases. Although current NCCN guide-
able or barely detectable above the background liver FDG lines state that the role of FDG PET/CT in management of
avidity. For HCC, obtaining FDG PET/CT before any particu- biliary cancers has overall “not yet been established,” these
lar form of treatment would be necessary to establish that the same guidelines also state “emerging evidence from retrospec-
HCC tumor is sufficiently FDG avid for PET-response evalu- tive studies indicates that it may be useful for the detection of
ation, although the cost-effectiveness of such an approach regional lymph node metastases and distant metastatic disease
might be questionable, with a single FDG PET/CT scan rela- in [biliary] cancer patients with otherwise potentially resectable
tively expensive (several thousand 2015 US dollars). Because disease” (NCCN, 2015).
HCC tumors are frequently PET scan negative, many baseline
PET scans with no potential value for evaluating subsequent Focal Liver Lesions
tumor response would be obtained. A cost-effective approach The differential diagnosis of a focal liver lesion, or collection of
may be to find a combination of clinical, histologic, and labora- liver lesions, generally includes tissue entities such as focal
tory blood test biomarkers that reliably predict that a particular nodular hyperplasia (FNH), hepatic adenoma, HCC, fibrola-
patient’s HCC tumor will be sufficiently FDG avid before PET/ mellar carcinoma, regenerative nodules, hemangioma, and
CT. This approach has been demonstrated in FDG PET in metastases (see Chapter 90A). Characterization and usually
prostate cancer, in which blood test biomarkers (specifically, initial detection of liver lesion(s) usually begin with structural
serum PSA-doubling time) predict tumor FDG avidity and imaging modalities: CT, US, and MRI (Agency for Healthcare
PET scan positivity, allowing a more patient-specific, cost- Research and Quality [AHRQ], 2014). Structural imaging
effective application of FDG PET in cancer management characteristics in proper clinical context may be sufficient for
(Schoder et al, 2005). diagnosis (e.g., as with simple hepatic cysts or liver hemangio-
Despite the frequent PET negativity of HCC, a few groups mas). For more solid liver lesions, blood tests may refine the
have demonstrated the ability of PET to evaluate the efficacy diagnosis (e.g., serum α-fetoprotein for suspected HCC), or
of certain types of treatments against FDG-avid HCC tumors, tissue-biopsy may be indicated, if structural imaging character-
such as transcatheter arterial (transarterial) chemoembolization istics are suspicious for cancerous lesions. The role of scintig-
(TACE) (Torizuka et al, 1994). raphy in characterizing liver lesions is typically limited and
confined to second-line imaging (AHRQ, 2014), but it can be
Recurrence helpful in selected cases when the differential diagnosis has
Hepatocellular carcinoma usually arises within a background been narrowed by initial evaluation of the liver lesion by CT/
of liver cirrhosis (of viral or other etiology), and metastatic US/MRI, clinical history, and blood test findings.
disease tends to be entirely or predominantly intrahepatic (see American College of Radiology (ACR) Appropriateness
Chapters 76 and 89). Similarly, recurrence typically manifests Criteria for characterization of newly discovered lesions (AHRQ,
as tumor regrowth at a prior site of treatment (e.g., ablation) 2014) suggest a limited role for nuclear scintigraphy, with CT
or as tumor appearing at a new site in the liver. Hayakawa and or MRI always the preferred initial imaging modality (see Chap-
colleagues evaluated FDG PET/CT for detecting recurrent ters 18 and 19). For a patient with a known extrahepatic malig-
HCC postoperatively in patients with either suspected recur- nancy, if CT or MRI characterizes a liver lesion as indeterminate,
rence on CT or MRI (group 1) or suspected recurrence because FDG PET/CT is considered potentially helpful for further
of abnormal serum tumor markers, but with no disease evident characterization, especially if the patient had a prior FDG PET/
on CT or MRI (group 2). FDG PET/CT had a 53 and 41% CT demonstrating the extrahepatic cancer was FDG avid (pre-
sensitivity and 100% specificity for recurrent tumor in both dicting that the liver lesion, if a metastasis, also would be FDG
groups 1 and 2, respectively. The data from group 1 support avid). Alternatively, radiolabeled SSRT-targeted scintigraphy is
the idea that FDG PET/CT cannot replace CT or MRI as the considered potentially useful if the extrahepatic cancer was a
first-line imaging modality for detection of recurrence; the data known neuroendocrine tumor (NET), particularly if the NET
from group 2 invite the hypothesis that FDG PET/CT may be was viable (e.g., pretreatment) but demonstrated a non–FDG-
of value as second-line imaging if CT or MRI fails to detect avid/PET-negative scan phenotype. FDG PET–negative NETs
recurrence (Hayakawa et al, 2014). Wang and colleagues (2013) tend to be better differentiated and associated with a more
reported a remarkably high sensitivity of FDG PET/CT for concentrated accumulation of SSTR-targeted tracer (see
HCC detection of 97%, with a specificity of 83% (n = 32). FDG Chapter 93).
Diagnostic nuclear medicine imaging for cancerous liver than 1.5 cm, which is smaller than the spatial resolution of
tumors most importantly includes the common PET imaging SPECT camera systems; this increases partial volume–averaging
agent 18F FDG, and for hepatic metastases of NETs, SSTR effects that diminish the detectability of tracer uptake by
scintigraphy with the SPECT agent 111In pentetreotide or colloid-avid FNH lesions, amid surrounding hepatic uptake.
68
Ga-radiolabeled octreotide analogues (e.g., DOTA). However, Some FNH lesions, including some greater than 1.5 cm, appear
benign tumors and nonneoplastic lesions can be FDG avid, to accumulate colloid tracer poorly, less than in surrounding
mimicking FDG-avid tumors. Similarly, note that a variety of normal liver, appearing as a focal paucity of tracer uptake,
non-NET cancers can also be avid for SSTR-targeted tracers sometimes referred to as a “cold spot.” Other FNH lesions
(e.g., HCC, prostate cancer, certain types of lung cancer) and accumulate SC equivalent to surrounding normal-liver colloid
possibly certain relevant benign liver tumors as well (see later). uptake, appearing indistinguishable from surrounding liver
Thus, the ACR guidelines advise FDG PET/CT as potentially (i.e., invisible) on 99mTc SC SPECT. The indistinguishable
useful for patients with known extrahepatic primary cancers and uptake of such invisible lesions is sometimes called “warm,”
SSTR-targeted scintigraphy (e.g., OctreoScan) for patients meant to imply that it is neither a hot spot nor a cold spot.
with a known history of NET, both to increase pretest likeli- “Warm” is a vague term likely to be misinterpreted (e.g., as
hood that any FDG avidity or SSTR-targeted tracer avidity meaning uptake that is not “hot” but still faintly detectable as
detected in liver lesions can be ascribed as more likely attribut- above liver-background lesion uptake). We prefer the scinti-
able to tracer-avid hepatic metastases. The improve the posttest graphic term isointense to describe lesion uptake that is equiva-
likelihood of FDG-avid liver lesions representing true-positive lent to and indistinguishable from surrounding normal-liver
liver metastases in patients with a colorectal primary, correla- uptake.
tion with serum CEA levels is useful. The specificity of 99mTc SC SPECT for FNH, using a visible
Certain benign, focal liver lesions (neoplastic and nonneo- hot spot as the diagnostic criterion, is very high; to our knowl-
plastic) are well recognized to concentrate 18F FDG, mimicking edge, no other non-FNH liver lesions (e.g., hepatic adenoma)
the appearance of an FDG-avid cancerous liver tumor (see concentrate SC sufficiently to produce a hot spot on 99mTc
Chapter 90A). Other types of tumors, not of neuroendocrine SPECT. Various non-FNH lesions also can accumulate SC
cell origin, can avidly concentrate the octreotide radiotracer, equivalent to normal-liver colloid uptake. Again, the indistin-
including HCC (variably) (Dimitroulopoulos et al, 2007). The guishable uptake of such “invisible” lesions refers to uptake by
octreotide-analogue radiotracers target SSTRs, particularly 2, lesions that was isointense with surrounding normal-liver
3, and 5. It must be remembered that although SSTRs are uptake.
expressed at high tissue concentrations in NETs, the SSTRs As an alternative to 99mTc SC for scintigraphic detection of
are not mutant, cancer-specific “oncoproteins” or even neuro- FNH, others have tested cholescintigraphy, the use of biliary
endocrine tissue specific. Whether SSTR expression is also high tracers such as 99mTc-radiolabeled disofenin or mebrofenin (see
for the common differential of a focal liver lesion (e.g., FNH) Hepatobiliary Scintigraphy, later). Hepatocytes in the FNH
is unclear. Few reports detail abnormal concentrations of lesion produce bile, as usual, and thus will uptake and secrete
radiolabeled octreotide analogues in benign liver lesions; one these biliary tracers. However, biliary canaliculi within the
described an absence of radiolabeled octreotide binding to FNH lesion are often characterized by disordered drainage
hepatic adenoma, but it would be a non sequitur to infer that connections with the remainder of the intrahepatic biliary tree.
false-positive results do not occur only because no literature As such, the classic (chole)scintigraphic appearance of FNH is
states that they do occur (Reubi et al, 1999). Other data, if abnormal focal accumulation and prolonged focal retention of
available, could hypothetically overturn the conclusions of the a biliary radiotracer. Small studies suggest cholescintigraphy
scant small studies discovered. Therefore, the clinician who has promising diagnostic sensitivity (e.g., 92%) (Boulahdour
wants to order an octreotide tracer scintigraphy scan to char- et al, 1993). Well-differentiated HCC can demonstrate a similar
acterize a hepatic lesion as an NET should consider pretest cholescintigraphic phenotype, as an abnormal focal, prolonged
likelihood and possible use of scintigraphy to guide tissue retention of biliary tracer. Some propose FNH and well-
biopsy confirmation. differentiated HCC differ in cholescintigraphic phenotype,
If structural imaging studies have narrowed the differential suggesting that HCC will usually clear biliary tracer more
diagnosis down to a suspicion of FNH (see Chapter 90A) but quickly than will FNH (e.g., when assessed at 5-10 minutes
diagnostic uncertainty remains, scintigraphy with technetium after injection). Whether cholescintigraphy can reliably differ-
99m–radiolabeled sulfur colloid (99mTc SC) may be useful. entiate FNH from HCC is subject to question, given limited
SPECT imaging with 99mTc SC visualizes FNH as a focus of clinical data. The FDG PET appearance of FNH is variable but
abnormally high SC concentration, attributable to a high typically “mild” or indistinct scintigraphically. FNH can appear
concentration of phagocytic Kupffer cells accumulating 99mTc mildly hypermetabolic (or FDG avid), mildly hypometabolic,
SC within a scintigraphically detectable FNH lesion, with rela- or isointense with surrounding hepatic FDG uptake. In our
tively less colloid accumulation in surrounding normal liver experience, FNH lesions are rarely, if ever, intensely FDG avid
parenchyma. Thus, the FNH lesion manifests as a hot spot on (Kurtaran et al, 2000).
99m
Tc SC SPECT. The sensitivity of 99mTc SC 3D SPECT is Scintigraphic characteristics of a hepatic hemangioma (see
superior to 2D planar scintigraphy (thus SPECT is standard). Chapter 90A), another notable entity on the differential diag-
However, the sensitivity of 99mTc SC SPECT for FNH, using nosis of a hepatic lesion, includes its absence of sulfur colloid
a visible “hot spot” as the diagnostic criterion, is poor, probably uptake, appearing as a cold spot on 99mTc-labeled SC scintig-
approximately 50% to 60%, suggesting that not all FNH lesions raphy, and tracer activity on FDG PET that is equivalent or
have relatively high concentrations of phagocytic Kupffer cells less than cardiac blood pool activity and also generally less than
compared with surrounding liver (Chung et al, 2010). The surrounding liver FDG uptake. Hemangioma also demonstrates
sensitivity is even poorer if the hepatic lesion in question is less a scintigraphic hot-spot appearance on radiolabeled red blood
cell (RBC) studies. These scintigraphic phenotypes are usually contract, can help identify and separate coexistent pancreatic
distinctly evident for lesions greater than 1.0 to 1.5 cm; smaller disease processes.
lesions tend to blend into surrounding hepatic tracer uptake, While the term intense as a descriptor of hypermetabolic
due to partial volume averaging associated with the limited activity is subjective, the use of a quantitative parameter such
spatial resolutions of PET and particularly SPECT. The liver as the conventional SUV does not appear to improve diagnostic
hot-spot appearance, on 99mTc-labeled RBC scintigraphy, is accuracy and may actually lower it. Moreover, any particular
considered almost pathognomonic for a hemangioma (specific- SUV threshold defined by one medical center’s experience, as
ity ~100%), although case reports do describe other pathologic best separating cancer from noncancer, will not necessarily be
entities (e.g., liver angiosarcoma) also demonstrating a liver applicable to another medical center, considering the technical
hot-spot appearance. Hemangiomas larger than 1.4 cm can differences and lack of mandatory calibrations among the
be detected by 99mTc-labeled RBC scintigraphy with high numerous models of PET/CT scanners. Also, some noncancer-
sensitivity. ous pancreatic lesions (e.g., infectious) will appear more FDG
avid than some cancerous pancreatic lesions (e.g., predomi-
nantly cystic). Assessment of a pancreatic lesion requires con-
DIAGNOSIS AND FOLLOW-UP sideration of the CT appearance of the lesion, for which a CE
OF PANCREATIC DISEASES CT is preferred, for optimal anatomic characterization. If the
lesion is predominantly cystic, one expects that even a cancer-
Pancreatic Cancer (See Chapters 59 and 62) ous tumor may be FDG PET negative, although even solid
Detection noncancerous pancreatic lesions can be as FDG avid as cancer-
For identification of pancreatic cancer, a 2012 meta-analysis of ous solid pancreatic lesions (e.g., benign solid pseudopapillary
16 studies with 804 patients found FDG PET/CT offered a tumor) (Kim et al, 2014). In fact, if a solid tumor has activity
diagnostic sensitivity of 87% and specificity of 83% (Wu et al, close to blood pool activity (i.e., minimal FDG-avidity), it is
2012). The likelihood ratio for a suspicious lesion on PET/CT more likely to be a pancreatic adenocarcinoma or NET than a
was 5.84 (95% CI, 4.59, 7.42) and, if PET was negative, the solid pseudopapillary tumor, which are usually FDG avid
ratio was 0.24 (95% CI, 0.17, 0.33). A 2014 meta-analysis (Guan et al, 2013).
(including much of same data as 2012 analysis) obtained similar Optimal PET/CT analysis therefore appears to be predomi-
results: FDG PET/CT for evaluation of suspected pancreatic nantly a qualitative assessment of CT appearance (cystic or
cancer was calculated to offer pooled sensitivity of 90% and solid), extent on PET (focal vs. diffuse pancreatic FDG-avid
specificity of 76% (Rijkers et al, 2014). disease), and relative intensity on PET (greater FDG avidity
Clinical reports consistently indicate that FDG PET/CT generally increasing suspicion for cancer). However, the diag-
that includes a standard-quality, CE CT scan offers superior nostic accuracy of FDG PET/CT for characterizing a pancreatic
diagnostic accuracy compared with PET/CT performed with a mass as cancer (or not) is limited.
noncontrast, low-mA-dose, “nondiagnostic” CT protocol, with
diagnostic sensitivity of 91% (95% CI, 86%-96%) versus 88% Staging (See Chapter 62)
(95% CI, 78%-100%), and specificity of 88% (95% CI, 73%- Clinical data on the potential added value of FDG PET/CT in
100%) versus 81% (95% CI, 69%-94%), respectively, although staging of pancreatic cancer, beyond that offered by standard
these were not statistically significant differences (P > .05) (Wu radiologic imaging (CT, MRI, US), is considered by standard
et al, 2012). This type of nondiagnostic CT protocol, as men- guidelines to be insufficient to recommend FDG PET/CT as a
tioned, is a standard approach for fusion PET/CT imaging, standard component of pancreatic cancer staging. However,
where the study is primarily performed for the sake of obtaining 2015 NCCN guidelines state “[FDG] PET/CT may be con-
PET data, and a standard CT will be obtained or has already sidered after formal pancreatic CT protocol in high-risk patients
been obtained separately. to detect extra-pancreatic metastases;” high-risk status includes
A variety of noncancerous pancreatic diseases can cause patients with borderline resectable disease, “markedly” elevated
hypermetabolic pancreatic disease, visualized by PET variably serum CA19-9, large primary tumors, or large regional lymph
as focal hypermetabolic lesions or relatively diffusely hyper- nodes (on CT) (NCCN, 2015). The main impact of FDG
metabolic pancreatic disease (Santhosh et al, 2015). Literature PET/CT on pancreatic cancer is to upstage patients by accurate
reports of false-positive FDG PET include noncancerous enti- detection of distant metastases not detected by previous con-
ties such as focal acute and chronic pancreatitis, mass-forming ventional CT or MRI. FDG PET/CT offers an overall diagnos-
pancreatitis, autoimmune pancreatitis, and infectious pancre- tic specificity for metastases typically reported as 90% or higher
atic disease (see Chapters 55 and 57). (Wang et al, 2014) and changed resectability status in more
A focus of qualitatively “intense” hypermetabolic activity in than 20% to 30% of patients in two studies (Bang et al, 2006;
the pancreas, however, has often been reported to have a rela- Chang et al, 2014).
tively high diagnostic accuracy for pancreatic cancerous lesion, One potential limitation of PET/CT for pancreatic cancer
with a sensitivity and specificity of 90% and 93%, respectively, staging is the companion CT, often a low-dose, noncontrast,
in one study (Santhosh et al, 2013). Following the “focal” “nondiagnostic” image. The nondiagnostic companion CT is
criterion strictly can reduce diagnostic sensitivity, if applied often used because it is suitable for providing anatomic localiza-
indiscriminately; for example, a diffuse or extensive hyper tion and attenuation correction of PET data, while exposing
metabolic pancreatitis (e.g., from pancreatic duct obstruction) the patient to less radiation dose (mA) than conventional CT.
can occur with a focal hypermetabolic pancreatic cancerous However, the relatively low mA dose and lack of IV contrast
tumor, with the hypermetabolic disease appearing inseparable make the companion CT truly nondiagnostic for pancreatic
and nonfocal overall on FDG PET. Close examination of the cancer tumor (T) staging (i.e., T designation of standard TNM
companion CT or a recent CT, standard quality with IV classification) (AJCC, 2010). Nonenhanced companion CT
often fails to detect vascular invasion (Strobel et al, 2008), and PET/CT for detecting pancreatic cancer response indicates that
PET alone is typically inferior to CT for assessment of size it is superior in general to conventional CT or MRI for predict-
because of its inferior spatial resolution. ing favorable histopathologic response and for predicting
The potential advantage of PET/CT for staging is in detect- patient prognosis (i.e., stratifying patient outcomes). Multiple
ing nodes and metastases (N and M). Available studies to date clinical trials have shown that changes in pancreatic tumor
report a wide range of sensitivity values for FDG PET/CT for hypermetabolic activity, measured by comparing a pretreat-
detection of pancreatic cancer nodal metastases, ranging ment PET scan with a posttreatment or on-treatment PET
from 0% to 57% in a 2014 literature review by Wang and scan, correlated with (predicted) histologic response, radiologic
colleagues. The variation is probably indicative of the small size (CT or MRI) response, or patient response (e.g., overall survival
of the studies cited (n = 14 or less, per study), but suggests or progression-free survival) (Wang et al, 2014). Notably, in
that PET/CT nodal (N) sensitivity is much too low to guide patients with locally advanced unresectable pancreatic carci-
surgical decision making on the necessity and extent of noma (n = 32), Topkan and colleagues (2011) showed that
lymphadenectomy. those with more than average decreases in tumor hypermeta-
For metastasis (M) staging, FDG PET/CT appears to have bolic activity (i.e., SUV parameter) after chemoradiotherapy
relatively high specificity for detection of distant metastases, with 50.4 Gy (1.8 Gy per fraction), 5-fluorouracil (5-FU), and
with specificity values as high as 91% to 100% (Wang et al, gemcitabine enjoyed a superior median overall survival (17.0
2014). For liver metastases, a common site for pancreatic vs. 9.8 months; P = .001), progression-free survival (8.4 vs 3.8
metastases, the reported sensitivity of FDG PET/CT varies months; P = .005), and locoregional progression-free survival
widely between medical centers, from 22% to 88%. Liver (12.3 vs. 6.9 months; P = .02). The predictive value of a favor-
metastases can be subcentimeter, and PET has lesser sensitivity able “metabolic response” on FDG PET/CT remained statisti-
for subcentimeter tumors because of partial volume averaging cally significant for each of these three outcomes on multivariate
(discussed earlier). Also, the normal liver accumulates FDG analysis. Others have reported statistically significant correla-
physiologically. While normal hepatic FDG uptake is usually tions between favorable FDG PET/CT tumor metabolic
considered relatively low in its signal intensity by PET readers response to chemoradiotherapy and favorable survival outcome
(e.g., compared with the intense physiologic accumulations of in unresectable, locally advanced pancreatic cancer as well
FDG in the normal brain or sometimes the myocardium), the (Bang et al, 2006; Chang et al, 2014; Schellenberg et al, 2010).
hepatic uptake is sufficiently high that it can reduce visibility of One study showed similar predictive power of FDG PET/CT
liver metastases, in terms of tumor/background contrast on for treating unresectable disease with chemotherapy (without
PET imagery, particularly for pancreatic tumors that also radiation) (Maisey et al, 2000).
demonstrate relatively low-level FDG avidity. Despite the vari- For resectable pancreatic cancer, FDG PET/CT accurately
ability in reported percentages, clinical studies predominantly predicts histologic response to neoadjuvant chemotherapy with
report that FDG PET/CT seems to perform better than con- gemcitabine and cisplatin, although neither PET response nor
ventional CT for detection of liver metastases, although most histologic response correlated with survival outcomes (Heinrich
were small studies, with fewer than 20 patients, limiting statisti- et al, 2008) (see Chapters 62 and 68). This particular neoad-
cal confidence (Wang et al, 2014). MRI appears to offer juvant treatment was clearly able to eradicate a majority of
superior diagnostic sensitivity for liver metastases compared tumor burden in some patients, but one can hypothesize that
with FDG PET/CT, on a per-lesion basis, particularly for the effect was insignificant compared to the impact of the
detection of subcentimeter liver metastases. subsequent surgery. The neoadjuvant therapy might have had
The sensitivity of FDG PET/CT for detection of distant an impact on residual microscopic tumor burden; both neoad-
metastases in the bones is clearly superior to standard CT. Bone juvant responders and nonresponders had microscopic disease
metastases may be associated with only subtle osteolytic changes after surgery, but the responders had less microscopic disease.
on CT or may be entirely invisible; CT-occult metastases pre- Differences in microscopic amounts of residual disease likely
dominantly involve those infiltrating the bone marrow. MRI is produce survival benefits not easily detected by such a small
at least as diagnostically accurate as FDG PET/CT for detec- study (n = 24). The study results strongly support use of FDG
tion of osseous metastases. Body MRI is not a standard workup PET/CT as an investigational clinical response biomarker in
for pancreatic cancer, whereas FDG PET/CT typically offers a pancreatic cancer. This study demonstrates that FDG PET/CT
torso field of view. At other sites, such as lungs and peritoneum, can noninvasively detect pancreatic cancer histologic response
the diagnostic accuracy of FDG PET/CT appears similar to at the end of neoadjuvant therapy. Future pancreatic cancer
that of standard CT; a dedicated CE CT scan improves PET/ studies should explore whether tumor FDG PET/CT response
CT detection of peritoneal metastases significantly compared early during neoadjuvant therapy predicts subsequent histologic
to PET with a nondiagnostic companion CT. response at the end of neoadjuvant therapy, as demonstrated in
For radiation therapy planning with an image-guided other cancer types. FDG PET/CT appears to be an excellent
intensity-modulated approach, FDG PET/CT assists in optimal noninvasive biomarker for identifying ineffective pancreatic
contouring of the radiation field, better defining pancreatic cancer neoadjuvant therapy and might be used to identify poor
gross tumor volume and pathologic adenopathy, in conjunction tumor response early during treatment, avoiding needless
with dedicated structural imaging (CT or MRI), than possible neoadjuvant toxicity and enabling alternative neoadjuvant regi-
with structural imaging alone (NCCN, 2015). mens to be pursued.
Pancreatic cancer recurrence can be detected by elevation
Pancreatic Tumor Response and Recurrence of serum levels of the tumor marker with high diagnostic sen-
NCCN guidelines do not currently advise a role for FDG sitivity; but serum assays do not distinguish the anatomic
PET/CT in evaluating tumor response to chemotherapy or location of the disease recurrence (e.g., in a tumor resection
radiotherapy. However, a review of the potential role of FDG bed or as new metastases). A few small studies indicate that
FDG PET/CT after conventional CT in the setting of suspected

recurrence will change therapeutic management in as many as
Gastroenteropancreatic Neuroendocrine Tumors
44% of patients (Sperti et al, 2010). (See Chapter 65)
FDG PET/CT identifies local pancreatic cancer recurrence Major progress in the evolving field of “personalized” precision
in the surgical bed better than standard CT or MRI (Sperti medicine of neuroendocrine tumor has resulted in improved
et al, 2010). FDG PET/CT detected local recurrence with 67% diagnostic and treatment strategies and improved 5 and 10 year
to 96% sensitivity, versus 39% to 50% sensitivity for CT and/ survival in patients with gastroenteropancreatic neuroendocrine
or MRI (Kitajima et al, 2010; Wang et al, 2014). tumors (GEP-NETs). Although rather rare tumor entities,
FDG PET/CT appears superior to conventional CT for GEP-NETs underwent an increase in incidence and prevalence
detection of recurrence in lymph nodes, peritoneal implants, from 1.0 per 100,000 population in 1973 to 3.65 per 100,000
and bones. Diagnostic sensitivity and specificity for FDG PET/ from 2003 to 2007 in the United States (Lawrence et al, 2011).
CT are typically greater than 90% for suspicious FDG-avid GEP-NETs are a heterogeneous group and can be divided into
lymph nodes, peritoneal lesions, and osseous hot spots. Stan- NETs arising in the pancreas and those arising elsewhere in the
dard CT has sensitivity of approximately 60% for nodal body. This heterogeneity is reflected in the multimodality of
metastases, 50% for peritoneal recurrence, and less than 5% treatment approaches.
for recurrence as bone metastatic disease (Kitajima et al, 2010). Currently, there is disparity in standards of practice for
As discussed, FDG PET/CT in general has higher sensitivity diagnostic NMI of NETs in the United States compared with
for nodal metastatic disease than conventional CT in pancreatic Europe, with varied practice in Asia and elsewhere. Both U.S.
cancer and other HPB cancers, because it can identify subcen- and European scintigraphy of NETs use radiotracers targeting
timeter nodal metastases in nonenlarged lymph nodes (whereas the SSTRs usually overexpressed on NETs. These radiotracers
CT relies on abnormal nodal enlargement for tumor detection). are radiolabeled derivative analogues of the hormone soma-
For detection of recurrence as liver metastases, MRI has supe- tostatin, usually almost identical in molecular structure to the
rior sensitivity to FDG PET-CT. As discussed, FDG PET/CT therapeutic somatostatin derivative octreotide. Octreotide and
may be superior to standard CT for detection of liver metastases other somatostatin analogues are used in treatment of SSTR-
at initial presentation. Studies evaluating FDG PET/CT versus overexpressing NETs.
CT for detection of recurrence typically included few patients In the United States, the FDA-approved radiotracer in
with liver metastases, although often many patients with recur- standard use for NET scintigraphy is an 111In-radiolabeled
rent disease but without liver involvement; Kitajima and associ- analogue of the SSTR-binding octapeptide, octreotide (111In
ates (2010) had four patients with and 41 without liver pentetreotide). Indium 111 is a (non-PET) radiotracer allowing
metastases. Such studies indicate that FDG PET/CT and CT planar and SPECT/CT scintigraphy. In Europe and elsewhere,
both have almost perfect specificity (i.e., no false-positives for NET scintigraphy is performed with structurally similar 68Ga-
recurrence in liver) but provide too little data to allow confident radiolabeled somatostatin analogues for PET/CT, notably
68
impressions (much less statistically meaningful analysis) regard- Ga-DOTA-TOC, 68Ga-DOTA-NOC, and 68Ga-DOTA-
ing comparative diagnostic sensitivities. TATE (Virgolini et al, 2010). The different clinical standards
are important, because NET scintigraphy with the 68Ga SSTR
PET/CT tracers has repeatedly been shown to offer significantly
Assessment of Pancreatic Function superior diagnostic sensitivity for SSTR-overexpressing NETs,
(See Chapter 4) in multiple clinical trials that compared 111In pentetreotide
Clinical research into development of tracer-based PET assays scintigraphy versus 68Ga radiotracer, on a per-lesion and per-
of endocrine and exocrine pancreatic functional mass is patient basis, with comparable high diagnostic specificity.
ongoing. Diabetes mellitus (both type 1 and type 2) is associ- Because SSTR-targeted scintigraphy is often used to predict
ated with loss of β-cell mass. To prevent diabetes progression, NET sensitivity to SSTR-targeted therapy (e.g., with octreotide
preservation of β-cell mass is an investigational therapeutic or other somatostatin analogues), using the diagnostic assay
goal, and a reliable noninvasive biomarker of β-cell mass would with the highest diagnostic sensitivity and accuracy for detect-
be an important aid to diabetes research. One candidate radio- ing NET SSTR overexpression is vital for optimal therapeutic
tracer of β-cell mass has entered human clinical trials, dihydro- decision making.
tetrabenazine (DTBZ), a novel PET imaging agent (Blomberg In many parts of Europe, NET scintigraphy with 68Ga SSTR
et al, 2013). DTBZ appears promising, but further validation PET/CT tracers is current standard practice (Virgolini et al,
studies are needed. 2010). The isotope gallium 68 is produced by a tabletop genera-
Pancreatic exocrine function can be measured with direct tor (no cyclotron required), and the radiolabeling process
and indirect tests. Certain surgical procedures are associated requires no special radiochemistry techniques (Eppard et al,
with potential complications of pancreatic exocrine and endo- 2014); this has facilitated dissemination of the 68Ga radiotracers
crine insufficiency. Currently, direct tests have the best diag- into medical centers across Europe. Despite the clear diagnostic
nostic accuracy but are relatively invasive (e.g., requiring advantage of 68Ga SSTR PET/CT tracers versus the FDA-
tube placement). Even relatively noninvasive indirect tests approved 111In pentetreotide, we have seen adoption of 68Ga
can be clinically cumbersome (e.g., requiring patient stool col- SSTR PET/CT at only a few major US medical centers, con-
lection and associated delay). Researchers have evaluated ducted as part of a research trial under FDA-sponsored IND
pancreatic uptake of the PET radiotracers, such as 11C methio- application. Recent US adoption of relatively stringent FDA
nine and 11C acetate, as noninvasive, real-time biomarkers of guidelines on synthesis of PET radiotracers for human use
exocrine pancreatic function (Hyun et al, 2014). These PET essentially precludes the practice of some European countries
assays remain investigational, although preliminary data are that allow medical centers to decide to obtain 68Ga generators
encouraging. and begin SSTR PET radiotracer synthesis.
Peptide Receptor Radionuclide Imaging and Therapy

Peptide receptor radionuclide imaging and therapy is a rapidly
evolving, promising field. The molecular profile of NETs, with
frequent overexpression of certain markers (e.g., SSTRs),
represents the platform for the development of emerging meta-
bolic imaging and treatment approaches in patients with
advanced disease (see Chapter 93).
The somatostatin receptor family consists of different sub-
types and belongs to the group of G-protein–coupled receptors.
Approximately 70% to 90% of GEP-NETs express different
SSTRs, predominantly subtype 2 and at lower levels, subtypes
1 and 5 receptors as well(Reubi et al, 1984, 2001). Consider-
able advances have been made in determining the optimal
radiotracer for diagnosis and treatment.
In the early 1980s, functional imaging of GEP-NETs with
different radiotracers, initially iodine 123 (123I) and 131I MIBG,
was developed (Kaltsas et al, 2001). Metaiodobenzylguanidine
(MIBG) shares similar structural characteristics with norepi-
nephrine and is therefore taken up and stored by catecholamine-
secreting cells of the sympathetic and parasympathetic system.
Radioiodinated MIBG serves mainly as an imaging tool for A B
functional pheochromocytoma, paraganglioma, and medullo- FIGURE 17.3. Abnormal indium 111 octreotide scan. A, Left supra-
blastoma and shows only limited sensitivity (10%-50%) for clavicular node in a woman with a known neuroendocrine tumor.
GEP-NET. B, Gallium 68 (68Ga) octreotate positron emission tomography (PET)
With octreotide, the first somatostatin analogue (SSA) was scan was performed less than 1 week later and revealed abnormalities
introduced in clinical practice. In 1994, 111In–diethylenetriamine in multiple bones and intraabdominal nodes in addition to the same
pentaacetic acid (DTPA)–octreotide (Octreoscan, Mallinck- supraclavicular node that was seen on the octreotide study. The greater
rodt, St. Louis, MO) was approved by the FDA as a clinical sensitivity of the 68Ga octreotate scan is well demonstrated. (Images
courtesy Dr. Michael Hofman.)
imaging agent (Kaltsas et al, 2001; Mukherjee et al, 2001),
showing preferential renal excretion and a high affinity for
SSTR2. 111In-DTPA-octreotide serves as gold standard in NMI
for patients with GEP-NET (Reubi et al, 2001). The sensitivity include 111In, 90Y, and lutetium 77 (77Lu) (Kaltsas et al, 2004;
of SSTR scintigraphy for well-differentiated (grades 1 and 2) Kwekkeboom et al, 2010). However, for combining the SSTR
tumors ranges to higher than 80%. Guidelines for 111In-DTPA- ligand and radioisotopes other than 111In, a universal chelator
octreotide scintigraphy have been published by the Society of molecule needed to be developed, which resulted in introduc-
Nuclear Medicine, European Association of Nuclear Medicine, tion of 1,4,7,10-tetraazacyclodecane-1,4,7,10-tetraacetic acid
and European Neuroendocrine Tumor Society. Typically, planar (DOTA) (Wester et al, 2002). Development of different ligands
and SPECT images are obtained 24 and 48 hours after injec- with higher affinity and preferential targeting of selected soma-
tion. Using multimodality techniques such as SPECT/CT, tostatin subtypes lead to the introduction of agents such as 177Lu
functional and anatomic localization can be correlated simul- and 90Y DOTA-TOC, DOTA-TATE, and DOTA-NOC, which
taneously. Applications of SSTR scintigraphy include localiza- are currently used in the clinical trials for imaging and therapy
tion of the primary GEP-NETs and their metastases for initial (Krenning et al, 1993; Rufini et al, 2006). 90Y is a β emitter with
diagnosis and follow-up evaluation and assessment of treatment maximal tissue penetration of 12 mm and a half-life of 2.7 days.
response. It serves as a tool for predicting therapy response to 177
Lu is a medium-energy β emitter with a maximal tissue
receptor-based targeted therapies. penetration of 2 mm. 177Lu also emits low-energy gamma rays,
Gallium-labeled SSAs for imaging with PET (PET/CT) enabling scintigraphy and subsequent dosimetry with the same
offer higher sensitivity in tumor detection (>90%) and earlier therapeutic compound (see Chapter 93).
imaging time points (1-3 hours after IV injection), for increased Complete and partial responses obtained after treatment
patient convenience (Mukherjee et al, 2001). This is expected with 90Y-DOTA-TOC are in the same range as after treatment
to be the future standard, replacing 111In-DTPA-octreotide with 177Lu-DOTA-TATE. Most studies report objective
scintigraphy (Fig. 17.3). response rates in 20% to 35% of patients. Outcomes in terms
Physiologic uptake of SSAs in pituitary, thyroid, kidney, of progression-free survival and overall survival (16-33 months
liver, and spleen may interfere with the evaluation of SSTR and 22-46 months, respectively) compare very favorably with
scintigraphy, as well as small lesion size (<1 cm). For NETs that for SSAs, chemotherapy, or other new “targeted” thera-
with a high Ki-67 index, indicating loss of SSTR expression pies (Kwekkeboom et al, 2008; Valkema et al, 2006; van Essen
and dedifferentiation, as well as insulinomas that express only et al, 2007; Waldherr et al, 2001, 2002). In addition, a long-
low levels of SSTR2, 18F-FDG PET should be considered for term follow-up study in 265 patients with NET receiving
whole-body imaging. treatment with 177Lu-DOTA-TATE demonstrated an improve-
Peptide receptor radionuclide therapy (PRRT) relies on the ment in quality of life and symptomatology of 40% to 70%
combination of the SSTR ligand and the selected radioisotope, (Khan et al, 2011). Long-term serious side effects of PRRT
preferentially a medium- to high-energy β emitter, which include renal failure or myelodysplastic syndrome (MDS)/
generates radiation-induced DNA damage. Select radioisotopes leukemia. Therefore, kidney protection using coinfusion of
positively charged amino acids is mandatory in PRRT (Bodei BOX 17.1 Clinical Indications for Hepatobiliary Scintigraphy
et al, 2003).
Recent approaches to improve antitumoral efficacy include 1. Functional biliary pain syndromes in adults
combining treatment approaches, such as PRRT with radiosen- 2. Functional biliary pain syndromes in pediatric patients
3. Acute cholecystitis
sitizing chemotherapeutic agents (e.g., camptothecin, gem-
4. Right upper quadrant pain variants
citabine, 5-FU or its prodrug capecitabine) (Barber et al, 2012; 5. Biliary system patency
van Essen et al, 2008) or other targeted therapies, such as 6. Bile leak
everolimus or sunitinib. Application of other radionuclides, 7. Neonatal hyperbilirubinaemia (biliary atresia versus neonatal
such as α-emitters or radionuclide combinations, and intraarte- hepatitis “syndrome”)
rial administration (hepatic artery) are under evaluation. 8. Assessment of biliary enteric bypass (e.g., Kasai procedure)
Applications of PRRT in the neoadjuvant setting to achieve 9. Assessment of liver transplant
tumor size reduction and thus allow curative surgery are in 10. Afferent loop syndrome
preparation. This was successfully performed in two patients 11. Assessment of biliary dilation (e.g., choledochal cyst)
with pancreatic NETs (Kaemmerer et al, 2009; Stoeltzing et al, 12. Calculation of gallbladder ejection fraction
13. Functional assessment of the liver before partial hepatectomy
2010).
14. Demonstration of anomalous liver lobulation
Another approach in targeting somatostatin relies on the 15. Enterogastric (duodenogastric) reflux assessment
implementation of a binding-receptor antagonist instead of a 16. Esophageal bile reflux after gastrectomy
receptor agonist. Whereas binding of SSTR agonists leads to 17. Sphincter of Oddi dysfunction
receptor-mediated internalization, degradation, and intracel- 18. Differentiation of focal nodular hyperplasia from other hepatic
lular trapping, the somatostatin antagonist is not internalized, neoplasms
offering the possibility to bind to more receptor sites and with
From Tulchinsky M, et al: SNM practice guideline for hepatobiliary scintigraphy 4.0. J Nucl Med
a longer retention time (Ginj et al , 2006; Hanyaloglu & von Technol 38:210-218, 2010; and American College of Radiology, Society for Pediatric Radiology:
Zastrow, 2008; Wang et al, 2012). Antagonists such as DOTA- ACR-SPR practice parameter for the performance of hepatobiliary scintigraphy. Amended 2014
BASS and DOTA-JR11 are under evaluation (Wild et al, 2011, (Resolution 39). www.acr.org.
2014).
Recent advances in the field of nuclear medicine play pivotal
roles not only for diagnosing and staging but also for identifica- cholestasis, with MR cholangiopancreatography (MRCP) and
tion of biologic characteristics of NET, thereby guiding and endoscopic ultrasound (EUS) as potential next diagnostic
validating novel treatment approaches and increasing the scope imaging assays, for unexplained cholestasis.
of personalized medicine. In current state-of-the-art practice, hepatobiliary scintigra-
phy uses an iminodiacetic acid (IDA)–derivative pharmaceuti-
HEPATOBILIARY SCINTIGRAPHY cal radiolabeled with 99mTc; most commonly, 99mTc-radiolabeled
disofenin or mebrofenin. Many clinicians still loosely refer to
Hepatobiliary scintigraphy (or cholescintigraphy) refers to hepatobiliary scintigraphy with these radiotracers as a “HIDA
diagnostic imaging of radiotracers that assess hepatic perfusion, scan,” although HIDA is actually the name of a specific 99mTc-
hepatocellular function, and hepatobiliary drainage (see radiolabeled IDA compound, lidofenin, which was the preemi-
Chapter 3). The most common clinical application of hepato- nent radiopharmaceutical in use at the time that hepatobiliary
biliary scintigraphy is for detection of cholecystitis, including scintigraphy disseminated into widespread clinical medical
acute cholecystitis, typically induced by cystic duct obstruction, practice. Lidofenin is no longer in widespread use because the
and chronic cholecystitis, typically associated with impaired subsequent agents, notably disofenin and mebrofenin, offer
gallbladder ability to contract normally. State-of-the-art hepa- more favorable in vivo pharmacokinetics (Fig. 17.4 and Table
tobiliary scintigraphy has high diagnostic sensitivity and speci- 17.1).
ficity for detection of both acute and chronic cholecystitis, The radiolabeled IDA compounds share common PK
typically 90% to 95% or higher, when optimal methodology is properties, differing predominantly in degrees or particular
used. Other notable but less common clinical indications for characteristics. All are injected intravenously and bind loosely
hepatobiliary scintigraphy are listed in Box 17.1; the diagnostic to plasma proteins, mainly albumin; exiting the bloodstream,
accuracy varies for each. the IDA tracers dissociate from plasma proteins in the space of
In general, despite the wide range of hepatobiliary diseases Disse and then undergo hepatic uptake by hepatocytes, fol-
that exist, most hepatobiliary medical society guidelines ascribe lowed by hepatic excretion into the intrahepatic biliary tree and
a role for diagnostic hepatobiliary scintigraphy in only a very subsequent clearance by biliary flow into the intestines via the
few specific clinical indications, typically as a second-line sphincter of Oddi.
imaging modality. This is because the information about hepatic These compounds also share PK properties with bilirubin.
function and biliary drainage obtained by scintigraphy can Bilirubin, the end product of heme catabolism, is taken up from
often confirm the existence of hepatobiliary dysfunction, which the blood circulation into hepatocytes by a high-affinity,
can usually be diagnosed without imaging by clinical examina- sodium-independent, organic anion transport protein shared
tion and blood tests, but typically does not clarify the underlying with IDA radiotracers. Inside hepatocytes, bilirubin is metabo-
etiology in many types of disease. For example, cholestasis can lized to a more water-soluble conjugated form; the IDA radio-
usually be positively identified by cholescintigraphy, but tracers, however, do not change form. Conjugated bilirubin and
cholestasis syndromes in adults have a variety of etiologies. the IDA radiotracers both exit hepatocytes by energy-dependent
EASL (2009) guidelines make no mention of hepatobiliary active transport (particularly transporter ABCC2; Bhargava
scintigraphy; rather, ultrasound is the primary noninvasive et al, 2009), then are excreted into the biliary tree along with
imaging procedure to distinguish intrahepatic from extrahepatic other components of bile (e.g., cholesterol, bile acids).
H
O
H
H O H O
O O O O
N O O
N H
O O
N
H
O
H
O N O
H H
N N O
Br
FIGURE 17.4. Iminodiacetic acid (IDA) molecular components of discussed biliary radiotracers. IDA molecules shown here without attachment to
technetium 99m (99mTc) isotope radiolabel. Left, Diisopropyl IDA, or disofenin. Middle, Bromotriethyl IDA, or mebrofenin. Right, Dimethyl IDA, or
lidofenin, better known as HIDA (hepatic IDA tracer). The compounds differ primarily in constituents along the phenyl ring. When radiolabeled, each
biliary radiotracer comprises a bis structural complex of two IDA ligands bound to one 99mTc atom (not shown).
TABLE 17.1 Pharmacokinetic Properties of Common Hepatobiliary Scintigraphy Radiotracers

Liver Clearance Renal Excretion by 2-3 hr Renal Excretion by 2-3 hr (Average
IDA Tracer Liver Uptake (%) Half-Time (min) (Average and Range %) and Range %) in Hyperbilirubinemia*
99m
Tc disofenin 88 19 ±3 <9 —
99m
Tc mebrofenin 98 17 ±1 1 (0.4-2.0) 3 (0.2-11.5)
*Mebrofenin, package insert. Mean elevated serum bilirubin levels of 9.8 mg/dL (1.7-46.3 mg/dL).
IDA, Iminodiacetic acid analogue; 99mTc, technetium 99m.
From Krishnamurthy GT, Turner FE, 1990: Pharmacokinetics and clinical application of technetium 99m-labeled hepatobiliary agents. Semin Nucl Med 20:130-149.
In marked hyperbilirubinemia, the hepatic uptake of IDA minutes after injection (Fig. 17.5) (Kwatra et al, 2013).
radiotracers can be inhibited by competition in carrier-mediated However, uptake of IDA radiotracers can remain relatively
transport of IDA radiotracers, due to high circulating levels of intact in the acute phase of hepatic injury, noting again that
bilirubin, which uses the same transporters. In severe hyperbili- hepatocellular uptake of IDA radiotracers is not ATP depen-
rubinemia, a higher activity (MBq) of IDA radiotracer may be dent. In the early acute phase of hepatic injury, the cessation
injected, but not in hopes of reducing competitive inhibition by of ATP-mediated hepatocellular excretion of IDA radiotracers
“mass effect,” because the higher radiotracer dose still involves may precede any significant (i.e., scintigraphically detectable)
only nanogram (trace) amounts of IDA compound; rather, the decrease in overall hepatic uptake of IDA radiotracer. The dif-
liver will likely absorb the same percentage of the radiotracer ferential analysis and clinical scenarios where this may be
dose, regardless of the activity. It is hoped, however, that this encountered are discussed later. Essentially, however, a marked
absorbed percentage, by using a higher administered activity decrease in IDA radiotracer and bile excretion and an associ-
(MBq), will yield more photons and more signal from the liver ated lack of bile (tracer) flow through the biliary tree is a sign
and biliary tree, for better-quality scintigraphic imagery. In of a severe hepatic dysfunction, which may be caused by a
moderate to severe hyperbilirubinemia, 99mTc-radiolabeled primary hepatic dysfunction (e.g., severe hepatitis) rather than
mebrofenin is probably preferable to disofenin, because mebro- a mechanical obstruction of biliary tree drainage (and thereby
fenin has higher hepatic extraction (Tulchinsky et al, 2010). secondary hepatic dysfunction).
Hepatobiliary scintigraphy is first an assay of hepatic paren- Bile production depends on normal hepatocellular function.
chymal function (see Chapter 3). Hepatic excretion of the If hepatic parenchymal dysfunction is severe, bile production
radiolabeled IDA compounds is an energy-dependent, active and bile flow will slow down, perhaps even stop, until hepato-
transport process, dependent on intact hepatocellular function. cellular recovery. Similarly, with severe hepatic parenchymal
An absent or abnormally low hepatic uptake of IDA radiotracer dysfunction and associated scant hepatic IDA radiotracer
is indicative of hepatic parenchymal dysfunction, often a diffuse uptake, hepatic excretion of the IDA radiotracer and its drain-
hepatic pathology (e.g., hepatitis). One proposed definition of age through the biliary tree will also be scant or undetectable
“poor hepatic uptake” is hepatic tracer uptake that is visually on scintigraphy.
equivalent, in scintigraphic intensity, to that of tracer present
in the cardiac blood pool, 5 minutes after tracer-injection. Normal Hepatobiliary IDA Radiotracer Scan
Normally, hepatic tracer concentration should be greater than Again, note that “HIDA scan” often referred specifically to
99m
that of circulating tracer in the cardiac blood pool tracer by 5 Tc-lidofenin, which is no longer commonly used. This
FIGURE 17.5. This study shows severe hepatocellular dysfunction, with the kidneys exhibiting increased clearance of the tracer. Note the parallel
reduction in activity in the cardiac blood pool and the liver. The two bright foci beneath the liver are caused by activity in the pelvis of the right and
left kidneys. The straight arrow points to the left renal pelvis; the curved arrow points to the right ureter. Images were acquired at 2, 15, 25, and 60
minutes after injection.
section discusses state-of-the-art imaging with the commonly BOX 17.2 Hepatobiliary Scintigraphy: Key Points
used 99mTc-labeled IDA tracers mebrofenin and disofenin,
unless otherwise specified. Scintigraphy with both tracers A “normal” hepatobiliary scan appearance is defined by the
yields similar findings in healthy individuals after standard amount of tracer present in the blood pool, liver, biliary tree,
and bowels, at different times after injection.
preparation.
Patient preparation varies, depending on the specific clinical Normal Patient Preparation
indication, but usually includes fasting (unless the patient has Fasting (only if gallbladder present):
no gallbladder; e.g., after cholecystectomy) and avoiding or • ≥2 hours for infants
counteracting recent use of opioids (Box 17.2) (Tulchinsky • ≥2-6 hours for adults
et al, 2010). Fasting is advised in patients with a gallbladder, • But not >24 hours fasting
because one important parameter of normalcy in hepatobiliary Opioids: wait ≥4 half-lives, or counteract (e.g., naloxone).
scintigraphy is detection of the excreted-radiotracer within the
gallbladder lumen, especially if the patient has clinical signs and
symptoms suggestive of cholecystitis. If excreted radiotracer is receiving a bolus of nutrients from the stomach, releases the
not detected within the gallbladder lumen, this can be a sign of hormone cholecystokinin (CCK); CCK causes contraction of
cystic duct obstruction, as seen in acute cholecystitis in adults. the gallbladder and relaxes the sphincter of Oddi, allowing
For excreted radiotracer in the extrahepatic biliary tree to enter concentrated bile from the gallbladder stores to be pumped into
the gallbladder via the cystic duct, there must be a pressure the duodenal lumen. Thus, in the recently fed patient with rela-
gradient moving bile in that direction. When the gallbladder tively high levels of circulating CCK, the gallbladder often will
contracts, the pressure gradient is in the opposite direction, be contracting, and excreted-tracer may not enter the gallblad-
moving bile from gallbladder lumen out the cystic duct into the der lumen. The absence of excreted tracer in the gallbladder
common bile duct (CBD) lumen. This occurs physiologically lumen caused by physiologic contractions of a healthy gallblad-
during feeding, as the bile passes into the duodenum to aid der mimics the appearance of a (pathologic) cystic duct
digestion of food entering from the stomach. The duodenum, obstruction, potentially being misdiagnosed as consistent with
acute cholecystitis on scintigraphy (see Chapter 33). Guidelines acquisition of a series of scintigraphic images of the liver region
typically advise a fasting period of at least 2 hours, but prefer- during time, often anterior and posterior planar images, which
ably 6 hours, for adults, likely based on the normal gastric (and can then be displayed as frames in a motion picture for visual
duodenal) clearance times after a typical solid meal. For infants, and quantitative assessment of whether tracer uptake and
a fasting period of only 2 hours is typical, likely because the excretion occur in a normal amount and in the normal time
infant diet is liquid, and gastroduodenal clearance of liquid frame (Fig. 17.6). For example, hepatobiliary scintigraphy often
meals is faster than that of solid meals. In the fasting state, begins with a series of consecutive planar images of the liver
neural and hormonal impulses cause the sphincter of Oddi to region, each planar image being a frame 1 minute in duration,
be predominantly closed and the cystic duct of the gallbladder acquired for a total of 1 hour, starting at the time of radiotracer-
to remain predominantly open; this halts bile flow into the injection. Nuclear medicine societies provide guidance docu-
intestines (via sphincter of Oddi) and creates a pressure gradi- ments (available online) with detailed technical suggestions for
ent directing bile (and IDA tracer) into the gallbladder. During image acquisition protocols (Tulchinsky et al, 2010).
fasting, some phasic contractions of gallbladder and relaxation Analysis of hepatobiliary scintigraphy imagery is based on
of the sphincter of Oddi still occur, possibly to churn the bile visual detection of tracer in expected locations at expected time
and avoid precipitation of some bile constituents. Therefore, a points and often on semiquantitative visual estimation of
normal variability (i.e., range of normal values) occurs in the whether the amount of tracer present in blood pool, liver, biliary
time required to observe excreted tracer in the gallbladder and tree, and bowels is appropriate for the time point after injection.
bowels in well-prepared, healthy individuals. Quantitative analysis may be facilitated by digital region-of-
Prolonged fasting for longer than 24 hours (including during interest (ROI) analysis to generate numeric indices.
hyperalimentation) is also associated with excreted IDA tracer Both 99mTc-labeled IDA tracers clear rapidly from the
failing to enter the gallbladder lumen, creating a potentially bloodstream. Typically, less than 15% of the radiotracer dose
misleading appearance scintigraphically. This is likely attribut- remains in the blood circulation by 10 minutes after injection;
able to accumulation of biliary sludge in the gallbladder lumen, the amount of tracer in the cardiac blood pool, if in the scinti-
obstructing the cystic duct in a nonpathologic manner. Clinical graphic field of view, drops greatly during that time. Both
research demonstrates that for patients who have undergone tracers are eliminated from the body predominantly by hepa-
prolonged fasting, it is beneficial to administer a synthetic CCK tobiliary excretion, usually with minor urinary excretion and
octapeptide (sincalide) to “clean out” the gallbladder as part of with no tracer accumulation in other organs. The tracers achieve
patient preparation for hepatobiliary scintigraphy. Pretreatment maximal hepatic concentration approximately 10 minutes after
with sincalide is typically 0.02 µg/kg intravenously during 30 to injection and have hepatic uptake that is visually distinct typi-
60 minutes by slow infusion. One guideline suggests that hepa- cally by 5 minutes. Both tracers are detectable in the extrahe-
tobiliary tracer injection can occur 15 to 30 minutes after slow patic biliary tree in 100% of normal individuals by 1 hour after
sincalide infusion pretreatment (Tulchinsky et al, 2010). injection, and in approximately 90%, excreted tracer is detect-
Attempts to administer sincalide more rapidly are associated able in the CBD by 30 minutes. The intrahepatic biliary tree is
with unpleasant patient symptoms, particularly gastrointestinal often detectable earlier, sometimes 10 minutes after injection.
(e.g., crampy abdominal pain), and can induce a contractile Frequently, the intrahepatic biliary tree is more distinctly
spasm of the gallbladder neck, especially if given as a bolus, visualized in the lateral region of the left hepatic lobe than in
creating a cystic duct obstruction that the sincalide pretreat- the right lobe, probably because of the normally greater mass
ment is intended to relieve; the spasm lasts approximately 1 of the right hepatic lobe causing attenuation of the scintigraphic
hour (Mesgarzadeh et al, 1983). signal emanating from within it.
If the patient has no gallbladder, no fasting is required for The amount of excreted tracer present in the extrahepatic
hepatobiliary scintigraphy. biliary tree is an important variable, as well the timeliness of its
Opioid medications (e.g., morphine) acutely induce strong, appearance after injection. After achieving maximal liver uptake,
prolonged contraction of the sphincter of Oddi. If a patient has the liver clears tracer (via excretion) with a typical half-time of
recently received opioid medications, passage of excreted about 30 (±15) minutes (Drane & Johnson, 1990; Williams
radiotracer from the biliary tree into the duodenum can be et al, 1984). Typically, one-third of the injected dose has passed
delayed as a side effect. If biliary-to-bowel transit is delayed into or through the common bile duct by 30 to 60 minutes after
beyond 3 to 4 hours after injection from lingering opioid effects, injection. Excreted tracer clears from the CBD with a half-time
this can mimic the appearance of CBD obstruction (e.g., (from peak CBD activity) of approximately 30 (±20) minutes
caused by choledocholithiasis). To avoid this potential diagnos- (Drane & Johnson, 1990).
tic pitfall, patients should avoid opioid medications before Excreted tracer passes from the biliary tree into the duode-
hepatobiliary scintigraphy; one guideline suggests waiting for num via the sphincter of Oddi in healthy persons with normal
four times the half-life of the particular opioid drug (Tulchinsky biliary anatomy. This is evident scintigraphically to the trained
et al, 2010). Alternatively, naloxone may be administered to eye of the imaging specialist, becoming most evident if intestinal
attempt to reverse the opioid effects (Patch et al, 1991). peristalsis sweeps the tracer through the duodenum into distal
Additional forms of patient preparation for less common bowel loops during imaging. Excreted tracer is detectable in the
indications of hepatobiliary scintigraphy are discussed later. intestines in most healthy people (80%) by 1 hour after injec-
The in vivo hepatobiliary parameters typically assessed on tion. Biliary-to-bowel transit takes longer in some healthy
scintigraphy are hepatic uptake and excretion of the adminis- individuals, particularly those who receive sincalide before
tered IDA tracer and subsequent biliary drainage of the treatment (a paradoxical phenomenon) (Kim et al, 1990). How
excreted-radiotracer through the biliary tree into small bowel. to distinguish physiologic delay from pathologic CBD obstruc-
These are time-dependent biologic processes. Thus, hepatobili- tion is discussed later. Excreted tracer that enters the intestines
ary scintigraphy typically involves dynamic imaging, with the remains in the intestines, with no significant intestinal tracer
FIGURE 17.6. Normal hepatobiliary iminodiacetic acid study. Serial images of the liver, in anterior projection, acquired beginning immediately after
tracer injection; each image is 5 minutes in duration. The serial images show prompt systemic clearance of radiotracer (e.g., disappearing cardiac
blood pool, gray arrowhead) via hepatic uptake, followed by prompt hepatobiliary excretion of tracer. Excreted tracer flows promptly through the
extrahepatic biliary tree, including into the gallbladder lumen (large arrow). Excreted tracer passes from biliary tree into the duodenum, carried away
into distal bowel loops by peristalsis (black arrowhead). One atypical incidental finding is enterogastric reflux of the biliary tracer (small arrow).
reabsorption or enterohepatic recycling of radiotracer. Entero- (half-time ~40 minutes) is much slower than that of the 99mTc-
gastric reflux of tracer can occasionally be observed; its clinical radiolabeled disofenin and mebrofenin compounds (both
significance is not well defined. ~15-20 minutes). Older literature and even recent hepatobiliary
Lastly, the gallbladder should be visualized within 1 hour scintigraphy guidelines often suggest obtaining delayed imaging
after injection. Gallbladder nonvisualization after 1 hour in the at 3 to 4 hours or later after radiotracer injection. With 99mTc-
proper clinical setting can diagnose cystic duct obstruction, a lidofenin, imaging at 3 to 4 hours or longer after injection was
hallmark of acute cholecystitis, with high diagnostic accuracy. possible because of the slow hepatobiliary excretion of the
In certain patients, confirmatory measures (e.g., morphine tracer; if gallbladder filling or biliary-to-bowel transit had not
augmentation) may be pursued. yet been visualized by the expected 1 hour after injection,
Pathology in hepatobiliary scintigraphy is thus indicated by delayed imaging was possible because a significant amount of
one or more of the following: radiotracer would still be in the liver undergoing hepatobiliary
• An abnormally low amount of hepatic tracer uptake transit. However, with 99mTc-radiolabeled disofenin and mebro-
• An abnormally prolonged hepatic retention of tracer fenin, the hepatobiliary clearance is relatively rapid. The liver is
• An abnormally delayed appearance of excreted tracer in often essentially “empty” of radiotracer by 2 to 4 hours after
the biliary tree injection, precluding delayed imaging. At this time, one cannot
• An abnormally delayed appearance of excreted tracer in expect delayed filling of the gallbladder with radiotracer if it has
the intestines already essentially cleared from the liver and the remainder of
Again, the radiotracer 99mTc-lidofenin is now rarely used the biliary tree. On the other hand, delayed imaging with the
compared with current agents such as 99mTc-radiolabeled dis- newer agents remains a logical approach in certain cases. In
ofenin and mebrofenin. The hepatic clearance of 99mTc-lidofenin patients with severe hepatocellular dysfunction (whether
secondary to biliary tract obstruction or from a separate hepatic to the operating room for cholecystectomy without preoperative
disease), uptake and secretion of biliary tracer by dysfunctional imaging, or an urgent abdominal US evaluation may be
hepatocytes may be abnormally delayed and associated biliary obtained. If US visualizes cystocholelithiasis and signs of cho-
flow through the biliary tree likewise greatly slowed; delayed lecystitis, cholecystectomy might be pursued. If gallstones and
imaging at 4 or even 24 hours may be useful to evaluate biliary sonographic signs of cholecystitis are absent, hepatobiliary
flow in these patients (Ziessman, 2014). scintigraphy might be pursued to identify cystic duct obstruc-
tion, which is the typical etiology of acute calculous or acalcu-
Augmented Hepatobiliary Scintigraphy lous cholecystitis in adults. For chronic or recurrent abdominal
Hepatobiliary scintigraphy can employ pharmacologic inter- pain in adults, evaluation must exclude nonbiliary etiologies as
ventions, most notably with phenobarbital, morphine, and well as biliary. Hepatobiliary scintigraphy with CCK might be
sincalide. used to confirm a diagnosis of gallbladder functional disorder
Phenobarbital is used in cholescintigraphic evaluation of (Zakko et al, 2015). For children, cholecystitis is uncommon
neonatal jaundice (see Chapters 1 and 40). Patient pretreat- and is usually associated with a predisposing risk factor, such
ment with phenobarbital will stimulate bile production and as sickle cell disease or cystic fibrosis; the scintigraphic findings
biliary flow and increase the likelihood of visualizing hepatobili- can also differ from adult scintigraphic signs.
ary radiotracer excretion, if biliary atresia is absent. It reduces
the likelihood of false-positive scintigraphic appearance of Acute Cholecystitis (See Chapter 33)
biliary flow obstruction. The reported sensitivity and specificity of cholescintigraphy for
Morphine is used in in cholescintigraphic evaluation of suspected acute calculous cholecystitis ranges from 87% to
gallbladder cystic duct obstruction (see Chapters 33 and 49). 98% and 81% to 100%, respectively, in the original clinical
If accumulation of excreted biliary tracer in gallbladder lumen studies up to 2000 (Ziessman, 2003). Reviewing their experi-
is not observed in the expected time frame, morphine admin- ence with patients seen in the emergency department with
istration normally contracts the sphincter of Oddi and relaxes acute upper abdominal pain (n = 406) from 2010 to 2012,
any physiologic gallbladder contraction; this increases the likeli- Kaoutzanis and colleagues (2014) reported that patients vari-
hood that biliary radiotracer can accumulate within the gall- ably underwent abdominal US alone (n = 132), scintigraphy
bladder, if no pathologic obstruction of the gallbladder cystic alone (n = 46), or both (n = 228), according to the clinician’s
duct (e.g., associated with calculous or acalculous cholecystitis) judgment at the time. Their sensitivity for acute cholecystitis
is present. It reduces the likelihood of false-positive scintigraphic was US alone, 73%; scintigraphy alone, 92%; and both, 98%.
appearance of gallbladder cystic duct obstruction. Because Although selection bias must be suspected to impact the results,
morphine contracts the sphincter of Oddi, scintigraphically, the findings indicate that hepatobiliary scintigraphy (1) contin-
entry of biliary tracer into small intestines is typically confirmed ues to have a standard role in the evaluation of the patient with
visually before use of morphine, first to exclude CBD obstruc- acute abdominal pain, often combined with abdominal US,
tion before attempting to exclude a cystic duct obstruction. If which remains the first-line imaging modality, and (2) offers
the decision to use morphine is made based on nonvisualization high sensitivity for preoperative diagnosis of acute cholecystitis,
of the gallbladder, but the injected biliary tracer has already alone or combined with US, in the proper clinical context.
essentially cleared from the liver and biliary tree into the bowels, While the radiotracer is injected, hepatobiliary scintigraphy
it is often necessary for patients to receive a “booster” injection typically begins with collection of an uninterrupted series of
of biliary tracer before administration of morphine, to renew a brief (1 to 3 second), consecutive images of the liver for 1 to 2
flow of radiotracer through the biliary tree. minutes, called the flow phase. These images can be displayed
Sincalide is used in cholescintigraphy for evaluation of gall- in dynamic, cinematic fashion to visualize the passage of the
bladder cystic duct or CBD obstruction or gallbladder dyski- tracer through the bloodstream, including its initial passage
nesia (see Chapter 33). Patient treatment with sincalide causes through the liver. If cholecystitis is present, the surrounding
contraction of the gallbladder and relaxation of the sphincter hepatic parenchyma may also be inflamed. This tissue inflam-
of Oddi. Sincalide can be given before radiotracer to evacuate mation is often associated with relative tissue hyperemia that
the gallbladder lumen if gallbladder accumulation of a biliary can be visualized as an abnormal “blush” of activity around the
sludge is possible (e.g., from patient hyperalimentation or gallbladder during the flow phase, called the “rim sign.” The
prolonged fasting) or evident sonographically; such sludge can rim sign, indicating inflammation of liver around the gallblad-
impede gallbladder accumulation of biliary tracer, creating a der, is typically associated with severe inflammation of the
scintigraphic appearance mimicking a cystic duct obstruction gallbladder itself. The rim sign has been associated with an
(false-positive). If the gallbladder lumen accumulates biliary increased risk of complicated cholecystitis, specifically gangrene
tracer (no cystic duct obstruction), but gallbladder dyskinesia and gallbladder perforation. The rim sign is visualized in
is suspected (e.g., associated with cholecystitis), sincalide can approximately 25% to 35% of patients with acute cholecystitis
then be administered to assess the normalcy of gallbladder (Ziessman, 2014).
sincalide response in terms of its ejection fraction—what per- After the flow phase, hepatocellular uptake and excretion of
centage of radiotracer is evacuated from the gallbladder lumen the radiotracer is monitored scintigraphically, typically as an
in response to sincalide. uninterrupted series of 1 to 5 minute consecutive images of the
liver for at least 60 minutes. During this period, the imaging
Right Upper Quadrant Pain specialist evaluates hepatobiliary tracer clearance for the
The approach to an adult patient with abdominal pain is expected normal scintigraphic findings (described earlier).
detailed elsewhere in this textbook (see Chapter 33). In brief, Acute cholecystitis in adults is almost always associated with a
however, for acute clinical signs and symptoms strongly sugges- pathologic obstruction of the cystic duct, whether cholelithiasis
tive of acute cholecystitis, the patient might be taken directly impacted in the cystic duct or acalculous cholecystitis (i.e.,
inflammation and edema in the walls of the gallbladder neck, Most often, chronic cholecystitis is associated with cysto
thereby occluding the duct). Again, in most healthy patients, cholelithiasis, detected sonographically. The US presence of
gallbladder accumulation of excreted biliary radiotracer occurs gallbladder stones and classic clinical signs and symptoms are
within 30 to 60 minutes after injection. The imaging specialist the main indications for cholecystectomy. Cholescintigraphy
must be wary of the “cystic duct sign,” which mimics gallblad- might be obtained for patients with questionable chronic cho-
der tracer accumulation. The cystic duct sign is a nonelongated, lecystitis (e.g., atypical clinical signs and symptoms, absent
spotlike focus of tracer accumulation in the expected location cystocholelithiasis) or who might poorly tolerate cholecystec-
of the cystic duct, along the common hepatic and bile duct tomy (e.g., because of comorbidity). In chronic cholecystitis,
course. The cholescintigraphic cystic duct sign represents a calculous or acalculous, the cystic duct is typically patent, and
focal accumulation of biliary radiotracer within the proximal the gallbladder will be visualized as accumulating radiotracer.
cystic duct (closest to and communicating with the common The scintigraphic hallmark of cholecystitis is gallbladder dyski-
duct) that can occur in cystic duct obstruction when choleli- nesia, manifest as an abnormally low gallbladder ejection frac-
thiasis obstructing bile passage from the gallbladder is lodged tion (GBEF) of less than 38%, on sincalide-stimulated
more distally within the cystic duct. The cystic duct sign should cholescintigraphy (Tulchinsky et al, 2010; Ziessman, 2014).
be suspected when biliary tracer accumulation in the gallblad- Gallbladder dyskinesia (low GBEF) can be observed with acute
der region is not elongated, but rather is more spotlike and cholecystitis as for chronic cholecystitis, because of dysfunction
smaller than usual. In cases of diagnostic uncertainty, a sus- of inflamed gallbladder smooth musculature in both acute and
pected cystic duct sign can be evaluated further by SPECT/CT, chronic cases. The GBEF study begins after the expected tracer
to confirm an absence of excreted tracer in the gallbladder accumulation in the gallbladder lumen has been visualized.
lumen. Again, if morphine augmentation was required, sincalide GBEF
As previously discussed, in a minority of healthy patients testing can still be pursued, and normal GBEF results reliably
with no pathologic obstruction of the cystic duct, the gallblad- exclude gallbladder dyskinesia and cholecystitis, although low
der lumen does not accumulate excreted tracer within the first GBEF results are nondiagnostic. When the gallbladder is
hour because of a physiologic obstruction of inward biliary flow “filled” sufficiently with excreted tracer (typically 30-60 minutes
associated with physiologic contractions of the gallbladder. To after tracer injection), sincalide is then administered.
minimize these contractions, patients are advised to fast before Besides morphine, a variety of medications have been associ-
the study. Even with proper fasting, however, this contraction ated with low GBEF results that could lead to a false diagnosis
can occur. In these healthy patients, gallbladder relaxation and of cholecystitis; medications include (but are not limited to)
tracer accumulation will usually occur by 3 to 4 hours after atropine, calcium channel blockers, octreotide, progesterone,
injection. To increase the diagnostic specificity of gallbladder indomethacin, theophylline, benzodiazepines, and histamine-2
nonvisualization as a scintigraphic biomarker of cystic duct receptor antagonists. As such, patient medications should be
obstruction, guidelines suggest that the imaging specialist reviewed at the time of the study. A variety of conditions other
obtain either delayed imaging or, without delay, administer than cholecystitis have been associated with low(er) GBEF
morphine (Tulchinsky et al, 2010). If the gallbladder remains results as well, including obesity, celiac disease, cirrhosis, myo-
nonvisualized after morphine, cystic duct obstruction can be tonic dystrophy, and diabetes.
diagnosed with greater specificity. Thus, standard guidelines
advise routine use of morphine augmentation or, alternatively, Extrahepatic Bile Duct Obstruction
delayed imaging, if the gallbladder does not accumulate excreted As discussed, in healthy volunteers, cholescintigraphy tracers
tracer in the first hour, but only after biliary tracer transit undergo hepatobiliary drainage with detectable passage of
though the sphincter of Oddi into bowel is confirmed, to excreted tracer into the duodenum within 1 hour after injection
exclude CBD obstruction. in most cases. In approximately one-third of healthy volunteers,
Regarding false-negative results, cholescintigraphy appears passage of excreted tracer into the duodenum is not evident on
to have less diagnostic sensitivity for acalculous acute cholecys- scintigraphy until the second hour or so; this is probably attrib-
titis, perhaps 70% to 80%, compared with more than 90% for utable to transient/phasic physiologic contraction of the sphinc-
calculous acute cholecystitis (Tulchinsky et al, 2010; Ziessman, ter of Oddi. In these healthy volunteers, with relatively delayed
2014). A significant number of patients with acalculous acute passage of tracer into bowels, excreted tracer is visible in the
cholecystitis will demonstrate gallbladder tracer accumulation extrahepatic biliary tree, awaiting passage into bowel once the
(no cystic duct obstruction). Acalculous cholecystitis can be sphincter of Oddi relaxes. Additionally, patients pretreated with
diagnosed despite gallbladder tracer accumulation in the pres- sincalide before cholescintigraphy paradoxically seem to be
ence of the rim sign or abnormally low gallbladder ejection more likely to demonstrate a nonpathologic delay in biliary
fraction, to increase diagnostic sensitivity (Ziessman, 2014). In transit into bowel, at least if sincalide is infused relatively rapidly
adults, cholecystitis is typically calculous cholecystitis, whereas (Kim et al, 1990).
acalculous cholecystitis is relatively uncommon. In pediatric Pathologic obstruction of the common hepatic duct or CBD
patients, however, acute cholecystitis is frequently acalculous. can be diagnosed by cholescintigraphy. Extrahepatic bile duct
For suspected acalculous cholecystitis, both adult and pediatric, obstruction can be caused by cholelithiasis (choledocholithiasis
our standard practice is to assay gallbladder ejection fraction, or rarely Mirizzi syndrome) (see Chapters 36 and 37) or tumor
if cystic duct obstruction is absent scintigraphically, to optimize (e.g., pancreatic head mass compressing adjacent duct) (Chap-
diagnostic sensitivity. ters 51, 59, and 62). Tumor-associated biliary obstruction is
usually identifiable by structural imaging (CT, US, or MRI),
Chronic Cholecystitis (See Chapter 33) because biliary duct dilation is often already present at diagno-
Chronic cholecystitis patients typically report biliary colic–type sis. Obstructive cholelithiasis, in contrast, often presents clini-
symptoms that have been recurrent over a prolonged period. cally as an acute biliary-type pain; in the first 24 to 72 hours
FIGURE 17.7. Complete biliary obstruction. Images were obtained immediately after injection and at 30 minutes, 60 minutes, and 3 hours. No
severe hepatocellular dysfunction is evident; systemic tracer clearance is somewhat delayed (cardiac blood pool still detectable at 30 minutes), but
the liver demonstrates prompt, visually distinct tracer uptake. No excreted tracer is detectable in the extrahepatic biliary tree at 1 hour after tracer
injection. This constellation of scintigraphic findings is suspicious for an acute high-grade extrahepatic bile duct obstruction, in the proper clinical
setting. The continued absence of detectable excreted tracer at 3 to 4 hours after tracer-injection increases diagnostic confidence, when hepatocellular
function is impaired.
after bile duct obstruction by cholelithiasis, bile duct distension If these two findings are present, extrahepatic bile duct
may not yet be marked enough to be suspicious on structural obstruction is the likely diagnosis, with other etiologies relatively
imaging. Cholescintigraphy can confirm a clinical suspicion of rare (Ziessman, 2003). If hepatic function is poor, however,
acute extrahepatic bile duct obstruction in patients with nega- passage of excreted tracer into the extrahepatic bile ducts can
tive structural imaging. Additionally, patients with previously be delayed because hepatic bile production is slow. Hepatic
treated bile duct obstruction can have persistent bile duct dila- function may be poor because of a true bile duct obstruction
tion after treatment; if these patients return to the clinic with (in which case it was more likely chronic than acute) or because
suspicious biliary colic, structural imaging may demonstrate a of a separate nonobstructive etiology (e.g., hepatitis, cirrhosis).
chronically distended bile duct of uncertain significance. Cho- If hepatic function is poor, delayed imaging will improve test
lescintigraphy can confirm a clinical suspicion of acute recur- specificity (i.e., help avoid false diagnosis of bile duct obstruc-
rent extrahepatic bile duct obstruction in such patients. tion (Fig. 17.7) (Ziessman, 2003).
Acute extrahepatic bile duct obstruction has been described A partial obstruction of the extrahepatic bile ducts is associ-
as having cholescintigraphic appearances that favor either a ated with the following cholescintigraphic appearance:
“high-grade” or a lesser, “partial” degree of biliary obstruction • Hepatic uptake is normally prompt (unless the patient also
(Ziessman, 2014). An acute high-grade obstruction of the has hepatic dysfunction unrelated to biliary obstruction).
extrahepatic biliary tree can be diagnosed if the following find- • Excreted-tracer appears promptly within the extrahepatic
ings are present: biliary tree.
• The patient has good hepatic function, as indicated by blood • Clearance of excreted tracer from the extrahepatic biliary
tests and on cholescintigraphy by rapid hepatic tracer-uptake tree is absent or scant during the first hour after injection.
and rapid tracer clearance from the blood pool. • Further clearance of excreted tracer from the extrahepatic
• Excreted biliary tracer is not detectable or scantly present biliary tree into bowels, after sincalide administration or at
in the extrahepatic biliary tree by 1 hour after injection. delayed time points, is less than expected.
Normally, approximately half of biliary tracer has passed Biliary Tract Complications After Surgery
from the hepatobiliary tree into the bowels by 1 hour after Cholescintigraphy can positively impact diagnostic evaluation
tracer injection, and a progressive decline in the net amount of of patients for biliary tract complications associated with a
tracer in the extrahepatic biliary tree is usually visibly evident variety of procedures, such as laparoscopic cholecystectomy
scintigraphically in that first hour (Drane & Johnson, 1990). (see Chapters 38 and 42), partial hepatic resections (Chapters
However, an absence of bowel activity at 1 hour after tracer 27 and 103), or liver transplantation (Chapter 120). Bile duct
injection is sensitive for partial obstruction (~98%) but not injuries are complications associated with laparoscopic chole-
highly specific, because of the normal variation in drainage of cystectomy, for example, with a reported rate of occurrence
bile into bowel associated with phasic physiologic contractions ranging from 0.4% to 0.6% (McMahon et al, 2000). Bile duct
of the sphincter of Oddi (Williams et al, 1984). Delayed imaging injuries may lead either to biliary duct obstruction or bile
can be obtained to allow for detection of biliary passage into leakage. CT and US are first-line imaging modalities to evalu-
bowel, allowing for the possibility of a physiologic contraction ate for suspected complications; nonspecific findings include
of the sphincter of Oddi. Alternatively, relaxation of the sphinc- fluid collection in the gallbladder fossa or peritoneal cavity.
ter of Oddi can be induced by sincalide treatment, to avoid Cholescintigraphy allows the noninvasive evaluation of the
delay. The gallbladder, if present, must accumulate excreted biliary tree for obstruction and leakage and can be used to
tracer normally, before use of sincalide, to exclude cystic duct characterize fluid collections detected by US or CT, particularly
obstruction. If clearance of excreted tracer from the extrahe- benefiting from the use of SPECT/CT to visualize both the
patic biliary tree into bowels, after sincalide administration or fluid collection (on CT) and the abnormal tracer accumulation
at delayed time points, remains less than expected, a partial bile within the collection (on fusion SPECT/CT). Most leaks
duct obstruction can be diagnosed with a specificity of approxi- are observed at the cystic duct stump, and characteristic
mately 85% (Ziessman, 2014). For clarity, cholelithiasis is not scintigraphy findings include tracer accumulation at the
a contraindication to use of sincalide. gallbladder fossa, near the porta hepatis or in the subphrenic
space. SPECT/CT is also advantageous for patients with com-
plicated postsurgical hepatobiliary anatomy (e.g., status post-
Postcholecystectomy Syndrome and Sphincter of Oddi hepaticojejunostomy), to help clarify biliary flow patterns.
Dysfunction (See Chapter 38)
Postcholecystectomy syndrome refers to the recurrence of similar Liver Transplantation
abdominal pain as experienced before cholecystectomy, mainly Biliary imaging objectives for postoperative liver transplant
upper abdominal pain (right upper quadrant) and dyspepsia, patients include detection of biliary anastomotic strictures
with or without jaundice. It is based on a variety of possible caused by thrombosis or stenosis of the hepatic artery or by
hepatobiliary findings and occurs in 10% to 20% of patients recurrent primary sclerosing cholangitis or primary biliary cir-
who have had surgery for chronic cholecystitis (Black et al, rhosis, if applicable, per standard liver transplantation guidelines
1994). One study reported an incidence of even 40%, more (see Chapter 120). For these objectives, the standard guidelines
often occurring in women (Macaron et al, 2011). Onset of focus on endoscopic retrograde cholangiopancreatography,
symptoms can occur from 2 days to as long as 25 years after MRCP, and US. Little to no mention of scintigraphic imaging
surgery. Hepatobiliary causes include retained or recurrent is included, although much clinical research has been conducted
stones, biliary strictures, bile leakage, formation of chronic into the potential value of scintigraphy in the liver transplant
biloma or abscess, long cystic duct remnant, obstruction of the population. Hepatobiliary scintigraphy with IDA-type tracers
sphincter of Oddi, and bile salt–induced diarrhea or gastritis. has been shown to provide accurate diagnosis of biliary com-
In the majority of patients, however, other extrahepatic disor- plications in adult and pediatric patients (Gelfand et al, 1992;
ders (e.g., reflux esophagitis, gastritis, irritable bowel syndrome, Kim et al, 2002; Kurzawinski et al, 1997). However, studies
chronic pancreatitis) are responsible for the postcholecystec- reporting assessment of organ rejection (usually diagnosed by
tomy pain syndrome (Jaunoo et al, 2010). liver biopsy) are limited, particularly in reliable differentiation
Sphincter of Oddi dysfunction is a condition that has been between cholestasis and organ rejection (Brunot et al, 1994;
implicated in 1% to 14% of this patient group (Baillie, 2010; Kim et al, 2002).
Bar-Meir, 1984). The dysfunction can be caused by true ste-
nosis or secondary to spasm of the sphincter. The establishment PEDIATRIC IMAGING
of a diagnosis as well as treatment approaches is challenging.
Sphincterotomy, advocated as effective therapy, is associated Neonatal Jaundice
with bleeding and pancreatitis. Cholescintigraphy and sphincter The incidence of neonatal jaundice in the United States
of Oddi manometry have been explored as assays to confirm has been increasing, currently affecting approximately two-
diagnosis in support of therapeutic intervention. thirds of newborns (Schwartz et al, 2011). In most infants, the
Various protocols assessing qualitative image analysis based etiology of jaundice is physiologic. However, high concentra-
on cholescintigraphy have been published recommending an tions of unconjugated bilirubin may cause permanent neuro-
infusion of sincalide during 3 to 10 minutes, completed 15 logic damage, known as chronic bilirubin encephalopathy or
minutes before radiotracer injection (Sostre et al, 1992). kernicterus. Even moderately elevated levels of bilirubin may
However, another study indicated that sphincter of Oddi induce permanent neurodevelopmental impairment (NDI)
manometry is superior to scintigraphy (Cicala et al, 1991). or bilirubin-induced neurologic dysfunction (BIND). The
Subsequently, another report indicated that although the sensi- various etiologies include infections, isoimmunization, inherited
tivity of scintigraphy is less compared with superior manometry, disorders of bilirubin conjugation and transport, and biliary
the higher accuracy in predicting treatment success compares malformation. Biliary atresia, inspissated bile syndrome and
favorably (Cicala et al, 2002). choledochal malformation are associated with a relatively good
clinical outcome after surgery. Early establishment of the

underlying cause is a key factor to improve associated morbidity
ROLE OF GENERAL NUCLEAR MEDICINE IN
and mortality rates. In pathologic jaundice, neonatal hepatitis ASSESSMENT OF ADULTS WITH MALIGNANT
and biliary atresia account for 70% to 80% of cases; Alagille INVOLVEMENT OF THE LIVER AND
syndrome and α1-antitrypsin deficiency account for 10% to HEPATOBILIARY TREE
15% (Nadel, 1996). The clinical dilemma is that the diseases
have similar clinical presentations and blood test findings. Yet Sulfur Colloid Imaging
biliary atresia, if present, requires urgent surgical intervention In decades past, scintigraphy with radiolabeled sulfur colloid
for optimal pediatric outcomes. 99mTc-IDA cholescintigraphy (SC) provided a useful means for gross structural evaluation of
plays a valuable role in evaluation of neonatal jaundice. Hepa- liver and spleen. Modern CT, MRI, and US provide far superior
tobiliary scintigraphy is performed with 99mTc-labeled mebro- structural/anatomic delineation of organs, almost entirely sup-
fenin as the preferred agent because of its high hepatic planting colloid scintigraphy in anatomic imaging, except in a
extraction. few very specific hepatic indications. For example, colloid
scintigraphy remains in use as second- or third-line imaging for
Biliary Atresia characterization of hepatic lesions (e.g., FNH discussion earlier)
Biliary atresia is a common cause of neonatal cholestasis and and, at some medical centers, as a diagnostic adjunct for hepatic
occurs in 1 in 8000 to 15,000 newborns (Sokol et al, 2003) arterial perfusion scintigraphy (see next). Colloid scintigraphy
(see Chapter 40). Biliary atresia develops either during embryo- has other clinical indications, for conditions other than HPB
genesis or the perinatal period, and different subgroups have diseases (e.g., for lymphoscintigraphy, as a diagnostic adjunct
been postulated, with ongoing research efforts to establish in radiolabeled white blood cell imaging of osteomyelitis, for
underlying causes (Davenport, 2005; Mack & Sokol, 2005) (see detection of splenosis), although a comprehensive discussion is
Chapter 1). Pathologic characteristics include progressive beyond the scope of this chapter.
inflammatory changes, subsequently leading to obliteration of Colloid scintigraphy is an umbrella term, because a variety
intrahepatic and extrahepatic biliary ducts. Treatment is based of radiolabeled colloid particulates have been developed for
on surgical procedures, mostly Kasai, and often liver transplan- human studies. In the United States, 99mTc-labeled SC is the
tation. Early diagnosis within the first 2 months after birth is only radiocolloid in common use for clinical hepatic scintigra-
crucial. phy. 99mTc-labeled SC is a particulate compound. 99mTc-labeled
Cholescintigraphy is valuable as a means to exclude biliary SC particulates administered to humans range from 0.1 to
atresia, but not as a means to positively diagnose biliary 1.0 µm in size. After IV injection, 99mTc-labeled SC normally
atresia. If excreted biliary tracer is detected scintigraphically in localizes at highest concentration in the liver, followed by lesser
the gallbladder or bowels, biliary atresia is effectively excluded uptake in the spleen, and relatively low uptake in bone marrow
(Ziessman, 2014). However, if no passage of biliary tracer into (Fig. 17.8). This liver-spleen-marrow biodistribution reflects
bowels is detectable, the diagnostic interpretation of the scin- the removal of 99mTc-labeled SC particulates from the blood-
tigraphic result is unclear. Biliary atresia does manifest on stream by phagocytes in the reticuloendothelial system (includ-
scintigraphy as a lack of tracer passage into gallbladder or ing hepatic Kupffer cells). Blood clearance is normally rapid
bowels, so the result might be caused by biliary atresia. (2 to 3 minute half-life), allowing scintigraphy to begin 20 to
However, abnormal lack of excreted biliary tracer passage into 30 minutes after injection. In a few hours, catabolism of
the gallbladder or bowels also might be caused by severe
hepatocellular dysfunction from hepatitis or other etiology
(without biliary atresia), or with an associated absence of bile
production and (nonobstructive) halt of biliary flow, or from
interlobular bile duct paucity. Given this possibility of severe
hepatocellular dysfunction leading to nondiagnostic cholescin-
tigraphy, two methodologic approaches are common: (1)
pediatric patients are normally pretreated with phenobarbital
(5 mg/kg/day) for as long as 5 days before cholescintigraphy,
and (2) cholescintigraphy imaging often continues at multiple
time points as long as 24 hours after tracer injection, to allow
for possible markedly slow biliary flow. Phenobarbital before
the scan results in activation of liver enzymes and increased
bile production, enhancing specificity. A 2013 meta-analysis
reported a pooled sensitivity and specificity of 98.7% (range,
98.1%-99.2%) and 70.4% (68.5%-72.2%), respectively, for
diagnosis of biliary atresia. Its near-perfect sensitivity for
biliary atresia makes cholescintigraphy an effective assay
for excluding biliary atresia. Its limited specificity demon-
strates that a lack of biliary tracer transit into gallbladder or
bowel, as a scintigraphic biomarker, has insufficient specificity FIGURE 17.8. Normal appearance of technetium 99m (99mTc)–labeled
to positively diagnose atresia and guide decision making on sulfur colloid scintigraphy. An anterior projection of the abdomen is
whether surgical intervention is necessary (Kianifar et al, shown. The larger mass at left is the liver. The smaller mass is the spleen.
2013). Cholescintigraphy is diagnostically useful as a means Sulfur colloid tracer accumulation in vertebral marrow is sometimes
to “rule out” biliary atresia. faintly detectable, although not in this patient.
99m
Tc-labeled SC can become evident by scintigraphic signs of lipiodol). HAIS applications may be used, when applicable, (1)
catabolic production of free 99mTc pertechnetate; signs of 99mTc to ensure the integrity and function of the infusion system (e.g.,
pertechnetate include (delayed–time point) accumulation of a hepatic pump reservoir-catheter system) and that the biodis-
radioisotope in stomach and thyroid and urinary radioisotope tribution of the infusate meets expectations (discussed next),
excretion. and/or (2) to predict or measure hepatic radiation dosimetry,
Focusing the discussion on liver considerations, a notable before or after infusion of therapeutic radioactive microspheres
abnormality in colloid biodistribution is colloid shift, a term or ethiodized oil (lipiodol).
denoting hepatic 99mTc-labeled SC concentration that is visu- Placement of the arterial catheter used for HAIT and HAIS
ally less than splenic concentration, which is an abnormal involves different possible techniques. The catheter may be
scintigraphic biodistribution. Splenic concentration of 99mTc- surgically implanted (e.g., Watkin technique) or guided into the
labeled SC normally is less than the scintigraphic-intensity of hepatic artery using a minimally invasive approach (e.g., IR)
hepatic concentration. This abnormality can be caused by (see Chapters 99 and 102). Catheters may be placed intermit-
abnormally low hepatic uptake or abnormally high splenic tently (e.g., as part of sequential or cyclic treatments) or longer
uptake of 99mTc-labeled SC. Broad examples of etiologies term (e.g., connected to an implanted drug-delivery device or
associated with low hepatic uptake include diffuse hepatic infusion port). Catheters may connect to a subcutaneously
diseases such as advanced cirrhosis and severe hepatitis. With implanted port or an infusion pump system. Pumps enable
such associations observed, investigators have explored whether slow, continuous drug infusion, typical of hepatic arterial che-
hepatic colloid uptake might offer a semiquantitative biomarker motherapy; such systems are more often found in the United
of hepatic function. In at least one study (Esmaili et al, 2008), States. In a port-catheter system, the implanted port may
no correlation was found between scintigraphic semiquantita- accessed for connection to an external pump system. In Europe,
tive uptake measurements and Child-Pugh classification, an transiently placed catheters and port-catheter systems are more
index of liver function. The finding of diminished colloid uptake common.
in the right and left liver lobes, along with sparing of the caudate The radiotracer used for HAIS, when performed before
lobe, has been associated with Budd-Chiari syndrome (hepatic treatment, is 99mTc macroaggregated albumin (MAA). A radio-
vein thrombosis) (Oppenheim et al, 1988). Various causes of labeled particulate, 25 to 50 µm in mean diameter, the infused
99m
hypersplenism can manifest as abnormally high splenic uptake Tc MAA physically occludes the first microvascular lumen
(Rutland, 1992). it encounters. Several hundred thousand MAA particles are
The differential for “cold defects” or photopenic defects— injected, labeled with a trace amount of activity. Radiolabeled
foci of absent or relatively low colloid uptake—within the liver MAA particles remain intact in vivo for several hours. The
parenchyma is extensive(Oppenheim et al, 1988), including infusate tracer, suspended in a small volume of saline (e.g.,
hepatoma, metastases, and many other pathologic entities. 2 cc), is injected into the pump or transiently placed catheter
Because of limited resolution of the planar and SPECT scans, by a nurse or physician trained in the technique and equipment
only hepatic lesions greater than 1.5 to 2 cm can be reliably necessary for accessing the device. The tracer infusate, followed
detected as colloid-tracer cold-spot or hot-spot lesions. As by a saline flush, traverses the connected catheter into the
previously discussed, to our knowledge, the only hepatic lesion bloodstream of the hepatic artery or accessed branch. Injection
entity reported to produce a focally increased concentration of an infusate can be slow (e.g., <1 cc/min), which mimics
(hot spot) of 99mTc-labeled SC, compared to surrounding liver somewhat the slow rate typical of pump-infusions in HAI
uptake, is focal nodular hyperplasia. Whereas a 99mTc-labeled chemotherapy. Injection of an infusate can be fast (e.g. ≥1 cc/
SC liver hot spot appears to have very high specificity in diag- sec; bolus), which introduces infusate rapidly into the hepatic
nosis of FNH, possibly pathognomonic (Herman et al, 2000; artery, creating turbulence in the bloodstream that mixes
Shamsi et al, 1993), the hot-spot sign has relatively poor sen- infusate and blood more-homogeneously and more rapidly.
sitivity for detection of FNH. Hepatic arterial blood flow is laminar, particularly along
straighter lengths of the artery. Bolus infusions of tracer infusate
Hepatic Arterial Perfusion Studies are typically not advised for HAIS, because the bolus pressure
Hepatic arterial infusion therapy (HAIT) is direct infusion of a may cause retrograde flow of tracer into other celiac branches
therapeutic compound (e.g., chemotherapeutic or radioembolic and associated extrahepatic tracer accumulations (Kaplan et al,
agents) to treat cancerous liver tumors (primary or secondary) 1978).
(see Chapters 96, 99, and 102). Clinical studies demonstrate Extrahepatic tracer accumulations, when tracer is infused
that arterial infusion improves tumor uptake of certain chemo- slowly, are usually attributable to variant hepatic arterial
therapeutic agents compared with portal venous or systemic anatomy, such as branches supplying extrahepatic viscera (e.g.,
venous infusion (Callahan & Kemeny, 2010). Cancerous liver stomach, pancreas, gastrointestinal tract). When the tracer is
tumors derive/stimulate a nutrient blood supply from the arte- bolused, however, such extrahepatic tracer accumulations can
rial system (by tumor neoangiogenesis). Healthy liver receives be mere artifact of bolus technique. True aberrant arterial
blood mostly from the portal venous system. branches typically must be embolized before HAIT, to avoid
Hepatic arterial infusion scintigraphy (HAIS), also called extrahepatic organ toxicity. One of the purposes of HAIS is
hepatic arterial perfusion scintigraphy, is the imaging of the therefore to exclude extrahepatic infusion caused by variant
biodistribution of a radionuclide delivered directly into liver hepatic arterial anatomy.
through an arterial catheter, typically in the (proper) hepatic To identify (1) whether the infused MAA is properly dis-
artery, for delivery of chemotherapy, or potentially into hepatic tributed throughout the downstream liver or targeted hepatic
lobar or segmental arterial branches, when HAIS is conducted subregion and (2) that no extrahepatic infusion is present, a
associated with HAIT radioembolization (e.g., treatment with correlation with abdominal anatomy is needed. SPECT/CT
90
Y-radiolabeled resin or glass microspheres or 131I-radiolabeled provides excellent anatomic localization of infused 99mTc MAA.
MAA (NET)
ANTERIOR 1 2 3 4
R L
5 6 7 8
9 10 11 12
MAA NET FLOW 5SE 5 SEC/FR
FIGURE 17.9. Normal appearance of hepatic arterial infusion scintigraphy. Serial abdominal images were obtained during slow bolus injection of
technetium 99m (99mTc) macroaggregated albumin (MAA) into the common hepatic artery via an implanted hepatic pump-catheter system. Each
image is 5 seconds in duration. On the first image, before the 99mTc MAA has been infused into the liver, faint activity is detectable in the liver and
spleen regions (arrowheads); this faint activity is from 99mTc-labeled sulfur colloid tracer, used immediately before to obtain a gross anatomic outline
(see text). The 99mTc MAA is visualized transiting through the pump catheter, as it is slowly infused (arrow). Progressive accumulation of the 99mTc-MAA
infusate throughout the liver is visualized as the infusion is completed. No abnormal extrahepatic accumulation of 99mTc-MAA infusate is evident.
Unlike the liver, the splenic activity remains low and constant across the frames; thus, dynamic imaging allows confirmation that the splenic activity
is only 99mTc-labeled sulfur colloid.
Alternatively, 99mTc-labeled SC scintigraphy can immediately residual 99mTc-SC or free-pertechnetate signal will remain
precede HAIS, to provide a gross outline of hepatic (and unchanged during dynamic imaging of the 99mTc-MAA infu-
splenic) anatomic contours (see Fig. 17.8). Because 99mTc- sion (Fig. 17.9). 99mTc-MAA SPECT/CT without 99mTc SC
MAA HAIS and 99mTc-SC scintigraphy use the same 99mTc has the advantage of having SPECT imagery that solely rep-
isotope, a larger amount (activity, MBq) of 99mTc MAA is resents 99mTc-MAA biodistribution, although CT typically is
administered than 99mTc SC, so that the signal on the subse- associated with a higher radiation dose to the patient than
quent HAIS scan predominantly represents 99mTc MAA. 99m
Tc-SC scintigraphy.
However, confusion can arise regarding overlap between Abnormal HAIS findings include subtotal hepatic infusion,
99m
Tc-SC and 99mTc-MAA signals, as in evaluation of the extrahepatic organ infusion, catheter obstruction, and catheter
homogeneity of 99mTc-MAA hepatic distribution. Additionally, leakage (Fig. 17.10). Abnormal HAIS findings such as these
if any delay occurs after 99mTc SC is injected, before 99mTc- typically undergo further workup by x-ray angiography (fluo-
MAA scintigraphy is performed, 99mTc-SC catabolic breakdown roscopy) or CT hepatic arteriography, to guide subsequent
and release of free pertechnetate can yield gastric tracer-uptake, management. In a clinical study of patients with implanted
potentially mimicking extrahepatic infusion of the stomach on hepatic arterial pump-catheter systems, HAIS studies revealed
99m
Tc-MAA scintigraphy. Obtaining dynamic images of 99mTc- abnormalities in 9% of patients after pump implantation; pre-
MAA uptake as it is infused slowly during 2 minutes or more, dominantly extrahepatic infusion (63% of abnormal studies)
permits detection of increases in hepatic or extrahepatic tracer (Sofocleous et al, 2006). Abnormal subtotal intrahepatic infu-
uptake indicative of true 99mTc-MAA signal, whereas any sion (i.e., infusion distributed to only a portion of the liver,
when infusion of the entire liver was expected) accounted for

12% of abnormal cases. Abnormal subtotal intrahepatic infu-
sion probably should be distinguished from intrahepatic distri-
bution of infusate that is heterogeneous but has no subregions
devoid of infusate uptake. Intrahepatic distribution of MAA
infusate may be heterogeneous, probably in part because of the
laminar flow phenomenon; in some cases, the presence of a
large hepatic tumor mass appears to draw hepatic arterial flow
away from other hepatic regions. However, if MAA infusate is
detectable throughout the liver, heterogeneities per se have no
clear clinical significance, from available literature. In our
experience, mere heterogeneity of intrahepatic uptake is rela-
tively common and not clearly associated with adverse out-
comes. However, if MAA infusate is clearly, unexpectedly absent
(not merely relatively low) in one or more hepatic subregions,
the finding is potentially clinical significant, possibly the result
of aberrant intrahepatic arterial anatomy; a misplaced catheter;
2 or an occluded arterial branch (e.g., stenotic or thrombosed).
Anecdotal cases, including our own experience, have found that
abnormal subtotal hepatic infusion on HAIS can predict poor
tumor response in those subregions without detectable receipt
of infusate, while tumors in the same patient that appear well
perfused on HAIS respond favorably (Borzutzky & Turbiner,
1985). As such, abnormal subtotal hepatic infusion warrants
further investigation.
X-ray pump angiography performed to evaluate hepatic
pump HAIS abnormalities fails to find corresponding abnor-
malities in approximately 25% of abnormal HAIS studies
(Sofocleous et al, 2006). In those cases, HAIS was repeated,
and the previously identified HAIS abnormality was no longer
evident scintigraphically in almost all cases. In the few cases in
which the HAIS abnormality persisted, radiographic or CT
studies were again repeated, and a corresponding aberrant
vessel was successfully identified. The investigators suggested a
policy of evaluating abnormal HAIS findings by x-ray correla-
tive imaging; if radiographic correlation reveals no concern,
HAIS is repeated in 2 to 3 weeks, seeking spontaneous normal-
ization of HAIS.
6
FIGURE 17.10. Abnormal hepatic arterial infusion scintigraphy. Two

selected anterior abdominal images are shown from a series of
5-second-per-frame images acquired during a 1-minute infusion of
technetium 99m (99mTc) macroaggregated albumin (MAA) into the
common hepatic artery via a hepatic pump-catheter system. The top
image was obtained at 5 seconds and the bottom image at 30 seconds
into the infusion. Abnormal progressive accumulation of 99mTc MAA in
the spleen (arrowhead) and in the gastropancreatic region (arrow) is
visualized. Hepatic activity remains low and constant (indicating it is
merely residual activity from preceding 99mTc-labeled sulfur colloid study;
see text). Subsequently, x-ray pump arteriography identified a severely
stenotic hepatic artery, with arterial blood flow redirected into the left
gastric artery and the splenic artery.
Chapter 17 Role of nuclear medicine in diagnosis and management of hepatopancreatobiliary disease 315.e1
REFERENCES Drane WE, Johnson DA: Sincalide-augmented quantitative hepatobili-

ary scintigraphy (QHBS): definition of normal parameters and
Abdel-Nabi H, et al: Staging of primary colorectal carcinomas with preliminary relationship between QHBS and sphincter of Oddi (SO)
fluorine-18 fluorodeoxyglucose whole-body PET: correlation with manometry in patients suspected of having SO dysfunction, J Nucl
histopathologic and CT findings, Radiology 206:755–760, 1998. Med 31:1462–1468, 1990.
Agency for Healthcare Research and Quality (AHRQ): 2014: National Dunphy MP, Lewis JS: Radiopharmaceuticals in preclinical and clinical
guideline. ACR Appropriateness Criteria: liver lesion initial charac- development for monitoring of therapy with PET, J Nucl Med
terization. Rockville, Md. 50(Suppl 1):106S–121S, 2009.
Akhurst T, et al: Recent chemotherapy reduces the sensitivity of [18F] Eppard E, et al: Ethanol-based post-processing of generator-derived
68
fluorodeoxyglucose positron emission tomography in the detection Ga toward kit-type preparation of 68Ga-radiopharmaceuticals,
of colorectal metastases, J Clin Oncol 23:8713–8716, 2005. J Nucl Med 55:1023–1028, 2014.
American Joint Committee on Cancer (AJCC): AJCC cancer staging Esmaili J, et al: Correlation of liver-spleen scan findings with
handbook, Chicago, 2010. modified Child-Pugh classification, Rev Esp Med Nucl 27:99–102,
Annunziata S, et al: Diagnostic accuracy of fluorine-18-fluorodeoxyglu- 2008.
cose positron emission tomography in the evaluation of the primary European Association for the Study of the Liver (EASL): EASL Clini-
tumor in patients with cholangiocarcinoma: a meta-analysis, Biomed cal Practice Guidelines: management of cholestatic liver diseases,
Res Int 2014:247693, 2014. J Hepatol 51:237–267, 2009.
Baillie J: Sphincter of Oddi dysfunction, Curr Gastroenterol Rep 12: Even-Sapir E, et al: Role of (18)F-FDG dual-head gamma-camera
130–134, 2010. coincidence imaging in recurrent or metastatic colorectal carcinoma,
Bang S, et al: The clinical usefulness of 18-fluorodeoxyglucose positron J Nucl Med 43:603–609, 2002.
emission tomography in the differential diagnosis, staging, and Fernandez FG, et al: Five-year survival after resection of hepatic
response evaluation after concurrent chemoradiotherapy for pancre- metastases from colorectal cancer in patients screened by positron
atic cancer, J Clin Gastroenterol 40:923–929, 2006. emission tomography with F-18 fluorodeoxyglucose (FDG-PET),
Barber TW, et al: The potential for induction peptide receptor chemo- Ann Surg 240:438–447, discussion 447-450, 2004.
radionuclide therapy to render inoperable pancreatic and duodenal Fong Y, et al: Utility of 18F-FDG positron emission tomography scan-
neuroendocrine tumours resectable, Eur J Surg Oncol 38:64–71, ning on selection of patients for resection of hepatic colorectal
2012. metastases, Am J Surg 178:282–287, 1999.
Bar-Meir S, et al: Frequency of papillary dysfunction among cholecys- Gelfand MJ, et al: Hepatobiliary scintigraphy in pediatric liver trans-
tectomized patients, Hepatology 4:328–330, 1984. plant recipients, Clin Nucl Med 17:542–549, 1992.
Bhargava KK, et al: Adenosine triphosphate-binding cassette subfamily Ginj M, et al: Radiolabeled somatostatin receptor antagonists are
C member 2 is the major transporter of the hepatobiliary imaging preferable to agonists for in vivo peptide receptor targeting of tumors,
agent (99m)Tc-mebrofenin, J Nucl Med 50:1140–1146, 2009. Proc Natl Acad Sci USA 103:16436–16441, 2006.
Black NA, et al: Symptoms and health status before and six weeks Guan ZW, et al: Hyperaccumulation of (18)F-FDG in order to dif-
after open cholecystectomy: a European cohort study. European ferentiate solid pseudopapillary tumors from adenocarcinomas and
Collaborative Health Services Study Group, Gut 35:1301–1305, from neuroendocrine pancreatic tumors and review of the literature,
1994. Hell J Nucl Med 16:97–102, 2013.
Blomberg BA, et al: beta-Cell mass imaging with DTBZ positron emis- Hanyaloglu AC, von Zastrow M: Regulation of GPCRs by endocytic
sion tomography: is it possible? Mol Imaging Biol 15:1–2, 2013. membrane trafficking and its potential implications, Annu Rev
Bodei L, et al: Receptor-mediated radionuclide therapy with Pharmacol Toxicol 48:537–568, 2008.
90Y-DOTATOC in association with amino acid infusion: a Phase I Hayakawa N, et al: Clinical utility and limitations of FDG PET in
study, Eur J Nucl Med Mol Imaging 30:207–216, 2003. detecting recurrent hepatocellular carcinoma in postoperative
Borzutzky CA, Turbiner EH: The predictive value of hepatic artery patients, Int J Clin Oncol 19(6):1020–1028, 2014.
perfusion scintigraphy, J Nucl Med 26:1153–1156, 1985. Heinrich S, et al: Neoadjuvant chemotherapy generates a significant
Boulahdour H, et al: The hot spot hepatobiliary scan in focal nodular tumor response in resectable pancreatic cancer without increasing
hyperplasia, J Nucl Med 34:2105–2110, 1993. morbidity: results of a prospective Phase II trial, Ann Surg 248:1014–
Boykin KN, et al: The use of FDG-positron emission tomography for 1022, 2008.
the evaluation of colorectal metastases of the liver, Am Surg 65: Herman P, et al: Hepatic adenoma and focal nodular hyperplasia:
1183–1185, 1999. differential diagnosis and treatment, World J Surg 24:372–376,
Brunot B, et al: Biopsy and quantitative hepatobiliary scintigraphy in 2000.
the evaluation of liver transplantation, J Nucl Med 35:1321–1327, Hwang KH, et al: Evaluation of patients with hepatocellular carcinomas
1994. using [(11)C]acetate and [(18)F]FDG PET/CT: A preliminary
Callahan MK, Kemeny NE: Implanted hepatic arterial infusion pumps, study, Appl Radiat Isot 67:1195–1198, 2009.
Cancer J 16:142–149, 2010. Hyun OJ, et al: An exocrine pancreatic stress test with 11C-acetate PET
Chang JS, et al: Clinical usefulness of (1)(8)F-fluorodeoxyglucose- and secretin stimulation, J Nucl Med 55:1128–1131, 2014.
positron emission tomography in patients with locally advanced Jaunoo SS, et al: Postcholecystectomy syndrome (PCS), Int J Surg
pancreatic cancer planned to undergo concurrent chemoradiation 8:15–17, 2010.
therapy, Int J Radiat Oncol Biol Phys 90:126–133, 2014. Jemal A, et al: Cancer statistics, 2006, CA Cancer J Clin 56:106–130,
Chung EM, et al: From the archives of the AFIP. Pediatric liver masses: 2006.
radiologic-pathologic correlation. Part 1. Benign tumors, Radio- Joyce DL, et al: Preoperative positron emission tomography to evaluate
graphics 30:801–826, 2010. potentially resectable hepatic colorectal metastases, Arch Surg 141:
Cicala M, et al: Quantitative cholescintigraphy in the assessment of 1220–1226, discussion 1227, 2006.
choledochoduodenal bile flow, Gastroenterology 100:1106–1113, Kaemmerer D, et al: Neoadjuvant peptide receptor radionuclide
1991. therapy for an inoperable neuroendocrine pancreatic tumor, World J
Cicala M, et al: Outcome of endoscopic sphincterotomy in post chole- Gastroenterol 15:5867–5870, 2009.
cystectomy patients with sphincter of Oddi dysfunction as predicted Kaltsas GA, et al: The value of radiolabelled MIBG and octreotide in
by manometry and quantitative choledochoscintigraphy, Gut 50: the diagnosis and management of neuroendocrine tumours, Ann
665–668, 2002. Oncol 12(Suppl 2):S47–S50, 2001.
Davenport M: A challenge on the use of the words embryonic and Kaltsas G, et al: Recent advances in radiological and radionuclide
perinatal in the context of biliary atresia, Hepatology 41:403–404, imaging and therapy of neuroendocrine tumours, Eur J Endocrinol
author reply 404-405, 2005. 151:15–27, 2004.
Delbeke D, Martin WH: PET and PET-CT for evaluation of colorectal Kaoutzanis C, et al: Abdominal ultrasound versus hepato-imino diace-
carcinoma, Semin Nucl Med 34:209–223, 2004. tic acid scan in diagnosing acute cholecystitis: what is the real
Dimitroulopoulos D, et al: Long acting octreotide in the treatment of benefit? J Surg Res 188:44–52, 2014.
advanced hepatocellular cancer and overexpression of somatostatin Kaplan WD, et al: Intra-arterial radionuclide infusion: a new technique
receptors: randomized placebo-controlled trial, World J Gastroenterol to assess chemotherapy perfusion patterns, Cancer Treat Rep 62:
13:3164–3170, 2007. 699–703, 1978.
Kemeny NE, et al: KRAS mutation influences recurrence patterns in Moulton CA, et al: Effect of PET before liver resection on surgical
patients undergoing hepatic resection of colorectal metastases, management for colorectal adenocarcinoma metastases: a random-
Cancer 120:3965–3971, 2014. ized clinical trial, JAMA 311:1863–1869, 2014.
Kennedy A, et al: Radioembolization for the treatment of liver tumors: Mukherjee JJ, et al: Treatment of metastatic carcinoid tumours, phaeo-
general principles, Am J Clin Oncol 35:91–99, 2012. chromocytoma, paraganglioma and medullary carcinoma of the
Khan S, et al: Quality of life in 265 patients with gastroenteropancreatic thyroid with (131)I-meta-iodobenzylguanidine [(131)I-mIBG], Clin
or bronchial neuroendocrine tumors treated with [177Lu- Endocrinol (Oxf) 55:47–60, 2001.
DOTA0,Tyr3]octreotate, J Nucl Med 52:1361–1368, 2011. Nadel HR: Hepatobiliary scintigraphy in children, Semin Nucl Med
Kianifar HR, et al: Accuracy of hepatobiliary scintigraphy for differen- 26:25–42, 1996.
tiation of neonatal hepatitis from biliary atresia: systematic review Nahas CS, et al: Positron emission tomography detection of distant
and meta-analysis of the literature, Pediatr Radiol 43:905–919, metastatic or synchronous disease in patients with locally advanced
2013. rectal cancer receiving preoperative chemoradiation, Ann Surg Oncol
Kim CK, et al: Delayed biliary-to-bowel transit in cholescintigraphy 15:704–711, 2008.
after cholecystokinin treatment, Radiology 176:553–556, 1990. National Comprehensive Cancer Network (NCCN): 2015: NCCN
Kim JS, et al: The usefulness of hepatobiliary scintigraphy in the Clinical Practice Guidelines in Oncology. Colon cancer. Hepatobiliary
diagnosis of complications after adult-to-adult living donor liver cancers. Pancreatic adenocarcinoma. Versions 2.2015, NCCN.org.
transplantation, Eur J Nucl Med Mol Imaging 29:473–479, 2002. Oppenheim BE, et al: Liver imaging. In Gottschalk A, et al, editors:
Kim YI, et al: Comparison of F-18-FDG PET/CT findings between Diagnostic nuclear medicine, Baltimore, 1988, Williams & Wilkins, pp
pancreatic solid pseudopapillary tumor and pancreatic ductal adeno- 538–565.
carcinoma, Eur J Radiol 83:231–235, 2014. Patch GG, et al: Naloxone reverses pattern of obstruction of the distal
Kinkel K, et al: Detection of hepatic metastases from cancers of the common bile duct induced by analgesic narcotics in hepatobiliary
gastrointestinal tract by using noninvasive imaging methods (US, imaging, J Nucl Med 32:1270–1272, 1991.
CT, MR imaging, PET): a meta-analysis, Radiology 224:748–756, Reubi JC, Landolt AM: High density of somatostatin receptors in
2002. pituitary tumors from acromegalic patients, J Clin Endocrinol Metab
Kitajima K, et al: Performance of integrated FDG-PET/contrast- 59:1148–1151, 1984.
enhanced CT in the diagnosis of recurrent pancreatic cancer: com- Reubi JC, et al: Regulatory peptide receptors in human hepatocellular
parison with integrated FDG-PET/non-contrast-enhanced CT and carcinomas, Gut 45:766–774, 1999.
enhanced CT, Mol Imaging Biol 12:452–459, 2010. Reubi JC, et al: Somatostatin receptor SST1-SST5 expression
Krenning EP, et al: Somatostatin receptor scintigraphy with in normal and neoplastic human tissues using receptor autoradiog-
[111In-DTPA-D-Phe1]- and [123I-Tyr3]-octreotide: the Rotterdam raphy with subtype-selective ligands, Eur J Nucl Med 28:836–846,
experience with more than 1000 patients, Eur J Nucl Med 20:716–731, 2001.
1993. Rijkers AP, et al: Usefulness of F-18-fluorodeoxyglucose positron emis-
Kurtaran A, et al: 18F-fluorodeoxyglucose (FDG)-PET features of sion tomography to confirm suspected pancreatic cancer: a meta-
focal nodular hyperplasia (FNH) of the liver, Liver 20:487–490, analysis, Eur J Surg Oncol 40:794–804, 2014.
2000. Ruers TJ, et al: Value of positron emission tomography with [F-18]fluo-
Kurzawinski TR, et al: Prospective study of hepatobiliary scintigraphy rodeoxyglucose in patients with colorectal liver metastases: a pro-
and endoscopic cholangiography for the detection of early biliary spective study, J Clin Oncol 20:388–395, 2002.
complications after orthotopic liver transplantation, Br J Surg 84: Rufini V, et al: Imaging of neuroendocrine tumors, Semin Nucl Med
620–623, 1997. 36:228–247, 2006.
Kwatra N, et al: Phenobarbital-enhanced hepatobiliary scintigraphy in Rutland MD: Correlation of splenic function with the splenic uptake
the diagnosis of biliary atresia: two decades of experience at a tertiary rate of Tc-colloids, Nucl Med Commun 13:843–847, 1992.
center, Pediatr Radiol 43:1365–1375, 2013. Ryan ER, et al: Split-dose technique for FDG PET/CT-guided percu-
Kwekkeboom DJ, et al: Treatment with the radiolabeled somatostatin taneous ablation: a method to facilitate lesion targeting and to
analog [177 Lu-DOTA 0,Tyr3]octreotate: toxicity, efficacy, and provide immediate assessment of treatment effectiveness, Radiology
survival, J Clin Oncol 26:2124–2130, 2008. 268:288–295, 2013.
Kwekkeboom DJ, et al: Somatostatin-receptor-based imaging and Santhosh S, et al: Role of (18)F-fluorodeoxyglucose positron emission
therapy of gastroenteropancreatic neuroendocrine tumors, Endocr tomography/computed tomography in the characterization of pan-
Relat Cancer 17:R53–R73, 2010. creatic masses: experience from tropics, J Gastroenterol Hepatol
Lai DT, et al: The role of whole-body positron emission tomography 28:255–261, 2013.
with [18F]fluorodeoxyglucose in identifying operable colorectal Santhosh S, et al: Metabolic signatures of malignant and non-malignant
cancer metastases to the liver, Arch Surg 131:703–707, 1996. mass-forming lesions in the periampulla and pancreas in FDG
Lawrence B, et al: The epidemiology of gastroenteropancreatic neuro- PET/CT scan: an atlas with pathologic correlation, Abdom Imaging
endocrine tumors, Endocrinol Metab Clin North Am 40:1–18, vii, 40(5):1285–1315, 2015.
2011. Schellenberg D, et al: 18Fluorodeoxyglucose PET is prognostic of
Lee JD, et al: Analysis of gene expression profiles of hepatocellular progression-free and overall survival in locally advanced pancreas
carcinomas with regard to 18F-fluorodeoxyglucose uptake pattern cancer treated with stereotactic radiotherapy, Int J Radiat Oncol Biol
on positron emission tomography, Eur J Nucl Med Mol Imaging Phys 77:1420–1425, 2010.
31:1621–1630, 2004. Schoder H, et al: 2-[18F]Fluoro-2-deoxyglucose positron emission
Liu IJ, et al: Evaluation of fluorodeoxyglucose positron emission tomography for the detection of disease in patients with prostate-
tomography imaging in metastatic transitional cell carcinoma with specific antigen relapse after radical prostatectomy, Clin Cancer Res
and without prior chemotherapy, Urol Int 77:69–75, 2006. 11:4761–4769, 2005.
Lu YY, et al: Use of FDG-PET or PET/CT to detect recurrent colorec- Schwartz HP, et al: Hyperbilirubinemia: current guidelines and emerg-
tal cancer in patients with elevated CEA: a systematic review and ing therapies, Pediatr Emerg Care 27:884–889, 2011.
meta-analysis, Int J Colorectal Dis 28:1039–1047, 2013. Shamsi K, et al: Focal nodular hyperplasia of the liver: radiologic find-
Macaron C, et al: Recurrent abdominal pain after laparoscopic chole- ings, Abdom Imaging 18:32–38, 1993.
cystectomy, Cleve Clin J Med 78:171–178, 2011. Shiomi S, et al: Usefulness of positron emission tomography with
Mack CL, Sokol RJ: Unraveling the pathogenesis and etiology of biliary fluorine-18-fluorodeoxyglucose for predicting outcome in patients
atresia, Pediatr Res 57:87r–94r, 2005. with hepatocellular carcinoma, Am J Gastroenterol 96:1877–1880,
Maisey NR, et al: FDG-PET in the prediction of survival of patients 2001.
with cancer of the pancreas: a pilot study, Br J Cancer 83:287–293, Sofocleous CT, et al: Arterial embolization for salvage of hepatic artery
2000. infusion pumps, J Vasc Interv Radiol 17:801–806, 2006.
McMahon AJ, et al: Impact of laparoscopic cholecystectomy: a Sokol RJ, et al: Pathogenesis and outcome of biliary atresia: current
population-based study, Lancet 356:1632–1637, 2000. concepts, J Pediatr Gastroenterol Nutr 37:4–21, 2003.
Mesgarzadeh M, et al: Filling, postcholecystokinin emptying, and refill- Sostre S, et al: A noninvasive test of sphincter of Oddi dysfunction in
ing of normal gallbladder: effects of two different doses of CCK on postcholecystectomy patients: the scintigraphic score, J Nucl Med
refilling: concise communication, J Nucl Med 24:666–671, 1983. 33:1216–1222, 1992.
Chapter 17 Role of nuclear medicine in diagnosis and management of hepatopancreatobiliary disease 315.e3
Sperti C, et al: Tumor relapse after pancreatic cancer resection is Waldherr C, et al: The clinical value of [90Y-DOTA]-D-Phe1-Tyr3-
detected earlier by 18-FDG PET than by CT, J Gastrointest Surg octreotide (90Y-DOTATOC) in the treatment of neuroendocrine
14:131–140, 2010. tumours: a clinical Phase II study, Ann Oncol 12:941–945, 2001.
Stoeltzing O, et al: Staged surgery with neoadjuvant 90Y-DOTATOC Waldherr C, et al: Tumor response and clinical benefit in neuroendo-
therapy for down-sizing synchronous bilobular hepatic metastases crine tumors after 7.4 GBq (90)Y-DOTATOC, J Nucl Med
from a neuroendocrine pancreatic tumor, Langenbecks Arch Surg 43:610–616, 2002.
395:185–192, 2010. Wang X, et al: Comprehensive evaluation of a somatostatin-based
Strobel K, et al: Contrast-enhanced 18F-FDG PET/CT: 1-stop-shop radiolabelled antagonist for diagnostic imaging and radionuclide
imaging for assessing the resectability of pancreatic cancer, J Nucl therapy, Eur J Nucl Med Mol Imaging 39:1876–1885, 2012.
Med 49:1408–1413, 2008. Wang XY, et al: Value of (1)(8)F-FDG-PET/CT in the detection of
Topkan E, et al: Predictive value of metabolic 18FDG-PET response recurrent hepatocellular carcinoma after hepatectomy or radiofre-
on outcomes in patients with locally advanced pancreatic carcinoma quency ablation: a comparative study with contrast-enhanced
treated with definitive concurrent chemoradiotherapy, BMC Gastro- ultrasound, J Dig Dis 14:433–438, 2013.
enterol 11:123, 2011. Wang XY, et al: Utility of PET/CT in diagnosis, staging, assessment of
Torizuka T, et al: Value of fluorine-18-FDG-PET to monitor hepatocel- resectability and metabolic response of pancreatic cancer, World J
lular carcinoma after interventional therapy, J Nucl Med 35:1965– Gastroenterol 20:15580–15589, 2014.
1969, 1994. Wester HJ, et al: Comparison of radioiodinated TOC, TOCA and Mtr-
Torizuka T, et al: In vivo assessment of glucose metabolism in hepato- TOCA: the effect of carbohydration on the pharmacokinetics, Eur J
cellular carcinoma with FDG-PET, J Nucl Med 36:1811–1817, Nucl Med Mol Imaging 29:28–38, 2002.
1995. Wild D, et al: First clinical evidence that imaging with somatostatin
Tulchinsky M, et al: SNM practice guideline for hepatobiliary scintig- receptor antagonists is feasible, J Nucl Med 52:1412–1417, 2011.
raphy 4.0, J Nucl Med Technol 38:210–218, 2010. Wild D, et al: Comparison of somatostatin receptor agonist and antago-
Valkema R, et al: Survival and response after peptide receptor radionu- nist for peptide receptor radionuclide therapy: a pilot study, J Nucl
clide therapy with [90Y-DOTA0,Tyr3]octreotide in patients with Med 55:1248–1252, 2014.
advanced gastroenteropancreatic neuroendocrine tumors, Semin Williams W, et al: Scintigraphic variations of normal biliary physiology,
Nucl Med 36:147–156, 2006. J Nucl Med 25:160–165, 1984.
Van Cutsem E, et al: Metastatic colorectal cancer: ESMO Clinical Wise DR, Thompson CB: Glutamine addiction: a new therapeutic
Practice Guidelines for diagnosis, treatment and follow-up, Ann target in cancer, Trends Biochem Sci 35:427–433, 2010.
Oncol 25(Suppl 3):iii1–iii9, 2014. Wu LM, et al: Diagnostic value of diffusion-weighted magnetic reso-
Van de Velde CJH, et al: EURECCA. Multidisciplinary management: nance imaging compared with fluorodeoxyglucose positron emission
European consensus conference, Eur J Cancer 50:1.e1–1.e34, 2014. tomography/computed tomography for pancreatic malignancy: a
Van Essen M, et al: Peptide receptor radionuclide therapy with meta-analysis using a hierarchical regression model, J Gastroenterol
177
Lu-octreotate in patients with foregut carcinoid tumours of bron- Hepatol 27:1027–1035, 2012.
chial, gastric and thymic origin, Eur J Nucl Med Mol Imaging 34: Yamamoto Y, et al: Detection of hepatocellular carcinoma using
1219–1227, 2007. 11C-choline PET: comparison with 18F-FDG PET, J Nucl Med
Van Essen M, et al: Report on short-term side effects of treatments 49:1245–1248, 2008.
with 177Lu-octreotate in combination with capecitabine in seven Zakko SF, Afdhal NH: Acute cholecystitis: Pathogenesis, clinical fea-
patients with gastroenteropancreatic neuroendocrine tumours, Eur J tures, and diagnosis. In Chopra S, editor: UpToDate, Waltham, Mass.
Nucl Med Mol Imaging 35:743–748, 2008. (Accessed March 3 2015).
Virgolini I, et al: Procedure guidelines for PET/CT tumour imaging Zakko SF, Zakko WF: Functional gallbladder disorder in adults. In
with 68Ga-DOTA-conjugated peptides: 68Ga-DOTA-TOC, Chopra S, editor: UpToDate, Waltham, Mass. (Accessed March 3
68Ga-DOTA-NOC, 68Ga-DOTA-TATE, Eur J Nucl Med Mol 2015).
Imaging 37:2004–2010, 2010. Ziessman HA: Acute cholecystitis, biliary obstruction, and biliary
Wahl RL, et al: From RECIST to PERCIST: evolving considerations leakage, Semin Nucl Med 33:279–296, 2003.
for PET response criteria in solid tumors, J Nucl Med 50(Suppl Ziessman HA: Hepatobiliary scintigraphy in 2014, J Nucl Med 55:
1):122S–150S, 2009. 967–975, 2014.
CHAPTER 18
Computed tomography of the liver, biliary tract,
and pancreas
Seth S. Katz
IMAGING OVERVIEW AND HISTORY

distribution and review of images. Electronic review consider-
Computed tomography (CT) is a relatively mature cross- ably improved radiologists’ ability to interpret the numerous
sectional diagnostic imaging modality in which delineation of images acquired in each examination. Scrolling rapidly through
anatomy and characterization of disease rely upon differences overlaid adjacent image slices (“cine” viewing) aids apprecia-
in the ability of various tissues (modified by administered tion and evaluation of tubular and other structures whose
contrast agents) to attenuate X-ray beams. Relative to magnetic course is skewed to the axial plane, such as dilated bile ducts
resonance imaging (MRI), a newer and more rapidly evolving and hepatic vessels. The ability to magnify images and alter
medical imaging modality, CT remains superior at in-plane and their contrast and brightness on the fly also greatly facilitates
z-axis (interslice) resolution. image interpretation. The many types of postprocessing
The usage of CT in clinical diagnosis has undergone great (Salgado et al, 2003) aided by voxel isotropy and digital display
expansion since its inception in the 1970s, spurred particularly include maximum intensity projection (MIP), where processed
over the past 2 decades by improvements in scan quality, speed, images depict only the brightest structures within the specified
and ease of display and review. Key innovations to quality data volume (Johnson et al, 1996, 2003; Magnusson et al,
improvement were helical scanning in the late 1980s and new 1991; Rubin, 2003; Salgado et al, 2003; Zeman et al, 1994).
technology (MDCT) beginning in the late 1990s (Goldman, Using portal venous phase source images, MIP images produce
2008). Helical scanning uses a slip ring to allow continuous tube a more robust depiction of portal venous and hepatic venous
gantry rotation while the patient moves smoothly through the anatomy (Fig. 18.1). Volume-rendered reconstructions provide
bore, increasing scanning speed. The advent of MDCT over- a lifelike three-dimensional (3D) model of complex anatomy
came a major obstacle—limited tube cathode heat dissipation that can be rotated in any direction (Fig. 18.2) (Johnson et al,
capacity that constrained imaging to either wide collimation 1996, 2003; Rubin, 2003; Zeman et al, 1994) transforming the
(thick slices) or small-volume coverage. Increasing the number numerous axial images into an interactive 3D model that may
of thin z-axis detector rows captured more of the X-rays emitted be more intuitive to the surgeon. Hepatic arterial anatomy can
by the tube cathode (effectively acquiring several axial slices at be similarly displayed (Fig. 18.3), and the relationship of tortu-
a time) permitting even faster scanning and much smaller slice ous vessels to the hepatic parenchyma and lobar anatomy can
thickness (z-axis resolution). The reduction in scanning times of be depicted clearly (Fig. 18.4). It is also possible to view a
even large body segments to several seconds, well below the rotating model that depicts the relationship of a mass to the
threshold for a typical breathhold, greatly decreased motion adjacent vessels from any angle to help the surgeon conceptual-
artifacts and allowed multiphasic evaluation (hepatic arterial, ize preoperatively what would be encountered at surgery (Fig.
portal venous, and delayed phases) of the liver and pancreas. 18.5) (see Chapters 66, 103, and 104).
More than one acquisition through the liver is now possible in a Although MRI offers superior tissue contrast and capability
single breathhold. Thinner collimation improves resolution and for functional imaging (and thus superior lesion characteriza-
increases detection of small hepatic lesions by up to 46% (Weg tion), the superior spatial resolution, ease of interpretation,
et al, 1998). At current array sizes (most modern scanners now greater availability, and lower cost of CT maintain it as the most
use at least 64 to 128 detector rows), resolution substantially frequently requested modality in evaluating the liver and pan-
below 1 mm is now available in all planes. Although the thinnest creas. A major drawback remains usage of ionizing radiation,
possible slices are not routinely reconstructed and displayed, with a trend in past decades toward increasing exposure of the
any of several desired slice thicknesses down to the width of the patient population to nontrivial doses, an issue attracting
thinnest detector rows in the array (usually 0.625 to 0.31 mm at increasing popular scrutiny and spurring substantial ongoing
128 rows, depending on the model and vendor) may be recon- attention among technical innovators.
structed after the study is acquired, as long as the raw scan data
remain available. Scan resolution equal in all planes—perfect RECENT TECHNICAL ADVANCES
cubes of imaging data or “voxel isotropy” (which was achieved
by the generation of 16-detector scanners in 2002)—yields Several recent advances and current avenues of research
smoother images when the data acquired in the axial plane are promise to have considerable positive impact on hepatopan-
reformatted in the coronal, sagittal, and oblique planes. creatobiliary CT: addressing the image quality/radiation dose
The introduction of picture archiving and communication axis, conspicuity of lesions, ability to evaluate iron and fat
systems (PACS) technology revolutionized CT image display. deposition, and capacity to characterize lesions and identify
PACS facilitates storage, rapid retrieval, and electronic underlying diffuse disease.
316
Chapter 18 Computed tomography of the liver, biliary tract, and pancreas 317
A B
FIGURE 18.1. A, Thick-slab maximum intensity projection (MIP) through the portal veins reveals standard bifurcation of the main portal vein into
the right (arrow) and left (arrowhead) portal veins. The anterior and posterior sectoral branches of the right portal vein are evident, as are the medial
and lateral sectoral branches of the left portal vein. B, Thick-slab MIP image through the hepatic veins reveals the three major hepatic veins
(arrowheads) and an accessory hepatic vein (arrow) draining into the inferior vena cava.
Fissure for
falciform ligament
A B
FIGURE 18.2. A, Volume-rendered three-dimensional (3D) image of the external contour of the liver created from contrast-enhanced axial computed
tomographic data. Note the fissure for the falciform ligament seen en face (arrow). B, Volume-rendered 3D image of the same patient rotated upward
reveals the undersurface of the liver. Note the surgical clips in the gallbladder fossa and the bare area of the liver (short arrow).
The increasing use of the modality, the propensity of early control systems (McCollough et al, 2006), until relatively
MDCT to generate higher radiation doses per examination, and recently, the traditional algorithm by which raw acquired data
increasing use of multiphasic studies in liver and pancreas CT are transformed into images, called filtered back projection
have made dose reduction a priority (Brenner & Hall, 2007). (FBP), had remained untouched throughout almost 40 years
Although some headway (20% to 40% dose reduction) had of CT progress. Around 2011, an improved “partially iterative”
been made early in the past decade with automatic exposure reconstruction algorithm called adaptive statistical iterative
FIGURE 18.3. Computed tomography (CT) angiogram. Volume-

rendered three-dimensional image created from axial contrast-enhanced
CT data reveals classic (or standard) arterial branching anatomy. The
celiac axis (small solid arrow) trifurcates into the splenic artery, common
hepatic artery (CHA; long solid arrow), and left gastric artery. The CHA
gives rise to the gastroduodenal artery (open arrowhead) and the proper
hepatic artery (solid arrowhead). The right hepatic artery (long open
arrow) and the left hepatic artery (small open arrow) originate from the
proper hepatic artery.
FIGURE 18.5. Volume-rendered three-dimensional image of the liver

from above reveals a large colorectal metastasis involving the right lobe
with extension into the medial sector of the left lobe. The mass is in
continuity with the middle hepatic vein (long arrow) at its confluence with
the inferior vena cava (IVC; open arrow). The middle hepatic vein and
left hepatic vein (short arrow) drain separately into the IVC.
rsDECT) or two radiation sources of differing energies (“dual

source” or dsDECT) are used to produce images. For the liver,
biliary system, and pancreas, applications include “material-
specific imaging” and “virtual monochromatic imaging”
(Morgan, 2014). In the liver, “virtual iron images” may be used
for quantitation of hepatic iron deposition without interference
from coexisting steatosis (Joe et al, 2012), whereas “virtual
non-iron images” can be created for fat quantitation indepen-
dent of concomitant siderosis (Zheng et al, 2013). “Virtual
FIGURE 18.4. Three-dimensional reformatted image reveals the iodine mapping” may also aid in identifying tumor thrombus
course of the left portal vein, which ascends in the umbilical fissure and distinguishing it from bland thrombus in the setting of
before looping anteriorly and inferiorly (short arrow) to branch into the hepatocellular carcinoma (HCC) (Qian et al, 2012). In the
medial (arrowhead) and lateral (long arrow) sectoral branches. pancreas, lower kilovolt peak images from “virtual monochro-
matic” dsDECT (Macari et al, 2010) and lower kiloelectron
volt images from virtual monochromatic rsDECT (Patel et al,
reconstruction (ASIR) became available in clinical scanners 2013) have been shown to improve conspicuity of hypovascular
producing lower noise and better-quality images per dose than tumors, such as typical ductal adenocarcinoma, an application
FBP, resulting in additional dose reduction benefits in the possibly extensible to hypervascular liver lesions. Lower-energy
20% to 40% range. With more powerful computer hardware virtual monochromatic images may also provide more robust
currently available, newer, previously prohibitively processor- surface-rendered 3D arterial images, whereas virtual higher-
intensive iterative reconstruction algorithms are now approach- energy images have less apparent metallic artifact around biliary
ing clinical availability. One such method called model-based stents and clips (Morgan, 2014). These images are in some
iterative reconstruction may yield dose reductions of 70% to institutions already being routinely sent to PACS for clinical
90% relative to classic FBP (Pickhardt et al, 2012) and 59% use, and it is possible to enable display of a continuum of virtual
relative to ASIR while maintaining a reasonable quality of liver monochromatic energies on the fly at the time of image inter-
evaluation (Shuman et al, 2013, 2014). pretation to aid lesion detection. There is also early indication
Another ongoing development over the past decade is dual that material-specific iodine imaging improves conspicuity of
energy CT (DECT), a technology in which either a single neuroendocrine lesions and increases sensitivity for complexity
radiation source alternates beam energies (“rapid switching” or within cystic lesions (Chu et al, 2012).
An exciting methodology called texture analysis, not yet in (see Chapter 2) and to assess the proximity of lesions to the
clinical use, is being brought to bear in interpretation. Not inflow and outflow vessels. This information determines the
specific to CT, the analysis of pixel intensity variation can be extent of potential hepatic resection required to achieve clear
made using data produced by any modality. Preliminary results surgical margins (Figs. 18.7 and 18.8). If a proposed operation
suggest that certain hepatic texture features can predict survival will be extensive, postprocessing of CT images with 3D volume
in colon cancer patients (Miles et al, 2009), can identify the rendering can also identify segmental or lobar hepatic atrophy
presence of colon cancer hepatic micrometastases not yet visible (Fig. 18.9) and compensatory hypertrophy. Manual or auto-
by standard visual image review (Rao et al, 2014), and can mated segmentation with volumetric calculation can help predict
preoperatively stratify risk for postoperative hepatic failure in whether postoperative liver volume will be sufficient to safely
candidates for major hepatic surgery (Simpson et al, 2015).
Texture analyses also show promise in discriminating among
hypervascular liver lesions, potentially increasing specificity in
the noninvasive diagnosis of HCC, hepatocellular adenoma,
and focal nodular hyperplasia (FNH).
ROLE OF COMPUTED TOMOGRAPHY IN LIVER

AND BILIARY IMAGING
CT imaging had traditionally primarily been used for hepatic
lesion detection and characterization and is still frequently used
in this capacity and in the posttreatment follow-up setting.
Because of its superior tissue contrast, MRI is now probably
better suited to hepatic lesion characterization (see Chapter
19). However, the broader and more rapid body coverage of
CT allows it to excel at identifying peritoneal disease involve-
ment, distant lymphadenopathy, and other distant metastatic
disease, and its superior spatial resolution best suits evaluation
of key vessel involvement by tumor. Therefore CT is best
positioned for a role in initial staging of patients with malignant
disease and preprocedural vascular evaluation.
In planning for surgery and other locoregional therapy, portal
and hepatic veins can be identified as anatomic landmarks
to localize tumors to specific hepatic segments (Fig. 18.6)
FIGURE 18.7. Contrast-enhanced computed tomography reveals a

small malignant lesion in segment I abutting the middle hepatic vein
(open arrow) and left hepatic vein (solid arrow) at their confluence with
the inferior vena cava (arrowhead). The location of this small mass
mandates a large hepatic resection to achieve clear surgical margins.
7 2
8 4a
3
4b
5
6
FIGURE 18.6. Volume-rendered three-dimensional model of the liver

created from axially acquired computed tomographic data. The superior
portion of the hepatic parenchyma has been removed to reveal the
course of the hepatic veins through the liver substance as they drain FIGURE 18.8. Contrast-enhanced computed tomography reveals a
into the inferior vena cava. The gallbladder is depicted in green, and segment I mass that abuts the inferior vena cava (IVC; arrowhead). Two
hepatic segments are indicated. accessory hepatic veins (arrows) are draining into the IVC.
FIGURE 18.9. Right lobe atrophy as a result of an infiltrative tumor

involving the right portal pedicle (arrow). Note the counterclockwise
rotation of the porta hepatis.
FIGURE 18.10. Arterial phase image reveals the right hepatic artery
(arrow) coursing anterior to the common hepatic duct, which contains
avoid hepatic failure (see Chapters 100 and 108). Steatosis and
air from a papillotomy.
cirrhosis, processes that affect the remnant liver’s ability to
regenerate, can also be recognized (although not as clearly as by
MRI).
Early arterial-phase high-resolution images (CT angiogra- angiography. Because most neoplasms involving the liver are
phy) are obtained to assess tumor involvement of arteries and hypovascular relative to adjacent normal liver parenchyma,
evaluate variant celiac axis anatomy, which occurs in more than routine abdominal imaging generally uses only the portal
half of the population (Michels, 1955). Because visualization of venous phase, in which these are most conspicuous. Hypervas-
the surgical field is often limited in patients with prior surgery, cular neoplasms such as HCC and certain types of metastasis,
obesity, hepatobiliary malignancy, and local inflammation, including neuroendocrine tumors, renal cell carcinoma, mela-
preoperative knowledge of variant celiac anatomy can assist noma, and thyroid cancer, are generally evaluated using a tri-
in surgical planning (see Chapters 2, 103, and 104), facilitate phasic protocol that includes precontrast and late arterial-phase
dissection, and help the surgeon avoid iatrogenic injury images as well as portal venous phase images. Delayed phases
(Winston et al, 2007). Identification of variant anatomy is also (see later) are now deemed important to add in evaluating for
crucial to decisions regarding hepatic arterial embolization and possible HCC (Bruix & Sherman, 2011) and in follow-up of
placement of hepatic arterial chemoinfusion pumps; gains in known HCC (Lencioni et al, 2010). Five-millimeter axial slices
image resolution during the MDCT era have allowed CT through the liver are routinely reviewed. Thinner images can be
angiography to largely supplant invasive direct catheter angiog- requested by the radiologist on any type of examination in
raphy for planning placement of intraarterial chemoinfusion problem-solving situations, provided the raw image data are still
pumps (Kapoor et al, 2003) (see Chapter 99). Vascular mapping extant on the scanner, usually for a day or two after the acquisi-
may now also be performed with CT. The relationship of arter- tion. Coronal and sagittal reformatted images (2.5 mm) are
ies to the bile ducts can be identified, which is also valuable for generally routinely generated for review in addition to axial
surgical planning (Fig. 18.10) (see Chapters 2 and 103). Late images on all examinations, and more advanced 3D postpro-
arterial-phase images (included in the “triphasic” examination, cessing is generally made available on certain studies, such as
unlike early arterial images) are obtained for detection and CT angiographic examinations.
delineation of hypervascular hepatic lesions.
Noncontrast Computed Tomography of the Liver
Technique and Protocols An unenhanced scan is generally included in the triphasic
State-of-the-art CT evaluation of the liver for hepatic resection dedicated liver evaluation. The normal unenhanced liver typi-
now may include, depending on the indication, any combina- cally has attenuation values between 54 and 60 Hounsfield
tion of image acquisition in the precontrast, IV contrast- units (HU), approximately 8 HU greater than the spleen due
enhanced early (CT angiography) and late arterial phase, portal to hepatic glycogen and iron stores. Abnormally increased
venous phase, and delayed phases. Oral hyperattenuating “posi- attenuation may be seen in hemochromatosis (Fig. 18.11),
tive” contrast medium (water-soluble iodinated agent or dilute glycogen storage diseases, Wilson disease, β-thalassemia, sickle
barium) is typically administered for routine examinations, cell disease, and after administration of certain drugs (amioda-
although some institutions now advocate regular use of oral rone, cisplatin), whereas abnormally decreased hepatic attenu-
water as a negative contrast agent (Kammerer et al, 2015; ation is most frequently attributable to steatosis (Fig. 18.12).
Makarawo, 2013), as is the standard practice with CT These processes may be obscured by the administration of
FIGURE 18.13. Unenhanced computed tomography reveals numer-

ous calcifications in a patient with amyloidosis.
FIGURE 18.11. Nonenhanced computed tomography through the

liver of a patient with hemochromatosis shows diffuse increased attenu-
ation of the liver compared with the spleen. 75% to 80% of its blood supply from the portal venous system
and only 20% to 25% from the hepatic artery, but the majority
of primary liver tumors receive most of their blood supply from
the hepatic arterial system (Fournier et al, 2004; Honda et al,
1992). Differential enhancement of various types of lesion rela-
tive to background hepatic parenchyma dictates which timing
or “phase(s)” to use in the evaluation.
ARTERIAL PHASE IMAGING. Different methods are available to

determine the time of maximal arterial enhancement for each
patient (which depends on factors such as cardiac function
and state of hydration) (Sica et  al, 2000), including fixed
timing, timing bolus (Kalra et  al, 2004), and commercially
available autoattenuation detection. Higher injection rates (4
to 6  mL/sec) are required for the early arterial imaging used
for CT angiography. Images are acquired at about 20 to 30
seconds after injection initiation to provide maximal contrast
enhancement of the arteries, keeping veins and abdominal
organs nearly unopacified. Because the same volume of IV
FIGURE 18.12. Nonenhanced computed tomography through the contrast agent is administered as for routine abdominal CT,
liver of a patient with steatosis shows hypoattenuation of the hepatic this phase may be conveniently added to routine or triphasic
parenchyma compared with the hepatic blood vessels, liver capsule, and evaluations, although oral positive contrast interferes with
spleen. creation of segmented 3D arterial images, so oral water is
substituted. Images can be reviewed at 1.25  mm collimation
in the axial plane, at 2.5  mm in the orthogonal planes, and
intravascular contrast medium. Calcifications (Fig. 18.13) and 3D manually segmented volume-rendered images of the arter-
hemorrhage are also better visualized on unenhanced CT. ies are made available on PACS.
Hepatic neoplasms usually have high water content and are The late arterial phase used to evaluate hypervascular
thus slightly hypoattenuating relative to normal liver paren- lesions is usually acquired at about 40 seconds after the start
chyma, as are the unenhanced portal veins and bile ducts. The of injection. Certain lesions, such as FNH and hemangiomata
attenuation difference between liver and tumor may be subtle, (discussed in greater detail later in this chapter) (see Chapter
and although conspicuity can be improved by review at very 90), have specific patterns of enhancement on arterial and
narrow window widths, much higher lesion-to-liver attenuation portal venous phase images (Fig. 18.14). Malignant hypervas-
differential can be achieved by the administration of IV contrast cular neoplasms, such as HCC (see Chapter 91) and metastatic
medium (Tidebrant et al, 1990). neuroendocrine carcinoma (see Chapter 93), often enhance
prominently at this time with increased lesion conspicuity but
Contrast-Enhanced Computed Tomography of the Liver may become isoattenuating to the remaining liver and thus
The IV administration of an iodinated contrast agent improves undetectable in the portal venous phase (Fig. 18.15). Conse-
lesion detection and characterization and enables vascular quently, when HCC or metastatic disease from hypervascular
mapping. Steady continuous injection is essential to image primary tumors is suspected, the triphasic liver evaluation
quality, requiring a dedicated power injector. The liver receives (containing late arterial-phase images) should be requested.
A B
FIGURE 18.14. A, Arterial phase image through the inferior right hepatic lobe reveals branches of the right hepatic artery. Faint enhancement of
a hypervascular mass is seen in segment VI, which proved to be focal nodular hyperplasia (arrow). B, Portal venous phase image through the same
level reveals enhancement of the right portal vein (long arrow) and the sectoral branches. The mass (short arrow) is nearly hypointense to the liver.
A B
FIGURE 18.15. Arterial and portal venous phase images as part of a triphasic examination in a patient with metastatic renal cell carcinoma.
A, Late arterial-phase image reveals multiple hypervascular enhancing foci of metastatic renal cell carcinoma (arrows). B, Portal venous phase image
at the same level does not reveal the numerous masses. The low-attenuation lesion represents a partially treated hypoattenuating metastasis.
PORTAL VENOUS PHASE IMAGING. Normal noncirrhotic hepatic “equilibrium” phase) may be added to the traditional triphasic
parenchyma enhances maximally at approximately 70 seconds examination—a four-phase CT. Some intrahepatic cholangio-
after injection initiation. At this time, referred to as the portal carcinomas may be inconspicuous on the traditional components
venous phase of enhancement, maximal contrast differential of the triphasic examination, and considerably delayed imaging
between typical hypovascular liver lesions and the surrounding may be added when this entity is suspected (see Chapter 50). No
parenchyma is achieved, as well as clear delineation of the specific optimal timing for this indication has been firmly estab-
portal and hepatic veins. Routine evaluation of the liver should lished, but a delay of 5 to 12 minutes is typically used.
be performed using this phase timing.
Anatomy
DELAYED IMAGING PHASES. Malignant hypervascular tumors, An extensive and detailed discussion of hepatic and biliary
particularly HCC (see Chapter 91), typically become hypoat- anatomy, some of it using CT images, can be found in Chapter
tenuating to adjacent normal hepatic parenchyma at some point 2. Axial CT images are displayed by convention as if the patient
after the late arterial phase. Although this is often seen in the were lying face up with feet toward the viewer and head away
routinely acquired portal venous phase, it may only be appreci- from the viewer into the viewing screen, so that the patient’s
ated on further delayed imaging (Iannaccone et al, 2005; right side, including the liver is to the interpreter’s left—the
Liu et al, 2012). Therefore, when HCC is suspected, a 3- to viewer is effectively looking cephalad at the caudal surface of
5-minute delayed phase (variably described as “late venous” or each slice.
Gross Morphology 3D reformatting of axial CT data can assist in localization of

Although three fissures interrupt the surface of the liver and the hepatic segments (see Fig. 18.6).
help define its segmentation (see Chapter 2), only two are
readily appreciated on review of axial CT images. Vascular Anatomy
The umbilical fissure (also known as the fissure for the liga- After following the portal scissurae tracing the boundary
mentum teres) notches the anteroinferior portion of the dia- between the hepatic sectors, the hepatic veins have a short
phragmatic surface of the liver and is located in the left hepatic extrahepatic course (Blumgart et al, 2001). The middle and left
lobe slightly to the right of the patient’s midline, oriented hepatic veins often join before draining into the IVC, whereas
roughly in the sagittal plane (Fig. 18.16D). Carrying the liga- the right hepatic vein joins the IVC directly. A variable number
mentum teres and some surrounding fat, it defines the inferior of retrohepatic veins drain from the posterior aspect of the
aspect of the border between the medial and lateral segments liver into the IVC, and an accessory right hepatic vein, a large
of the left liver. Because it contains the left portal pedicle, the inferior right hepatic vein, may be present in some patients (see
umbilical fissure is not synonymous with the left portal scissura Fig. 18.1B) (see also Chapter 2).
because the portal scissurae separate portal pedicles rather than The portal vein is formed by the junction of the splenic vein
containing them. Rather, the left portal scissura runs along the and superior mesenteric vein immediately posterior to the neck
course of the left hepatic vein, defining the superior aspect of of the pancreas. It courses cephalad toward the right as it enters
the border between left medial and lateral segments. the hepatoduodenal ligament, within which it courses parallel
The fissure for the ligamentum venosum, the other surface- to the proper hepatic artery, lying just anteromedial, and the
interrupting fissure well seen on axial CT (Fig. 18.16C), sepa- common bile duct (CBD), located just anterolateral. After
rates segments II and III from the caudate lobe and is continuous entering the hepatic parenchyma, the portal vein branches to
with the umbilical fissure. supply each portal sector (see Fig. 18.1A) and subsequently
The main hepatic or interlobar scissura, an incomplete each segment. Portal venous variants, well delineated on CT
structure forming the inferior separation of the right and left (Fig. 18.17), are not uncommon. Covey and colleagues (2004)
hepatic lobes, is not easily appreciable on axial CT slices. reported variant portal venous anatomy in 35% of a group of
Oriented in a vertical plane oblique to the sagittal and coronal 200 patients.
planes, its inferior extent interrupts the inferior liver surface Standard or classic hepatic arterial supply arises from the
following the gallbladder fossa (Fig. 18.16E), whereas its celiac trunk via the common hepatic artery, which gives rise to
superior extent approximates the course of the middle hepatic the gastroduodenal and proper hepatic arteries (see Fig. 18.3).
vein (Fig. 18.16B). The proper hepatic artery divides into right and left hepatic
arteries before entering the liver, but the precise location of the
Segmental Anatomy bifurcation varies. Variant hepatic and celiac arterial anatomy,
A surgically useful functional scheme corresponding to the liver reported in 55% of patients based on initial cadaveric dissec-
anatomy has been described (Bismuth, 1982), and the 1957 tions by Michels (1955), is demonstrable by CT (Figs. 18.18
Couinaud description is schematically reproduced in Chapter and 18.19) and has significance for operative and other vascular
2. Hepatic segmentation by CT is based primarily on venous procedural management, particularly insertion of hepatic arte-
anatomy (hepatic venous outflow and the portal inflow pedicles), rial chemoinfusion pumps.
also using some morphology features. The four sectors, each
receiving its own separate portal venous and hepatic arterial Biliary Anatomy
supply and excreting to a separate bile duct, are separated by Ductal anatomy and its variants are discussed quite extensively
the portal scissurae following the three main hepatic veins. in Chapter 2. Briefly, the intrahepatic biliary tree ramifies in
Because contrast-enhanced axial CT readily shows the a fashion nearly identical to the portal system (Figs. 18.20
course of all three major hepatic veins on superior slices near and 18.21). Along the anterior aspect of the main portal
the diaphragm, the hepatic sectors are most easily identified at venous bifurcation, the left and right hepatic ducts join to
this level (see Chapter 2). The interlobar scissura, following the form the common hepatic duct, which courses caudally and
middle hepatic vein, divides the liver into right and left lobes. posteriorly toward the left within the hepatoduodenal liga-
The right lobe is further divided into two sectors by the right ment, always maintaining its position anterolateral to the
portal scissura following the right hepatic vein, and the left liver portal vein. It becomes retroperitoneal at the level of the
is divided similarly by the left portal scissura following the left pancreatic head coursing along the gland’s posterolateral
hepatic vein. surface until it joins the pancreatic duct just before entering
More caudally, a line drawn between the inferior vena cava the ampulla of Vater.
(IVC) and the gallbladder divides the right and left liver, cor- On axial CT, the common duct is seen as a circular structure
responding to an imaginary inferior extrapolation of the course of fluid attenuation (0 to 20 HU) at the posterolateral aspect
of the middle hepatic vein (see Fig. 18.16E). The fissure for the of the pancreatic head on a contrast-enhanced scan. It is con-
ligamentum teres forms the boundary between the medial and sidered dilated at greater than or equal to 9-mm caliber under
lateral sectors of the left liver and generally follows an inferior normal conditions, although a diameter of 7 to 10 mm is com-
projection of the course of the left hepatic vein. The inferior monly observed in elderly patients, and this caliber is typical
boundary between the right anterior and posterior sectors has postcholecystectomy. Intrahepatic bile ducts were usually not
no clear anatomic or vascular landmark and must be estimated well seen on older CT scans unless abnormally dilated and are
by extrapolating an inferior projection of the course of the right still better evaluated by MRI due to its superior tissue contrast,
hepatic vein onto lower sections. The axial planes containing particularly on T2-weighted images. However, with thin colli-
each main portal vein mark the boundary between superior (II, mation available on modern scanners and IV contrast enhance-
IVa, VIII, and VII) and inferior (III, IVb, V, and VI) segments. ment, peripheral bile ducts measuring 1 to 3 mm may be seen
M L
M L
R ivc R ivc
A B
pvl
pvr
c c
ivc
ivc ivc
C D
GB
pv c
ivc
E
FIGURE 18.16. Normal hepatic segmental and vascular anatomy on computed tomography. A, On cephalad sections, the three hepatic veins (R,
right hepatic vein; M, middle hepatic vein; L, left hepatic vein) are seen to enter the inferior vena cava (ivc). B, The lines of the hepatic veins divide
the liver into sectors. The right hepatic vein divides the right liver into anterior and posterior sectors, the middle hepatic vein separates the right from
the left liver following the line of the interlobar scissura, and the left hepatic vein separates segments II and III of the left lobe. C, Approaching the
midpoint of the liver, the left portal vein (pvl) is seen. The fissure for the ligamentum venosum (arrows) separates the caudate lobe (c) from the left
lateral segments. D, The next section shows the horizontal portion of the right portal vein (pvr). The fissure for the ligamentum teres is now seen
anteriorly (large arrows) separating segment IV from the left lateral segments. E, On the most caudad section, the gallbladder (GB) has come into
view; a line drawn along its axis to the IVC represents the inferior extent of the interlobar scissura.
FIGURE 18.17. Variant portal venous anatomy. Thick-slab maximum

intensity projection reveals that the portal venous branch perfusing the
anterior sector of the right hepatic lobe (arrow) shares a common trunk FIGURE 18.19. Three-dimensional volume-rendered computed tomo-
with the left portal vein (arrowhead). graphic angiogram reveals the right hepatic artery (solid arrow) originates
from the celiac axis and courses posterior to the gastroduodenal artery
(open arrow).
benign and malignant lesions may coexist, each lesion should

be evaluated independently when there are multiple lesions
(Fig. 18.23). Mistaking an incidental benign lesion for a
malignant mass has important implications in patient manage-
ment. When assessing the size of lesions for treatment response
on serial CT scans, it is important to routinely compare the
lesions on similar phases of contrast enhancement (or precon-
trast images). Nazarian and colleagues (1999) demonstrated
that hepatic metastases from colorectal carcinoma tend to
appear significantly smaller after IV contrast administration
than on precontrast imaging.
Benign Tumors and Tumor-Like Conditions of the Liver

Cyst (See Chapters 75 and 90B)
Hepatic cysts are common, occurring in at least 2% to 7% of
the population (Horton et al, 1999), and are typically discov-
ered incidentally with no malignant potential. They may be
FIGURE 18.18. Three-dimensional volume-rendered computed tomo- either congenital or acquired. The more common congenital
graphic angiogram reveals a replaced right hepatic artery (arrows) origi- variety may represent malformed bile ducts that have lost com-
nating from the superior mesenteric artery. The right hepatic artery munication with the remainder of the biliary tree; their single-
courses posterior to a stent in the common bile duct (arrowheads). layered cuboidal or columnar epithelial lining fills them by
secreting serous fluid (Blumgart et al, 2001). The acquired type
of hepatic cyst usually arises as a sequela of inflammation,
in normal subjects (Liddell et al, 1990). Delineation of trauma, or parasitic disease.
variant biliary anatomy is also possible if the biliary tree is On CT, hepatic cysts are round or ovoid, sharply-marginated,
dilated (Fig. 18.22). bounded by an imperceptible wall, and measure fluid attenua-
tion; they do not enhance after administration of IV contrast
Lesion Evaluation agent. Sometimes adjacent enhancing liver parenchyma com-
The liver is typically reviewed using the standard abdomen pressed by a cyst may mimic the appearance of an enhancing
soft tissue window width and level preset values (width 350 wall, causing diagnostic ambiguity. Multiple hepatic cysts may
to 400  HU, level 35 to 50  HU) supplied with PACS vendor be associated with polycystic kidney disease (Fig. 18.24),
software. Additional review using narrower window width and polycystic liver disease, or von Hippel-Lindau disease.
lower brightness (width 150  HU, window 50 to 100  HU) can Although small cysts generally do not affect adjacent hepatic
be helpful (Sabouri et  al, 2008), often increasing the conspicu- parenchyma, very large cysts may cause lobar atrophy with
ity of both hypovascular lesions in the portal venous phase compensatory contralateral hypertrophy (Blumgart et al, 2001).
and hypervascular lesions on the late arterial phase. Because Occasionally, patients may present with discomfort or biliary
ha
pv
A B
FIGURE 18.20. Contrast-enhanced computed tomography in a patient with a minimally dilated biliary system. A, The intrahepatic bile ducts are
seen as branching, low-attenuation structures adjacent to the portal veins. At the porta hepatis, the hepatic artery (arrow) is seen to pass between
the main portal vein (pv) and the common hepatic duct. B, A scan caudal to the porta hepatis shows the common hepatic duct and the cystic duct
running adjacent to each other in the hepatoduodenal ligament (arrows). The hepatic artery (ha) is in a more medial position.
CBD
*
FIGURE 18.21. Coronal reformatted image of the confluence of the
right (solid arrow) and left (open arrow) hepatic ducts reveals bilobar
intrahepatic biliary ductal dilation. Note subtle soft tissue thickening
and enhancement along the common bile duct (arrowheads), reflecting FIGURE 18.22. Coronal reformatted image created from axial com-
recurrent gallbladder carcinoma in a patient after cholecystectomy. puted tomographic data acquired after intravenous administration of
contrast medium. Bile is displayed as bright areas with marked dilation
of the intrahepatic bile ducts and the common bile duct (CBD). Note the
distended gallbladder (asterisk) and the aberrant biliary drainage of
obstruction secondary to a large cyst (Fig. 18.25); in such cases, segment VI (arrowheads) directly into the CBD.
therapeutic intervention may be indicated.
Hemangioma (See Chapter 90A) On CT, hemangiomata are hypoattenuating to surrounding

Hemangiomata are the commonest solid tumors of the liver liver; after administration of IV of contrast, they have a fairly
(Blumgart et al, 2001). Usually detected incidentally, their characteristic enhancement pattern. On arterial-phase images,
prevalence has been reported at 7% overall in an autopsy series peripheral discontinuous nodular enhancement (Fig. 18.26A)
(Karhunen, 1986), more common in women (Horton et al, is often seen; the attenuation of these peripheral nodules
1999). Hemangiomata are variable-sized lesions (<1 cm to is similar to that of the aorta (Quinn & Benjamin, 1992).
>40 cm); larger examples are referred to as giant hemangiomata Portal venous and later-phase images demonstrate progressive
(Blumgart et al, 2001). Composed of endothelium-lined vascu- centripetal enhancement—a “filling-in” of enhancement (Fig.
lar spaces separated by fibrous septa, they derive their blood 18.26B-D) toward the center of the lesion (Leslie et al, 1995;
supply from the hepatic artery (Horton et al, 1999). Nelson & Chezmar, 1990).
FIGURE 18.23. Metastatic disease (M) adjacent to a hemangioma.

Note the discontinuous peripheral nodular, clumplike enhancement
(arrow) within the hemangioma. FIGURE 18.25. Giant simple cyst arising in right lobe of liver. It is well
circumscribed and has homogeneous, near-water attenuation.
is indicated, CT angiography can delineate the arterial supply

to these tumors, facilitating enucleation or hepatic resection.
Focal Nodular Hyperplasia (See Chapter 90A)

FNH is the second most common benign liver tumor, with a
reported prevalence ranging from 0.9% (Nguyen et al, 1999)
to 3% (Karhunen, 1986). More common in women, they
occur in relatively young patients (Nguyen et al, 1999). Usually
incidentally detected when patients are scanned for other
reasons, they are being detected more frequently due to increas-
ing requests for dedicated cross-sectional abdominal evaluation
(Carlson et al, 2000). At histopathologic analysis, classic FNH
lesions comprise nodular hyperplastic parenchyma at least
partially bounded by fibrous septa and contain a central scar
composed of fibrous connective tissue, cholangiolar prolifera-
tion surrounded by an inflammatory infiltrate, and malformed
vessels (Nguyen et al, 1999; Wanless et al, 1985).
On CT, FNH is isoattenuating or slightly hypoattenuating
to adjacent liver. Although the lesion may be invisible without
FIGURE 18.24. Multiple simple hepatic cysts in a patient with poly-
cystic kidneys. IV contrast, the value of unenhanced images is to assess for
absence of intralesional hemorrhage, intratumoral fat, and
calcifications—findings atypical of FNH (Carlson et al, 2000).
Giant hemangiomata (>6 to 10 cm) (Bouras et al, 1996; After IV contrast administration, its typical appearance is
Mitsudo et al, 1995; Ros et al, 1987) have a heterogeneous diffuse homogeneous hyperenhancement in the late arterial/
appearance with a hypoattenuating central scar on unenhanced early portal phase due to the rich arterial supply (Fig. 18.28),
imaging attributed to thrombosis, hyalinization, and fibrosis followed by fading to background hepatic attenuation by the
(Ros et al, 1987). The typical arterial-phase peripheral, nodular, late portal venous phase (Fig. 18.29) (Carlson et al, 2000). A
clumplike enhancement is observed, but giant hemangiomata helpful sign when present, the central scar—initially hypoat-
often do not completely fill in with contrast material on delayed tenuating and gradually enhancing on portal venous and
images (Fig. 18.27) (Vilgrain et al, 2000). The larger lesions are delayed phases—is visualized on CT in only about one third of
more likely to cause symptoms such as abdominal pain, increas- patients, although it is thought to be present histopathologically
ing abdominal girth, early satiety, and nausea and vomiting in almost all cases (Shamsi et al, 1993; Welch et al, 1985).
(Blumgart et al, 2001). Other complications include Kasabach-
Merritt syndrome, a coagulopathy with systemic fibrinolysis Adenoma (See Chapter 90A)
and thrombocytopenia (Maceyko & Camisa, 1991), intratu- Hepatocellular adenomata (HCA) are uncommon benign
moral hemorrhage, spontaneous rupture resulting in hemoperi- neoplasms most often discovered in women. They are usually
toneum (Vilgrain et al, 2000), and volvulus of a pedunculated solitary, with multiple lesions in only 30% of patients (Foster
hemangioma (Tran-Minh et al, 1991). If surgical intervention & Berman, 1977). A distinct entity referred to as hepatic
A B
C D
FIGURE 18.26. Cavernous hemangioma of the right liver showing typical enhancement starting at the periphery of the lesion (A to C) and finally
showing complete opacification (D).
A B
FIGURE 18.27. A, Volume-rendered three-dimensional image of a giant hemangioma. The image is oriented such that the patient is facing forward,
with the observer looking down from above. Note the peripheral nodular, clumplike enhancement (black arrow) and the central low-attenuation scar
(white arrow). The hemangioma is in contact with the left hepatic vein (arrowhead) and with the inferior vena cava (IVC). B, Volume-rendered coronal
image of the same patient. The giant hemangioma exerts mass effect on a long segment of the IVC (arrows), which is displaced to the left of midline.
The right kidney (arrowhead) is displaced inferomedially by the mass. Resection was performed with a good result. (Courtesy L.H. Blumgart, MD.)
A B
FIGURE 18.28. Focal nodular hyperplasia. A, Contrast-enhanced computed tomographic image acquired in the late arterial phase reveals the
arterial supply to a well-circumscribed hypervascular mass in the right hepatic lobe. B, Oblique coronal, volume-rendered three-dimensional reformat-
ted image of the same patient reveals early venous drainage from the mass via a large vein (arrow) into the inferior vena cava (arrowhead). The major
hepatic veins have not yet opacified with contrast material.
A B
FIGURE 18.29. Focal nodular hyperplasia. A, Arterial phase image shows a well-circumscribed, brightly enhancing lesion in segment IV/V. B, During
the portal phase, the lesion is isoattenuating to liver parenchyma and cannot be identified.
adenomatosis, a term coined in 1985, refers to multiple (>10) immunohistochemistry have allowed subtyping of HCA into
adenomas in a patient without known risk factors (Flejou et al, four groups: HNF1α-mutated, β-catenin–mutated, inflamma-
1985). A strong association with oral contraceptive use has long tory (most common), and unclassified (Zucman-Rossi et al,
been established (Blumgart et al, 2001), and other risk factors 2006). The inflammatory type appears most prone to hemor-
for the development of HCA include type I glycogen storage rhage, and the β-catenin–mutated type confers the most risk of
disease (Labrune et al, 1997) and use of androgen-containing developing HCC (Katabathina et al, 2011).
steroids (Soe, 1992). The CT appearance of adenomata often overlaps with FNH
Adenomas contain large plates of hepatocytes separated by and HCC. Adenomata tend to have similar attenuation to
dilated sinusoids perfused solely by peripheral arterial feeding normal liver on unenhanced scans, enhance homogeneously
vessels. Intracellular and intercellular lipids manifest as macro- after contrast administration in the arterial phase with sharply
scopic fat within the tumor (Ichikawa et al, 2000). The combi- marginated hypervascularity and occasional (30%) capsulated
nation of poor connective tissue support and central necrosis appearance, and fade on portal venous phase images (Grazioli
predisposes to hemorrhage (Levy & Ros, 2001), particularly et al, 2001). Larger adenomas are more heterogeneous in
the large lesions (Leese et al, 1988). Recent advances in appearance than smaller lesions, presumably because of the
predilection to hemorrhage. Peripheral enhancement reflects bubbles may be seen if anaerobic bacteria are present, and
the presence of large subcapsular feeding vessels, with a resul- the periphery and internal septa may enhance after contrast
tant centripetal pattern of enhancement. Hemorrhage and fat agent administration. Abscesses often have a thick wall, which
within lesions can be best detected on unenhanced imaging as enhances with contrast administration (Fig. 18.30). In addition,
well as calcifications, which are present in a small minority there may be hepatic enhancement peripheral to the enhancing
(10%) of lesions (Grazioli et al, 2000, 2001). Although these wall, secondary to increased capillary permeability. This is
features are nonspecific, recent studies attempting to correlate referred to as the double-target sign (Mendez et al, 1994; Murphy
recently designated subtypes with MRI imaging features (van et al, 1989).
Aalten et al, 2011) suggest that presence of homogeneous fat
within HCA is associated with the HNF1α-mutated subtype. Fungal Abscesses (See Chapter 72)
We are unaware of any similar large studies correlating MDCT Fungal hepatic abscesses generally are disseminated microab-
features with subtype (Katabathina et al, 2011). scesses that occur in immunosuppressed patients. The most
common fungal organism implicated is Candida albicans. Other
Biliary Hamartoma (See Chapters 48 and 90A) fungal infections that cause microabscesses include cryptococ-
Bile duct hamartomata, also known as the von Meyenburg cus, histoplasmosis, and mucormycosis.With contrast-enhanced
complex, are common benign tumors composed of disorganized CT, hepatic microabscesses appear as small, hypodense lesions,
bile ducts and ductules with a fibrocollagenous stroma (Horton ranging from several millimeters to 1.5 cm in size. A bull’s-eye
et al, 1999). The tumors usually range from 1 to 15 mm, do appearance may be identified, with a small, high-attenuation
not communicate with the biliary tree, and are scattered focus centrally surrounded by a low-attenuation zone. Authors
throughout the liver (Wei et al, 1997); although they are often have reported significant increase in the sensitivity and lesion
multiple (Blumgart et al, 2001), they may be solitary. They have conspicuity using arterial-phase CT, compared with portal
a nonspecific imaging appearance most often confused with venous phase CT, when evaluating liver lesions in immunocom-
cysts although they may also be mistaken on CT for metastases promised patients suspected to have hepatosplenic fungal
or microabscesses with greater clinical consequence (Eisenberg infections (Metser et al, 2005).
et al, 1986; Lev-Toaff et al, 1995; Sada & Ramakrishna, 1994).
On MRI, they are more readily recognized as cystic in Echinococcus (See Chapter 74)
appearance. In endemic areas, involvement of the liver by hydatid disease is
a common finding. Echinococcus granulosus presents with a large
Bile Duct Adenoma (See Chapters 48 and 90A) solitary mass or multiple well-defined cystic lesions, which
Bile duct adenomata are usually incidentally detected, benign, often contain internal “daughter” cysts (Fig. 18.31). Coarse
asymptomatic lesions (Welch et al, 1985). They are usually calcifications of the wall are present in 50% of patients
solitary, subcapsular, and small, ranging in size from 0.1 to (Fig. 18.32) and internal calcification in the matrix may also
1 cm. On CT, bile duct adenomata are well-defined hypoat- be seen. Dense or complete lesion calcification implies parasite
tenuating or isoattenuating lesions, enhancing little or not at all death, though partial cyst calcification does not necessarily
after contrast administration. No specific imaging findings are indicate inactivity (Pedrosa et al, 2000). Daughter cysts are
known, however, and definitive diagnosis can be made only at identified in approximately 75% (de Diego Choliz et al, 1982;
histologic analysis (Horton et al, 1999). Murphy et al, 1989). Communication between the cysts
and the biliary tree is found in approximately 25% of patients
Inflammatory Conditions
Pyogenic Abscess (See Chapter 72)
Recent surgery, biliary disease, diverticulitis, Crohn disease,
and alcoholism all predispose to pyogenic hepatic abscess,
commonly caused by Clostridium species and gram-negative
bacteria such as Escherichia coli and Bacteroides, which enter the
liver via the biliary tree or portal venous system (Mergo & Ros,
1997). Ascending cholangitis and portal phlebitis are the most
common causes of pyogenic hepatic abscesses (Murphy et al,
1989).
As a result of the widespread use of broad-spectrum antibi-
otics, the clinical presentation and CT appearance vary widely
(Halvorsen et al, 1984). Pyogenic hepatic abscesses may be
classified as either microabscesses (<2 cm) or macroabscesses,
larger confluent lesions. Microabscesses appear as small, well-
defined, hypodense lesions on contrast-enhanced CT. They
may be widely scattered or clustered with a tendency to coalesce
(Jeffrey et al, 1988). Larger pyogenic abscesses range in appear-
ance from unilocular cavities with smooth outer margins to
highly complex and septated structures with internal debris and
irregular contours. Most large pyogenic abscesses appear as a
well-defined mass of low attenuation on unenhanced CT. The FIGURE 18.30. Contrast-enhanced computed tomography of a
attenuation value of the abscess cavity depends on the age of pyogenic liver abscess shows enhancement of the walls and internal
the abscess; it becomes lower as the abscess matures. Gas septa. A sympathetic right-sided pleural effusion is seen.
A B
FIGURE 18.31. Contrast-enhanced axial computed tomography reveals a multiseptated echinococcal cyst in the right hepatic lobe. Note eccentric
calcifications along the wall of one of the cysts (arrow).
et al, 1988). The CT findings are not specific, however, and

serologic tests are necessary to confirm the diagnosis.
Diffuse Hepatocellular Diseases

Fat Deposition (See Chapter 71)
Fat deposition, also known as steatosis, as seen in patients on
chemotherapy or in patients with systemic disorders such as
diabetes mellitus, cystic fibrosis, or malnourishment, causes a
decrease in hepatic attenuation on CT. Mild degrees of fatty
change may be difficult to detect unless liver attenuation is
compared carefully with that of the spleen. The liver is normally
6 to 12 HU higher in attenuation than the spleen, and reversal
of this relationship is the earliest CT indication of fat deposi-
tion. With more advanced changes, the hepatic parenchyma
becomes less dense than the intrahepatic vessels. In most
FIGURE 18.32. Unenhanced computed tomography reveals a calci-
instances, steatosis is diffuse and uniform (see Fig. 18.12), but
fied rim of echinococcal cyst (arrow). Air within the cyst was from recent a nonuniform focal distribution also can occur and may mimic
instrumentation. a mass (Fig. 18.33). Steatosis should be suspected if blood
vessels can be seen passing through the focal area in a normal
pattern (Fig. 18.34).
An awareness of the common locations of focal fat deposi-
(de Diego Choliz et al, 1982), and frank rupture of cyst con- tion and focal sparing from steatosis is important, so as not
tents into the bile ducts occurs in 5% to 10%, accompanied by to mistake these for tumor. Common locations for focal fat
clinical features of cholangitis. In contrast, Echinococcus alveo- deposition include segment IV adjacent to the falciform liga-
laris infection resembles an infiltrating hepatic tumor clinically ment (Ohashi et  al, 1995; Paulson et  al, 1993) and around
and radiographically with its irregular margins and heteroge- the gallbladder (Yoshimitsu et  al, 1997). A common location
neous attenuation (Didier et al, 1985). for focal fatty sparing is in the posterior aspect of segment IV
(Matsui et  al, 1995; White et  al, 1987) or along the gallbladder
Amebic Abscess (See Chapter 73) fossa. It has been suggested that areas of focal fatty sparing
The protozoan Entamoeba histolytica infects 10% of the world’s are due to aberrant venous drainage to the liver because these
population. Hepatic abscess is the most common extraintestinal areas do not receive portal venous flow from the main portal
complication of amebiasis and occurs in approximately 8.5% trunk (Gabata et  al, 1997; Marchal et  al, 1986; Matsui et  al,
of all patients with amebic infection (Ralls, 1998). On CT, 1995).
amebic abscesses appear as well-defined masses of near-water It has been shown that focal areas of fat sparing in a steatotic
attenuation with a thick enhancing rim. One or more internal liver can result from compression of an adjacent portal venous
septations may be present. A feature of amebic liver abscess that branch by a mass, causing regionally decreased portal blood
may aid in distinguishing it from other focal hepatic lesions is flow. This fat sparing can appear as a wedge-shaped, peripheral,
its tendency to extend beyond the surface of the liver (Radin segmental, or lobar region of relative hyperattenuation,
A B
FIGURE 18.33. Numerous foci of fat deposition at two different levels through the liver (left and right panels). Contrast-enhanced computed
tomography reveals numerous hypoattenuating lesions (white arrows) lesions in a patient with breast cancer. Note lack of mass effect or distortion
of the hepatic vessels. Lesions were stable over time and were shown to represent focal fatty deposition on magnetic resonance imaging.
et al, 1998). If it is unclear based on the CT findings whether

an area of focal hypoattenuation represents a mass, MRI of the
liver, including routine chemical shift imaging, may make a
definitive diagnosis.
Cirrhosis (See Chapters 76 and 103D)

CT is capable of demonstrating the morphologic changes
associated with advanced hepatic cirrhosis. Typical CT features
include bands or regions of confluent fibrosis, relative left
lateral segment and caudate lobe hypertrophy with right lobe
(particularly posterior sector) and left medial segment atrophy
(Fig. 18.35) (Torres et al, 1986), and a nodular hepatic contour,
particularly in the hypertrophied segments, due to regenerative
nodules (Martí-Bonmatí & Delgado, 2010). One of the first
imaging signs of cirrhosis developed was elevation of the
caudate-to–right lobe width ratio to greater than 0.65 (Harbin
FIGURE 18.34. Focal fat deposition. A wedge-shaped area of low et al, 1980). A subsequent larger study using MRI (Awaya et al,
attenuation is visible in the right liver. The hepatic vessels pass through 2002) showed the accuracy of this ratio is improved to 74% by
the area normally with no deviation. designating the right portal vein bifurcation as the lateral
boundary of the caudate lobe rather than the main portal vein.
Other typical imaging features of cirrhosis include enlargement
depending on the site of the portal venous compression (Gross- of the space anterior to the main portal vein at the hepatic
holz et al, 1998). hilum, an expanded gallbladder fossa space, and generalized
Unenhanced CT is superior to contrast-enhanced CT for widening of the interlobar fissures (Brancatelli et al, 2007).
assessing steatosis, because factors such as the rate of contrast Secondary evidence of cirrhosis includes signs of portal hyper-
injection and timing of measurements significantly influence tension: presence of ascites, splenomegaly, and portosystemic
the optimal threshold value of liver-to-spleen attenuation varices, particularly recanalization of the umbilical vein. MDCT
comparison for distinguishing fatty liver from normal. The combined with 3D reformatting can help determine the extent
required ratios are protocol-specific, which limits the clinical of portosystemic collateral vessels, such as the left gastric vein,
usefulness of such measurements (Jacobs et al, 1998; Johnston short gastric vein, esophageal and paraesophageal varices,
A B
FIGURE 18.35. Cirrhosis. A, The liver has an irregular outline. B, The interlobar fissure, delineated by the gallbladder, is deviated toward the right
as a result of atrophy of the right liver with hypertrophy of the left liver and the caudate lobe.
splenorenal and gastrorenal shunts, and paraumbilical and and hepatitis B and C viruses. Approximately 70% of HCCs
abdominal wall veins (Kang et al, 2002). develop in cirrhotic livers (Blumgart et al, 2001) in which the
parenchyma is often distorted by nodular regeneration and
Budd-Chiari Syndrome (See Chapter 88) fibrosis, and in which alterations in arterial and portal venous
Budd-Chiari syndrome is caused by chronic hepatic venous blood flow are common. These changes can both simulate
congestion secondary to obstruction to hepatic venous outflow. disease and obscure malignant lesions (Baron & Peterson,
The CT appearance is characterized by hepatomegaly and 2001), making detection of HCC in the cirrhotic liver an even
ascites; the hepatic veins either contain thrombus (Fig. 18.36) greater diagnostic challenge than in noncirrhotic livers.
or are obliterated. Precontrast scans show decreased attenua- A limitation to noninvasive diagnosis of HCC has been its
tion of hepatic parenchyma because of congestion; postcontrast variable imaging appearance. On CT, HCCs are typically
scans show normal enhancement of the caudate lobe, which is hypervascular in the arterial phase, subsequently becoming
hypertrophied in chronic cases, and patchy enhancement of the hypoattenuating to surrounding liver in later phases, a pattern
remainder of the liver (Vogelzang et al, 1987). Concomitant termed “washout appearance” (Fig. 18.37). However, some
portal vein thrombosis is present in 20% of patients with Budd- tumors exhibit washout appearance only on delayed imaging or
Chiari syndrome. A similar CT appearance of the liver, with not at all, and others are not hypervascular. Another classic
patchy enhancement, is seen in patients with congestive heart feature is the presence of an enhancing tumor capsule hyperat-
failure; in contrast to patients with true Budd-Chiari syndrome, tenuating to the lesion and the surrounding parenchyma on
these patients have patent and enlarged hepatic veins (Moulton portal venous or delayed phases. HCC has a propensity to
et al, 1988). invade adjacent portal (Fig. 18.38) or hepatic veins and IVC
(Fig. 18.39) (Freeny et al, 1992) particularly in cirrhotic livers
Hepatic Infarction (Smalley et al, 1988); identification of enhancing tumor in the
Hepatic infarction is rare and implies compromise to the arte- portal vein is a highly specific sign of HCC in cirrhotic livers.
rial and portal venous supply to a segment or lobe of the liver. Larger lesions (>5 cm) tend to be heterogeneous with a mosaic
CT findings include a well-defined, wedge-shaped area of pattern of enhancement (Stevens et al, 1996) and may contain
low attenuation in a segmental or subsegmental distribution, necrosis and fatty metamorphosis. Approximately 5% of HCCs
extending to the liver surface. After contrast administration, the are associated with spontaneous rupture and hemoperitoneum
affected zone enhances minimally, often with a peripheral thin (Blumgart et al, 2001), and tumor may also invade bile ducts,
subcapsular enhancing rim. Gas formation within sterile infarcts causing obstruction. Whereas HCC in a noncirrhotic liver typi-
is a recognized finding and does not imply superimposed infec- cally presents as a large hypervascular solitary or dominant mass
tion. A late complication of hepatic infarction is formation of with small satellite nodules and central necrosis (Fig. 18.40)
bile lakes in the affected segments as a result of disruption of (Brancatelli et al, 2002; Winston et al, 1999b), in a cirrhotic
biliary radicles. liver, HCC commonly appears as multifocal small lesions.
In the context of cirrhosis, a large study at an experienced
Malignant Lesions of the Liver liver transplant center demonstrated a false-positive HCC
Hepatocellular Carcinoma (See Chapter 91) diagnosis rate of 8% on the basis of CT findings in 1329
HCC is the most common primary hepatic malignancy world- patients referred for hepatic transplantation (Brancatelli et al,
wide, with highest incidence in Asia, Southeast Asia, and sub- 2003). Benign lesions often mistaken for HCC include focal
Saharan Africa. Although less common in the West, the incidence confluent fibrosis, transient hepatic attenuation difference,
of HCC in the United States has been increasing (Arakawa flash-filling hemangiomata, peliosis, and regenerative nodules
et al, 1986). The most significant risk factors include cirrhosis with or without dysplastic change. Regenerative nodules are
CL
A B
C
FIGURE 18.36. A, Contrast-enhanced computed tomography (CT) in a patient with acute hepatic venous thrombosis leading to Budd-Chiari
syndrome shows nonenhancement of the thrombosed hepatic veins. There also is patchy peripheral parenchymal enhancement. B, In a patient with
chronic Budd-Chiari syndrome resulting from a congenital stricture of the intrahepatic inferior vena cava, enlargement of the caudate lobe (CL) and
marked homogeneity of parenchymal enhancement are seen. Splenomegaly also is seen. C, Contrast-enhanced CT reveals mottled parenchymal
enhancement and enlargement of the caudate lobe (arrows) in a patient with occlusion of the hepatic veins secondary to paroxysmal nocturnal
hemoglobinuria.
only detectable on CT when they contain iron (and are thus reduce variability and errors in imaging interpretation. Subse-
hyperattenuating on unenhanced scans) or when they distort quently, it has undergone two updates, the latest in 2014,
the hepatic contour. They typically do not enhance during the (American College of Radiology. LI-RADS version, 2014)
arterial phase, an important characteristic distinguishing them bringing it into alignment with the Organ Procurement and
from typical HCC, and during the portal venous phase, they Transplantation Network (OPTN) classification. The most
enhance homogeneously to the same degree as surrounding central aspect of the system is to consider a lesion “definitely
fibrotic parenchyma, rendering them invisible (Baron & Peter- HCC” if either (1) enhancing tumor is observed in the portal
son, 2001). vein, (2) it is greater than or equal to 1 and less than 2 cm in
Dysplastic nodules are precursor lesions to frank HCC with diameter, hypervascular, and has at least two of three major
cellular atypia and abnormal “nontriadal” arterial supply that features, or (3) it is greater than or equal to 2 cm in diameter
distinguish them from regenerative nodules. Although the and hypervascular, with at least one major feature. Major fea-
majority are invisible on CT, like regenerative nodules, some tures are washout appearance, enhancing tumor capsule, and
particularly higher-grade lesions do enhance in the arterial “threshold growth” defined as diameter increase of greater than
phase and may mimic HCC (Baron & Peterson, 2001). or equal to 50% in less than or equal to 6 months, greater than
The elevated pretest probability of HCC in patients with or equal to 100% increase in greater than 6 months, or any new
cirrhosis or other risk factors and the difficulties associated with lesion greater than or equal to 1 cm. Lesions only partly meeting
diagnosis by CT and MRI prompted development in 2011 of these criteria are considered “probably HCC” or “intermediate
the Liver Imaging Reporting and Data System (LI-RADS). probability.” Although an initial evaluation of the system using
Advocating use of consistent terminology, its goal was to catego- MRI (Davenport et al, 2014) suggested that LI-RADS creates
rize liver findings, specifically in high-risk patients, aiming to too many strata of HCC probability for robust interreader
A B
C
FIGURE 18.37. Precontrast, arterial-phase, and portal venous phase images in a patient with a large hepatocellular carcinoma. A, Unenhanced
imaging reveals a hypoattenuating mass. B, Arterial-phase images reveal the mass is heterogeneously hypervascular with supply from small arteries
(arrow). C, Portal venous phase image reveals the “washout appearance” with tumor now enhancing to a lesser degree than surrounding
parenchyma.
A B
FIGURE 18.38. A, Contrast-enhanced computed tomography reveals an exophytic hepatocellular carcinoma (arrow) with tumor thrombus extend-
ing up the main portal vein (arrowheads). B, Coronal maximum intensity projection image in the same patient reveals extensive collateral venous flow
(arrows) about the mass.
A B
FIGURE 18.39. Hepatocellular carcinoma (HCC). Contrast-enhanced computed tomography reveals HCC involving segment I of the liver (arrow).
Tumor thrombus extends into the inferior vena cava and right atrium (arrow).
On unenhanced CT scans, fibrolamellar carcinoma usually

appears as a hypoattenuating large solitary mass with well-
defined lobular margins (McLarney et al, 1999) and a central
scar. Calcification, a finding used to help differentiate this lesion
from FNH, has been reported in 35% to 55% of tumors
(Brandt et al, 1988; Friedman et al, 1985; Soyer et al, 1991).
During the arterial and portal phases of enhancement, the
nonscar portion of the tumor enhances heterogeneously
(McLarney et al, 1999), unlike FNH in which homogeneous
arterial enhancement is more typical around the central scar.
On delayed images, increasing homogeneity of fibrolamellar
carcinoma and occasional scar enhancement may simulate the
appearance of FNH, again underscoring the importance of
multiphasic evaluation. Although fibrolamellar carcinoma does
not typically have a true capsule, the liver surrounding the mass
may be compressed and may form a pseudocapsule (McLarney
et al, 1999). Associated findings of vascular invasion, lymph-
adenopathy, pulmonary metastases, or peritoneal implants
strongly suggest fibrolamellar carcinoma (Do et al, 2014).
FIGURE 18.40. Hepatocellular carcinoma in a noncirrhotic liver.
Angiosarcoma (See Chapter 89)
Risk factors for angiosarcoma of the liver include hemochro-
matosis, radiotherapy, and exposure to compounds such as
repeatability, its principles for designating “definite” and “prob- polyvinyl chloride, arsenic, or Thorotrast. Angiosarcoma is a
able” HCC have begun to find clinical traction. vascular hepatic neoplasm that appears as an infiltrating,
brightly enhancing mass on CT (Silverman et al, 1983). High-
Fibrolamellar Carcinoma (See Chapter 91) attenuation thorium deposition may be seen in a patchy distri-
Originally considered within the HCC spectrum and formerly bution in the liver, perihepatic lymph nodes, and spleen.
termed fibrolamellar HCC, this malignant hepatocellular Histologic “cavernous” angiosarcoma may enhance in a way
tumor’s distinct clinical, pathologic, and imaging features have similar to that of hemangioma, with the tumor becoming wholly
led to the newer nomenclature. Fibrolamellar carcinoma typi- or partly isodense with liver on delayed scanning (Itai & Teraoka,
cally affects adolescents or young adults with no evidence of 1989; White et al, 1993).
cirrhosis and no associated risk factors, and its prevalence is
only 1% to 9% of that of HCC (Berman et al, 1980; Craig et al, Hepatic Lymphoma
1980; Goodman et al, 1985; Pinna et al, 1997). The main Primary lymphoma of the liver is rare, but the liver is a common
diagnostic challenge is to distinguish it from FNH, a benign secondary site of lymphomatous involvement in patients with
lesion occurring in the same age group (see Chapter 90A). Hodgkin disease and non-Hodgkin lymphoma. At autopsy,
hepatic involvement is seen in 50% of patients with non- Diffuse or infiltrative forms of the disease result in irregular
Hodgkin lymphoma and in 60% of patients with Hodgkin infiltration in the portal areas.
disease. Approximately 6% to 20% of patients with Hodgkin
disease have hepatic involvement at presentation (Castellino, Hepatic Metastases (See Chapters 92 to 94)
1982; Fishman et al, 1991; Sandrasegaran et al, 1994). Liver The second most common site for metastases behind only
involvement usually is associated with lymph node disease and lymph nodes, the liver provides a receptive environment for
almost invariably is associated with splenic disease. The more metastases not only because of its dual blood supply by systemic
extensive the splenic involvement, the greater the likelihood of and portal systems but also because of humoral factors that
hepatic involvement (Figs. 18.41 and 18.42). promote cell growth. The CT appearance of metastases varies,
On CT, hepatic involvement usually appears diffuse; discrete depending on the size and vascularity of the primary tumor, the
nodular lesions are seen in only 10% of cases. The sensitivity degree of necrosis, and the CT technique used (see “Liver and
of CT in the detection of hepatic disease is low, presumably Biliary Technique and Protocol” section). Lesion morphology
because liver involvement is diffuse. It is thought that extensive may be round, ovoid, or irregular, and borders may be sharp,
disease must be present before becoming apparent at imaging, poorly defined, or nodular. Attenuation is usually lower than
and CT may reveal only hepatomegaly with no focal lesion. that of surrounding hepatic parenchyma on unenhanced CT.
IV contrast enhancement during the appropriate phase
increases the sensitivity of lesion detection. Hypovascular
metastases enhance less and are hypoattenuating compared
with surrounding enhancing hepatic parenchyma during portal
venous phase scanning, whereas hypervascular metastases may
be obscured during the portal venous phase and are usually
most clearly visualized during the arterial phase (see Fig.
18.15). Metastases from mucinous carcinomas tend to contain
punctate or amorphous calcification. Cystic metastases often
contain mural nodules, fluid-fluid levels, or septations, whereas
benign cysts tend to have a less complex appearance. The
appearance of metastases may change in response to treatment.
The attenuation of lesions may decrease as they undergo
necrosis (Figs. 18.43 and 18.44); for instance, hepatic metas-
tases from gastrointestinal stromal tumors (GISTs) that respond
to imatinib become near cystic on contrast-enhanced CT (Choi
et al, 2007). Therefore, when evaluating the liver on CT in a
patient with a history of GIST, it is important to know whether
the patient has undergone treatment, and comparison with
pretreatment images becomes essential. Metastases may cause
focal biliary ductal dilation by compressing or invading an
FIGURE 18.41. Contrast-enhanced computed tomography reveals adjacent bile duct (Fig. 18.45).
multiple hepatic masses in a patient with lymphomatous involvement of
the liver and spleen. Note the low-attenuation mass in the spleen (arrow). Biliary System
Advances made earlier in the MDCT era expanded the role of
CT in evaluating the biliary tree. Acquisition of high-resolution
thin-collimation images obtained in a single breathhold and
image review in cine mode and multiple planes facilitate tracing
dilated bile ducts to the point of obstruction. Biliary obstruc-
tion is diagnosed by CT on the basis of dilated intrahepatic or
extrahepatic bile ducts (see Fig. 18.21) that appear as tubular
branching or circular structures of fluid attenuation that enlarge
as they approach the ductal confluence in the porta hepatis.
However, a direct correlation between ductal caliber and the
presence of clinically significant obstruction is not always found.
Normal-caliber bile ducts may be observed despite the presence
of intermittently obstructing calculi, subtle strictures of the
extrahepatic duct, and surgically correctable causes of jaundice.
Conversely, the biliary tree may remain permanently dilated
after surgical or spontaneous relief (e.g., stone passage) of
biliary obstruction; in such patients, the CT appearance may
falsely suggest biliary obstruction. When a discrepancy exists
between the CT findings and clinical or biochemical evidence,
clinical findings generally carry greater weight, and magnetic
resonance cholangiopancreatography (MRCP) or direct percu-
FIGURE 18.42. Large unifocal lymphomatous mass in the right taneous or endoscopic cholangiography may be performed to
hepatic lobe (arrow). Small splenic lesions are also seen. resolve the issue.
A B
FIGURE 18.43. Metastatic colon cancer. A, Before treatment. B, After treatment.
A B
FIGURE 18.44. Contrast-enhanced computed tomography in a patient with metastatic gastrointestinal stromal tumor. A, Before treatment, het-
erogeneous solid enhancing masses are seen in the right and left hepatic lobes (arrows). B, After treatment, lesions have lower attenuation and
appear more cystic (arrows).
Gallstones the gallbladder wall may be seen in acute cholecystitis and is a

CT is inferior to ultrasound for detection of gallstones, but it useful sign in an appropriate clinical setting, but it also can be
may reveal unsuspected gallstones during studies performed for seen in patients with ascites or hypoalbuminemia. The attenu-
other reasons. Gallstones are visible when calcified or contain- ation of the gallbladder bile is usually increased, and extremely
ing material of substantially lower attenuation than the sur- high attenuation is seen in patients with hemorrhagic cholecys-
rounding bile (such as trapped gas or high cholesterol content). titis, a rare complication.
Stones missed by CT include mixed cholesterol stones that are Although many of the CT findings in uncomplicated acute
isoattenuating to bile. cholecystitis are nonspecific, CT is valuable when complica-
tions such as pericholecystic abscess, emphysematous cholecys-
Cholecystitis (See Chapter 33) titis, or gallbladder perforation are present, and CT can identify
CT should not be used as a screening technique for the detec- patients in need of emergency surgery. CT findings most spe-
tion of acute cholecystitis, but occasionally, patients with cho- cific for acute gangrenous cholecystitis are gas in the gallbladder
lecystitis who display a confusing clinical picture may undergo wall or lumen, intraluminal membranes, irregular gallbladder
CT examination before the precise nature of the disease is clear. wall enhancement, and pericholecystic abscess (Bennett et al,
Gallbladder distension, wall thickening, and gallstones are 2002).
often present in acute cholecystitis, but these are nonspecific
signs that occur in most patients with chronic cholecystitis. Mirizzi Syndrome (See Chapter 33)
The presence of ill-defined pericholecystic lucency within the Mirizzi syndrome is an uncommon condition in which the
hepatic parenchyma adjacent to the gallbladder suggests gall- common hepatic duct is obstructed extrinsically by calculi
bladder inflammation (Fig. 18.46). Intramural edema within impacted in or extruded from a Hartmann pouch or adjacent
further evaluate the nature of the obstruction and to search for

the presence of a biliary fistula.
Choledocholithiasis (See Chapters 36 and 37)

The poor sensitivity of sonography in detecting choledocholi-
thiasis has been well documented, with a reported detection
rate of 18% to 22%, whereas CT has a detection rate of 76%
(Baron, 1987). Meticulous attention must be paid to technique,
and when a change in duct caliber is detected, thin collimation
images should be requested through the transition area to
detect the calculus. The only reliable indicator of choledocho-
lithiasis is the presence of dense intraluminal calcification or a
target sign representing a halo of bile surrounding the higher-
attenuation stone. Hyperattenuating calcified stones lying
within an obstructed duct present no challenge to the radiolo-
gist; most false-negative examinations are attributable to
cholesterol stones, which may blend imperceptibly with the
surrounding bile. Errors also occur when ductal dilation does
not accompany calculi, which occurs in about one third of
FIGURE 18.45. Contrast-enhanced computed tomography in a patients with choledocholithiasis.
patient with metastatic colorectal cancer reveals infiltrative tumor invad- Intrahepatic calculi (see Chapter 39) are rare in Western
ing the bile duct (arrowheads). Note the resultant biliary dilation (long countries but most frequently occur in association with iatro-
arrow).
genic bile duct strictures (Fig. 18.48). CT is of little value when
stones are small and noncalcified and bile duct dilation is
minimal or absent. Intrahepatic choledocholithiasis may give a
bizarre appearance, owing to segmental or subsegmental biliary
radicles filled with calculi. In Asian patients with recurrent
pyogenic cholangitis who subsequently form bile pigment
stones, the debris filling the biliary system generally has higher
attenuation than normal bile. Marked bile duct dilation is
present, and often the larger intrahepatic ducts are dilated
without side-branch dilation. Eccentric and diffuse extrahepatic
bile duct wall thickening is usually seen (Schulte et al, 1990).
Gallbladder Carcinoma (See Chapter 49)

Gallbladder carcinoma is the sixth most common gastrointes-
tinal malignancy in the United States. Risk factors include
female gender, age, postmenopausal status, and cigarette
smoking (Khan et al, 1999). Gallstones are present in 74% to
92% of patients with gallbladder carcinoma and represent
another well-established risk factor (Lowenfels et al, 1985;
Nagorney & McPherson, 1988). Porcelain gallbladder, a term
used to describe calcification within the gallbladder wall, places
a patient at increased risk for gallbladder carcinoma. Older
FIGURE 18.46. Acute cholecystitis. Computed tomography shows a studies suggested that 10% to 25% of patients with a porcelain
distended, thick-walled gallbladder with pericholecystic fluid. No gall- gallbladder develop gallbladder carcinoma (Berk et al, 1973),
stones are seen. but more recent reports indicate that the risk may be lower and
related to the type of calcification (lower risk with complete
calcification of the entire wall compared with selective calcifica-
cystic duct. Cholecystobiliary and cholecystoenteric fistulae are tion) (Kim et al, 2009; Stephen & Berger, 2001). Most patients
common complications. It is clinically important to recognize with gallbladder carcinoma are diagnosed with advanced
the diagnosis before surgery because failure to appreciate the disease. Early-stage gallbladder carcinoma typically is detected
extraluminal obstructing process results in unrewarding explo- incidentally because of inflammation related to coexistent
ration of the common duct and persistent obstruction. The cholelithiasis or cholecystitis. In 1% of patients undergoing
typical CT features of Mirizzi syndrome are identification of cholecystectomy for cholelithiasis, gallbladder carcinoma is
the impacted gallstone in an adjacent structure (eccentrically discovered incidentally (Wanebo & Vezeridis, 1993).
located relative to the bile duct) and associated dilation of the The CT imaging appearances of gallbladder carcinoma
proximal biliary system with a normal-caliber downstream include a mass replacing the gallbladder (seen in 40% to 65%
system. An irregular cavity with surrounding edema and inflam- of patients), focal or diffuse gallbladder wall thickening (seen
mation may be seen adjacent to the gallbladder neck (Fig. in 20% to 30%) (see Fig. 18.49), and an intraluminal polypoid
18.47). Because not all of the typical findings may be present mass (seen in 15% to 25%) (Fig. 18.50) (Franquet et al, 1991;
on CT, direct cholangiography or MRCP can be obtained to Lane et al, 1989; Yeh, 1979; Yum & Fink, 1980). Additional
A B
FIGURE 18.47. Mirrizi syndrome. A, Contrast-enhanced computed tomography reveals a large calcified gallstone associated with gallbladder wall
thickening and extensive pericholecystic inflammatory change (arrowheads). B, Extensive inflammatory change surrounds the internal biliary stent
(arrow).
FIGURE 18.48. Intrahepatic cholelithiasis. Intrahepatic duct dilation is

seen after recurrent anastomotic stricture formation at a hepaticojeju- FIGURE 18.49. Contrast-enhanced computed tomography reveals
nostomy. Several laminated, noncalcified calculi can be seen within the focal gallbladder wall thickening (arrows) in a patient with gallbladder
dilated ductal system. cancer.
imaging findings associated with gallbladder carcinoma reflect

the pattern of disease spread. The most common mode by
which gallbladder carcinoma spreads is direct invasion into the
adjacent organs. Direct tumor invasion is enhanced by the thin
gallbladder wall, which contains only a single muscle layer. The
perimuscular connective tissue of the gallbladder is continuous
with the interlobular connective tissue of the liver, facilitating
the direct spread of tumor into the hepatic parenchyma (Henson
et al, 1992). The liver is the organ most frequently invaded,
followed by the colon, duodenum, and pancreas (Sons et al,
1985).
The gallbladder lies in the intersegmental plane between
segments IVb and V (see Fig. 18.6), and tumors of the gallblad-
der fundus may invade directly into these segments at an early
stage (Blumgart et al, 2001). Air may be seen within the gall-
bladder lumen if tumor results in a fistula to the transverse
colon or duodenum. Tumors involving the infundibulum or FIGURE 18.50. Gallbladder carcinoma. The gallbladder is distended
cystic duct may invade directly and obstruct the CBD or portal and contains calcified stones. Nodular soft tissue emanates from the
vein, precluding surgical resection (Blumgart et al, 2001). gallbladder wall into the lumen (arrow).
Associated biliary dilation resulting from infiltrative tumor appearance: mass forming, periductal infiltrating, and intraductal
along the cystic duct to the extrahepatic bile duct, lymphade- growing, corresponding in older literature to the terms “nodular,”
nopathy compressing the bile duct, or intraductal spread of “sclerosing/infiltrating,” and either “exophytic” or “papillary,”
tumor is apparent on CT (Levy et al, 2001). Tumor growth respectively. A combination of periductal infiltrating and mass
through the gallbladder wall away from the liver results in forming types is also common. Cholangiocarcinoma has also
peritoneal carcinomatosis. been divided by the seventh edition of the American Joint
Laparoscopic cholecystectomy has become increasingly Committee on Cancer Cancer Staging Manual (Edge et al,
accepted as a treatment for symptomatic uncomplicated chole- 2010) into perihilar, distal, and intrahepatic types, which are
lithiasis (Southern Surgeon’s Club, 1991), and approximately distinct clinical and radiologic entities (Han et al, 2002). The
15% to 30% of gallbladder carcinomas are detected incidentally perihilar location is most common at 60% to 70%, distal sites
at microscopic review of cholecystectomy specimens (Clair are less common at 20% to 30%, and intrahepatic is least
et al, 1993). A serious potential complication of laparoscopic common at 5% to 15% (Matos et al, 2010; Patel, 2006).
cholecystectomy is the inadvertent dissemination of unsus-
pected gallbladder carcinoma (Winston et al, 1999a). Wib- PERIHILAR CHOLANGIOCARCINOMA (SEE CHAPTER 5). Imaging fea-
benmeyer and colleagues (1995) reported that gallbladder tures of perihilar tumors, in general, include intrahepatic seg-
carcinoma recurs more rapidly after laparoscopic cholecystec- mental biliary dilatation, ductal wall thickening, endoluminal
tomy than after open cholecystectomy. Tumor can recur in the lesions and tumor spread to adjacent liver and/or vessels. Seg-
laparoscopic port tracks, even if surgical resection of the port mental intrahepatic biliary dilation is often seen just peripheral
sites is performed at subsequent hepatic resection for attempted to an ill-defined perihilar mass with abrupt transition between
cure. Recurrent tumor may appear as a mass along the anterior dilated and nondilated ducts. Because this is a key diagnostic
abdominal wall with extension into the subjacent omental fat feature, diagnosis is impeded by biliary drainage prior to scan-
(Fig. 18.51) (Winston et al, 1999a). These observations under- ning (Valls C, et al, 2013). Other clues to the presence of a
score the need to correctly identify findings suggestive of gall- cholangiocarcinoma include crowding of bile ducts and lobar
bladder carcinoma on preoperative imaging when elective atrophy (Engelbrecht et al, 2015). CT (after the advent of
laparoscopic cholecystectomy is planned. helical scanning) is excellent at detecting and localizing these
tumors; one study reported accuracy of 91% (Feydy et al,
Cholangiocarcinoma (See Chapters 50 and 51) 1999).
Cholangiocarcinoma is a relatively rare adenocarcinoma of the In this location, the periductal infiltrating morphologic
bile duct epithelium presenting mostly after the sixth decade of subtype is most common (Valls et al, 2013), accounting for
life. Although most patients have no known risk factors, condi- more than 70% of cases (Matos et al, 2010) (Fig. 18.52). This
tions conferring increased risk include liver fluke infestation, tumor subtype often appears as focal duct wall thickening with
primary sclerosing cholangitis, choledochal cyst (including obliteration of the lumen (Han et al, 2002) but may only mani-
Caroli’s disease), hepatolithiasis, bile stasis, abnormal choled- fest as ductal enhancement or merely narrowing. Hypoenhanc-
ochopancreatic junction, hepatitis C viral infection, cirrhosis, ing soft tissue infiltration of adjacent periductal fat may also be
alcoholic liver disease, ulcerative colitis, type 2 diabetes, thyro- visible with delayed-phase hyperenhancement, although this
toxicosis, and pancreatitis (Valls et al, 2013). infiltration may sometimes be hypervascular, as is seen in the
The Liver Cancer Study Group of Japan (1997) categorized intrahepatic tumor type (Valls et al, 2013). Most strictures are
cholangiocarcinomas into three subtypes based on macroscopic malignant and are thus considered worrisome until proven
otherwise. Malignant strictures tend to be longer (≥18 to
22 mm) and have a thicker wall (≥2 mm) than benign strictures,
and they show arterial- or portal-phase bile duct hyperenhance-
ment (Choi et al, 2005).
Mass-forming perihilar tumors, like their intrahepatic coun-
terparts, appear as heterogeneous hypovascular masses with
peripheral rim enhancement in the arterial and portal phase
and central enhancement in delayed phases.
Tumors in the least common intraductal-growing subtype
(8% to 18%) predominantly have multiple lesions along various
segments of the bile ducts with papillary histologic features
differentiating them from the other types of cholangiocarcinoma
(Engelbrecht et al, 2015). The World Health Organization
endorses the term “intraductal papillary neoplasm of the bile
ducts” to encompass all variants of biliary intraductal neoplasia,
which has been recognized as a biliary counterpart of pancreatic
intraductal papillary mucinous neoplasm. These tumors are
more likely resectable and carry a more favorable prognosis
than other types (Pitt et al, 1995). On CT, asymmetric biliary
dilation is usually noted with enhancing, expansile, well-defined
FIGURE 18.51. Contrast-enhanced computed tomography in a intraductal soft tissue mass in the most dilated ducts. Unlike
patient who had previously undergone laparoscopic cholecystectomy. mass-forming or periductal-infiltrating subtypes, there is no
The enhancing mass (arrow) within the anterior abdominal wall reflects invasion of adjacent liver parenchyma (Han et al, 2002; Valls
recurrent gallbladder carcinoma within a laparoscopic port tract. et al, 2013).
C
FIGURE 18.52. Hilar cholangiocarcinoma. A, Axial contrast-enhanced computed tomography (CT) reveals an infiltrative hypoattenuating tumor
(arrow) surrounding an internal biliary stent (arrowhead) causing bilobar intrahepatic biliary duct dilation. B, Axial contrast-enhanced CT reveals conflu-
ent tumor (large arrows) encasing a replaced right hepatic artery (small arrow) that arises from the superior mesenteric artery and causing intrahepatic
biliary duct dilation. C, Oblique axial maximum intensity projection image reveals narrowing of the right portal vein (large arrow) by infiltrative tumor,
also causing bilobar intrahepatic biliary ductal dilation (smaller arrows).
Because long-term survival is possible only with en bloc hepatic arteries can also be identified with CT angiography. In
resection of the liver and the extrahepatic biliary ducts (Jarnagin a recent meta-analysis (Ruys et al, 2012), MDCT had sensitiv-
& Winston, 2005), preoperative high-resolution MDCT is typi- ity and specificity of 89% and 92%, respectively, for portal
cally used (see Fig. 18.52) to address the factors that determine venous involvement, and 84% and 93%, respectively, for hepatic
resectability per the staging system for hilar cholangiocarcinoma arterial involvement (as well as 86% accuracy in assessing
developed by Jarnagin and colleagues (2001): extent of ductal involvement). In another small study of 27
1. Level and extent of tumor within the biliary tree patients (Park et al, 2008), ductal staging accuracy was 87%
2. Vascular invasion for MDCT versus 85% for MRI, with somewhat observer-
3. Hepatic lobar atrophy dependent sensitivity and specificity for vascular involvement.
4. Distant metastatic disease Lobar atrophy should be readily apparent on CT. Although
On CT, individual dilated bile ducts can be traced to the long-standing obstruction of the ipsilateral bile duct may cause
point(s) of obstruction (Fig. 18.53), and the cause of biliary moderate atrophy, associated portal venous obstruction causes
obstruction can be correctly identified with good accuracy rapid and severe atrophy of the affected lobe (Fig. 18.54)
(Zandrino et al, 2002). Tumor extent along the portal veins and (Jarnagin & Winston, 2005). Atrophy should also be recognized
FIGURE 18.53. Cholangiocarcinoma. Contrast-enhanced computed tomography at two levels through the liver (left and right) reveals bilobar central
intrahepatic biliary duct dilation as a result of an obstructing enhancing mass at the hepatic hilum (arrow).
with or without hypoenhancing soft tissue infiltration of adja-

cent periductal fat. The assessment and treatment of distal bile
duct cancers is similar to that of pancreatic head carcinoma
(Blumgart et al, 2001).
Although the intraductal growing subtype also represents a
minority of cholangiocarcinomas in the distal bile duct, this
location is where this uncommon subtype is most often found
(Baer et al, 1990). Imaging features are identical to its appear-
ance in the perihilar region. Enhancement of the mass depends
on whether or not the tumor is fixed or detached to the duct
wall, and there may be segmental asymmetric thickening of
the bile duct wall with enhancement. Tumors are friable and
slough easily, which may result in intermittent or partial biliary
obstruction (Lim et al, 2003). Some papillary tumors produce
a profuse amount of mucus, resulting in severe biliary dilation
(Lim et al, 2004).
INTRAHEPATIC (PERIPHERAL) CHOLANGIOCARCINOMA (SEE CHAPTER

50). The mass-forming type is most common in this location at
60% (Kim et al, 2011). On contrast-enhanced CT, peripheral
cholangiocarcinoma typically appears as a hypoattenuating
mass with rim enhancement (Fig. 18.55) and with focal dilation
of intrahepatic bile ducts just peripheral. Contrast enhance-
FIGURE 18.54. Infiltrative hilar cholangiocarcinoma. Reformatted ment of the mass may be seen on quite delayed (5 to 15
oblique axial image. The infiltrative mass (long arrow) in the left hepatic
minutes) images (Keogan et al, 1997; Lacomis et al, 1997; Valls
lobe causes left intrahepatic biliary ductal dilation (short arrows). Tumor
et al, 2000), likely due to the increased stromal fibrosis seen
extends along the central portion of the right portal vein with mild atrophy
of the left lobe. within these tumors (Valls et al, 2000). However, an atypical
enhancement pattern has more recently been recognized in
≈30% of lesions (more common in cases of chronic liver disease
and reported when biliary decompression is under consider- and in smaller lesions) characterized by arterial-phase enhance-
ation because drainage of an atrophic lobe does not relieve ment of the majority of the lesion, predominantly without
jaundice and is only indicated to relieve biliary sepsis (Jarnagin washout appearance (Kim et al, 2011). These atypical lesions
& Winston, 2005). have less central fibrous stroma and more cholangiocellularity,
possibly due to more cellular differentiation, and are associated
DISTAL CHOLANGIOCARCINOMA (SEE CHAPTER 59). The periductal with longer disease-free survival.
infiltrating subtype is also the most common form involving the
extrahepatic bile duct. The radiologic and pathologic features Biliary Cystic Tumors (Cystadenoma and
are identical to those of periductal infiltrating perihilar cholan- Cystadenocarcinoma) (See Chapter 90B)
giocarcinoma: Tumor usually appears as a mass or thickening Biliary cystadenomas (BCA) are uncommon benign cystic
of the bile duct wall at the site of biliary obstruction/stricturing lesions that occur predominantly (90%) in women, particularly
A B
FIGURE 18.55. Late arterial-phase and portal venous phase images of a patient with a peripheral cholangiocarcinoma. A, Late arterial-phase image
reveals peripheral enhancement of a heterogeneous mass in the right hepatic lobe (arrows). B, Portal venous phase image reveals heterogeneous
enhancement of the central portion of the mass.
FIGURE 18.57. Recurrent biliary cystadenoma after resection. A

multiloculated cystic lesion is seen at the resection margin containing
enhancing septa and calcification.
FIGURE 18.56. Biliary cystadenoma. Contrast-enhanced computed depends on its content, which may be hemorrhagic, mucinous,
tomography reveals a large cystic mass with a very subtle internal septa- proteinaceous, or bilious (Buetow et al, 1995a). Multiple
tion (curved arrows). studies suggest BCTs are more often located in the left lobe.
Calcifications may be present within the wall of the cyst or
within septa, and septa may enhance with IV contrast (Fig.
between 42 and 55 years of age (Ishak et al, 1977). Although 18.57). Ultrasound is considered more sensitive than CT for
benign, these lesions may recur after excision and have the detection of these characteristic septa. Thickening of septa in
potential to develop into biliary cystadenocarcinomas (BCACs). particular is a predictor of BCT versus simple cysts (Choi et al,
BCACs at presentation are more evenly distributed between 2010). Nonetheless the accuracy of cross-sectional imaging in
men and women and generally occur a decade later (Soares distinguishing the different types of complex cystic lesions
et al, 2014). BCAs are most often intrahepatic (83%), but they (including hemorrhagic cysts) remains relatively low (Soares
may also occur within the extrahepatic bile ducts (13%) or et al, 2014). It had been reported in a small series that intra-
gallbladder (0.02%) (Devaney et al, 1994). cystic debris, associated bile duct dilation, and mural nodularity
On CT, biliary cystic tumors (BCT) tend to be multilocu- suggest BCAC over BCA (Seo et al, 2010). However, a recent
lated cystic lesions with internal septations; only rarely are large multiinstitutional analysis (primarily using CT) investi-
they unilocular (Fig. 18.56). The attenuation of the cyst fluid gating features, including multiloculation, mural nodularity,
and biliary ductal dilation (Arnaoutakis et al, 2015), has once The dedicated pancreas protocol uses 750 to 1000 mL of
again failed to demonstrate imaging findings reliably predictive oral water as a negative contrast agent administered before the
of BCAC, concordant with older reports (Buetow et al, 1995a). examination, to aid distinction of enhanced vessels from the
gastrointestinal tract (Lawler & Fishman, 2002; Soriano et al,
COMPUTED TOMOGRAPHY IN PANCREATIC 2004; Takeshita et al, 2002) and to facilitate identification of
IMAGING: HISTORY AND ROLE tumor invasion into adjacent bowel. Avoiding positive oral
contrast also facilitates analysis on the 3D workstation. Images
CT evaluation of the pancreas was revolutionized by the advent are typically reviewed at 2.5-mm collimation in all planes.
of helical scanning, which enabled rapid evaluation of the Routine evaluation for the follow-up of pancreatic cancer in
pancreas during different phases of contrast injection (Ferrucci, a postoperative patient or for evaluation of response to medical
1999; Ishiguchi et al, 2001; Mertz et al, 2000; Van Hoe & therapy uses 500 to 1000 mL of oral water-soluble iodinated
Baert, 1997). MDCT scanners improved pancreatic imaging contrast solution or equivalent 2% barium sulfate suspension
further by allowing extremely rapid data acquisition, retrospec- administered 45 minutes before scanning. This examination
tive change in slice thickness and location, and 3D image review type requires an injection rate of only 2.5 mL/sec, so a small
in multiple projections (Ferrucci, 1999; Horton & Fishman, 20-gauge catheter may be used. Only the routine portal venous
2000; Ishiguchi et al, 2001; Johnson et al, 2003; Mertz et al, phase is acquired. Five-millimeter axial slices are typically
2000; Rubin, 2003; Van Hoe & Baert, 1997). The role of CT reviewed, although thinner images may always be requested by
now includes initial evaluation and characterization of solid and the radiologist provided the raw image data are still extant.
cystic lesions, preoperative local staging of known and suspected
malignant disease, and follow-up of suspected benign, suspected Anatomy of the Pancreas
and known premalignant, and treated malignant lesions. CT is The normal pancreas (see Chapter 2) is well imaged on routine
also routinely used to evaluate suspected pancreatitis. As with CT of the abdomen. It is a retroperitoneal structure bounded
liver examination, the spatial resolution of the modality particu- anteriorly by the stomach and posteriorly and inferiorly by the
larly suits preoperative evaluation of vascular involvement by third portion of the duodenum, and it lies medial to the C-loop
malignant disease. of the duodenum and medial to the spleen. Crucial to the
evaluation of the pancreas and possible tumor resection is its
Technique and Protocols relationship to vessels. The body and tail of the pancreas lie
CT imaging of the pancreas relies on the differential IV contrast immediately anterior to the splenic vein, and the neck, head,
enhancement between tumor tissue and normal pancreatic and uncinate process of the pancreas are in close relationship
parenchyma. The use of IV contrast agent is mandatory for to the superior mesenteric artery (SMA) and vein and lie just
accurate diagnosis, and timing of the contrast injection and caudal to the celiac axis. The normal pancreas has a fat plane
accurate delivery of the appropriate volume requires use of a completely surrounding the SMA, whereas no fat plane lies
dedicated CT power injector. As with liver imaging, several between the superior mesenteric vein and the pancreatic head
methods are available to determine the time of maximal arterial (Fig. 18.58).
enhancement for each patient. These include fixed timing Patients may sometimes be seen with fatty replacement of a
(subject to suboptimal imaging due to variation in cardiac portion of the pancreas or involving the gland diffusely (Fig.
function and hydration state), timing bolus (Kalra et al, 2004), 18.59). In such instances, the gland is best identified by iden-
and commercially available solutions in which scanning may be tification of surrounding vascular structures and bowel. Severe
triggered automatically when a vascular structure reaches a fatty atrophy permits visualization of fine vascular detail within
predefined attenuation. the pancreas, but a potential pitfall is that a small island of
Initial scan protocols for dedicated pancreas evaluation/CT residual normal pancreatic parenchyma may be confused with
angiography recommended four phases, but more modern an intrapancreatic mass.
practice uses three acquisitions, omitting early arterial imaging
because pancreatic parenchyma and peripancreatic arterial
anatomy are both well evaluated in the late arterial phase
(Fletcher et al, 2003). The precontrast scan permits evaluation
of pancreatic calcifications and allows localization of the gland
and pertinent arteries in the z-axis for subsequent acquisition
of contrast-enhanced phases. IV contrast medium is injected
at a high flow rate of 4 to 6 mL/seconds. The late arterial or
pancreatic parenchymal phase, acquired roughly 30 to 40 seconds
after initiating contrast injection, is designed to maximize dif-
ferences in contrast enhancement between pancreatic neoplasms
and adjacent normal pancreatic tissue and is also useful in
evaluating hypervascular liver metastases seen in patients with
pancreatic endocrine neoplasms. Last, the routine portal venous
phase acquired at approximately 70 to 90 seconds from the start
of IV contrast injection provides the best evaluation of hepatic
metastases from pancreatic ductal adenocarcinoma (Bashir &
Gupta, 2012; Cantisani et al, 2003; Takeshita et al, 2002) and
in some cases offers the best contrast to identify the primary FIGURE 18.58. Normal pancreas. Note the gastroduodenal artery
pancreatic lesions themselves. along the anterior margin of the pancreas (arrow).
FIGURE 18.59. Extensive fatty replacement of the pancreas. Minimal

pancreatic acinar tissue is visible in the posterior aspect of the pancreatic FIGURE 18.61. Low-attenuation mass (arrow) in the head of the
head (arrow). pancreas.
SA
LHA
RHA
CHA
SMA
PHA
GDA
FIGURE 18.60. Volume-rendered computed tomographic angiogram

of the hepatic vasculature. CHA, Common hepatic artery; LHA, left FIGURE 18.62. Dilated pancreatic duct (arrow) secondary to carci-
hepatic artery; GDA, gastroduodenal artery; PHA, proper hepatic artery; noma in the head of the pancreas.
RHA, right hepatic artery; SA, splenic artery; SMA, superior mesenteric
artery.
life, pancreatic ductal adenocarcinoma (PDAC) is the most
common form of pancreatic cancer and is the fourth most
Vascular anatomy is depicted in exquisite detail on CT common cause of death from malignancy (Kim et al, 2004a).
angiography (Ferrari et al, 2007; Horiguchi et al, 2008; Ibukuro, The clinical presentation depends on the location within the
2001; Ibukuro et al, 1995; Yamada et al, 2000). In addition to gland; cancers within the tail are often clinically silent and may
assessment of the celiac and superior mesenteric arteries and not be discovered until metastatic, but cancers within the head
vein and their major branches, smaller vessels are routinely or periampullary region tend to become clinically evident
visualized and assessed, including the gastroduodenal artery, earlier, with biliary obstruction in 50% of patients. Operative
pancreaticoduodenal vessels, and pancreatic branches (Fig. management may be curative under certain circumstances, but
18.60). MDCT permits even more exquisite evaluation of these mortality and morbidity rates of 2.8% and 40%, respectively,
small vessels and better 3D evaluation of the data (Fishman & (Wagner et al, 2004) underscore the need for preoperative
Horton, 2001; Fishman et al, 2000; House et al, 2004; Johnson evaluation to select the subgroup of patients in whom resection
et al, 2003; Takeshita et al, 2002). is most likely to be beneficial.
On CT, PDAC often appears as a hypoattenuating poorly
Tumors of the Pancreas circumscribed mass, with secondary signs, depending on the
Pancreatic Ductal Adenocarcinoma tumor location (Figs. 18.61 and 18.62). Early findings may be
(See Chapters 59 and 66) extremely subtle, and smaller tumors can be isoattenuating to
Pancreatic ducts represent only 4% of the pancreatic tissue but normal pancreatic parenchyma (Yoon et al, 2011) and thus
are the source of 90% of pancreatic malignancies (Cubilla & difficult, if not impossible, to detect. In one study, only half of
Fitzgerald, 1975; Morohoshi et al, 1983). Presenting later in 62 CT scans obtained 2 to 6 months or 6 to 18 months before
the histologic diagnosis of pancreatic cancer were interpreted 2003; Lygidakis et al, 2004; Nakao et al, 2004; Yoshimi et al,
as definite or suspicious for pancreatic carcinoma upon retro- 2003). Because strong correlation has been demonstrated
spective review (Gangi et al, 2004). A mass in the head is fre- between the extent of vessel contact at imaging and histologic
quently associated with biliary or pancreatic ductal dilation, tumor vascular invasion, detailed evaluation of vascular involve-
with atrophy of gland parenchyma upstream of the obstruction ment is essential at imaging. Reformations of acquired axial
a helpful secondary indication of the tumor site in subtle images in the coronal and sagittal planes in the arterial and
nonadvanced cases. venous phases are now typically created for image review in
Detection at an advanced stage, however, is common and is dedicated pancreatic imaging. 3D workstations also permit
associated with involvement of major local arteries and veins, visualization of vascular involvement using maximum intensity
peripancreatic adenopathy, hepatic metastases, or ascites (Figs. projections or volume rendering (Prokesch et al, 2003; Tamm
18.63 and 18.64). Invasion of the superior mesenteric artery or & Charnsangavej, 2001; Tamm et al, 2001), and curved-plane
vein, common hepatic artery, or celiac axis has historically been reformatted images can offer a superb demonstration of vascu-
considered to preclude curative resection. Involvement of the lar involvement (Gong et al, 2004; Vargas et al, 2004).
main portal vein is not an absolute contraindication to surgery A recent consensus statement on standardized reporting of
because a portion of the main portal vein may be resected and ductal adenocarcinoma from the Society of Abdominal Radiol-
reanastomosed proximal to the jejunal branches (Chae et al, ogy and the American Pancreatic Association (Al Hawary et al,
A B
D
FIGURE 18.63. Pancreatic cancer with vascular encasement. A, Cancer in the body of the pancreas tracks posteriorly to encase the celiac axis,
proximal hepatic artery (black arrowhead), and splenic artery (white arrowhead). B, Curved multiplanar reconstruction showing the mass (arrow) in
the body of the pancreas. C, Coronal volume-rendered image shows a mass in the body of the pancreas occluding the splenic vein (arrowhead)
and encasing celiac axis branches (arrows). D, Sagittal volume-rendered image shows mass encasing the celiac axis (arrow).
C
FIGURE 18.64. Large pancreatic mass with vascular encasement and collateral vessels. A, Mass in the body and tail of the pancreas encases the
celiac artery and its branches. B, Coronal volume-rendered image shows the pancreatic mass and encasement of the celiac and superior mesenteric
arteries (arrows) by tumor. C, Coronal volume-rendered image shows extensive collateral vessels in the abdomen from encasement of the splenic
vein by pancreatic cancer.
2014) advocates consistent terminology in describing vascular neoplasms. They present most frequently in the fourth to sixth
involvement: “Encasement” of a vessel is defined as greater than decades of life without significant gender predilection. Most
180-degree circumferential contact, whereas “abutment” is PENs occur sporadically; only 1% to 2% are associated with
defined as less than or equal to 180 degrees of contact. These familial syndromes, most commonly multiple endocrine neo-
should always be reported, as well as narrowing and contour plasia type 1, von Hippel-Lindau syndrome, neurofibromatosis
deformity of a vessel in question. type 1, and tuberous sclerosis (Lewis et al, 2010). Presentation
Regarding metastatic disease, one study (Kitagawa H et al, often depends on the hormonal activity (or lack thereof) of
2008) found specific spread patterns from tumors of the head tumor products (Kulke, 2003). Hormonally active tumors may
depending upon the origin of the cancer in either the dorsal or present earlier than nonhormonally active tumors. As a whole,
ventral pancreatic embryologic anlage. Tumors confined to the these tumors generally have a much more favorable prognosis
ventral pancreas had metastases along the SMA and peripan- than PDAC.
creatic nodes, whereas tumors that arose from the dorsal pan- Multiphase CT is the first imaging choice for patients with
creas metastasized to common hepatic artery, hepatoduodenal suspected PEN, with a median sensitivity of ≈84%. Multiple
ligament, and peripancreatic nodes. Tumors involving both studies confirm the increased sensitivity of arterial-phase images
ventral and dorsal domains metastasized widely. compared with portal venous phase images (Megibow, 2012).
However, the two phases are complementary, and some lesions
Pancreatic Endocrine Neoplasm (See Chapter 65) may be seen only on venous phase images (Lewis et al, 2010).
Also called neuroendocrine cancer of the pancreas, pancreatic Although their imaging appearance may be quite variable,
endocrine neoplasms (PENs) are predominantly well- the most typically expected imaging appearance of PEN on CT
differentiated pancreatic or peripancreatic tumors with endo- is a well-circumscribed hypervascular focus best seen during
crine differentiation. They do not arise from the islets of the pancreatic phase of enhancement (Figs. 18.65 and 18.66).
Langerhans as once thought, but rather from ductal pluripotent Smaller lesions tend to be more homogeneous, whereas larger
stem cells. PENs are relatively rare, with a prevalence of ≈1 in lesions tend to enhance more heterogeneously due to areas
100,000 people, accounting for 1% to 2% of pancreatic of cystic degeneration, necrosis, fibrosis, and calcification,
sometimes yielding attenuation similar or less than background

pancreatic parenchyma. Central calcifications are extremely
helpful in suggesting the correct diagnosis in solid-appearing
pancreatic masses. Five percent to 10% of PENs are cystic, with
a characteristic somewhat thickened enhancing peripheral rim.
Lymph node and liver metastases also are hypervascular and
often more conspicuous on arterial-phase images.
Cystic Pancreatic Neoplasms (See Chapter 60)

Cystic pancreatic neoplasms as a group include a large number
of benign, premalignant, and malignant entities. The five most
common constituents comprising the vast majority of cases are
intraductal papillary mucinous neoplasm (IPMN), mucinous
cystic neoplasm (MCN), serous cystadenoma (SCA), cystic
PENs, and solid pseudopapillary neoplasm (SPN). In particular,
the mucinous entities IPMN and MCN, which have malignant
potential, may be difficult to distinguish from other benign
pancreatic cystic lesions. To address this issue, several guidelines
FIGURE 18.65. Pancreatic endocrine neoplasm appearing as a have been proposed within the past decade for imaging workup
hypervascular mass in the head (arrows) in the pancreatic parenchymal and management of cystic lesions. Two of the more influential
(late arterial) phase. are a section of the white paper produced by the ACR Incidental
A B
C
FIGURE 18.66. Pancreatic endocrine neoplasm (PEN). A, Arterial-phase scan shows hypervascular liver metastasis from PEN. B, Arterial phase
scan shows hypervascular PEN (arrow). C, Liver metastasis from PEN has a lower attenuation than hepatic parenchyma on portal venous phase.
Findings Committee in 2010 (Berland et al, 2010) and a con- that contain no invasive carcinoma, actuarial risk of future
sensus statement by the International Association of Pancreatol- carcinoma development is also much higher among main and
ogy updated in 2012 (Tanaka et al, 2012). The ACR white paper mixed types (Levy P, 2006).
proposes an algorithmic approach to all incidentally detected On CT, branch-duct IPMN may appear as a fluid attenua-
cystic lesions lacking solid elements in asymptomatic patients, tion (0 to 20 HU) cluster of round or tubular lesion(s) or as a
whereas the International Consensus Guidelines are aimed single lesion of greater than 5 mm (Fig. 18.67). Connection to
more specifically at handling suspected mucinous lesions in any a nondilated main duct may be seen (Pedrosa et al, 2010).
patient. Although there are differences in focus and recom- Presence of the lesion in the pancreatic head, its most common
mendations, both guidelines support, under certain circum- location, may compress the CBD with resulting dilation of the
stances, a role for imaging surveillance that had been suggested intrahepatic bile ducts (Fig. 18.68). Main-duct IPMN is char-
by previous work (Allen et al, 2003; Visser et al, 2004). Both acterized by diffuse or segmental dilation of the main duct to
guidelines indicate preference for MRI over CT in characteriz- greater than 5 mm in caliber without other identifiable causes
ing lesions, but because published reports suggest little differ- of obstruction. The mixed type is designated when features of
ence in accuracy between the two modalities in characterizing both branch- and main-duct types are present (Tanaka et al,
cystic pancreatic lesions (Visser et al, 2007), CT is frequently 2012).
requested. The white paper stratifies lesions into surveillance It is largely the mucin produced by the tumor that is identi-
frequencies mostly by size, indicating that among lesions lacking fied at imaging rather than the neoplastic epithelium itself.
solid elements in asymptomatic patients, only those greater than However, visible solid mural nodules and enhancing (in any
3 cm not clearly representing a benign SCA should undergo phase) solid components, when present, may actually represent
further invasive workup. The International Consensus Guide- the tumor. Presence of invasive cancer and high-grade dysplasia
lines puts more emphasis on two groups of mostly imaging has been associated with these features, as well as with
features: “high-risk stigmata,” including main duct caliber
greater than 9 mm and enhancing solid component; “worrisome
features,” including main-duct 5- to 9-mm caliber; abrupt
change in duct caliber with distal gland atrophy; nonenhancing
mural nodules; and size greater than 3 cm. Only lesions without
any of these imaging features in patients also without any of the
specified clinical features (obstructive jaundice, pancreatitis) are
placed into imaging surveillance. One potential problem in
applying this approach is that agreement on the presence of
certain of the imaging features may not be solidly repeatable
among interpreting radiologists (Do et al, 2014). Although the
two guidelines have made a strong contribution to thinking
about management of these lesions, neither has gained universal
acceptance as yet, and particularly large centers with specialty
expertise use their own combination of clinical judgment and
imaging data to guide management decisions.
INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM (SEE CHAPTER 60).

IPMNs, probably the most common of the cystic neoplasms, FIGURE 18.67. Intraductal papillary mucinous tumor of the pancreatic
are mucin-producing neoplasms arising from the main pancre- head (arrow).
atic duct or its branches, exhibiting a spectrum of dysplasia
from low grade through high grade to invasive carcinoma
(Katabi & Klimstra, 2008). They occur only slightly more fre-
quently in men than in women at a mean age of 68 years and
may represent up to 30% of all cystic pancreatic lesions. Patients
had classically presented with nonspecific abdominal symptoms
or pancreatitis, although as cross-sectional imaging has become
more common, IPMNs are now being discovered more fre-
quently as incidental small pancreatic cystic lesions on CT
scans obtained for other reasons (Basturk et al, 2009; Laffan
et al, 2008).
IPMN can be classified into main-duct, branch-duct, and
mixed-type varieties (Bassi et al, 2000; Freeman, 2008; Naga-
saka & Nakashima, 2001; Nakamura, 2002; Ogawa et al, 2008;
Tan et al, 2009) based either on imaging appearance or histo-
pathology. The importance of this distinction is highlighted in
an analysis of a group of studies from 2003 to 2010 (Tanaka
et al, 2012), demonstrating that among main-duct and mixed-
type IPMNs, prevalence of invasive carcinoma was 44% to 45% FIGURE 18.68. Intraductal papillary mucinous tumor (arrow) of the
versus only 16.6% in the branch-duct type (overall 30.8% pancreatic head with invasive adenocarcinoma at pathology examina-
prevalence among all types in 3568 specimens). Among IPMNs tion; note associated biliary ductal dilation.
A B
C D
FIGURE 18.69. Intraductal papillary mucinous neoplasm. A to D, Diffuse involvement of the main (white arrows) and branch (black arrows) pan-
creatic ducts with invasive adenocarcinoma. B, No evidence of vascular encasement is found, as shown on the normal-appearing portal vein (arrows),
and none was found at surgery or at pathology examination.
main-duct dilation (Anand et al, 2013; Kawamoto et al, 2005;

Kim et al, 2014; Sugiyama et al, 2003) (Fig. 18.69), prompting
their inclusion in the groups considered concerning in the
International Consensus Guidelines.
It is often difficult or impossible using either CT or MRI to
distinguish accurately among branch-duct IPMN, SCA, and
MCN (Visser et al, 2007), although IPMN is the only entity
among the group to consistently communicate with the main
duct and the only entity to commonly appear as multiple
lesions. They also typically have more “grapelike” clustered
morphology than MCN, which typically appears more
“orange-like.”
SEROUS CYSTADENOMA (SEE CHAPTER 60). Often discovered inci-

dentally on abdominal CT scans performed for other indica-
tions, pancreatic SCA are slow-growing tumors with a 2 : 1
female-to-male predominance. These tumors present at a mean
age of 63 years versus ≈50 years for mucinous cystadenomas/
carcinomas. Symptoms include vague abdominal pain or back
pain. Other symptoms, including nausea, weight loss, and FIGURE 18.70. Computed tomography showing low-attenuation
vomiting, are less common. Initially, these tumors were called pancreatic serous cystadenoma in the pancreatic tail (arrow) and hepatic
hemangioma.
microcystic adenomas, a term that has fallen into disfavor
because of reported macrocystic tumors (Anderson & Scheiman,
2002).
Serous cystadenomas may vary from 1 mm to greater than 2008; Warshaw et al, 1990), and the classic morphology is not
10 cm. Because of their slow-growing nature and the mildness seen in the majority of cases; therefore many lesions are still
of symptoms in the 50% of patients who exhibit symptoms, accurately diagnosed only upon resection (Chu et al, 2014).
these masses may be quite large at the time of discovery (Figs.
18.70 and 18.71). Lesions are generally surrounded by a fibrous MUCINOUS CYSTIC NEOPLASMS (SEE CHAPTER 60). Characterized
capsule and contain numerous tiny cysts, which are sometimes histologically by the presence of ovarian type stroma, MCNs
so small that the cystic nature of the lesion is difficult to appre- are now considered nearly exclusively (>95%) tumors of peri-
ciate. MRI is often useful to visualize the cystic nature more menopausal women (Basturk et al, 2009). Most frequently
clearly under such circumstances (Allen et al, 2003; Sahani located in the body and tail, they comprise a heterogeneous
et al, 2002; Visser et al, 2004). When present, a central stellate group pathologically, including benign mucin-producing epi-
scar with radiating enhancing septa around which the many tiny thelium, dysplasia, carcinoma in situ, and invasive tumor.
cysts are arranged should strongly suggest the diagnosis. Dif- On CT, the typical appearance of an MCN is a round or
ferentiation from mucinous entities (particularly IPMN) may loculated cystic lesion in the body or tail. Unlike IPMN, the
be difficult (Chaudhari et al, 2007; Kim et al, 2006; Visser et al, lesions only communicate with the main duct when there is
A
FIGURE 18.71. Serous cystadenoma of the pancreatic head shown
in the portal phase.
fistula formation and do not tend to be multiple. Morphologi-

cally, they are described as being “orange-like” rather than
having the more typical “grapelike” clustered appearance of
IPMN, with locules or individual cystic components larger than
the typical SCA and often with a pseudocapsule (Tanaka et al,
2006).
Because no imaging criteria definitely permit differentiation
of benign from malignant forms of the group, previous wisdom
was to consider all of these premalignant lesions for complete
excision (Sarr et al, 2000; Visser et al, 2004; Wilentz et al,
2000). Early attempts to correlate imaging features with malig-
nancy included a series of 52 female patients with pathologically
proven mucinous cystic tumors (Procacci et al, 2001) in which
the most highly associated features of malignancy were a thick B
wall (>2 mm), septations, and calcification within the wall or
FIGURE 18.72. Mucinous cystadenocarcinoma. A, Contrast-
septa (Fig. 18.72). More recently, the 2012 International
enhanced computed tomography (CT) shows thick-walled cystic mass
Consensus Guidelines were designed to include MCN in its with punctate mural calcification. B, Contrast-enhanced CT shows
management scheme, supporting a role for imaging surveillance thick-walled cystic mass with mural ulceration.
under appropriate conditions, and with solid enhancing com-
ponents, mural nodules, and size considered most important
among concerning features. pain and less frequently with nausea/vomiting, weight loss,
jaundice in three, and back pain (Butte et al, 2011). Typically,
SOLID PSEUDOPAPILLARY NEOPLASM (SEE CHAPTERS 59 AND 60). laboratory findings are normal, with no elevation of tumor
Formerly known by several other names, including solid and markers, such as cancer antigen 19-9 (CA 19-9) and carcino-
papillary epithelial neoplasm, SPN is a rare tumor with variable embryonic antigen (CEA), or serum amylase.
biologic behavior. Most consider the entity a low-grade malig- On CT, lesions (which may appear anywhere within the
nancy (Madan et al, 2004) and suggest aggressive, curative gland) are classically often larger than 3 cm, appearing as round
resection when discovered (Megibow, 2012). In a review of 292 encapsulated masses with heterogeneously attenuating compo-
cases reported in the world literature, 14.7% were malignant nents, including hemorrhage, cystic spaces, and soft tissue.
(Mao et al, 1995), the likelihood increasing with patient age Most lesions appear solid and cystic, some solid, and few
and tumor size (Lam et al, 1999). SPN reportedly represents entirely cystic (Butte et al, 2011). Enhancement is more
0.13% to 2.7% of all pancreatic malignancies and is most heterogeneous on earlier (arterial) phases, becoming more
commonly encountered in the first 3 decades of life (Crawford, homogeneous on later phases (Megibow, 2012). Peripheral
1998; Cubilla & Fitzgerald, 1980; Lam et al, 1999), although calcifications (Fig. 18.73) are not uncommon (Buetow et al,
in a recent series, the age range was 10 to 63 years (mean, 38), 1996), reportedly associated with more aggressive tumors. Pos-
with female-to-male predominance 84% (Butte et al, 2011). In sibly because of increased referral for abdominal cross-sectional
older literature, presenting complaints included abdominal imaging, smaller (<3 cm) tumors are being seen more fre-
pain, palpable mass, dyspepsia with bloating, and early satiety quently. Smaller tumors may have features distinct from the
(Buetow et al, 1996; Yoon et al, 2001); in a newer series, most classic SPN appearance, manifesting as sharply marginated
patients were symptomatic, most commonly with abdominal solid lesions with poor enhancement in the pancreatic (late
arterial) phase gradually increasing over time (Baek et al, to encase them. This is a major distinguishing feature useful for
2010). After resection, patients should be followed by imaging diagnosis (Fig. 18.76). Detection of metastatic disease to the
because late metastases can occur. pancreas requires a high index of suspicion, because primary
pancreatic carcinomas of ductal origin are far more common.
OTHER CYSTIC PANCREATIC LESIONS. Cystic masses may also be When clinical suspicion is present, the study can be tailored
secondary to degeneration of previously solid masses. Neuro- to optimize the likelihood of detection. In a patient with
endocrine cancers may become necrotic and present with cystic renal carcinoma, hypervascular metastases to the pancreas may
degeneration (Fig. 18.74) (see Chapter 65). Mucinous tumors be detected on routine scanning or triphasic examination
may invade adjacent structures and present as cystic masses (Fig. 18.77).
(Fig. 18.75).
Acinar Cell Carcinoma of the Pancreas (See Chapter 59)
Metastatic Disease to the Pancreas (See Chapter 64) Acinar cell carcinoma is predominantly a tumor of the elderly
A pathology review (Adsay et al, 2004) of surgical and autopsy man, with peak incidence in the 60s and many more men
cases showed the most common primary malignancies meta- affected than women. Accounting for approximately 1% of all
static to the pancreas were lung carcinomas, gastrointestinal pancreatic tumors, acinar cell carcinoma of the pancreas is
carcinomas, and lymphomas. These are usually found in defined as a carcinoma of the exocrine pancreas exhibiting
patients with widely disseminated disease. As at other sites in
the body, pancreatic lymphoma tends not to occlude vessels but
FIGURE 18.73. Papillary epithelial neoplasm of the pancreas with FIGURE 18.74. Pancreatic neuroendocrine neoplasm with cystic
calcification in portal venous phase. degeneration in the portal venous phase.
A B
FIGURE 18.75. Intraductal papillary mucinous tumor with invasive mucinous adenocarcinoma. A, Large multicystic tumor extends into the papilla.
B, Computed tomography shows that the mass also invades into the duodenum and stomach.
pancreatic enzyme production by the neoplastic cells (Cubilla islet-cell tumors contain calcification (Buetow et al, 1995b;
& Fitzgerald, 1975; Morohoshi et al, 1983). They have a char- Eelkema et al, 1984). In a more recent series (Chiou et al,
acteristic microscopic appearance of a relatively circumscribed 2004), tumor diameters ranged from 3.3 to 18 cm, with an
periphery and minimal desmoplastic stroma within the tumor, overall mean size of 7.2 cm. Most of these tumors were
with cells arranged in nests and cords. The tumor is most hypodense with respect to normal pancreatic parenchyma,
common in the head, neck, and uncinate process of the pancreas although some showed hypervascular enhancement on the early
but may arise in any portion of the gland (Holen et al, 2002; arterial phase, some of which persisted into the portal venous
Klimstra et al, 1992). Initial symptoms include abdominal pain phase.
or fullness, postprandial vomiting, and jaundice. Some patients
may be asymptomatic, and the mass may be detected inciden- Inflammatory Diseases of the Pancreas
tally. Serum tumor markers may be normal, or there may be Pancreatitis (See Chapters 54 to 58)
mild elevation of serum α-fetoprotein, CEA, and CA 19-9. The most common cystic mass of the pancreas is not a malig-
CT features include a well-defined tumor with a thin, nancy but rather a complication of acute or chronic pancreatitis.
enhancing capsule and central necrosis, which may contain Fluid collections occur in approximately 50% of patients with
areas of internal calcification (Fig. 18.78). Such calcification moderate to severe pancreatitis. Of these, approximately 50%
may be a useful feature in distinguishing this entity because only resolve spontaneously within 6 weeks (Baron & Morgan, 1997;
2% of ductal adenocarcinomas and 22% of nonfunctioning Byrne et al, 2002), and 10% to 15% may progress to pseudo-
cyst formation after developing a capsule. Pseudocyst formation
is more common in patients with chronic pancreatitis and may
occur in 20% to 40% of these patients.
FIGURE 18.76. Lymphoma of the pancreas. Note absence of vascular

narrowing. FIGURE 18.78. Acinar pancreatic cancer (arrow).
A B
FIGURE 18.77. Renal cell carcinoma metastatic to pancreas. A, Computed tomographic angiogram of the pancreas shows hypervascular pan-
creatic metastases (arrows) from renal carcinoma. B, Portal venous phase imaging shows far fewer masses.
Common causes of pancreatitis include alcohol, gallstones, retroperitoneum; this may dissect retroperitoneal planes and
idiopathic hypertriglyceridemia, drugs, pancreas divisum, resolve spontaneously 50% of the time. An acute pseudocyst is
trauma, and endoscopic retrograde cholangiopancreatography defined as a collection of pancreatic juice confined by a
(Baron & Morgan, 1997). Clinical presentation includes nausea, nonepithelialized wall of granulation tissue that occurs as a
vomiting, and epigastric pain. Serum amylase, frequently used result of acute pancreatitis (Baron & Morgan, 1997) and
to monitor pancreatitis, may be normal in acute pancreatitis requires at least 4 weeks to form (Fig. 18.81). It is character-
and elevated in nonpancreatitic conditions. Elevation of the ized by the presence of a well-defined wall and may contain
serum lipase may support the diagnosis of acute pancreatitis. internal septations.
Mild acute pancreatitis may appear on CT scan as enlarge-
ment of the pancreas with loss of definition of the borders of Lymphoplasmacytic (Autoimmune) Pancreatitis
the pancreas. Depending on the part of the pancreas involved, (See Chapter 59)
there may be thickening of the right or left anterior pararenal Lymphoplasmacytic pancreatitis, also known as autoimmune
fascia or diffuse thickening of adjacent fascial planes (Figs. pancreatitis (Kawaguchi et al, 1991; Kim et al, 2004b; Yoshida
18.79 and 18.80). Acute fluid collections associated with pan- et al, 1995), was first described as a “primary inflammatory
creatitis are enzyme-rich collections, which lack a defined wall. sclerosis of the pancreas” (Sarles et al, 1961). It is characterized
Pancreatic fluid collections may present as an acute by a mixed inflammatory infiltrate about pancreatic ducts and
fluid collection interspersed with pancreatic parenchyma ductules, combined with obliterative phlebitis (Hardacre et al,
without a defined wall. The fluid collection may be well cir- 2003). It is now recognized as the pancreatic manifestation of
cumscribed, or it may infiltrate throughout the pancreas and a more systemic disease process known as immunoglobulin G4
(IgG4)-related sclerosing disease (Vlachou et al, 2011). Men
are affected at least twice as frequently as women. The clinical
presentation can be identical to that of pancreatic adenocarci-
noma, with abdominal pain, jaundice, weight loss, and elevated
levels of CEA or CA 19-9. Elevation of pancreatic enzymes is
not a prominent feature of this entity, with mild elevation in
serum amylase or lipase levels most commonly reported. Ele-
vated levels of IgG and hypergammaglobulinemia are reported
in 37% to 76% of patients with autoimmune pancreatitis (Kim
et al, 2004b). IgG4 in particular is often elevated, as are other
autoantibodies, such as rheumatoid factor and antinuclear
antibody. Patients may also present with fluctuating obstructive
jaundice without pain (Kamisawa et al, 2007).
At imaging, three patterns of involvement have been recog-
nized: diffuse, focal, and more recently multifocal (Kajiwara
et al, 2008). Diffuse disease (Fig. 18.82) is most common, in
which a diffusely enlarged “sausage-like” gland is noted with
loss of the lobular contour and pancreatic clefts. Focal disease
is less common, manifesting as a focal mass, particularly in the
FIGURE 18.79. Pancreatitis involving the body and tail of the pan- head, often resembling ductal adenocarcinoma but with vari-
creas. Note the loss of fat planes to the spleen and in the left anterior ably present milder upstream main-duct dilation. In focal and
pararenal space. multifocal forms, the affected area(s) of the gland typically
FIGURE 18.80. Severe pancreatitis obliterates tissue planes in a FIGURE 18.81. Pancreatic pseudocyst in portal venous phase. The
patient after gastrectomy. patient also has adenopathy from an islet cell tumor.
appear hypoattenuating on CT, with decreased enhancement gallbladder wall (Kawamoto et al, 2004; Takahashi et al, 2008).
and sometimes moderate delayed enhancement likely due to Involvement of the kidneys is also common, manifesting as
fibrosis (Vlachou et al, 2011). A capsule-like hypoattenuating hypoattenuating round or wedge-shaped cortical nodules,
rind or “halo” about the gland is common at affected areas, as peripheral cortical lesions, and renal pelvic lesions; less com-
is narrowing of peripancreatic veins. The main pancreatic duct monly involved organs include lymph nodes and the retroperi-
typically is narrow and irregular in affected regions. Calcifica- toneum (Vlachou et al, 2011).
tion and peripancreatic stranding/fluid collections are rare. It is often not possible to distinguish by imaging the focal
Because the disease is part of a multisystem spectrum of form of lymphoplasmacytic pancreatitis (with localized stenosis
involvement, imaging also may demonstrate thickening and/or of the main pancreatic duct and upstream duct dilation) from
narrowing of the wall of the CBD, intrahepatic biliary tree, and pancreatic adenocarcinoma, but because the disease typically
responds well to a trial of corticosteroids, rapid clinical and
imaging response may help establish the diagnosis. If this step
is omitted and IgG4 levels are either not obtained or negative,
the diagnosis may require laparotomy (Learn et al, 2011).
Groove Pancreatitis
Groove pancreatitis affects the groove between the pancreatic
head, duodenum, and the CBD and was first described by
Becker and named by Stolte and colleagues (Becker, 1973;
Stolte et al, 1982). This can be extremely difficult to distinguish
from groove pancreatic carcinoma, as patients are seen initially
with abdominal pain and vomiting secondary to duodenal
stenosis. This represents chronic pancreatitis that occurs with a
platelike mass between the head of the pancreas and the duo-
denum, and it may cause intrahepatic and extrahepatic biliary
duct dilation. Duodenal stenosis may be associated with wall
thickening and cystlike lesions in the duodenal wall or in the
groove area (Gabata et al, 2003; Lauffer et al, 1999; Trianto-
poulou et al, 2009) (Fig. 18.83). Pancreatic duct dilation is
much less common than bile duct dilation, and distinguishing
FIGURE 18.82. Lymphoplasmacytic pancreatitis causing diffuse groove pancreatitis from groove pancreatic carcinoma is done
pancreatic enlargement. from a biopsy.
A B
C D
FIGURE 18.83. Groove pancreatitis. Images A to D show paraampullary duodenal wall cysts with extension to the pancreas along with cystic
dilation of the pancreatic duct. At pathology examination, no malignancy was found. The duodenal wall showed fibrosis and chronic inflammation,
and the adjacent pancreas showed focal chronic pancreatitis. The gallbladder showed changes of chronic cholecystitis.
Other Nonneoplastic Pancreatic Masses:

Ectopic Splenic Tissue
Ectopic splenic tissue is an unusual masslike entity that may be
encountered particularly in the pancreatic tail (Fig. 18.84).
Autopsy studies have shown that 10% of patients have acces-
sory spleens, and of these, nearly 20% are found in or near the
tail of the pancreas (Halpert & Alden, 1964; Halpert & Gyorkey,
1959; Lauffer et al, 1999). Accessory/ectopic splenic tissue foci
are generally well-marginated, round masses on CT, smaller
than 2 cm, enhancing homogeneously (Mortele et al, 2004),
but prospective diagnosis of this entity may be difficult on CT.
A Japanese report described identical enhancement of intrapan-
creatic splenic tissue and the proper spleen on CT angiography
(Miyayama et al, 2003). If the diagnosis is questionable on CT,
specificity can be added by performing a contrast-enhanced
MRCP, where there are more sequences in which appearance
FIGURE 18.84. Hypervascular mass in the pancreatic tail (arrow) from identical to the proper spleen can be established, or the gold-
ectopic splenic tissue. standard technetium-99m–tagged heat-damaged red blood cell
scan (Ota et al, 1997).

Chapter 18 Computed tomography of the liver, biliary tract, and pancreas 357.e1
REFERENCES Brancatelli G, et al: Helical CT screening for hepatocellular carcinoma

in patients with cirrhosis: frequency and causes of false-positive
Adsay NV, et al: Secondary tumors of the pancreas: an analysis of a interpretation, AJR Am J Roentgenol 180(4):1007–1014, 2003.
surgical and autopsy database and review of the literature, Virchows Brancatelli G, et al: Cirrhosis: CT and MR imaging evaluation, Eur J
Arch 444(6):527–535, 2004. Radiol 61:57–69, 2007.
Al Hawary MM, et al: Pancreatic ductal adenocarcinoma radiology Brandt DJ, et al: Imaging of fibrolamellar hepatocellular carcinoma,
reporting template: consensus statement of the Society of Abdominal AJR Am J Roentgenol 151(2):295–299, 1988.
Radiology and the American Pancreatic Association, Radiology Brenner DJ, Hall EJ: Computed tomography—an increasing source of
270:248–260, 2014. radiation exposure, N Engl J Med 357:2277–2284, 2007.
Allen PJ, et al: Cystic lesions of the pancreas: selection criteria for Bruix J, et al: Management of hepatocellular carcinoma: an update,
operative and nonoperative management in 209 patients, J Gastroin- Hepatology 53:1020–1022, 2011.
test Surg 7(8):970–977, 2003. Buetow PC, et al: Biliary cystadenoma and cystadenocarcinoma:
American College of Radiology, 2014: Liver Imaging Reporting and Data clinical-imaging-pathologic correlations with emphasis on the
System version 2014. Available at http://www.acr.org/quality-safety/ importance of ovarian stroma, Radiology 196(3):805–810, 1995a.
resources/LIRADS/. Buetow PC, et al: Islet cell tumors of the pancreas: pathologic imaging
Anand N, et al: Cyst features and risk of malignancy in intraductal correlation among size, necrosis and cysts, calcification, malignant
papillary mucinous neoplasms of the pancreas: a meta-analysis, Clin behavior, and functional status, AJR Am J Roentgenol 165:1175–1179,
Gastroenterol Hepatol 11:913–921, 2013. 1995b.
Anderson MA, Scheiman JM: Nonmucinous cystic pancreatic neo- Buetow PC, et al: Solid and papillary epithelial neoplasm of the pan-
plasms, Gastrointest Endosc Clin N Am 12(4):769–779, 2002. creas: imaging-pathologic correlation on 56 cases, Radiology
Anonymous: A prospective analysis of 1518 laparoscopic cholecystec- 199(3):707–711, 1996.
tomies. The Southern Surgeons Club, N Engl J Med 324(16):1073– Butte JM, et al: Solid pseudopapillary tumors of the pancreas. Clinical
1078, 1991. features, surgical outcomes, and long-term survival in 45 consecutive
Arakawa M, et al: Emergence of malignant lesions within an adenoma- patients from a single center, J Gastrointest Surg 15(2):350–357,
tous hyperplastic nodule in a cirrhotic liver: observations in five 2011.
cases, Gastroenterology 91:198–208, 1986. Byrne MF, et al: Pancreatic pseudocysts, Curr Treat Options Gastroen-
Arnaoutakis DJ, et al: Management of biliary cystic tumors: a multi- terol 5(5):331–338, 2002.
institutional analysis of a rare liver tumor, Ann Surg 261:361–367, Cantisani V, et al: MR imaging features of solid pseudopapillary tumor
2015. of the pancreas in adult and pediatric patients, AJR Am J Roentgenol
Awaya M, et al: Cirrhosis: modified caudate-right lobe ratio, Radiology 181(2):395–401, 2003.
224:769–774, 2002. Carlson SK, et al: CT of focal nodular hyperplasia of the liver, AJR Am
Baek JH, et al: Small (<or=3 cm) solid pseudopapillary tumors of the J Roentgenol 174(3):705–712, 2000.
pancreas at multiphasic multidetector CT, Radiology 257(1):97–106, Castellino RA: Imaging techniques for staging abdominal Hodgkin’s
2010. disease, Cancer Treat Rep 66(4):697–700, 1982.
Baer HU, et al: Management of the Mirizzi syndrome and the surgical Chae YS, et al: Preoperative chemoradiation and pancreaticoduodenec-
implications of cholecystocholedochal fistula, Br J Surg 77(7):743– tomy with portal vein resection for localized advanced pancreatic
745, 1990. cancer, Yonsei Med J 44(3):551–556, 2003.
Baron RL: Common bile duct stones: reassessment of criteria for CT Chaudhari VV, et al: Pancreatic cystic lesions: discrimination accuracy
diagnosis, Radiology 162(2):419–424, 1987. based on clinical data and high-resolution CT features, J Comput
Baron RL, Peterson MS: From the RSNA refresher courses: screening Assist Tomogr 31(6):860–867, 2007.
the cirrhotic liver for hepatocellular carcinoma with CT and MR Chiou YY, et al: Acinar cell carcinoma of the pancreas: clinical and
imaging: opportunities and pitfalls, Radiographics 21(Suppl):S117– computed tomography manifestations, J Comput Assist Tomogr
S132, 2001. 28(2):180–186, 2004.
Baron TH, Morgan DE: The diagnosis and management of fluid col- Choi H, et al: Correlation of computed tomography and positron emis-
lections associated with pancreatitis, Am J Med 102(6):555–563, sion tomography in patients with metastatic gastrointestinal stromal
1997. tumor treated at a single institution with imatinib mesylate: proposal
Bashir MR, Gupta RT: MDCT evaluation of the pancreas: nuts and of new computed tomography response criteria, J Clin Oncol
bolts, Radiol Clin North Am 50:365–377, 2012. 25:1753–1759, 2007.
Bassi C, et al: Intraductal papillary mucinous tumors of the pancreas. Choi HK, et al: Differential diagnosis for intrahepatic biliary cystade-
Verona University Pancreatic Team, Int J Pancreatol 27(3):181–193, noma and hepatic simple cyst: significance of cystic fluid analysis
2000. and radiologic findings, J Clin Gastroenterol 44:289–293, 2010.
Basturk O, et al: Pancreatic cysts: pathologic classification, differential Choi SH, et al: Differentiating malignant from benign common bile
diagnosis, and clinical implications, Arch Pathol Lab Med 133:423– duct stricture with multiphasic helical CT, Radiology 236:178–183,
438, 2009. 2005.
Becker V: Proceedings: fundamental morphological aspects of acute Chu AJ, et al: Dual-source, dual-energy multidetector CT for the evalu-
and chronic pancreatitis [author’s transl], Langenbecks Arch Chir ation of pancreatic tumours, Br J Radiol 85:e891–e898, 2012.
334:317–322, 1973. Chu LC, et al: Characterization of pancreatic serous cystadenoma on
Bennett G, et al: CT findings in acute gangrenous cholecystitis, AJR dual-phase multidetector computed tomography, J Comput Assist
Am J Roentgenol 178:275–281, 2002. Tomogr 38:258–263, 2014.
Berk R, et al: Carcinoma in the porcelain gallbladder, Radiology Clair DG, et al: Rapid development of umbilical metastases after lapa-
106:29–31, 1973. roscopic cholecystectomy for unsuspected gallbladder carcinoma,
Berland LL, et al: Managing incidental findings on abdominal ct: white Surgery 113(3):355–358, 1993.
paper of the acr incidental findings committee, J Am Coll Radiol Covey AM, et al: Incidence, patterns, and clinical relevance of variant
7:754–773, 2010. portal vein anatomy, AJR Am J Roentgenol 183(4):1055–1064,
Berman M, et al: Hepatocellular carcinoma: polygonal cell type with 2004.
fibrous stroma—an atypical variant with a favorable prognosis, Craig J, et al: Fibrolamellar carcinoma of the liver: a tumor of adoles-
Cancer 46:1448–1455, 1980. cents and young adults with distinctive clinicopathologic features,
Bismuth H: Surgical anatomy and anatomical surgery of the liver, World Cancer 46:372–379, 1980.
J Surg 6:3–9, 1982. Crawford BE 2nd: Solid and papillary epithelial neoplasm of the
Blumgart LH: Hepatobiliary cancer: atlas of clinical oncology, Hamilton, pancreas, diagnosis by cytology, South Med J 91(10):973–977,
Ontario, 2001, BC Decker. 1998.
Bouras T, et al: Imaging of symptomatic giant hemangioma, J Radiol Cubilla AL, Fitzgerald PJ: Morphological patterns of primary nonen-
77:1145–1148, 1996. docrine human pancreas carcinoma, Cancer Res 35(8):2234–2248,
Brancatelli G, et al: Hepatocellular carcinoma in noncirrhotic liver: CT, 1975.
clinical, and pathologic findings in 39 U.S. residents, Radiology Cubilla AL, et al: Cancer (non-endocrine) of the pancreas: a suggested
222(1):89–94, 2002. classification, Monogr Pathol 21:82–110, 1980.
Davenport MS, et al: Repeatability of diagnostic features and scoring Goodman ZD, et al: Combined hepatocellular-cholangiocarcinoma: a
systems for hepatocellular carcinoma by using MR imaging, Radiol- histologic and immunohistochemical study, Cancer 55(1):124–135,
ogy 272:132–142, 2014. 1985.
de Diego Choliz J, et al: Computed tomography in hepatic echinococ- Grazioli L, et al: Liver adenomatosis: clinical, histopathologic, and
cosis, AJR Am J Roentgenol 139(4):699–702, 1982. imaging findings in 15 patients, Radiology 216(2):395–402, 2000.
Devaney K, et al: Hepatobiliary cystadenoma and cystadenocarcinoma: Grazioli L, et al: Hepatic adenomas: imaging and pathologic findings,
a light microscopic and immunohistochemical study of 70 patients, Radiographics 21(4):877–892, discussion 892-894, 2001.
Am J Surg Pathol 18(11):1078–1091, 1994. Grossholz M, et al: Focal sparing in the fatty liver as a sign of an
Didier D, et al: Hepatic alveolar echinococcosis: correlative US and CT adjacent space-occupying lesion, AJR Am J Roentgenol 171(5):1391–
study, Radiology 154(1):179–186, 1985. 1395, 1998.
Do RKG, et al: Interobserver agreement for detection of malignant Halpert B, Alden ZA: Spleen in or at the tail of the pancreas: a survey
features of intraductal papillary mucinous neoplasms of the pancreas of 2700 additional necropsies, Arch Pathol 77:651–654, 1964.
on MDCT, AJR Am J Roentgenol 203:973–979, 2014. Halpert B, Gyorky F: Lesions observed in accessory spleens of 311
Edge S, et al: Perihilar bile ducts. In Edge S, et al, editors: American patients, Am J Clin Pathol 32(2):165–168, 1959.
joint committee cancer staging manual, ed 7, New York, 2010, Springer, Halvorsen RA, et al: The variable CT appearance of hepatic abscesses,
pp 219–227. AJR Am J Roentgenol 142(5):941–946, 1984.
Eelkema EA, et al: CT features of nonfunctioning islet cell carcinoma, Han JK, et al: Cholangiocarcinoma: pictorial essay of CT and cholan-
AJR Am J Roentgenol 143(5):943–948, 1984. giographic findings, Radiographics 22(1):173–187, 2002.
Eisenberg D, et al: CT and sonography of multiple bile-duct hamarto- Harbin WP, et al: Diagnosis of cirrhosis based on regional changes in
mas simulating malignant liver disease (case report), AJR Am J hepatic morphology: a radiological and pathological analysis, Radiol-
Roentgenol 147(2):279–280, 1986. ogy 135(2):273–283, 1980.
Engelbrecht MR, et al: Imaging of perihilar cholangiocarcinoma, AJR Hardacre JM, et al: Results of pancreaticoduodenectomy for lympho-
Am J Roengenol 204:782–791, 2015. plasmacytic sclerosing pancreatitis, Ann Surg 237(6):853–858, dis-
Ferrari R, et al: Anatomical variations of the coeliac trunk and the cussion 858-859, 2003.
mesenteric arteries evaluated with 64-row CT angiography, Radiol Henson DE, et al: Carcinoma of the gallbladder: histologic types, stage
Med 112(7):988–998, 2007. of disease, grade, and survival rates, Cancer 70(6):1493–1497, 1992.
Ferrucci JT: Biliopancreatic malignancy current diagnostic possibilities: Holen KD, et al: Clinical characteristics and outcomes from an insti-
an overview, Ann Oncol 10(Suppl 4):143–144, 1999. tutional series of acinar cell carcinoma of the pancreas and related
Feydy A, et al: Helical CT assessment in hilar cholangiocarcinoma: tumors, J Clin Oncol 20(24):4673–4678, 2002.
correlation with surgical and pathologic findings, AJR Am J Roent- Honda H, et al: Differential diagnosis of hepatic tumors (hepatoma,
genol 172:73–77, 1999. hemangioma, and metastasis) with CT: value of two-phase incre-
Fishman EK, Horton KM: Imaging pancreatic cancer: the role of mental imaging, AJR Am J Roentgenol 159(4):735–740,
multidetector CT with three-dimensional CT angiography, Pancre- 1992.
atology 1(6):610–624, 2001. Horiguchi A, et al: Multislice CT study of pancreatic head arterial
Fishman EK, et al: CT of lymphoma: spectrum of disease, Radiograph- dominance, J Hepatobiliary Pancreat Surg 15(3):322–326, 2008.
ics 11(4):647–669, 1991. Horton KM, Fishman EK: 3D CT angiography of the celiac and
Fishman EK, et al: Multidetector CT angiography in the evaluation of superior mesenteric arteries with multidetector CT data sets: pre-
pancreatic carcinoma: preliminary observations, J Comput Assist liminary observations, Abdom Imaging 25(5):523–525, 2000.
Tomogr 24(6):849–853, 2000. Horton KM, et al: CT and MR imaging of benign hepatic and biliary
Flejou JF, et al: Liver adenomatosis: an entity distinct from liver tumors, Radiographics 19(2):431–451, 1999.
adenoma?, Gastroenterology 89(5):1132–1138, 1985. House MG, et al: Predicting resectability of periampullary cancer
Fletcher JG, et al: Pancreatic malignancy: value of arterial, pancreatic, with three-dimensional computed tomography, J Gastrointest Surg
and hepatic phase imaging with multi-detector row CT, Radiology 8(3):280–288, 2004.
229(1):81–90, 2003. Iannaccone R, et al: Hepatocellular carcinoma: role of unenhanced and
Foster JH, Berman MM: Solid liver tumors, Major Probl Clin Surg delayed phase multi-detector row helical CT in patients with cir-
22:1–342, 1977. rhosis, Radiology 234:460–467, 2005.
Fournier LS, et al: Early modifications of hepatic perfusion Ibukuro K: Vascular anatomy of the pancreas and clinical applications,
measured by functional CT in a rat model of hepatocellular carci- Int J Gastrointest Cancer 30(1–2):87–104, 2001.
noma using a blood pool contrast agent, Eur Radiol 14:2125–2133, Ibukuro K, et al: Helical CT angiography with multiplanar reforma-
2004. tion: techniques and clinical applications, Radiographics 15(3):671–
Franquet T, et al: Primary gallbladder carcinoma: imaging findings in 682, 1995.
50 patients with pathologic correlation, Gastrointest Radiol Ichikawa T, et al: Hepatocellular adenoma: multiphasic CT and histo-
16(2):143–148, 1991. pathologic findings in 25 patients, Radiology 214(3):861–868,
Freeman HJ: Intraductal papillary mucinous neoplasms and other 2000.
pancreatic cystic lesions, World J Gastroenterol 14(19):2977–2979, Ishak KG, et al: Biliary cystadenoma and cystadenocarcinoma: report
2008. of 14 cases and review of the literature, Cancer 39(1):322–338,
Freeny PC, et al: Hepatocellular carcinoma: reduced frequency of 1977.
typical findings with dynamic contrast-enhanced CT in a non-Asian Ishiguchi T, et al: CT and MR imaging of pancreatic cancer, Hepato-
population, Radiology 182(1):143–148, 1992. gastroenterology 48(40):923–927, 2001.
Friedman AC, et al: Fibrolamellar hepatocellular carcinoma, Radiology Itai Y, Teraoka T: Angiosarcoma of the liver mimicking cavernous
157(3):583–587, 1985. hemangioma on dynamic CT, J Comput Assist Tomogr 13(5):910–912,
Gabata T, et al: Aberrant gastric venous drainage in a focal spared area 1989.
of segment IV in fatty liver: demonstration with color Doppler Jacobs JE, et al: Diagnostic criteria for fatty infiltration of the liver on
sonography, Radiology 203(2):461–463, 1997. contrast-enhanced helical CT, AJR Am J Roentgenol 171(3):659–664,
Gabata T, et al: Groove pancreatic carcinomas: radiological and patho- 1998.
logical findings, Eur Radiol 13(7):1679–1684, 2003. Jarnagin W, Winston C: Hilar cholangiocarcinoma: diagnosis and
Gangi S, et al: Time interval between abnormalities seen on CT and staging, HPB (Oxford) 7(4):244–251, 2005.
the clinical diagnosis of pancreatic cancer: retrospective review of Jarnagin WR, et al: Staging, resectability, and outcome in 225 patients
CT scans obtained before diagnosis, AJR Am J Roentgenol 182:897– with hilar cholangiocarcinoma, Ann Surg 234(4):507–517, discus-
903, 2004. sion 517-519, 2001.
Goldman LJ: Principles of CT: multislice CT. Nucl, Med. Technol. Jeffrey RB Jr, et al: CT of small pyogenic hepatic abscesses: the cluster
36:57–68, 2008. sign, AJR Am J Roentgenol 151(3):487–489, 1988.
Gong JS, Xu JM: Role of curved planar reformations using multidetec- Joe E, et al: Feasibility and accuracy of dual-source dual-energy CT for
tor spiral CT in diagnosis of pancreatic and peripancreatic diseases, noninvasive determination of hepatic iron accumulation, Radiology
World J Gastroenterol 10(13):1943–1947, 2004. 262:126–135, 2012.
Johnson PT, et al: CT angiography with volume rendering: advantages Labrune P, et al: Hepatocellular adenomas in glycogen storage disease
and applications in splanchnic vascular imaging, Radiology 200(2): type I and III: a series of 43 patients and review of the literature,
564–568, 1996. J Pediatr Gastroenterol Nutr 24(3):276–279, 1997.
Johnson PT, et al: Multidetector-row computed tomography with Lacomis JM, et al: Cholangiocarcinoma: delayed CT contrast enhance-
three-dimensional volume rendering of pancreatic cancer: a complete ment patterns, Radiology 203(1):98–104, 1997.
preoperative staging tool using computed tomography angiography Laffan TA, et al: Prevalence of unsuspected pancreatic cysts on MDCT,
and volume-rendered cholangiopancreatography, J Comput Assist AJR Am J Roentgenol 191:802–807, 2008.
Tomogr 27(3):347–353, 2003. Lam KY, et al: Pancreatic solid cystic papillary tumor: clinicopathologic
Johnston RJ, et al: Diagnosis of fatty infiltration of the liver on contrast features in eight patients from Hong Kong and review of the litera-
enhanced CT: limitations of liver-minus-spleen attenuation differ- ture, World J Surg 23(10):1045–1050, 1999.
ence measurements, Abdom Imaging 23(4):409–415, 1998. Lane J, et al: Primary carcinoma of the gallbladder: a pictorial essay,
Kajiwara M, et al: Autoimmune pancreatitis with multifocal lesions, Radiographics 9(2):209–228, 1989.
J Hepatobiliary Pancreat Surg 15(4):449–452, 2008. Lauffer JM, et al: Intrapancreatic accessory spleen: a rare cause of a
Kalra MK, et al: Sixteen-detector row CT of abdomen and pelvis: study pancreatic mass, Int J Pancreatol 25(1):65–68, 1999.
for optimization of Z-axis modulation technique performed in 153 Lawler LP, Fishman EK: Three-dimensional CT angiography with
patients, Radiology 233(1):241–249, 2004. multidetector CT data: study optimization, protocol design, and
Kamisawa T, et al: Clinical management of autoimmune pancreatitis, clinical applications in the abdomen, Crit Rev Comput Tomogr
Adv Med Sci 52:61–65, 2007. 43(2):77–141, 2002.
Kammerer S, et al: Abdominal and pelvic CT: is positive enteric con- Learn PA, et al: Pitfalls in avoiding operation for autoimmune pancre-
trast still necessary? Results of a retrospective observational study, atitis, Surgery 150:968–974, 2011.
Eur Radiol 25:669–678, 2015. Leese T, et al: Liver cell adenomas: a 12-year surgical experience from
Kang HK, et al: Three-dimensional multi-detector row CT portal a specialist hepato-biliary unit, Ann Surg 208(5):558–564, 1988.
venography in the evaluation of portosystemic collateral vessels in Lencioni R, et al: Modified RECIST (mRECIST) assessment for
liver cirrhosis, Radiographics 22(5):1053–1061, 2002. hepatocellular carcinoma, Semin Liver Dis 30:52–60, 2010.
Kapoor V, et al: Multidetector CT arteriography with volumetric three- Leslie DF, et al: Single-pass CT of hepatic tumors: value of globular
dimensional rendering to evaluate patients with metastatic colorectal enhancement in distinguishing hemangiomas from hypervascular
disease for placement of a floxuridine infusion pump, AJR Am J metastases, AJR Am J Roentgenol 165(6):1403–1406, 1995.
Roentgenol 181(2):455–463, 2003. Lev-Toaff AS, et al: The radiologic and pathologic spectrum of biliary
Karhunen PJ: Benign hepatic tumours and tumour-like conditions in hamartomas, AJR Am J Roentgenol 165(2):309–313, 1995.
men, J Clin Pathol 39(2):183–188, 1986. Levy AD, Ros PR: Hepatic adenomas: imaging findings with pathologic
Katabathina VS, et al: Genetics and imaging of hepatocellular adeno- correlation [invited commentary], Radiographics 21:892–894, 2001.
mas: 2011 update, Radiographics 31:1529–1543, 2011. Levy AD, et al: Gallbladder carcinoma: radiologic-pathologic correla-
Katabi N, Klimstra DS: Intraductal papillary mucinous neoplasms of tion, Radiographics 21(2):295–314, questionnaire, 549-555, 2001.
the pancreas: clinical and pathological features and diagnostic Lévy P, et al: Natural history of intraductal papillary mucinous tumors
approach, J Clin Pathol 61:1303–1313, 2008. of the pancreas: actuarial risk of malignancy, Clin Gastroenterol
Kawaguchi K, et al: Lymphoplasmacytic sclerosing pancreatitis with Hepatol 4:460–468, 2006.
cholangitis: a variant of primary sclerosing cholangitis extensively Lewis RB, et al: Pancreatic endocrine tumors: radiologic-
involving pancreas, Hum Pathol 22(4):387–395, 1991. clinicopathologic correlation, Radiographics 30(6):1445–1464, 2010.
Kawamoto S, et al: Lymphoplasmacytic sclerosing pancreatitis with Liddell RM, et al: Normal intrahepatic bile ducts: CT depiction,
obstructive jaundice: CT and pathology features, AJR Am J Roent- Radiology 176(3):633–635, 1990.
genol 183(4):915–921, 2004. Lim J, et al: Papillary neoplasms of the bile duct that mimic biliary
Kawamoto S, et al: Intraductal papillary mucinous neoplasm of the stone disease, Radiographics 23:447–455, 2003.
pancreas: can benign lesions be differentiated from malignant lesions Lim JH, et al: Intraductal papillary mucinous tumor of the bile ducts,
with multi-detector CT?, Radiographics 25(6):1451–1468, 2005. Radiographics 24(1):53–66, discussion 67, 2004.
Keogan MT, et al: Contrast-enhanced CT of intrahepatic and hilar Liver Cancer Study Group of Japan: Classification of primary liver cancer,
cholangiocarcinoma: delay time for optimal imaging, AJR Am J Tokyo, 1997, Kanehara, pp 6–8.
Roentgenol 169(6):1493–1499, 1997. Liu Y, et al: Multidetector computed tomography triphasic evaluation
Khan ZR, et al: Risk factors for biliary tract cancers, Am J Gastroenterol of the liver before transplantation: importance of equilibrium phase
94(1):149–152, 1999. washout and morphology for characterizing hypervascular lesions,
Kim DJ, et al: Validation of the Harvard Cancer Risk Index: a predic- J Comput Assist Tomogr 36:213–219, 2012.
tion tool for individual cancer risk, J Clin Epidemiol 57(4):332–340, Lowenfels AB, et al: Gallstones and risk of gallbladder cancer, J Natl
2004a. Cancer Inst 75(1):77–80, 1985.
Kim KP, et al: Autoimmune chronic pancreatitis, Am J Gastroenterol Lygidakis NJ, et al: Mono-bloc total spleno-pancreaticoduodenectomy
99(8):1605–1616, 2004b. for pancreatic head carcinoma with portal-mesenteric venous
Kim SY, et al: Macrocystic neoplasms of the pancreas: CT differentia- invasion: a prospective randomized study, Hepatogastroenterology
tion of serous oligocystic adenoma from mucinous cystadenoma and 51(56):427–433, 2004.
intraductal papillary mucinous tumor, AJR Am J Roentgenol Macari M, et al: Dual-source dual-energy MDCT of pancreatic adeno-
187(5):1192–1198, 2006. carcinoma: initial observations with data generated at 80 kVp and at
Kim JH, et al: Should we perform surgical management in all patients simulated weighted-average 120 kVp, AJR Am J Roengenol
with suspected porcelain gallbladder?, Hepatogastroenterology 56(93): 194:W27–W32, 2010.
943–945, 2009. Maceyko RF, Camisa C: Kasabach-Merritt syndrome, Pediatr Dermatol
Kim SA, et al: Intrahepatic mass-forming cholangiocarcinomas: 8(2):133–136, 1991.
enhancement patterns at multiphasic CT, with special emphasis on Madan AK, et al: Solid and papillary epithelial neoplasm of the pan-
arterial enhancement pattern—correlation with clinicopathologic creas, J Surg Oncol 85(4):193–198, 2004.
findings, Radiology 260:148–157, 2011. Magnusson M, et al: Evaluation of methods for shaded surface display
Kim KW, et al: Imaging features to distinguish malignant and benign of CT volumes, Comput Med Imaging Graph 15(4):247–256,
branch-duct type intraductal papillary mucinous neoplasms of the 1991.
pancreas: a meta-analysis, Ann Surg 259:72–81, 2014. Makarawo TP: Water as a contrast medium: a re-evaluation using the
Kitagawa H, et al: Carcinomas of the ventral and dorsal pancreas multidetector-row computed tomography, Am Surg 79:728–733,
exhibit different patterns of lymphatic spread, Front Biosci 13:2728– 2013.
2735, 2008. Mao C, et al: Papillary cystic and solid tumors of the pancreas: a
Klimstra DS, et al: Acinar cell carcinoma of the pancreas: a clinico- pancreatic embryonic tumor? Studies of three cases and cumulative
pathologic study of 28 cases, Am J Surg Pathol 16(9):815–837, 1992. review of the world’s literature, Surgery 118(5):821–828, 1995.
Kulke MH: Neuroendocrine tumours: clinical presentation and man- Marchal G, et al: “Skip areas” in hepatic steatosis: a sonographic-
agement of localized disease, Cancer Treat Rev 29(5):363–370, 2003. angiographic study, Gastrointest Radiol 11(2):151–157, 1986.
Martí-Bonmatí L, Delgado F: MR imaging in liver cirrhosis: classical Patel BN, et al: Single-source dual-energy spectral multidetector
and new approaches, Insights Imaging 1(4):233–244, 2010. CT of pancreatic adenocarcinoma: optimization of energy level
Matos C, et al: Magnetic resonance imaging of biliary tumors, Magn viewing significantly increases lesion contrast, Clin Radiol 68:148–
Reson Imaging Clin N Am 18:477–496, 2010. 154, 2013.
Matsui O, et al: Focal sparing of segment IV in fatty livers shown by Paulson EK, et al: Focal fatty infiltration: a cause of nontumorous
sonography and CT: correlation with aberrant gastric venous drain- defects in the left hepatic lobe during CT arterial portography,
age, AJR Am J Roentgenol 164:1137–1140, 1995. J Comput Assist Tomogr 17(4):590–595, 1993.
McCollough CH, et al: CT dose reduction and dose management Pedrosa I, et al: Hydatid disease: radiologic and pathologic features and
tools: overview of available options, Radiographics 26:503–512, 2006. complications, Radiographics 20:795–817, 2000.
McLarney JK, et al: Fibrolamellar carcinoma of the liver: radiologic- Pedrosa I, et al: Imaging considerations in intraductal papillary muci-
pathologic correlation, Radiographics 19(2):453–471, 1999. nous neoplasms of the pancreas, World J Gastrointest Surg 2:324–330,
Megibow AJ: Unusual Solid Pancreatic Tumors, Radiol Clin North Am 2010.
50:499–513, 2012. Pickhardt P, et al: Abdominal CT with model-based iterative recon-
Mendez RJ, et al: Hepatic abscesses: MR imaging findings, Radiology struction (MBIR): initial results of a prospective trial comparing
190(2):431–436, 1994. ultralow-dose with standard-dose imaging, AJR Am J Roentgenol
Mergo PJ, Ros PR: MR imaging of inflammatory disease of the liver, 199:1–9, 2012.
Magn Reson Imaging Clin N Am 5(2):367–376, 1997. Pinna AD, et al: Treatment of fibrolamellar hepatoma with subtotal
Mertz HR, et al: EUS, PET, and CT scanning for evaluation of hepatectomy or transplantation, Hepatology 26(4):877–883, 1997.
pancreatic adenocarcinoma, Gastrointest Endosc 52(3):367–371, Pitt HA, et al: Malignancies of the biliary tree, Curr Probl Surg
2000. 32(1):1–90, 1995.
Metser U, et al: Fungal liver infection in immunocompromised patients: Procacci C, et al: CT features of malignant mucinous cystic tumors of
depiction with multiphasic contrast-enhanced helical CT, Radiology the pancreas, Eur Radiol 11(9):1626–1630, 2001.
235(1):97–105, 2005. Prokesch RW, et al: Multidetector CT of pancreatic adenocarcinoma:
Michels N: Blood supply and anatomy of the upper abdominal organs with diagnostic advances and therapeutic relevance, Eur Radiol
a descriptive atlas, Philadelphia, 1955, Lippincott, p 581. 13(9):2147–2154, 2003.
Miles KA, et al: Colorectal cancer: texture analysis of portal phase Qian LJ, et al: Differentiation of neoplastic from bland macroscopic
hepatic CT images as a potential marker of survival, Radiology portal vein thrombi using dual-energy spectral CT imaging: a pilot
250:444–452, 2009. study, Eur Radiol 22:2178–2185, 2012.
Mitsudo K, et al: Nonenhanced hepatic cavernous hemangioma with Quinn SF, Benjamin GG: Hepatic cavernous hemangiomas: simple
multiple calcifications: CT and pathologic correlation, Abdom diagnostic sign with dynamic bolus CT, Radiology 182(2):545–548,
Imaging 20(5):459–461, 1995. 1992.
Miyayama S, et al: Intrapancreatic accessory spleen: evaluation by CT Radin DR, et al: CT of amebic liver abscess, AJR Am J Roentgenol
arteriography, Abdom Imaging 28(6):862–865, 2003. 150(6):1297–1301, 1988.
Morgan DE: Dual-energy CT of the abdomen, Abdom Imaging Ralls PW: Focal inflammatory disease of the liver, Radiol Clin N Am
39:108–134, 2014. 36(2):377–389, 1998.
Morohoshi T, et al: Exocrine pancreatic tumours and their histological Rao SX, et al: Whole-liver CT texture analysis in colorectal cancer:
classification: a study based on 167 autopsy and 97 surgical cases, Does the presence of liver metastases affect the texture of the remain-
Histopathology 7(5):645–661, 1983. ing liver?, United European Gastroenterol J 2:530–538, 2014.
Mortele KJ, et al: CT features of the accessory spleen/homogeneously, Ros PR, et al: Hemangioma of the liver: heterogeneous appearance on
AJR Am J Roentgenol 183:1653–1657, 2004. T2-weighted images, AJR Am J Roentgenol 149(6):1167–1170,
Moulton JS, et al: Passive hepatic congestion in heart failure: CT 1987.
abnormalities, AJR Am J Roentgenol 151(5):939–942, 1988. Rubin GD: 3-D imaging with MDCT, Eur J Radiol 45(Suppl 1):S37–
Murphy BJ, et al: The CT appearance of cystic masses of the liver, S41, 2003.
Radiographics 9(2):307–322, 1989. Ruys AT, et al: Radiological staging in patients with hilar cholangiocar-
Nagasaka T, Nakashima N: Problems in histological diagnosis of cinoma: a systematic review and meta-analysis, Br J Radiol
intraductal papillary-mucinous tumor (IPMT), Hepatogastroenterol- 85:1255–1262, 2012.
ogy 48(40):972–976, 2001. Sabouri S, et al: Adding liver window setting to the standard abdominal
Nagorney DM, McPherson GA: Carcinoma of the gallbladder and ct scan protocol: is it useful? Iran J Radiol 5:65–70, 2008.
extrahepatic bile ducts, Semin Oncol 15(2):106–115, 1988. Sada PN, Ramakrishna B: Computed tomography of von Meyenburg
Nakamura A: New classification of pancreatic intraductal papillary- complex simulating micro-abscesses, Australas Radiol 38(3):225–226,
mucinous tumour by mucin expression: its relationship with potential 1994.
for malignancy, J Pathol 197(2):201–210, 2002. Sahani D, et al: Cystic pancreatic neoplasms evaluation by CT and
Nakao A, et al: Extended radical resection versus standard resection for magnetic resonance cholangiopancreatography, Gastrointest Endosc
pancreatic cancer: the rationale for extended radical resection, Pan- Clin N Am 12(4):657–672, 2002.
creas 28(3):289–292, 2004. Salgado R, et al: CT angiography: basic principles and post-processing
Nazarian L, et al: Size of colorectal liver metastases at abdominal CT: applications, JBR-BTR 86(6):336–340, 2003.
comparison of precontrast and postcontrast studies, Radiology Sandrasegaran K, et al: Staging of lymphoma in adults, Clin Radiol
213:825–830, 1999. 49(3):149–161, 1994.
Nelson RC, Chezmar JL: Diagnostic approach to hepatic hemangio- Sarles H, et al: Chronic inflammatory sclerosis of the pancreas: an
mas, Radiology 176(1):11–13, 1990. autonomous pancreatic disease?, Am J Dig Dis 6:688–698, 1961.
Nguyen BN, et al: Focal nodular hyperplasia of the liver: a comprehen- Sarr M, et al: Clinical and pathologic correlation of 84 mucinous cystic
sive pathologic study of 305 lesions and recognition of new histologic neoplasms of the pancreas: can one reliably differentiate benign from
forms, Am J Surg Pathol 23(12):1441–1454, 1999. malignant (or premalignant) neoplasms?, Ann Surg 231:205–212,
Ogawa H, et al: Intraductal papillary mucinous neoplasm of the pan- 2000.
creas: assessment of the likelihood of invasiveness with multisection Seo JK, et al: Appropriate diagnosis of biliary cystic tumors: compari-
CT, Radiology 248(3):876–886, 2008. son with atypical hepatic simple cysts, Eur J Gastroenterol Hepatol
Ohashi I, et al: Hepatic pseudolesion in the left lobe around the falci- 22:989–996, 2010.
form ligament at helical CT, Radiology 196:245–249, 1995. Schulte SJ, et al: CT of the extrahepatic bile ducts: wall thickness and
Ota T, et al: Intrapancreatic accessory spleen diagnosed by technetium- contrast enhancement in normal and abnormal ducts, AJR Am J
99m heat-damaged red blood cell SPECT, J Nucl Med 38:494–495, Roentgenol 154(1):79–85, 1990.
1997. Shamsi K, et al: Focal nodular hyperplasia of the liver: radiologic find-
Park HS, et al: Preoperative evaluation of bile duct cancer: MRI com- ings, Abdom Imaging 18(1):32–38, 1993.
bined with MR cholangiopancreatography versus MDCT with direct Shuman WP, et al: Model-based iterative reconstruction versus adap-
cholangiography, AJR Am J Roentgenol 190:396–405, 2008. tive statistical iterative reconstruction and filtered back projection in
Patel T: Cholangiocarcinoma, Nat Clin Pract Gastroenterol Hepatol liver 64-MDCT: focal lesion detection, lesion conspicuity, and image
3:33–42, 2006. noise, AJR Am J Roentgenol 200:1071–1076, 2013.
Shuman WP, et al: Standard and reduced radiation dose liver CT Vargas R, et al: MDCT in pancreatic adenocarcinoma: prediction of
images: adaptive statistical iterative reconstruction versus model- vascular invasion and resectability using a multiphasic technique
based iterative reconstruction-comparison of findings and image with curved planar reformations, Am J Roentgenol 182(2):419–425,
quality, Radiology 273:793–800, 2014. 2004.
Sica GT, et al: CT and MR imaging of hepatic metastases, AJR Am J Vilgrain V, et al: Imaging of atypical hemangiomas of the liver with
Roentgenol 174(3):691–698, 2000. pathologic correlation, Radiographics 20(2):379–397, 2000.
Silverman PM, et al: CT appearance of abdominal Thorotrast deposi- Visser BC, et al: Diagnostic imaging of cystic pancreatic neoplasms,
tion and Thorotrast-induced angiosarcoma of the liver, J Comput Surg Oncol 13(1):27–39, 2004.
Assist Tomogr 7(4):655–658, 1983. Visser BC, et al: Characterization of cystic pancreatic masses: relative
Simpson AL, et al: Texture analysis of preoperative ct images for predic- accuracy of CT and MRI, AJR Am J Roentgenol 189:648–656,
tion of postoperative hepatic insufficiency: a preliminary study, J Am 2007.
Coll Surg 220:339–346, 2015. Visser BC, et al: Diagnostic evaluation of cystic pancreatic lesions, HPB
Smalley SR, et al: Hepatoma in the noncirrhotic liver, Cancer (Oxford) 10(1):53–69, 2008.
62(7):1414–1424, 1988. Vlachou PA, et al: IgG4-related sclerosing disease: autoimmune
Soares KC, et al: Cystic neoplasms of the liver: biliary cystadenoma pancreatitis and extrapancreatic manifestations, Radiographics
and cystadenocarcinoma, J Am Coll Surg 218:119–128, 2014. 31(5):1379–1402, 2011.
Soe K: Liver pathology associated with the use of anabolic-androgenic Vogelzang RL, et al: Budd-Chiari syndrome: CT observations, Radiol-
steroids, Liver 12:73–79, 1992. ogy 163(2):329–333, 1987.
Sons HU, et al: Carcinoma of the gallbladder: autopsy findings in 287 Wagner M, et al: Curative resection is the single most important factor
cases and review of the literature, J Surg Oncol 28(3):199–206, 1985. determining outcome in patients with pancreatic adenocarcinoma,
Soriano A, et al: Preoperative staging and tumor resectability assess- Br J Surg 91:586–594, 2004.
ment of pancreatic cancer: prospective study comparing endoscopic Wanebo HJ, Vezeridis MP: Carcinoma of the gallbladder, J Surg Oncol
ultrasonography, helical computed tomography, magnetic resonance Suppl 3:134–139, 1993.
imaging, and angiography, Am J Gastroenterol 99(3):492–501, 2004. Wanless IR, et al: On the pathogenesis of focal nodular hyperplasia of
Soyer P, et al: CT of fibrolamellar hepatocellular carcinoma, J Comput the liver, Hepatology 5(6):1194–1200, 1985.
Assist Tomogr 15(4):533–538, 1991. Warshaw AL, et al: Cystic tumors of the pancreas. New clinical,
Stephen AE, Berger DL: Carcinoma in the porcelain gallbladder: a radiologic, and pathologic observations in 67 patients, Ann Surg
relationship revisited, Surgery 129(6):699–703, 2001. 212(4):432–443, discussion 444-445, 1990.
Stevens WR, et al: Mosaic pattern of hepatocellular carcinoma: histo- Weg N, et al: Liver lesions: improved detection with dual-detector-array
logic basis for a characteristic CT appearance, J Comput Assist Tomogr CT and routine 2.5 mm thin collimation, Radiology 209(2):417–426,
20(3):337–342, 1996. 1998.
Stolte M, et al: A special form of segmental pancreatitis: “groove Wei SC, et al: Bile duct hamartomas: a report of two cases, J Clin
pancreatitis.”, Hepatogastroenterology 29(5):198–208, 1982. Gastroenterol 25(4):608–611, 1997.
Sugiyama M, et al: Predictive factors for malignancy in intraductal Welch TJ, et al: Focal nodular hyperplasia and hepatic adenoma:
papillary-mucinous tumours of the pancreas, Br J Surg 90(10):1244– comparison of angiography, CT, US, and scintigraphy, Radiology
1249, 2003. 156(3):593–595, 1985.
Takahashi N, et al: Dual-phase CT of autoimmune pancreatitis: a White EM, et al: Focal periportal sparing in hepatic fatty infiltration: a
multireader study, AJR Am J Roentgenol 190(2):280–286, 2008. cause of hepatic pseudomass on US, Radiology 162(1 Pt 1):57–59,
Takeshita K, et al: Multidetector row helical CT of the pancreas: value 1987.
of three-dimensional images, two-dimensional reformations, and White PG, et al: The computed tomographic appearances of angiosar-
contrast-enhanced multiphasic imaging, J Hepatobiliary Pancreat coma of the liver, Clin Radiol 48(5):321–325, 1993.
Surg 9(5):576–582, 2002. Wibbenmeyer L, et al: Laparoscopic cholecystectomy can disseminate
Tamm E, Charnsangavej C: Pancreatic cancer: current concepts in situ carcinoma of the gallbladder, J Am Coll Surg 181:504–510,
in imaging for diagnosis and staging, Cancer J 7(4):298–311, 1995.
2001. Wilentz RE, et al: Mucinous cystic neoplasms of the pancreas, Semin
Tamm E, et al: Advanced 3-D imaging for the evaluation of pancreatic Diagn Pathol 17(1):31–42, 2000.
cancer with multidetector CT, Int J Gastrointest Cancer 30(1–2):65– Winston C, et al: CT angiography for delineation of celiac and SMA
71, 2001. arterial variants in patients undergoing hepatobiliary and pancreatic
Tan L, et al: Imaging features of intraductal papillary mucinous neo- surgery, AJR Am J Roentgenol 189(1):W13–W19, 2007.
plasms of the pancreas in multi-detector row computed tomography, Winston CB, et al: Recurrent gallbladder carcinoma along laparoscopic
World J Gastroenterol 15(32):4037–4043, 2009. cholecystectomy port tracks: CT demonstration, Radiology 212(2):
Tanaka M, et al: International consensus guidelines for management of 439–444, 1999a.
intraductal papillary mucinous neoplasms and mucinous cystic Winston CB, et al: Hepatocellular carcinoma: MR imaging findings in
neoplasms of the pancreas, Pancreatology 6:17–32, 2006. cirrhotic livers and noncirrhotic livers, Radiology 210(1):75–79,
Tanaka M, et al: International consensus guidelines 2012 for the 1999b.
management of IPMN and MCN of the pancreas, Pancreatology Yamada Y, et al: CT assessment of the inferior peripancreatic veins:
12:183–197, 2012. clinical significance, Am J Roentgenol 174(3):677–684, 2000.
Tidebrant G, et al: Comparison of non-enhanced, bolus-enhanced, and Yeh HC: Ultrasonography and computed tomography of carcinoma of
delayed scanning techniques in computed tomography of hepatic the gallbladder, Radiology 133(1):167–173, 1979.
tumours, Acta Radiol 31(2):161–166, 1990. Yoon DY, et al: Solid and papillary epithelial neoplasms of the pancreas:
Torres WE, et al: Computed tomography of hepatic morphologic aggressive resection for cure, Am Surg 67(12):1195–1199, 2001.
changes in cirrhosis of the liver, J Comput Assist Tomogr 10(1):47–50, Yoon SH, et al: Small (≤ 20 mm) pancreatic adenocarcinomas: analysis
1986. of enhancement patterns and secondary signs with multiphasic
Tran-Minh VA, et al: Volvulus of a pedunculated hemangioma of the multidetector CT, Radiology 259:442–452, 2011.
liver, AJR Am J Roentgenol 156(4):866–867, 1991. Yoshida K, et al: Chronic pancreatitis caused by an autoimmune
Triantopoulou C, et al: Groove pancreatitis: a diagnostic challenge, Eur abnormality: proposal of the concept of autoimmune pancreatitis,
Radiol 19(7):1736–1743, 2009. Dig Dis Sci 40(7):1561–1568, 1995.
Valls C, et al: Intrahepatic peripheral cholangiocarcinoma: CT evalua- Yoshimi F, et al: Reconstruction of the portal vein and the splenic vein
tion, Abdom Imaging 25(5):490–496, 2000. in pancreaticoduodenectomy for pancreatic cancer, Hepatogastroen-
Valls C, et al: Radiological diagnosis and staging of hilar cholangiocar- terology 50(51):856–860, 2003.
cinoma, World J Gastrointest Oncol 5(7):115–126, 2013. Yoshimitsu K, et al: Anatomy and clinical importance of cholecystic
van Aalten SM, et al: Hepatocellular adenomas: correlation of MR venous drainage: helical CT observations during injection of contrast
imaging findings with pathologic subtype classification, Radiology medium into the cholecystic artery, Am J Roentgenol 169:505–510,
261:172–181, 2011. 1997.
Van Hoe L, Baert AL: Pancreatic carcinoma: applications for helical Yum HY, Fink AH: Sonographic findings in primary carcinoma of the
computed tomography, Endoscopy 29(6):539–660, 1997. gallbladder, Radiology 134(3):693–696, 1980.
Zandrino F, et al: Multislice CT cholangiography without biliary Zheng X, et al: Assessment of hepatic fatty infiltration using spectral
contrast agent: technique and initial clinical results in the assessment computed tomography imaging: a pilot study, J Comput Assist Tomogr
of patients with biliary obstruction, Eur Radiol 12(5):1155–1161, 37:134–141, 2013.
2002. Zucman-Rossi J, et al: Genotype-phenotype correlation in hepatocel-
Zeman RK, et al: Three-dimensional models of the abdominal vascu- lular adenoma: new classification and relationship with HCC,
lature based on helical CT: usefulness in patients with pancreatic Hepatology 43:515–524, 2006.
neoplasms, AJR Am J Roentgenol 162(6):1425–1429, 1994.
CHAPTER 19
Magnetic resonance imaging of the liver, biliary
tract, and pancreas
Scott R. Gerst and Richard Kinh Gian Do
Magnetic resonance imaging (MRI) is a cross-sectional multi- tissues. Differences in T1 and T2 times intrinsic to various soft
planar imaging technique (Fig. 19.1). MRI uses magnetic fields tissues can be exploited to improve image contrast and diag-
and radiofrequency pulses to generate images with outstanding nostic accuracy. For example, free water, which has a long T1
tissue contrast and excellent spatial resolution. The principles relaxation, is low signal on standard T1-weighted imaging and
of nuclear magnetic resonance were first described in the 1940s markedly high signal on T2-weighted imaging. Thus so-called
by Bloch and colleagues (1946) and Purcell and colleagues heavily T2-weighted imaging sequences, such as those used in
(1946) as a method for in vitro chemical analysis. Several MR cholangiopancreatography, highlight high water content
decades later, Damadian (1971) and Lauterbur (1973) applied structures, such as bile and pancreatic ducts, while reducing
some of these basic principles to design MRI to allow in vivo the signals of other organs.
imaging. Today, MRI is used extensively as a medical imaging Diffusion-weighted MRI (DWI) is an alternative tissue con-
tool throughout the body to visualize and distinguish normal trast mechanism that produces images dependent on the local
and pathologic tissue. magnitude of water diffusion. DWI of the liver may be used to
help detect focal hepatic lesions, especially metastases, and may
PRINCIPLES OF MAGNETIC also be used to assess changes in the liver parenchyma, such as
RESONANCE IMAGING liver fibrosis. Tremendous interest in this sequence first arose
due to its increased sensitivity to early changes in the brain after
MRI harnesses the signal available from the magnetic moment a cerebrovascular accident compared with more traditional T1-
or spin present in certain nuclei, including hydrogen atoms. or T2-weighted imaging. Applications for DWI in body imaging,
When placed in a large static magnetic field, the nuclei align initially hampered by hardware and software limitations, were
themselves in a parallel or antiparallel direction to the field and addressed in the mid-2000s. With new advances in receiver coil
also rotate or spin at a precise frequency termed the Larmor design, improved gradient coils used to acquire images, and
frequency. This frequency depends on the specific type of nuclei motion correction through respiratory or navigator techniques,
imaged and the strength of the magnetic field. The most com- DWI has flourished as a functional imaging sequence in hepa-
monly imaged nucleus in clinical practice is hydrogen (H1) topancreatobiliary imaging (Taouli et al, 2010).
because of its great abundance in the human body, mostly in
the form of water. Other nuclei that may be imaged by MRI MAGNETIC RESONANCE IMAGING SAFETY
include phosphorus (P31), sodium (Na23), and carbon (C13).
When unperturbed in a static magnetic field, nuclei with MRI is safe for most patients. However, physicians and
magnetic spins are in equilibrium: An almost an equal number patients should be aware of some important considerations,
of nuclei are in an “up” (parallel) or “down” (antiparallel) such as MR contrast safety, and the risk of interactions
alignment. The slight difference between the two states creates between the patient (including their implants) with the strong
a net magnetic moment. A measurable signal is generated from static magnetic field of the scanner as well as the RF field
the magnetic moment after excitation by a radiofrequency (RF) used to generate images. The magnetic field interactions are
pulse at the resonance Larmor frequency. This signal is gener- potentially dangerous by causing the motion of ferromagnetic
ated as the excited nuclei in the body return to equilibrium, aneurysm clips or causing electronic malfunctioning of cardiac
releasing energy in the form of an electromagnetic field that is pacing devices and defibrillators. Many MRI facilities have
captured by a receiver coil. The strength of this emitted signal screening programs in place to address potentially contraindi-
determines the signal intensity (SI) of a tissue. The precise cated devices, under new MRI safety guidelines proposed by
tissue SI depends on several factors, including longitudinal the American College of Radiology (Expert Panel on MR
relaxation (T1), transverse relaxation (T2), proton density Safety, 2013). However, referring physicians should be aware
(essentially the number of nuclei present), and flow. Some of of their local imaging centers’ approach to MRI safety, such
these factors can be manipulated (e.g., through T1 and T2 as their ability to scan patients with MR-conditional or
weighing) to create images highlighting various soft tissue MR-unsafe pacemakers, for example.
properties. MRI has a role in patients with minimal or mild renal insuf-
All tissues have an intrinsic T1 value for relaxation along the ficiency. Iodinated contrast used in CT is associated with a risk
longitudinal magnetic axis, which varies between 200 and 800 of contrast-induced nephropathy; however, administration of
milliseconds. T2 time, or transverse relaxation, is a measure of intravenous gadolinium contrast agents in patients with severe
signal loss perpendicular to the long axis of the magnetic field renal insufficiency carries a risk of nephrogenic systemic fibrosis
and typically varies between 20 and 200 millisecconds for most (NSF). NSF is a serious and potentially fatal complication
358
Chapter 19 Magnetic resonance imaging of the liver, biliary tract, and pancreas 359
C
FIGURE 19.1. Multiplanar T2-weighted images through the liver in a patient with multiple hepatic metastases. A, Axial image. B, Sagittal image.
C, Coronal image. A variety of techniques can be used to obtain these images. Note that fluid-containing structures, including small bowel (curved
arrow, A), gallbladder, biliary tree, and pancreatic duct (arrow, C), are bright.
related to free gadolinium deposition within soft tissues and an important role in imaging benign disorders, as well as com-
organs, with resulting scleroderma-like fibrosis. The exact prehensive evaluation of malignancies of the biliary system
causal mechanism of NSF remains unknown, but the risk of (Mandelia et al, 2013; Singh et al, 2014; Vaishali et al, 2004).
NSF increases in patients with a glomerular filtration rate Heavily T2-weighted images are used to provide an overview
(GFR) less than 30 mL/min, and particularly in patients with of biliary and pancreatic ductal anatomy, removing the signal
severe, end-stage disease (GFR < 15 mL/min). In addition, of surrounding structures. Excellent diagnostic-quality images
linear agents with lower thermodynamic stability carry a greater are obtainable, with high sensitivity and specificity for evalua-
risk than others. If available, more stable, lower-risk macrocy- tion of ductal dilatation, strictures, and intraductal abnormali-
clic agents can be used in patients with a GFR less than 30 mL/ ties (Hekimoglu et al, 2008; Palmucci et al, 2010a; Palmucci
min. Gadolinium contrast usage in patients with a GFR less et al, 2010b; Sandrasegaran et al, 2010). Cross-sectional
than 15 mL/min should only be performed if essential, after images and maximum intensity projection images (Fig. 19.2)
careful evaluation of risks versus benefits, and with consultation are produced with current MRCP techniques, and projection
with a nephrologist, if available. Since NSF was initially images are similar to direct contrast-enhanced cholangiograms
described, the rapid international investigation, dissemination obtained with either endoscopic retrograde cholangiopancrea-
of information, and swift adjustment to policy have led to a tography (ERCP) or percutaneous transhepatic cholangiogra-
precipitous drop of reported NSF cases. In multiple countries, phy (PTC; see Chapter 18). MRCP is noninvasive, eliminating
NSF has essentially disappeared since 2009 (Bennett et al, the potential morbidity associated with ERCP or PTC (Zhong
2012; Grobner, 2006; Idee et al, 2014). et al, 2004).
The basic principle of MRCP is to use T2-weighted imaging
MAGNETIC RESONANCE IMAGING to highlight stationary or slowly moving fluid, including bile, as
CHOLANGIOPANCREATOGRAPHY high in signal intensity; surrounding tissues, including retroperi-
toneal fat and the solid visceral organs, with shorter T2 values,
MR cholangiopancreatography (MRCP) is an imaging tech- are markedly reduced in signal. In addition to heavilyT2-weighted
nique used to evaluate the bile and pancreatic ducts and plays sequences, MRCP protocols also include routine intermediate
A B
C D
FIGURE 19.2. Mixed main- and branch-duct intraductal papillary mucinous neoplasm (IPMN) of the pancreatic tail. A, Coronal T2 image. A lesion
with high T2 signal indicating cystic contents involves the pancreatic tail and shows multiple thin septations (arrow). B, Coronal three-dimensional
maximum intensity projection T2-weighted magnetic resonance cholangiopancreatography image. The multiseptated cystic pancreatic tail mass is
apparent (arrow). C, Axial T2 image. Multiple septations are apparent with dilated main pancreatic duct (star). Incidental note is made of a right renal
cyst with thin septations (open circle). D, Axial T1-weighted fat saturation image postintravenous gadolinium contrast. The numerous septations
enhance within the lesion (arrow). Some components appeared nodular, and the lesion was resected. Histopathology revealed an IPMN with a focus
of noninvasive mucinous cancer.
T2-weighted sequences and T1-weighted sequences to evaluate enhancement patterns between benign and malignant liver
surrounding structures. MR-specific techniques for obtaining lesions have been well reported and are discussed in individual
cholangiographic images include breath-hold, as well as respira- entities later.
tory and navigator gated techniques, which can generate images Several new hepatobiliary-specific contrast agents have been
in a patient during free breathing. developed to add additional functional information compared
with traditional ECF contrast agents. These hepatobiliary spe-
MAGNETIC RESONANCE IMAGING cific contrast agents are taken up to varying degrees by function-
CONTRAST AGENTS ing hepatocellular tissue and are excreted in bile as well as
through the kidneys. Hepatobiliary specific agents include
MRI contrast agents work primarily by altering the intrinsic T1 mangafodipir trisodium, gadobenate dimeglumine, and gadox-
relaxation times of various soft tissues where contrast accumu- etic acid (Gd-EOB-DTPA). Both gadobenate dimeglumine
lates. In the hepatobiliary system, MRI contrast agents can (MultiHance; Bracco Imaging, Cranbury Township, NJ) and
improve the detection of liver lesions and improve the charac- gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid
terization of focal liver abnormalities. MRI contrast agents for (Eovist/Primovist; Bayer Healthcare, Wayne, NJ) have been
hepatobiliary imaging are divided into two basic categories: approved in the United States. These contrast agents are all
extracellular fluid contrast (ECF) agents, and those with addi- administered intravenously, although the dose and pharmaco-
tional hepatobiliary excretion. In the past, the most commonly kinetics are different. These two agents are sometimes referred
used contrast agents were the ECF agents, such as gadopen- to as combination agents because of their dual capacity for
tetate dimeglumine (gadolinium diethylenetriaminepentaacetic imaging both in the dynamic phase (as done routinely with ECF
acid [Gd-DTPA]), which is distributed within the intravascular agents) and in the delayed, hepatocyte-specific phase. These
compartment initially and rapidly diffuses through the extra- agents provide comprehensive information about the hepatic
vascular space, similar to the action of iodinated contrast agents parenchyma, bile ducts, and liver vasculature (Burke et al,
in computed tomography (CT; see Chapter 18). Differences in 2013; Seale et al, 2009).
NORMAL HEPATIC APPEARANCE ON or focal. The pattern of fatty liver is related to regional differ-
MAGNETIC RESONANCE IMAGING ences in perfusion. It may be difficult to distinguish focal fat
from focal hepatic lesions on CT because both appear low in
MRI is routinely used to evaluate diffuse and focal liver abnor- attenuation. Additionally, focal fatty sparing may appear similar
malities. With liver MRI, a combination of precontrast T1, T2, to an enhancing neoplasm on contrast-enhanced CT. Although
and DWI sequences are used. Normal hepatic parenchyma is geographic margins and location may aid in the diagnosis, this
brighter (hyperintense) than the spleen on T1-weighted images, is more readily distinguished on MRI due to different signal
whereas on T2-weighted images, the spleen is relatively brighter characteristics. T1-weighted in-phase and out-of-phase imaging
than the liver (Fig. 19.3). Although most hepatic lesions are allows differentiation of signals from fat and water when they
low in signal intensity on T1-weighted images, they have more are present in the same voxel, as it exploits minute differences
variable intensity on T2-weighted images, with cysts and hem- in resonance frequencies between fat and water protons (Fig.
angiomas having the highest T2 signal intensity in general. 19.4). Using fast gradient-echo techniques, fat and water
Precontrast and postcontrast T1-weighted imaging is also per- protons that share the same voxel will be imaged both “in phase”
formed to assess enhancement patterns of the liver parenchyma or 180 degrees “out of phase.” Tissues that have relatively equal
and liver lesions. If a hepatobiliary contrast agent is used, addi- quantities of fat and water will appear dark on the out-of-phase
tional delayed hepatobiliary phase images are acquired. sequence because the signals from fat and water are “opposed
in phase” and can cancel each other out (Boll et al, 2009;
DIFFUSE HEPATIC DISEASE Springer et al, 2010). Areas of fatty sparing will remain hyper-
intense relative to the surrounding fatty liver on out-of-phase
Fatty Liver imaging. On standard T1-weighted imaging, areas of focal fat
Fat may accumulate within hepatocytes for many reasons, can show increased signal. The appearance on T2-weighted
including alcohol intake, diabetes, medications, and obesity (see images depends on the type of sequence acquired and whether
Chapters 71 and 100). Hepatic steatosis may be diffuse, patchy, fat saturation is used. The ability of MRI to exploit those signal
A B
C
FIGURE 19.3. Normal hepatic magnetic resonance image. A, T1-weighted spin-echo image. The liver is brighter (hyperintense) relative to the signal
of the spleen. There are normal flow voids that appear as dark areas in the visualized vessels (straight arrow, intrahepatic portion of the inferior vena
cava; curved arrow, aorta). B, T2-weighted fast spin-echo image. Because of reversal of the liver/spleen contrast, the normal spleen is brighter than
the liver. Note the flow voids within the vessels (straight arrow, intrahepatic portion of the inferior vena cava; curved arrow, aorta). C, Postcontrast
T1-weighted gradient-echo image. Normal enhancement in the liver and spleen is evident, and the parenchyma of both is about equal in this phase
of the injection. All the vessels are bright as a result of the T1 shortening effect of gadolinium (straight arrow, intrahepatic portion of the inferior vena
cava; curved arrow, aorta).
A B
C D
E
FIGURE 19.4. Focal fat within the liver. A, Non–contrast-enhanced computed tomography image in a young adult with history of testicular cancer
shows heterogeneous regions of low attenuation in the right hepatic lobe (arrows). B, T1-weighted in-phase image at the same level now shows
signal abnormality in this region. C, T1-weighted out-of-phase image shows geographic areas (arrow) that have lost signal. On the out-of-phase
images, a black line “India ink effect” surrounds tissue interfaces. D, T1-weighted fat saturation precontrast image also shows mild signal loss in this
area (arrow). E, T1-weighted fat saturation postcontrast image. The area in question has normal vascularity coursing through it; no mass was
present—the typical appearance of fatty infiltration.
differences, as well as to apply diffusion-weighted and postcon-

trast imaging, offers an advantage over CT when evaluating for Cysts
focal lesions on a background of hepatic steatosis. Cysts are common hepatic lesions that may be sporadic (see
Newer, breath-hold three-dimensional (3D) T1 techniques Chapters 75 and 90B) or associated with polycystic disease of
with interpolation allow fast single breath-hold acquisition with the kidney (Fig. 19.6). Small cysts (<1 cm) may be difficult to
added fat and water separation (Bashir et al, 2012). Enhance- characterize on CT because of volume averaging but are easily
ment characteristics also aid in the diagnosis, as focal fatty liver diagnosed by MRI. The MRI criteria for a simple cyst is a
enhancement should parallel surrounding normal hepatic nonenhancing lesion that is homogeneously low in signal inten-
parenchyma. This is important because other lesions, including sity on T1-weighted images and homogeneously very bright on
hepatocellular carcinoma (HCC) and adenoma, may contain T2-weighted images, similar to cerebrospinal fluid. Cysts main-
small quantities of fat within them (see Chapters 90A and 91). tain their markedly high signal on heavily T2-weighted images,
similar to fluid in the cerebrospinal fluid, bile ducts, and pan-
Iron Deposition Disease creatic duct.
Hepatic iron deposition may occur from primary genetic hemo-
chromatosis or from secondary nongenetic systemic overload Hemangioma
(see Chapter 76). Primary hemochromatosis is an autosomal Hepatic hemangiomas (see Chapter 90A) are benign tumors of
recessive genetic disorder characterized by abnormal intestinal the liver with an estimated incidence of 20% (Albiin, 2012).
iron absorption, with accumulation of iron in the nonreticulo- Most hemangiomas are thus commonly found incidentally on
endothelial organ system. This predominantly occurs in hepa- other imaging studies, such as CT or ultrasound. MRI is ideal
tocytes later in the disease when other viscera, including the for characterizing hemangiomas and is the preferred modality
pancreas and heart, are more likely to be involved. The liver for imaging. On T1-weighted images, hemangiomas are hypoin-
shows abnormally low signal intensity compared with spleen on tense compared with the surrounding hepatic parenchyma,
T1-weighted in-phase sequences. T1-weighted gradient-echo have smooth, well-marginated borders, and are frequently lob-
sequences are very sensitive for detecting the presence of iron ulated. Hemangiomas are markedly hyperintense compared
within the hepatic parenchyma due to the local field inhomo- with normal liver on T2-weighted images. Hemangiomas also
geneities caused by the presence of iron. Primary, or genetic, tend to retain a high signal on more heavily T2-weighted
hemochromatosis is important to diagnose because this entity sequences, almost to the same degree as cysts, and become
may be unnoticed until late in the disease process, and its long- more conspicuous compared with surrounding liver paren-
term sequelae include fibrosis, cirrhosis, and HCC. chyma. After the administration of contrast, the presence of
Nongenetic systemic iron overload from secondary causes peripheral nodular discontinuous enhancement, with either
can sometimes be distinguished from primary hemochromato- partial or complete filling in on successive postcontrast phases,
sis based on the pattern of organ involvement. Nongenetic has been shown to be highly accurate for the diagnosis of hem-
systemic iron overload is typically associated with multiple angioma (Fig. 19.7; Silva et al, 2009).
blood transfusions, myelodysplastic syndrome (MDS), or Dynamic contrast-enhanced MRI after administration of
abnormal iron absorption in patients with cirrhosis (Queiroz- intravenous gadolinium is useful for increasing the specificity
Andrade et al, 2009). In secondary iron overload, there is pre- of diagnosis of hepatic hemangioma and for differentiating this
dominant accumulation of iron in the reticuloendothelial lesion from others, including metastases and HCCs (Albiin,
system (RES), with splenic and bone marrow iron deposition 2012). Larger hemangiomas tend to follow these criteria more
early and hepatic deposition later. If the iron deposition is frequently, although giant hemangiomas also may have variable
limited to the RES only, it is termed hemosiderosis. If non-RES signal intensity and a central scar (Prasanna et al, 2010). Prob-
organ involvement is seen, it is termed secondary hemochroma- lems arise with smaller hemangiomas, however, which may
tosis. These distinguishing characteristics and the patient’s “flash fill” and appear as hypervascular, homogeneous lesions.
history often allow correct differentiation. MRI is excellent for These smaller lesions (usually <1 cm) may be distinguished by
identifying the abnormal signal intensity due to iron deposition, both their T2 signal characteristics and enhancement pattern
and gradient-echo imaging has been shown to reliably quantify paralleling enhancing vessels (Albiin, 2012). In addition, hem-
hepatic iron load (Fig. 19.5) (Mavrogeni et al, 2013; Queiroz- angiomas have higher apparent diffusion coefficient values
Andrade et al, 2009). Iron quantification may have indirect obtained from DWI than metastases (Miller et al, 2010).
added uses; for example, limited data suggest it may help Caution must be used when assessing enhancement using the
predict nonrelapse mortality prior to allogeneic stem cell trans- hepatobiliary contrast agent gadoxetate disodium, however, as
plant in patients with MDS and acute myeloid leukemia the overlapping extracellular phase and hepatobiliary excretion
(Wermke et al, 2012). of this contrast may confound the typical hemangioma enhance-
ment and lead to “pseudowashout” in late dynamic phase,
FOCAL HEPATIC LESIONS making characterization more difficult (Goshima et al, 2010).
For this reason, we prefer not to use gadoxetate as the contrast
The advantage of MRI over CT is in the improved characteriza- agent for initial liver lesion characterization and use a tradi-
tion of liver lesions by improved soft tissue contrast combined tional ECF agent instead.
with novel tissue contrast mechanism. We will discuss common Sclerosed hemangiomas represent a diagnostic challenge,
benign and malignant liver lesions and emphasize distinguish- often mimicking malignancies in the liver; limited data suggest
ing imaging characteristics. Familiarity with typical lesion T1 that increasing sclerosis shows corresponding decreased T2
and T2 signal, as well as enhancement characteristics, will aid signal and less avid enhancement, probably corresponding to
the clinician in developing an approach to assess the most com- areas of fibrosis. Adjacent capsular retraction and adjacent
monly encountered hepatic lesions. wedge-shaped transient perfusional changes, as well as size
A B
C D
E F
G
FIGURE 19.5. Iron deposition. A, T1-weighted in-phase gradient-echo image (echo time [TE] 4.8) shows abnormal liver/spleen contrast. The liver
is homogeneously darker than the spleen from preferential deposition of iron within the hepatic parenchyma. This is a pediatric patient who had
undergone numerous transfusions while undergoing therapy for cancer. Note the difference from Figure 19.X. B, T1-weighted out-of-phase gradient-
echo image (TE 2.2) shows the liver/spleen contrast reverses, with the liver being brighter than spleen. This is opposite from Figure 19.X and shows
iron deposition disease within the liver. C, Axial T2-weighted fast spin-echo sequence shows marked low T2 signal within both liver and spleen. D-G,
Using T2-weighted multigradient-echo technique, there is progressive signal loss in the liver. This type of multiecho imaging allows estimation of iron
quantification using computer software.
A B
FIGURE 19.6. Hepatic cysts associated with polycystic kidney disease. A, Axial T2-weighted image at the level of the kidneys shows bilaterally
enlarged kidneys with multiple hyperintense cysts. Little normal renal parenchyma is present at this level, and multiple small hepatic cysts (arrow)
are seen. B, Coronal T2-weighted images through the kidneys show similar findings of enlarged kidneys containing multiple cysts (black arrows),
and small hepatic cysts are identified (white arrow).
decrease over time, may all offer clues to suggest the diagnosis, Hepatic Adenoma
but biopsy may be necessary for confirmation, depending on Hepatocellular adenomas (HCAs) (see Chapter 90A) also are
the clinical picture (Ridge et al, 2014). benign liver lesions, but unlike FNH, are associated with com-
plications, including abdominal pain, bleeding, and rarely,
Focal Nodular Hyperplasia malignant degeneration. HCAs are strongly associated with oral
Focal nodular hyperplasia (FNH; see Chapter 90A) is the contraceptive or steroid use and are more common in women
second most common benign tumor of the liver after heman- of childbearing age. Newer low-dose oral contraceptives have
gioma. Pathologically, FNH contains all the elements of normal a less strong association with HCAs. Although HCA may be
liver and may have a central fibrous scar, which shows delayed large at presentation, they are increasingly incidentally discov-
enhancement, and is surrounded by hepatocytes and small bile ered and often less than 4 cm in diameter (Grazioli et al, 2012).
ducts (see Chapter 89). FNH has been associated with oral Hemorrhage may be life threatening if it extends into the peri-
contraceptive use (Scalori, 2002) as well as chemotherapy in toneum. Although imaging characteristics vary, a combination
the pediatric age group (Do et al, 2011; Sudour et al, 2009). of features, including enhancement characteristics, intralesional
Most FNH are isointense to normal liver parenchyma on T1- lipid, and hypointensity on delayed hepatobiliary phase imaging,
and T2-weighted images, and a few are mildly T1 hypointense may help increase diagnostic confidence (Mohajer et al, 2012).
or minimally hyperintense on T2-weighted images (Fig. 19.8A Nonetheless, small adenomas without hemorrhage may be dif-
and B). When discussing relative T1 and T2 signal, a normal ficult to differentiate from FNH without a central scar and from
background liver parenchyma is assumed. In a liver with iron well-differentiated HCC.
deposition, which diffusely lowers the signal intensity of the Recently, distinct genetic subtypes of adenomas have been
liver parenchyma, T1 and T2 relative hyperintensity can be described and show differing behaviors, histopathology, and
expected. This may be seen, for example, in pediatric oncologic imaging features. These include inflammatory, hepatocyte
patients who have concomitant iron deposition (Do et al, nuclear factor-1α (HNF-1α–mutated), and β-catenin–mutated
2011). Frequently, FNH is visualized by displacement of the HCAs (see Chapter 89). Of note, some adenomas may be both
normal hepatic vasculature, and classic FNH shows strong β-catenin mutated and inflammatory. A group remains unclas-
arterial peak enhancement following administration of an intra- sified and encompasses other HCAs without any genetic abnor-
venous contrast agent, with lack of washout in portal venous or malities (Grazioli et al, 2013). Inflammatory HCA, the most
late dynamic phase imaging. common subtype (30% to 50% of HCAs) typically seen in
A common feature of FNH is the presence of a central scar. women with a history of oral contraceptive use, is generally
This distinguishing feature is hypointense on T1-weighted hypointense on T1 images, hyperintense on T2, and frequently
images and hyperintense on T2-weighted images with enhance- heterogeneous. They usually show moderate arterial enhance-
ment on delayed imaging beginning at approximately 3 minutes ment, persistent dynamic phase enhancement, and variable
by using ECF contrast agents (Fig. 19.8C) (Karam et al, uptake in delayed hepatobiliary phase (Grazioli et al, 2013).
2010). The presence of enhancement, similar to or higher- The enhancement pattern may reflect a mix of poorly function-
than-normal parenchyma, during the delayed hepatobiliary ing hepatocytes, arterioles without accompanying veins,
phase with a hepatocyte-specific contrast agent is expected for inflamed bile ducts, and dilated sinusoids. On MRI, small
the majority of FNHs. Of note, with the hepatobiliary agent amounts of intralesional lipid may be seen. A small percentage
gadoxetate disodium, a central scar, if present, does not usually of these HCA show contrast washout in late dynamic phases,
enhance relative to the hepatocytes present in this lesion. whereas others retain contrast similar to hemangiomas. Delayed
A B C
D E F
FIGURE 19.7. Hepatic hemangioma undergoing sclerosis. A, Heavily T2-weighted image (echo time [TE] 150) shows a mass (m) that is hyperintense
to hepatic parenchyma with well-defined margins. B, Precontrast T1-weighted fat saturation image at the same level shows the mass to be low
signal compared with background parenchyma. C, T1-weighted postcontrast image in early portal venous phase shows peripheral nodular enhance-
ment within this mass (arrows). Portions of the lesion fail to show enhancement during this phase. D, Equilibrium phase T1-weighted image post-
contrast shows the lesion has partially filled in toward the central portion. This constellation of findings is typical for hepatic hemangioma.
E-F, The same lesion 5 years later on T1-weighted postcontrast scans in portal venous and equilibrium phase imaging shows less avid enhancement
and has shown slight size decrease, consistent with developing sclerosis. Sclerosing hemangiomas may show size decrease with less avid enhance-
ment, associated capsular retraction, and lower T2 signal than nonsclerosed hemangiomas.
peripheral rim enhancement in hepatobiliary phase may also be malignant degeneration. They show moderate, often heteroge-
seen (Grazioli et al, 2013). neous arterial phase enhancement, which may persist but is
Hepatocyte nuclear factor-1α mutated HCAs (30% to 35% variable in late dynamic and delayed phases. Although T1
of HCAs) are almost exclusively seen in women and often show hypointensity and T2 hyperintensity are common, it is often
large amounts of intralesional lipid. They enhance less avidly heterogeneous. These lesions show diffuse glutamine synthetase
in arterial phase than inflammatory HCA and typically are expression from upregulation, although it may be heteroge-
homogeneously low in signal on delayed hepatobiliary phase. neous. This is in contradistinction to FNH, which shows map-
They may also remain low signal in portal venous or equilib- like glutamine synthetase expression distribution adjacent to
rium dynamic phases, but that may be due to the opposed- hepatic veins (Agarwal et al, 2014; Balabaud et al, 2013; Grazi-
phase property of postcontrast T1-weighted imaging (Grazioli oli et al, 2013).
et al, 2013). Unclassified HCA (10% of HCAs) have no specific genetic
β-Catenin–mutated HCAs (10% to 15% of HCAs) are asso- or histopathologic abnormalities. No specific imaging features
ciated with glycogen storage disease, androgen use, and have a have been identified for these lesions, although experience
male predilection; they also have a higher chance of undergoing remains limited due to their infrequent diagnosis.
A B
C D
FIGURE 19.8. Focal nodular hyperplasia (FNH). A, T2-weighted image shows a mass that is isointense to hepatic parenchyma (arrows). Centrally,
within this mass is a linear, hyperintense focus (arrowhead). B, A precontrast, T1-weighted gradient-echo image shows the mass (arrows) to be
mildly hypointense to liver. C, Arterial dominant phase T1-weighted gradient-echo sequence shows that the mass enhances intensely with respect
to hepatic parenchyma (arrows). The central focus, which was bright on the T2-weighted images, does not show enhancement on this phase of the
injection (arrowhead). D, Postcontrast equilibrium phase image shows the mass to be isointense to background hepatic parenchyma. The central
portion of scar shows delayed enhancement (arrowhead) characteristic of FNH.
Hepatic Abscess/Infection lesions, including hemangiomas or focal fat, may appear similar
CT is generally the modality of choice in recent postoperative to metastases when imaged with CT or ultrasound. In general,
patients in whom there is a clinical suspicion of a hepatic metastases are mildly hypointense on T1-weighted images and
abscess (see Chapter 72). Patients with more atypical symptoms are mildly hyperintense on T2-weighted images. Unless they are
may present a diagnostic dilemma. Although some patients necrotic, mucinous, or cystic, metastases are not as bright on
show symptoms suggesting abscesses, others do not, but T2-weighted images as hemangiomas or cysts and often become
patients suspicious for abscesses or bile leaks may be imaged less conspicuous with greater (longer echo time [TE]) T2
with MRI. Abscesses are usually hyperintense on T2-weighted weighting. A number of metastases are hypervascular and best
images with an irregular rim of intermediate signal intensity seen on arterial phase imaging, such as those arising from a
surrounding a hyperintense outer rim. Abscesses are generally primary neuroendocrine tumor, renal cell carcinoma, or intra-
hypointense on T1-weighted images, unless they have hemor- hepatic metastases from hepatocellular carcinoma. Most
rhage or proteinaceous debris within them. After administration metastases, however, are hypovascular and tend to have less
of contrast material, the rim enhances, whereas the central well-defined borders after contrast enhancement than other
portion does not. The imaging findings may overlap with malig- benign lesions (Fig. 19.9). During contrast administration,
nancies, including gallbladder cancer. Infected cystic hepatic metastases often show early peripheral marginal enhancement.
lesions also may occur, such as parasitic echinococcal infec- These peripheral rims may wash out and appear hypointense on
tions. Echinococcal cysts are typically multiloculated with inter- delayed images (Fig. 19.10). Larger metastases tend to have a
nal thin-walled daughter cysts, and typically show no internal thick irregular rim of enhancement representing viable tumor
enhancement. A peripheral low T1 and T2 signal rim, as well with areas of central necrosis. Multiple studies have demon-
as correlation with a history of possible exposure, may provide strated superior accuracy of hepatobiliary agent contrast-
added clues to the diagnosis (see Chapter 74) (Qian et al, enhanced MRI over contrast-enhanced CT (Lafaro et al, 2013).
2013). Both ECF and hepatobiliary agent contrast-enhanced MRI has
shown higher sensitivity and specificity than CT–positron-emis-
Hepatic Metastases sion tomography (PET) imaging as well, particularly for small
The liver is the most common site for the hematogenous spread (<1 cm) lesions, and hepatobiliary contrast-enhanced MRI
of malignant neoplasms (see Chapters 92 to 94). Other hepatic combined with PET imaging may offer added benefit (Donati
A B
C D
FIGURE 19.9. Hepatic metastasis from colorectal carcinoma. A, T1-weighted image shows a low signal intensity metastasis (arrow) within segment
IVB of the liver. B, T2-weighted image shows the mass is mildly brighter than background hepatic parenchyma. C, Post–contrast-enhanced,
T1-weighted fat saturation image shows the lesion as centrally hypointense, in contrast to the normal surrounding hepatic parenchyma, with minimal
areas of marginal enhancement. D, Fused computed tomography/fluorodeoxyglucose–positron-emission tomography image confirms avid tracer
uptake within the lesion, consistent with active tumor. Note the differences in appearance from hemangioma (see Fig. 19.7).
et al, 2010; Maegerlein et al, 2015; Seo et al, 2011). Arterial of disease, progressing from benign to malignant, which occurs
phase imaging may also provide anatomic detail preoperatively; in response to hepatic parenchymal damage and scarring (Lee
however, CT angiography offers improved spatial resolution et al, 2012). The advantage of MRI in the evaluation of HCC
over MRI. is that areas of carcinoma show differences in signal intensity,
diffusion, blood pool, and functional hepatocyte enhancement
Hepatocellular Carcinoma and growth compared with regenerative or dysplastic nodules,
The diagnosis of HCC (see Chapter 91) on CT and ultrasound or background cirrhotic parenchyma. MRI also has an advan-
can be challenging. It is often associated with cirrhosis and tage over other modalities in assessing vascular invasion.
parenchymal heterogeneity (Fig. 19.11), increasing the diffi- Although HCC generally is hypointense on T1-weighted
culty in distinguishing focal hepatic lesions. Cirrhotic livers and hyperintense on T2-weighted images, and shows arterial
show both nodular hepatic contour and parenchymal multiphase hyperenhancement with delayed dynamic phase washout,
nodular change. These nodules represent a spectrum, from signal intensity may be variable (Silva et al, 2009), particularly
regenerative or dysplastic nodules to HCC. These nodular for early HCC measuring less than 2 cm (Rhee et al, 2012).
lesions are generally thought to constitute a multistep spectrum On T2-weighted images, larger, more poorly differentiated
A B
C D
FIGURE 19.10. Hepatic metastasis from breast carcinoma. A, T2-weighted image shows a small lesion within the liver (straight arrow) that is mildly
hyperintense. Portal lymphadenopathy also is present (curved arrow). B, Pre-contrast-enhanced, T1-weighted gradient-echo image shows a well-
defined mass (arrow) that is hypointense to liver. C, T1-weighted gradient-echo image after the administration of contrast shows the peripheral portion
that enhanced irregularly during the arterial dominant phase (not shown) beginning to wash out, suggesting an early target appearance (arrow). D,
Delayed postcontrast gradient-echo image shows continued filling of the central portion of this lesion; the periphery has washed out (arrow).
A B
FIGURE 19.11. Liver with cirrhotic configuration. A, T1-weighted spin-echo image at the level of the portal vein shows hypertrophy of the left lobe
and caudate with atrophy of the right lobe of the liver. B, T2-weighted image of the liver at the same level shows mildly heterogeneous signal in the
atrophied right lobe, making exclusion of an underlying mass difficult. Note also the focal hepatic scarring (arrow).
A B
FIGURE 19.12. Large hepatocellular carcinoma showing the mosaic pattern. A, T1-weighted image shows a large heterogeneous mass (m) in the
liver. Note the hypertrophied left hepatic lobe and caudate. B, T2-weighted image also shows the mass (m) to be heterogeneous, with islands of
tissue that are hyperintense with respect to other portions of the same mass. This pattern is known as the mosaic pattern, and it can be seen in
hepatocellular carcinoma, especially when the lesion is large.
A B
FIGURE 19.13. Hepatocellular carcinoma containing intracellular lipid. A, T1-weighted in-phase gradient-echo image shows a mass in the liver
that is darker than normal parenchyma, with a small internal area of brighter signal (arrow). B, T1-weighted out-of-phase gradient-echo image shows
the same area (arrow) has lost signal, which is consistent with an admixture of fat and water-containing tissue.
HCCs may be heterogeneous due to areas of necrosis (Fig. hyperenhance, they tend to be more homogeneous and isoin-
19.12). Areas of heterogeneity may indicate moderate to poorly tense to background liver parenchyma during the equilibrium
differentiated tumors (Witjes et al, 2012). HCC may show high phase, often show low T2 signal, and are variable on delayed
T1 signal (Fig. 19.13) (Basaran et al, 2005), and lesions with hepatobiliary phase imaging (Cruite et al, 2010). In HCC,
intracellular lipid show signal loss on out-of-phase T1-weighted hepatocyte contrast agents generally show hypointensity in
imaging. Intracellular lipid may be an important clue to small delayed hepatobiliary phase imaging compared with surround-
HCCs without typical enhancement patterns, as early HCC are ing liver in the vast majority of cases, with a small percentage
sometimes hypovascular in arterial phase (Rhee et al, 2012). showing variable internal uptake, which may reflect either the
Contrast washout in portal venous or later dynamic phase grade of tumor differentiation or specific enzyme production
MRI, as well as heterogeneous enhancement, has been associ- and tumor receptors. (Cruite et al, 2010) Rarely, well-
ated with moderate to poorly differentiated tumors; a higher differentiated HCC may show tracer concentration in delayed
percentage of well-differentiated tumors may not demonstrate hepatobiliary phase with areas of increased signal intensity.
washout (Okamoto et al, 2012; Tan et al, 2011; Witjes et al, Due to the complexity of imaging features and overlap, as
2012). In addition to washout, peritumoral capsules, which are well as multimodality availability, there have been organized
low in signal intensity on the arterial dominant phase and efforts to improve report standardization and communication
enhance later, are also associated with microvascular invasion, regarding imaging findings, as well as recommendations regard-
an important feature with clinical significance (Witjes et al, ing surveillance imaging and screening. The American Associa-
2012). These are distinguishing features in contrast to benign tion for the Study of Liver Diseases (AASLD) issued revised
or dysplastic nodules. Although dysplastic nodules may also recommendations in 2010, recommending HCC screening for
patients at risk with ultrasound (US) every 6 months combined Although extracellular lipid is typical in HAMLs, they often
with serum α-fetoprotein (AFP). MRI imaging has an emphasis show brisk, avidly enhancing soft tissue components with
on arterial phase hyperenhancement and washout, with biopsy washout, which overlaps with other liver malignancies, such as
improving sensitivity for indeterminate lesions (Tan et al, hepatocellular carcinoma. HAML is generally considered a
2011). The Liver Imaging Reporting and Data System (LI- benign, solitary tumor comprising three elements: smooth
RADS), an initiative supported by the American College of muscle, thick-walled blood vessels, and mature adipose tissue
Radiology, as well as the Liver and Intestinal Organ Transplant (Cai et al, 2013), although they can also present without fat
Committee (OPTN) have also issued guidelines regarding (Butte et al, 2011). Limited data suggest a well demarcated,
HCC classification. LI-RADS was revised in 2014 to better lipid-containing tumor with a thin peripheral enhancing rim,
incorporate OPTN and AASLD guidelines regarding growth lack of peritumoral capsule, and an early draining vein that may
and lesion detection on screening ultrasound. Ancillary fea- offer clues to the diagnosis, particularly in patients without cir-
tures, including nodule-in-nodule or mosaic appearance, intra- rhosis or elevated serum AFP (Fig. 19.14) (Cai et al, 2013; Du
lesional fat, diffusion restriction, and perilesional or coronal et al, 2012). Surgical excision remains preferable, as there have
enhancement, for example, are also incorporated into the been reports of HAML hemorrhage and capsular rupture, as
LI-RADS imaging interpretation algorithm (http://nrdr.acr.org/ well as venous obstruction or Budd-Chiari syndrome from mass
lirads/). Although continued revisions are inevitable, consensus effect. Furthermore, pathologic analysis remains essential for
guidelines will encourage report standardization, improve com- diagnostic confirmation; immunohistochemical stains for
munication, and ultimately improve decision making. Agree- markers may offer increased accuracy (Du et al, 2012).
ment among readers using newer algorithms tends to be
moderate to substantial for expert readers but lower among Mesenchymal Tumors
novices, suggesting implementation of criteria may require a Rarely, tumors of mesenchymal origin, including hepatic angio-
learning curve (Davenport et al, 2014). sarcoma (HAS), leiomyosarcoma, and fibrous histiocytoma,
may be present within the liver (Anderson et al, 2009). The
Fibrolamellar Carcinoma MRI appearance of these tumors is nonspecific; they are gener-
Fibrolamellar HCC (FLHCC), a rare variant, occurs in young ally of low intensity on T1-weighted images and hyperintense
adults, and is distinct from conventional HCC on multiple on T2-weighted images, with heterogeneous enhancement.
levels, including molecular and clinical differences (see Chapter HAS typically shows hyperenhancement, often peripheral, but
91). The average age at presentation is 25 years, and the pres- sometimes central and irregular, with progressive fill-in on more
ence of nodal metastases is common (Ganeshan et al, 2014). delayed phases following blood pool. Although similar to cav-
FLHCC may have a central scar, and these scars tend to be ernous hemangiomas, HAS may also show centrifugal or
low in signal intensity on T1-weighted and T2-weighted images “reverse” enhancement patterns and show greater degrees of
due to fibrous changes. FNH also may have a central scar, variance and heterogeneity, with more disordered peripheral
although FNH scars generally show increased T2 signal; enhancement than nodular, discontinuous, clump-like enhance-
however, it is not always a reliable differentiating feature (Gane- ment typically seen in benign hemangiomas. HAS is typically
shan et al, 2014; Kim et al, 2009). FLHCC may show early multiple, of varying size, involving both hepatic lobes, and
heterogeneous enhancement with variable washout and may rapidly growing, as median survival has been reported as less
show partial hepatocyte agent concentration in delayed hepa- than 6 months (Huang et al, 2014; Pickhardt et al, 2015).
tocyte phase imaging (Meyers et al, 2011). Metastatic disease
is also common at presentation, both in the abdomen and chest, Epithelioid Hemangioendothelioma
with adenopathy being prominent (>2 cm) in a majority of Epithelioid hemangioendothelioma (EH) is a rare tumor of
cases (Do et al, 2014). adults that should be distinguished from infantile hemangioen-
dothelioma, which occurs in infants and children. The progno-
LESS COMMON HEPATIC TUMORS sis is better for EH than for other parenchymal tumors, such
as sarcoma, although extrahepatic metastases do occur. EH is
Lymphoma generally low in signal intensity on T1-weighted images and
Non-Hodgkin lymphoma accounts for most hepatic lympho- hyperintense on T2-weighted images but not as hyperintense
mas. MRI shows masses of varying size, which are generally as hemangiomas. They also may show a distinct target sign on
low in signal intensity on T1-weighted images and hyperintense T1- and T2-weighted images (Economopoulos et al, 2008).
on T2-weighted images (Coenegrachts et al, 2005; Rizzi et al, After Gd-DTPA administration, an irregular nodular and con-
2001), but not as great as with other benign hepatic lesions, centric enhancement may be seen. Capsular retraction also may
such as hemangiomas or cysts. This difference in signal inten- be seen with EH, helping to elucidate the diagnosis (Lin et al,
sity makes it relatively easy to determine the malignant nature 2010) (see Chapter 89).
of the lesion but not the specific tumor type. Lymphoma may
be relatively hypoenhancing in early dynamic phases, with Biliary Cystadenoma and Biliary Cystadenocarcinomas
isointensity in later phases (Coenegrachts et al, 2005). The Biliary cystadenomas and biliary cystadenocarcinomas are rare
imaging characteristics of lymphoma overlap with other malig- tumors of the liver that present as a predominantly cystic mass
nant hepatic lesions. containing enhancing septations (Fig. 19.15; see Chapters 89
and 90B). These lesions may be of low, intermediate, or
Angiomyolipoma increased signal on T1-weighted images, depending on the
Hepatic angiomyolipoma (HAML) is an uncommon tumor, protein or hemorrhagic content within the mass, and they are
more frequent in females, which currently presents a diagnostic usually hyperintense on T2-weighted sequences. Enhancing
challenge for imaging diagnosis (see Chapters 89 and 90). mural or internal irregular soft tissue nodularity, as well as
A B
C D
FIGURE 19.14. Angiomyolipoma. A, T1-weighted in-phase image shows a high-signal right hepatic mass (arrow). B, T1-weighted out-of-phase
image shows low signal at its margin due to chemical shift artifact between the mass and surrounding liver, confirming the presence of bulk fat.
C, Axial T2-weighted image with fat saturation shows the mass is dark. D, Axial T1-wieghted fat saturation image shows the mass is dark as well.
internal hemorrhage, has been shown in to be more often asso- high T2 signal thickening of the gallbladder wall, with early and
ciated with cystadenocarcinomas; however, differentiation may prolonged heterogeneous enhancement, often in patients with
not be possible (Arnaoutakis et al, 2015; Qian et al, 2013; multiple gallstones (Tan et al, 2013). Cholelithiasis is a predis-
Wang et al, 2012). Communication with the biliary ductal posing condition. Evidence of liver invasion and spread to
system on delayed hepatobiliary phase imaging with hepatocyte regional lymph nodes also may be identified with MRI, which
contrast agents has been reported, helping to differentiate these is optimal for evaluation of these lesions because of its out-
tumors from nonneoplastic simple hepatic cysts. (Billington standing tissue contrast, ability to image directly in multiple
et al, 2012; Marrone et al, 2011). Infectious cysts may also planes (Dai et al, 2009; Sikora et al, 2006), with DWI offering
communicate with the biliary tree, however, such as with added sensitivity and specificity for the primary tumor as well
complex cysts from echinococcal disease. as nodal or hepatic metastatic disease (Tan et al, 2013). Fluo-
rodeoxyglucose (FDG) PET imaging may also provide
BILIARY TUMORS improved sensitivity and specificity for nodal or distant meta-
static disease (Lee et al, 2010).
Gallbladder Carcinoma
MRI offers improved characterization of gallbladder cancer Bile Duct Cancer
compared with other modalities. However, its differentiation Bile duct tumors (see Chapters 50, 51, and 59) are well
from inflammatory conditions, which may occur concurrently, depicted on MRI and may be intrahepatic, hilar, or extrahe-
may be difficult (see Chapters 33 and 49). As opposed to the patic, with a mass-forming appearance or as periductal infil-
inflammatory associated wall thickening with maintained trating or papillary intraductal morphologies (Chung et al,
mucosal and submucosal layers, with differential enhancement, 2009). Tumors may show focal nodular thickening along
gallbladder cancer typically shows irregular intermediate to ductal walls or be discrete intrahepatic masses. Central hilar
A B
C
FIGURE 19.15. Biliary cystadenoma. A, Coronal T2-weighted image shows a lobulated mass (m) in close proximity to the left hepatic duct (arrow-
head). B, Projection image in the coronal oblique plane shows the well-defined hyperintense mass to be in continuity with a mildly dilated left-sided
biliary radicle (straight arrow). The normal right-sided biliary radicle (curved arrow) is easily seen with this technique. C, Postcontrast, T1-weighted
gradient-echo image shows no enhancement within this mass (m).
cholangiocarcinomas are the most common and are generally capsular retraction as well as vascular encasement, typically
associated with biliary ductal dilation. Bile duct tumors may without tumor thrombus (Chung et al, 2009). Smaller tumors
be seen as intermediate, mildly increased T2 signal (Fig. may show early, more uniform arterial enhancement, particu-
19.16), however, less so than the surrounding biliary dilata- larly when less than 4 cm in size (Kim et al, 2011). Intraductal
tion. The level of obstruction and continuity of the tumor with papillary neoplasm of the bile duct is an uncommon subtype of
the vasculature are well evaluated with MRI (Peporte et al, bile duct cancer with characteristics similar to pancreatic intra-
2013), although CT angiography with multiphasic imaging ductal mucinous neoplasms. The majority of tumors occur near
offers improved spatial resolution and similar accuracy to MRI the hilum or in extrahepatic ducts, and the clinical outcomes
to assess vascular involvement, and exceeds MRI accuracy are better than conventional bile duct cancers (Rocha et al,
when assessing for nodal and distant metastases (Aljiffry et al, 2012).
2009). Focused assessment of ductal, portal venous, and
hepatic arterial involvement is performed in staging and pre- BENIGN DISEASES OF THE BILIARY TRACT
operative imaging.
Peripheral, mass-forming intrahepatic cholangiocarcinomas Cholelithiasis and Choledocholithiasis
occur in approximately 10% of cases and generally have high Gallstones are well identified on T2-weighted images and on
T2 signals, lobulated heterogeneous masses with peripheral MRCP sequences (Fig. 19.18; see Chapters 32, 33, and 36).
enhancement, and gradual centripetal fill-in (Fig. 19.17) They usually appear as low signal intensity structures in a
(Hennedige et al, 2014, Kang et al, 2012) (see Chapter 50). fluid-filled gallbladder. Ultrasound is usually the modality of
Satellite lesions are common within the liver, and these lesions choice in the evaluation of uncomplicated cholelithiasis or cho-
tend to show low signal intensity on delayed hepatocyte imaging lecystitis due to its high sensitivity and lower cost, but choledo-
phase (Chung et al, 2009; Kim et al, 2011; Peporte et al, cholithiasis is more effectively imaged by MRI (Johnson et al,
2013). Intrahepatic cholangiocarcinomas may show adjacent 2010; Samaraee et al, 2009). Coronal T2-weighted imaging,
A B
C
FIGURE 19.16. Hilar mass. A, Axial T2-weighted image shows dilatation of the left hepatic duct and a low to intermediate–intensity mass expand-
ing the right hepatic duct (arrow). B, T1-weighted late arterial phase postcontrast image shows early enhancement of the tumor (arrow). C, Portal
venous phase postcontrast T1-weighted imaging shows the mass (arrow) is hypoenhancing to liver. Biopsy of a more superior intrahepatic mass in
this patient with cirrhosis showed poorly differentiated hepatocellular carcinoma. Extension into the biliary ducts with ductal expansion and obstruc-
tion is an uncommon finding in hepatocellular carcinoma (HCC). Although HCC and cholangiocarcinoma may exist simultaneously as two distinct
tumors, a distinct subtype of mixed HCC-cholangiocarcinoma is increasingly recognized.
performed routinely with MRI, readily identifies common duct have been described (Lee et al, 2009; Todani et al, 1977): type
stones (Fig. 19.19). MRCP can also be obtained to evaluate I has been subdivided into A, B, and C, in addition to types II,
whether retained stones are present after cholecystectomy, III, IVa, IVb, and V. Recently, there have been advocates for
although clips may limit examinations in the perioperative time dropping the numeric classification system, instead using a
period. If no stones are present on MRCP, this may obviate the more descriptive, clinically meaningful nomenclature (Visser
need for ERCP. et al, 2004). MRI is well suited not only to diagnose these cysts,
but also to classify the type of cyst. Not only can 3D MR chol-
Choledochal Cysts angiograms depict the normal and abnormal anatomy, but
Choledochal cysts (see Chapter 46) represent dilatation of the direct coronal imaging and delayed scans post–hepatocyte-
extrahepatic bile ducts with possible associated intrahepatic specific gadolinium-based contrast agents can be obtained for
biliary duct dilatation. This entity is a relatively uncommon further evaluation (Gupta et al, 2010).
congenital anomaly that usually presents before 10 years of age.
The classic triad includes a palpable mass, abdominal pain, and Postoperative Biliary Complications
jaundice. Cysts may be associated with chronic inflammation Complications from surgical procedures include bile leaks,
and increase the risk for cholangiocarcinoma. Five types of cysts abscess formation, and biliary strictures (see Chapters 27 and
A B C
D E F
FIGURE 19.17. Mass forming intrahepatic cholangiocarcinoma. A, T1-weighted in-phase gradient-echo image shows a peripheral hypointense
mass (m) with capsular retraction. B, A T2-weighted fat saturation image shows increased signal intensity within the mass. C, Diffusion-weighted
image shows restricted diffusion within the mass, which is brighter than liver. D, T1-weighted fat saturation post–contrast-enhanced image in late
arterial phase shows peripheral enhancement. E, T1-weighted post–contrast-enhanced image in late portal venous phase shows only partial filling
of the mass. F, T1-weighted post–contrast-enhanced image in late portal venous phase at a separate level shows an additional satellite lesion (arrow),
smaller in size with marginal enhancement, typical of intrahepatic metastasis.
A B
FIGURE 19.18. Choledocholithiasis. A, Coronal T2-weighted single-shot fast spin-echo image through the common duct in a patient after chole-
cystectomy shows multiple stones within the common bile duct. Note the distal stone impacted at the level of the ampulla (arrow). B, Axial T2-weighted
image with the same technique also shows a stone, surrounded by bile, in the distal common bile duct (arrow).
biliary phase scans by opacifying the biliary tree, which may

improve diagnostic confidence to evaluate for active biliary leak,
anatomic variants, and choledocholithiasis (Boraschi & Donati,
2014; Gupta et al, 2010).
PANCREAS
Solid Tumors of the Pancreas
Pancreatic solid tumors include both neoplastic and nonneo-
plastic origins (see Chapter 59). Although MRI characteristics
alone may help predict etiologies, there may be imaging-feature
overlap between entities. Imaging characteristics combined
with clinical presentation and demographic data often guide
therapy and predict the diagnosis. Neoplastic etiologies include
tumors arising in the pancreas, such as adenocarcinoma, solid
pseudopapillary tumor, neuroendocrine tumors, or pancreatic
lymphoma. Metastases, lymphomas, and other rare tumors
may involve the pancreas (Low et al, 2011). In addition to
enhancement characteristics, such as early arterial enhance-
ment in neuroendocrine tumors, precontrast signal character-
istics, such as hemorrhage with increased T1 signal in solid
A pseudopapillary tumors, may provide clues to the diagnosis.
Nonneoplastic etiologies include focal pancreatitis, accessory
spleen, focal fat, congenital anomalies, and other rare etiolo-
gies. Intrapancreatic splenic tissue shows signal intensity and
enhancement characteristics paralleling that of spleen on all
MRI sequences. Focal pancreatitis may mimic pancreatic ade-
nocarcinoma; however, restricted diffusion on DWI is typically
seen in adenocarcinoma (Fattahi et al, 2009; Kartalis et al,
2009), compared with focal pancreatitis. Focal pancreatitis may
also cause irregular narrowing of the main pancreatic duct, as
opposed to the abrupt cutoff and upstream dilatation with
associated parenchymal atrophy noted in adenocarcinoma
(Low et al, 2011; Siddiqi et al, 2007).
Pancreatic Cancer
Pancreatic adenocarcinoma (see Chapters 59 and 62) is well
B evaluated on MRI and MRCP. T2, diffusion-weighted, and
dynamic contrast-enhanced T1-weighted MRI sequences are
FIGURE 19.19. Afferent loop syndrome. A, Coronal T2-weighted particularly helpful to characterize pancreatic cancer and stage
sequence in a patient after pancreaticoduodenectomy for pancreatic patients (Tamm et al, 2012). MRI is generally thought to be
carcinoma. The patient has a hepaticojejunostomy with a patent anas-
highly accurate in assessing for vascular invasion or encasement
tomosis (white arrow). The afferent loop is markedly dilated (black arrow)
compared with other loops of bowel. No mechanical obstruction was
(tumor surrounds >50% of the vessel circumference), local-
noted. B, Axial T2-weighted image through the level of the anastomosis regional tumor extension, and liver metastases (Tamm et al,
(arrow) shows the site to be patent without evidence of a stricture. The 2012), although recent data have suggested MRI may under-
afferent loop is dilated. estimate tumor size (Hall et al, 2013; Legrand et al, 2015).
Similar to other modalities, accuracy is limited for assessment
of lymph node metastases with MRI techniques. Pancreatic
30). Although using other imaging modalities in the immediate cancer subtypes may be predicted based on imaging features,
postoperative setting to assess for bile leaks and abscesses may patient demographics, and clinical presentation, particularly
be prudent, MRI is uniquely suited for the assessment of the neuroendocrine tumors; however, there is overlap in the imaging
postoperative biliary tract. Bile duct injuries after surgery can appearance and location of solid and cystic neoplasms. Prelimi-
be caused by a number of things, but these may all lead to the nary data suggest grade of tumor enhancement may correlate
same result—a stricture at the anastomotic site. MRCP is with tumor grade in adenocarcinomas (Lauenstein et al, 2010).
uniquely suited for addressing this problem. Surgical clips near
the anastomosis may create streak artifacts during CT scan- Cystic Lesions of the Pancreas
ning, but currently used clips are less problematic for MRI. Detection of pancreatic cystic lesions (see Chapter 60) has
MRI has multiplanar capabilities and superior tissue contrast, increased in recent years with increased imaging and improved
and MRCP sequences can minimize susceptibility to artifacts spatial resolution. They are increasingly found incidentally.
to allow improved visualization of the region of the anastomosis Benign lesions may also progress to malignant lesions over
and improved diagnostic confidence (see Fig. 19.19). More time. Diagnostic possibilities include benign lesions, such as
recently, hepatocyte contrast agents also allow delayed hepato- pseudocysts, serous cystadenomas, and true epithelial cysts; to
benign lesions with potential for malignant degeneration, such to have a higher association with malignancy; their presence
as intraductal mucinous or parenchymal mucinous cystic neo- may warrant surgical excision (Kim et al, 2014). Evaluation
plasms; to malignant lesions, such as rare cystic or necrotic after contrast administration helps assess cyst wall irregularity
adenocarcinomas or cystic neuroendocrine tumors (Kucera or enhancement of soft tissue components. Molecular analysis
et al, 2012). Multiple imaging modalities may be used to iden- of cyst contents to assess for genetic mutations or gene silencing
tify and characterize pancreatic cystic lesions, including ultra- associated with IPMN may offer improved accuracy in diag-
sound, CT, and MRI. nostic workup (Freeny et al, 2014).
Intraductal Papillary Mucinous Neoplasms Autoimmune Pancreatitis

Cystic lesions arising from intraductal papillary mucinous neo- An increasing incidence of autoimmune pancreatitis is most
plasms (IPMNs) of either branch-duct, main-duct, or mixed likely due to increased awareness and thereby increased detec-
type have been increasingly discovered (Campbell et al, 2015), tion (Vlachou et al, 2011) (see Chapter 59). Imaging features
and the International Association of Pancreatology has pub- may suggest the diagnosis and are correlated with clinical
lished consensus guidelines for their management, which were features, including elevated serum immunoglobulin G4, rela-
last revised in 2012 (see Chapter 60). The International Asso- tively indolent clinical presentation, and response to corticoste-
ciation of Pancreatology stratifies patients based on “high risk” roid therapy. Autoimmune pancreatitis may be diffuse, focal,
or “worrisome” features, with guidelines regarding resection, or multifocal, and often shows heterogeneous T1 hypointensity,
endoscopic ultrasound, and imaging follow-up (Tanaka et al, slight T2 hyperintensity, and delayed enhancement due to asso-
2012). A recent revision includes main duct dilatation greater ciated fibrosis. Fibrosis may also show diffuse T2 hypointensity
than 5 mm (without obvious cause) as a worrisome feature, but (Vlachou et al, 2011). Diffuse main pancreatic ductal narrow-
also downgraded size greater than or equal to 3 cm from high ing or irregularity, and upstream common bile duct dilatation
risk to worrisome (Tanaka et al, 2012). Soft tissue contrast is or wall thickening, may be present. There may also be a peri-
improved with MRI imaging, which permits optimal depiction pancreatic halo, which appears hypointense on T1- and
of features deemed worrisome or high risk, such as septations, T2-weighted imaging postcontrast (Heyn et al, 2012). Focal
thick enhancing walls, nodularity, or soft tissue components. disease, particularly when involving the head with associated
The single-shot fast spin-echo T2-weighted sequence is par- upstream pancreatic and main bile duct dilatation, may mimic
ticularly useful, as cystic lesions are high in T2 signal and bright ductal adenocarcinoma. Extrapancreatic manifestations may
on these sequences, and communication with the main duct also be present within the spectrum of immunoglobulin G G4-
may be more evident (Acar et al, 2011; Campbell et al, 2015). related sclerosing disease, including bile duct, renal, and retro-
Although branch-duct IPMNs have a lower incidence of malig- peritoneal involvement (Kim et al, 2013).
nancy compared with main- or mixed-duct subtypes, recent
meta-analysis has shown mural nodules in branch-duct IPMNs References are available at expertconsult.com.
Chapter 19 Magnetic resonance imaging of the liver, biliary tract, and pancreas 377.e1
REFERENCES CT and Gd-EOB-DTPA-enhanced MRI, J Nucl Med 51(5):692–

699, 2010.
Acar M, Tatli S: Cystic tumors of the pancreas: a radiological perspec- Du S, et al: Diagnosis and treatment of hepatic angiomyolipoma, Hepa-
tive, Diagn Interv Radiol 17(2):143–149, 2011. tobiliary Surg Nutr 1(1):19–24, 2012.
Agarwal S, et al: Inflammatory hepatocellular adenomas can mimic Economopoulos N, et al: Bright-dark ring sign in MR imaging of
focal nodular hyperplasia on gadoxetic acid–enhanced MRI, AJR hepatic epithelioid hemangioendothelioma, J Magn Reson Imaging
203:W408–W414, 2014. 27(4):908–912, 2008.
Albiin N: MRI of Focal Liver Lesions, Curr Med Imaging Rev 8(2):107– Expert Panel on MR Safety, et al: ACR guidance document on MR
116, 2012. safe practices: 2013, J Magn Reson Imaging 37(3):501–530, 2013.
Aljiffry M, et al: Advances in diagnosis, treatment and palliation of Fattahi R, et al: Pancreatic diffusion-weighted imaging (DWI): com-
cholangiocarcinoma: 1990-2009, World J Gastroenterol 15(34):4240– parison between mass-forming focal pancreatitis (FP), pancreatic
4262, 2009. cancer (PC), and normal pancreas, J Magn Reson Imaging 29(2):350–
Anderson SW, et al: Benign hepatic tumors and iatrogenic pseudotu- 356, 2009.
mors, Radiographics 29:211–229, 2009. Freeny PC, Saunders MD: Moving beyond morphology: new insights
Arnaoutakis DJ, et al: Management of biliary cystic tumors: a multi- into the characterization and management of cystic pancreatic
institutional analysis of a rare liver tumor, Ann Surg 261(2):361–367, lesions, Radiology 272(2):344–362, 2014.
2015. Ganeshan D, et al: Imaging features of fibrolamellar hepatocellular
Balabaud C, et al: Focal nodular hyperplasia and hepatocellular carcinoma, AJR Am J Roentgenol 202(3):544–552, 2014.
adenoma around the world viewed through the scope of the immu- Goshima S, et al: Hepatic hemangioma and metastasis: differentiation
nopathological classification, Int J Hepat Article ID 268625, 2013. with gadoxetate disodium-enhanced 3-T MRI, AJR Am J Roentgenol
Basaran C, et al: Fat-containing lesions of the liver: cross-sectional 195(4):941–946, 2010.
imaging findings with emphasis on MRI, AJR Am J Roentgenol Grazioli L, et al: Hepatocellular adenoma and focal nodular hyperpla-
184:1103–1110, 2005. sia: value of gadoxetic acid-enhanced MR imaging in differential
Bashir MR, et al: Hepatic MR imaging for in vivo differentiation of diagnosis, Radiology 262(2):520–529, 2012.
steatosis, iron deposition and combined storage disorder: single-ratio Grazioli L, et al: MR imaging of hepatocellular adenomas and differ-
in/opposed phase analysis vs. dual-ratio Dixon discrimination, Eur J ential diagnosis dilemma, Int J Hepatol 2013:374170, 2013.
Radiol 81(2):e101–e109, 2012. Grobner T: Gadolinium—a specific trigger for the development of
Bennett CL, et al: Gadolinium-induced nephrogenic systemic fibrosis: nephrogenic fibrosing dermopathy and nephrogenic systemic fibro-
the rise and fall of an iatrogenic disease, Clin Kidney J 5(1):82–88, sis? Nephrol Dial Transplant 21(4):1104–1108, 2006.
2012. Gupta RT, et al: Dynamic MR imaging of the biliary system using
Billington PD, et al: Diagnosis of a biliary cystadenoma demonstrating hepatocyte-specific contrast agents, AJR Am J Roentgenol 195(2):405–
communication with the biliary system by MRI using a hepatocyte- 413, 2010.
specific contrast agent, Br J Radiol 85(1010):e35–e36, 2012. Hall WA, et al: Tumor size on abdominal MRI versus pathologic speci-
Bloch F, et al: The nuclear induction experiment, Physical Rev 70:474– men in resected pancreatic adenocarcinoma: implications for radia-
485, 1946. tion treatment planning, Int J Radiat Oncol Biol Phys 86(1):102–107,
Boll DT, Merkle EM: Diffuse liver disease: strategies for hepatic CT 2013.
and MR imaging, Radiographics 29(6):1591–1614, 2009. Hekimoglu K, et al: MRCP vs. ERCP in the evaluation of biliary
Boraschi P, Donati F: Postoperative biliary adverse events following pathologies: review of current literature, J Dig Dis 9:162–169,
orthotopic liver transplantation: assessment with magnetic resonance 2008.
cholangiography, World J Gastroenterol 20(32):11080–11094, 2014. Hennedige TP, et al: Imaging of malignancies of the biliary tract—an
Burke C, et al: The role of hepatocyte-specific contrast agents in hepa- update, Cancer Imaging 14:14, 2014.
tobiliary magnetic resonance imaging, Semin Ultrasound CT MR Heyn C, et al: MRI of the pancreas: problem solving tool, J Magn Reson
34(1):44–53, 2013. Imaging 36(5):1037–1051, 2012.
Butte JM, et al: Liver angiomyolipomas: a clinical, radiologic, and Huang H, et al: Hepatic angiosarcoma with multiple metastases in a
pathologic analysis of 22 patients from a single center, Surgery young man, Quant Imaging Med Surg 4(4):291–293, 2014.
150(3):557–567, 2011. Idee J, et al: The role of gadolinium chelates in the mechanism of
Cai PQ, et al: Hepatic angiomyolipoma: CT and MR imaging findings nephrogenic systemic fibrosis: a critical update, Crit Rev Toxicol
with clinical-pathologic comparison, Abdom Imaging 38(3):482–489, 44(10):895–913, 2014.
2013. Johnson C, et al: Detection of gallstones in acute pancreatitis: when
Campbell NM, et al: Imaging patterns of intraductal papillary muci- and how? Pancreatology 10(1):27–32, 2010.
nous neoplasms of the pancreas: an illustrated discussion of the Kang Y, et al: Intrahepatic mass-forming cholangiocarcinoma: enhance-
International Consensus Guidelines for the Management of IPMN, ment patterns on gadoxetic acid-enhanced MR images, Radiology
Abdom Imaging 40(3):663–677, 2015. 264(3):751–760, 2012.
Chung YE, et al: Varying appearances of cholangiocarcinoma: radiologic- Karam AR, et al: Focal nodular hyperplasia: central scar enhancement
pathologic correlation, Radiographics 29(3):683–700, 2009. pattern using gadoxetate disodium, J Magn Reson Imaging 32(2):341–
Coenegrachts K, et al: MRI findings in primary non-Hodgkin’s lym- 344, 2010.
phoma of the liver, JBR-BTR 88(1):17–19, 2005. Kartalis N, et al: Diffusion-weighted magnetic resonance imaging of
Cruite I, et al: Gadoxetate disodium-enhanced MRI of the liver: part pancreas tumours, Eur Radiol 19(8):1981–1990, 2009.
2, protocol optimization and lesion appearance in the cirrhotic liver, Kim T, et al: Liver masses with central or eccentric scar, Semin Ultra-
AJR Am J Roentgenol 195(1):29–41, 2010. sound CT MR 30:418–425, 2009.
Dai M, et al: Treatment of T3 gallbladder cancer, J Gastrointest Surg Kim SA, et al: Intrahepatic mass-forming cholangiocarcinomas:
13:2040–2042, 2009. enhancement patterns at multiphasic CT, with special emphasis on
Damadian R: Tumor detection by nuclear magnetic resonance, Science arterial enhancement pattern—correlation with clinicopathologic
171(3976):1151–1153, 1971. findings, Radiology 260(1):148–157, 2011.
Davenport MS, et al: Repeatability of diagnostic features and scoring Kim JH, et al: Atypical manifestations of IgG4-related sclerosing
systems for hepatocellular carcinoma by using MR imaging, Radiol- disease in the abdomen: imaging findings and pathologic correla-
ogy 272(1):132–142, 2014. tions, AJR Am J Roentgenol 200(1):102–112, 2013.
Do RK, et al: Variable MR imaging appearances of focal nodular Kim KW, et al: Imaging features to distinguish malignant and benign
hyperplasia in pediatric cancer patients, Pediatr Radiol 41(3):335– branch-duct type intraductal papillary mucinous neoplasms of the
340, 2011. pancreas: a meta-analysis, Ann Surg 259(1):72–81, 2014.
Do RK, et al: CT and MRI of primary and metastatic fibrolamellar Kucera JN, et al: Cystic lesions of the pancreas: radiologic-
carcinoma: a case series of 37 patients, Br J Radiol 87(1040):20140024, endosonographic correlation, Radiographics 32(7):E283–E301, 2012.
2014. Lafaro KJ, et al: Gd-EOB-DTPA-enhanced MRI for detection of liver
Donati OF, et al: Value of retrospective fusion of PET and MR images metastases from colorectal cancer: a surgeon’s perspective! Int J
in detection of hepatic metastases: comparison with 18F-FDG PET/ Hepatol 2013:572307, 2013.
Lauenstein TC, et al: Pancreatic adenocarcinoma tumor grade deter- Ridge CA, et al: Sclerosed hemangioma of the liver: concordance of
mination using contrast-enhanced magnetic resonance imaging, MRI features with histologic characteristics, J Magn Reson Imaging
Pancreas 39(1):71–75, 2010. 39(4):812–818, 2014.
Lauterbur PC: Image formation by induced local interactions: exam- Rizzi EB, et al: Non-Hodgkin’s lymphoma of the liver in patients with
ples employing nuclear magnetic resonance, Nature 242:190–191, AIDS: sonographic, CT, and MRI findings, J Clin Ultrasound
1973. 29:125–129, 2001.
Legrand L, et al: Pancreatic adenocarcinoma: MRI conspicuity and Rocha FG, et al: Intraductal papillary neoplasm of the bile duct: a
pathologic correlations, Abdom Imaging 40(1):85–94, 2015. biliary equivalent to intraductal papillary mucinous neoplasm of the
Lee HK, et al: Imaging features of adult choledochal cysts: a pictorial pancreas? Hepatology 56(4):1352–1360, 2012.
review, Korean J Radiol 10(1):71–80, 2009. Samaraee AI, et al: Preoperative diagnosis of choledocholithiasis: the
Lee SW, et al: Clinical usefulness of 18F-FDG PET-CT for patients role of MRCP, Br J Hosp Med 70(6):339–343, 2009.
with gallbladder cancer and cholangiocarcinoma, J Gastroenterol Sandrasegaran K, et al: State-of-the-art pancreatic MRI, AJR Am J
45(5):560–566, 2010. Roentgenol 195(1):42–53, 2010.
Lee JM, et al: Diagnosis of hepatocellular carcinoma: newer radiologi- Scalori A, et al: Oral contraceptives and the risk of focal nodular hyper-
cal tools, Semin Oncol 39(4):399–409, 2012. plasia of the liver: a case-control study, Am J Obstet Gynecol
Lin J, et al: CT and MRI diagnosis of hepatic epithelioid hemangioen- 186(2):195–197, 2002.
dothelioma, HBPD INT 9(2):154–158, 2010. Seale MK, et al: Hepatobiliary-specific MR contrast agents: role in
Low G, et al: Multimodality imaging of neoplastic and nonneo- imaging the liver and biliary tree, Radiographics 29(6):1725–1748,
plastic solid lesions of the pancreas, Radiographics 31(4):993–1015, 2009.
2011. Seo HJ, et al: Gadoxetate disodium-enhanced magnetic resonance
Maegerlein C, et al: Detection of liver metastases in patients with imaging versus contrast-enhanced 18F-fluorodeoxyglucose positron
adenocarcinomas of the gastrointestinal tract: comparison of F-FDG emission tomography/computed tomography for the detection of
PET/CT and MR imaging, Abdom Imaging 40(5):1213–1222, colorectal liver metastases, Invest Radiol 46(9):548–555, 2011.
2015. Siddiqi AJ, Miller F: Chronic pancreatitis: ultrasound, computed
Mandelia A, et al: The value of magnetic resonance cholangio- tomography, and magnetic resonance imaging features, Semin Ultra-
pancreatography (MRCP) in the detection of choledocholithiasis, sound CT MR 28(5):384–394, 2007.
J Clin Diagn Res 7(9):1941–1945, 2013. Sikora SS, et al: Surgical strategies in patients with gallbladder cancer:
Marrone G, et al: Biliary cystadenoma with bile duct communication nihilism to optimism, J Surg Oncol 93(8):670–681, 2006.
depicted on liver-specific contrast agent-enhanced MRI in a child, Silva AC, et al: MR imaging of hypervascular liver masses: a review of
Pediatr Radiol 41(1):121–124, 2011. current techniques, Radiographics 29(2):385–402, 2009.
Mavrogeni S, et al: The reproducibility of cardiac and liver T2* mea- Singh A, et al: Diagnostic accuracy of MRCP as compared to
surement in thalassemia major using two different software packages, ultrasound/CT in patients with obstructive jaundice, J Clin Diagn
Int J Cardiovasc Imaging 29(7):1511–1516, 2013. Res 8(3):103–107, 2014.
Meyers AB, et al: Characterization of pediatric liver lesions with gadox- Springer F, et al: Liver fat content determined by magnetic resonance
etate disodium, Pediatr Radiol 41(9):1183–1197, 2011. imaging and spectroscopy, World J Gastroenterol 16(13):1560–1566,
Miller FH, et al: Utility of diffusion-weighted MRI in distinguishing 2010.
benign and malignant hepatic lesions, J Magn Reson Imaging 32(1): Sudour H, et al: Focal nodular hyperplasia of the liver following hema-
138–147, 2010. topoietic SCT, Bone Mar Trans 43:127–132, 2009.
Mohajer K, et al: Characterization of hepatic adenoma and focal Tamm EP, et al: Imaging of pancreatic adenocarcinoma: update on
nodular hyperplasia with gadoxetic acid, J Magn Reson Imaging staging/resectability, Radiol Clin North Am 50(3):407–428, 2012.
36(3):686–696, 2012. Tan CH, et al: APASL and AASLD consensus guidelines on imaging
Okamoto D, et al: Enhancement pattern analysis of hypervascular diagnosis of hepatocellular carcinoma: a review, Int J Hepatol
hepatocellular carcinoma on dynamic MR imaging with histopatho- 2011:519783, 2011.
logical correlation: validity of portal phase imaging for predicting Tan CH, Lim KS: MRI of gallbladder cancer, Diagn Interv Radiol
tumor grade, Eur J Radiol 81(6):1116–1121, 2012. 19(4):312–319, 2013.
Palmucci S, et al: Magnetic resonance cholangiopancreatography and Tanaka M, et al: International consensus guidelines 2012 for the man-
contrast-enhanced magnetic resonance cholangiopancreatography agement of IPMN and MCN of the pancreas, Pancreatology 12(3):
versus endoscopic ultrasonography in the diagnosis of extrahepatic 183–197, 2012.
biliary pathology, Radiol Med 115(5):732–746, 2010a. Taouli B, Koh DM: Diffusion weighted MR imaging of the liver, Radi-
Palmucci S, et al: Evaluation of the biliary and pancreatic system with ology 254(1):47–66, 2010.
2D SSFSE, breathhold 3D FRFSE and respiratory-triggered 3D Todani T, et al: Congenital bile duct cysts: classification, operative
FRFSE sequences, Radiol Med 115(3):467–482, 2010b. procedures, and review of thirty-seven cases including cancer arising
Peporte AR, et al: Imaging features of intrahepatic cholangiocarcinoma from choledochal cyst, Am J Surg 134:263–269, 1977.
in Gd-EOB-DTPA-enhanced MRI, Eur J Radiol 82(3):e101–e106, Vaishali MD, et al: Magnetic resonance cholangiopancreatography in
2013. obstructive jaundice, J Clin Gastroenterol 38:839–840, 2004.
Pickhardt PJ, et al: Primary hepatic angiosarcoma: multi-institutional Visser BC, et al: Congenital choledochal cysts in adults, Arch Surg
comprehensive cancer centre review of multiphasic CT and MR 139(8):855–860, 2004.
imaging in 35 patients, Eur Radiol 25(2):315–322, 2015. Vlachou PA, et al: IgG4-related sclerosing disease: autoimmune pan-
Prasanna PM, et al: Best cases from the AFIP: giant cavernous hem- creatitis and extrapancreatic manifestations, Radiographics 31(5):
angioma, Radiographics 30(4):1139–1144, 2010. 1379–1402, 2011.
Purcell EM, et al: V. Resonance absorption by nuclear magnetic Wang C, et al: Intrahepatic biliary cystadenoma and cystadenocarci-
moments in a solid, Physical Rev 69:37, 1946. noma: an experience of 30 cases, Dig Liver Dis 44(5):426–431, 2012.
Qian LJ, et al: Spectrum of multilocular cystic hepatic lesions: CT and Wermke M, et al: MRI-based liver iron content predicts for nonrelapse
MR imaging findings with pathologic correlation, Radiographics mortality in MDS and AML patients undergoing allogeneic stem cell
33(5):1419–1433, 2013. transplantation, Clin Cancer Res 18(23):6460–6468, 2012.
Queiroz-Andrade M, et al: MR imaging findings of iron overload, Witjes CD, et al: Histological differentiation grade and microvascular
Radiographics 29(6):1575–1589, 2009. invasion of hepatocellular carcinoma predicted by dynamic contrast-
Rhee H, et al: Gadoxetic acid-enhanced MRI findings of early hepato- enhanced MRI, J Magn Reson Imaging 36(3):641–647, 2012.
cellular carcinoma as defined by new histologic criteria, J Magn Zhong L, et al: Magnetic resonance cholangiopancreatography, Chin J
Reson Imaging 35(2):393–398, 2012. Dig Dis 5:139–148, 2004.
CHAPTER 20
Direct cholangiography: approaches, techniques, and
current role
Robert H. Siegelbaum and Robin B. Mendelsohn
DIRECT CHOLANGIOGRAPHY OVERVIEW the gallbladder. The first report of PTC was by Huard and
Do-Xuan-Hop in 1937, who performed cholangiography with
Direct cholangiography, the introduction of contrast medium the suboptimal contrast agent Lipiodol. The use of transhe-
into the biliary system, can be performed under fluoroscopic patic cholangiography as a diagnostic tool gained popularity
guidance percutaneously, endoscopically, or intraoperatively 15 years later, after Carter and Saypol’s (1952) discussion of
via surgically placed catheters. The nonoperative techniques are PTC with the use of a water-soluble contrast agent. In the
percutaneous transhepatic cholangiography (PTC) and endo- ensuing years, many investigators (Flemma & Shingleton,
scopic retrograde cholangiopancreatography (ERCP). Cur- 1966; Glenn et al, 1962; Mujahed & Evans, 1966; Seldinger,
rently, noninvasive magnetic resonance imaging (MRI) and 1966; Shaldon et al, 1962; Weicher, 1964) described a variety
contrast-enhanced computed tomography (CECT) have virtu- of different methodologic details, including the use of sheathed
ally eliminated the indications and need for direct cholangiog- or unsheathed needles of various sizes and different puncture
raphy (Bakhal et al, 2009; Miller et al, 2007; Stroszczynski & sites with varying directions of puncture. The procedure
Hünerbein, 2005) (see Chapters 18 and 19). Today, ERCP and remained associated with a significant risk of bile peritonitis,
PTC are typically performed only as part of a planned inter- especially in obstructed biliary systems, and less frequently,
ventional procedure, such as biliary drainage, stent placement, hemorrhage. The use of fine needle technique was developed
stone extraction, or stricture dilation (see Chapters 29 and 30). at Chiba University and was first presented by Ohto and
Imaging evaluation of the biliary tree with magnetic reso- Tsuchiya (1969), and later by Tsuchiya (1969). Numerous
nance cholangiopancreatography (MRCP) or CECT is nonin- additional reports have also described decreased complications
vasive, does not require sedation, and is generally regarded as with fine needle transhepatic access, and the fine needle tech-
safe. Imaging protocols with multiplanar reconstruction allow nique has been generally accepted as the standard (Benjamin
a large field of view that enables visualization of the entire et al, 1978; Ferrucci et al, 1976; Fraser et al, 1978; Gold &
biliary tree and pancreatic duct (Fig. 20.1). Reconstructed Price, 1980; Harbin et al, 1980; Hinde et al, 1977; Jain et al,
three-dimensional data sets can be displayed in an appearance 1977; Pereiras et al, 1977).
similar to a diagnostic cholangiogram, with the added benefit Transhepatic cholangiography has essentially been sup-
of being able to rotate the images to further delineate the planted by noninvasive imaging techniques for the evaluation
anatomy of the bile duct and pancreatic duct. In comparison, of the biliary tree. In 1985, Kadir reported that less than 5%
injection of contrast dye under fluoroscopic guidance is limited of patients referred for evaluation of biliary disease required
to visualization of structures in direct continuity with the opaci- concomitant drainage procedures. With continued advance-
fied, nonisolated segments of the biliary tree (Fig. 20.2). While ment of cross-sectional imaging techniques, in particular
performing direct cholangiography with PTC, simultaneous MRCP (see Chapter 19) (Park et al, 2004; Romagnuolo et al,
cross-sectional imaging may be required to ensure that the 2003; Simone et al, 2004; Vaishali et al, 2004; Zhong et al,
entire biliary tree has been opacified, even with multiple per- 2003), percutaneous fluoroscopic imaging of the biliary tree is
cutaneous puncture sites and injections. Cross-sectional typically only performed as part of a planned radiologic inter-
imaging of the liver can be performed in a fluoroscopy room ventional procedure.
with rotational angiography or by direct CT imaging if there is Imaging evaluation of the biliary tree with MRCP or CECT
an in-line CT scanner. In addition, MRCP and CECT typically is noninvasive, does not require sedation, and is generally
provide superior diagnostic accuracy when compared with PTC regarded as safe (see Chapters 18 and 19). With direct puncture
or ERCP, allowing visualization of the bile duct wall, structures and contrast injection in the biliary tree, only bile ducts in
contiguous to the biliary tree (such as cysts and neoplasms), continuity with the punctured duct are visualized. Multiple
and structures adjacent to the liver. percutaneous punctures may be required to visualize the entire
biliary tree in the presence of bile duct isolation (see Fig. 20.2).
Even with multiple transhepatic bile duct punctures, CT
PERCUTANEOUS TRANSHEPATIC imaging performed at the time of the procedure can be helpful
CHOLANGIOGRAPHY to ensure that the biliary tree has been completely assessed.
Alternatively, MRI and CECT can image the entire biliary
History system and can demonstrate the presence and severity of bile
The first report of fluoroscopic imaging of the biliary tree was duct isolation (see Fig. 20.1). Furthermore, by also visualizing
presented by Burckhardt and Müller in 1921, who performed the bile duct wall, liver, and surrounding structures, additional
cholecystocholangiography through percutaneous puncture of anatomic information is provided with MRI or CECT.
378
Chapter 20 Direct cholangiography: approaches, techniques, and current role 379
Coagulation parameters and platelet count should be

checked prior to the procedure and corrected if necessary.
Informed consent is obtained in accordance with institutional
policy, including a discussion of the risks and benefits of the
procedure. Patients generally receive conscious sedation con-
sisting of an anxiolytic, such as midazolam, and an opiate, such
as fentanyl. All patients should receive broad-spectrum intra-
venous antibiotic coverage (according to institutional prefer-
ence), which should be administered approximately 1 hour
before the procedure to ensure appropriate blood levels at the
time of procedure.
Procedure
Although not always present in modern fluoroscopy suites,
PTC is ideally performed on a tilting fluoroscopic table. As the
specific gravity of contrast material is greater than that of bile,
less contrast dye is typically required to fully delineate the
biliary tree when the table is tilted.
FIGURE 20.1. Magnetic resonance cholangiopancreatography. Three- Right-Sided Puncture
dimensional maximum intensity projection image showing marked dilata-
tion of the intrahepatic and extrahepatic biliary tree, as well as the After the patient has been prepared and draped in a sterile
pancreatic duct, in a patient with a head of pancreas mass. fashion, a puncture site is selected in the right midaxillary line,
typically one or two interspaces below the costophrenic angle.
Puncture sites are generally kept below the ninth intercostal
space to avoid inadvertent entry into the pleural space. One
percent lidocaine is infused subcutaneously, and a small derma-
totomy is made with a No. 11 scalpel. A 21- or 22-gauge, 15- to
20-cm Chiba-style needle is advanced under fluoroscopic guid-
ance superior to the closest rib to the target puncture site, to
avoid laceration of an intercostal vessel. Real-time ultrasound
guidance can be helpful for the initial puncture in the presence
of bile duct dilatation (Covey & Brown, 2008). The puncture
site and direction are chosen based on evaluation of the patient’s
cross-sectional imaging studies (Fig. 20.3A and B).
Care should be taken to avoid puncture of the gallbladder
or extrahepatic bile duct. A small amount of water-soluble
contrast agent is injected while slowly withdrawing the needle
until a bile duct is identified. Injection of contrast agent into a
bile duct has the appearance of oil being dropped in water.
Inadvertent opacification of a vascular structure is recognized
by the rapid clearance of the contrast agent. If a bile duct is not
entered during withdrawal of the needle, the needle is reintro-
duced in a slightly different direction. It is generally recom-
mended not to withdraw the needle completely from the liver,
to minimize the number of punctures through the liver capsule.
When opacification of the biliary tree is challenging, particu-
larly in the case of a nondilated biliary tree, the same puncture
FIGURE 20.2. Hilar cholangiocarcinoma involving first-order right and site may be used multiple times prior to selecting a new percu-
left hepatic ducts. Separate needle punctures were required to fill the taneous access site. If successful puncture of a bile duct is
right and left ducts. unlikely from that site after multiple attempts, a new site should
be chosen. In patients with biliary obstruction, gentle injection
of 10 mL of iodinated contrast is typically sufficient for bile
Preprocedural Preparation duct opacification. A larger volume of contrast material is typi-
Ideally, all patients should undergo cross-sectional imaging cally required in nonobstructed systems. When the bile ducts
(CECT or MRI) prior to PTC. This imaging is particularly are visualized, fluoroscopic images should be acquired and
important in patients who have undergone prior liver resection. stored in multiple projections to identify specific portions of the
MRI or CECT are the imaging modalities of choice due to their biliary tree and to completely delineate abnormal findings, if
diagnostic accuracy in depicting the level of obstruction, posi- present.
tion of the liver, portal vein patency, the relationship of the liver
and bile ducts to other structures, and presence of tumors or Left-Sided Puncture
lobar atrophy, all of which could greatly facilitate successful There is considerable variation in size and anatomic position
puncture. of the left lateral segments of the liver, and careful review of
LOGIQ
preprocedural cross-sectional imaging is recommended to
E9 select an optimal puncture site. Although left-sided punctures
can be performed through the right liver via an anterior axillary
line approach to the segment IV bile ducts, the generally pre-
21 ferred method is a subxiphoid approach to a bile duct in
segment II or segment III. As in punctures of right sided bile
ducts, percutaneous access to the biliary tree can be simplified
by using real-time ultrasound guidance in the presence of
dilated bile ducts (Covey & Brown, 2008).
–21 Success Rate and Accuracy

cm/s Percutaneous opacification of the biliary tree is successful in
95% to 100% of patients with biliary obstruction (Harbin et al,
A 1980; Jain et al, 1977; Mueller et al, 1981). A success rate of
60% to 95% is reported for nondilated biliary systems. The
LOGIQ likelihood of success in a nondilated system is increased by the
E9 number of needle passes performed (Jaques et al, 1980). No
current radiologic descriptions of obstructed bile ducts are
pathognomonic for differentiation of benign and malignant
disease (Hadjis et al, 1985; Wetter et al, 1991). Bile cytology
and review of cross-sectional imaging can be helpful in conjunc-
tion with fluoroscopic images in diagnosing the cause of the
obstruction.
Pitfalls in Interpretation
Lack of Opacification
Failure to inject adequate volume of contrast agent can result
in false interpretation of the level of obstruction. This pitfall
can occur in the setting of complete obstruction and can be
B
recognized by the presence of a hazy margin at the level of the
FIGURE 20.3. Real-time ultrasound guidance can be helpful for the apparent (false) obstruction (Fig. 20.4) (Kittredge & Baer,
initial puncture in the presence of bile duct dilatation. A, Color Doppler 1975). In high bile duct obstruction, especially when associated
image of the left liver revealing a dilated segment III bile duct. B, Static with variant anatomy, isolated segments of the biliary tree can
grayscale ultrasound image taken during puncture of the dilated bile duct be visualized only by direct puncture. If the wrong bile duct is
with an echogenic 21-gauge needle.
selected, or if puncture of an additional bile duct is needed but
A B
FIGURE 20.4. Ampullary carcinoma. A, With the patient supine, contrast pools proximally, giving a false impression of a high bile duct obstruction.
The spurious nature of the level is suggested by the hazy inferior margin to the contrast column. B, With the patient sitting semierect, the contrast
pools at the true point of obstruction, which is sharply defined.
not performed, the diagnosis of bile duct injuries and bile leaks that ERCP, MRCP, and endoscopic ultrasound have compa-
can be missed. rable sensitivity and specificity in the diagnosis of common bile
duct (CBD) stones. With newer diagnostic imaging technolo-
Ductal Dilatation gies, however, ERCP has evolved into a predominantly thera-
The absence of bile duct dilatation does not exclude the pres- peutic procedure. They stated that avoidance of unnecessary
ence of clinically significant biliary obstruction. Disorders such ERCPs is the best way to reduce the number of complications
as sclerosing cholangitis, acquired immunodeficiency syn- and that endoscopists performing ERCPs should have appro-
drome, and chemotherapy-induced biliary sclerosis can present priate training and expertise. In 2005, the American Society of
with bile duct fibrosis, impeding the ability of the bile ducts to Gastrointestinal Endoscopy published guidelines, stating that
become dilated. Conversely, bile duct dilatation does not based on expert opinion, ERCP is primarily a therapeutic pro-
always imply the presence of an obstructed biliary system. For cedure for the management of pancreaticobiliary disorders
example, dilatation that may be seen in patients with Caroli (Adler et al, 2005). Therefore ERCP is rarely performed
disease, or choledochal cysts, can have the radiographic appear- without a therapeutic component.
ance of bile duct dilation without the presence of obstruction.
Technique
Complications Most ERCPs are performed as outpatient procedures in a hos-
Serious complications of PTC are rare and occur in approxi- pital setting, although they can be performed in ambulatory
mately 3% of patients (Harbin et al, 1980). The most common centers as well. Given that ERCP is a complex procedure,
major complications are bile leak (1% to 2%), sepsis (2% to requiring special equipment and training, the risks and benefits
3%), and hemorrhage (0.2% to 0.4%). Other rare complica- of the procedure must be heavily considered prior to proceed-
tions include pneumothorax, biliothorax, injury to the colon ing. Patients should be alert, oriented, and able to give informed
related to the puncture, and abscess formation. Puncture below consent. In the rare cases where patients are unable to give
the ninth intercostal space will clearly decrease the incidence informed consent, the next of kin can provide consent. Patient
of pleural and lung complications. The risk of infectious com- age and clinical picture are important. Elderly patients can
plications such as sepsis and abscess formation can be mitigated safely undergo ERCP; they have the same risks of bleeding and
with proper antibiotic coverage. Care should be taken to not perforation as younger patients and actually have a lower risk
overdistend the biliary tree, as opacification and incomplete of pancreatitis (Badalov et al, 2009).
drainage of the biliary tree can be a source of cholangitis (Covey Once consent is obtained, the patient is brought into a room
& Brown, 2008), particularly in the presence of bile duct with a C-arm for availability of fluoroscopy during the proce-
isolation. dure. Intravenous sedation is given under monitored control.
In the past, drug combinations of narcotics, such as meperidine
and droperidol, and benzodiazepines, such as midazolam or
ENDOSCOPIC RETROGRADE diazepam, were used. More recently, this has been replaced by
CHOLANGIOPANCREATOGRAPHY propofol, a short-acting sedative and amnestic with a rapid
recovery profile. Studies have shown that propofol is more
History effective than sedation with midazolam, is safe, and is associ-
Endoscopic retrograde cholangiopancreatography (ERCP) (see ated with a faster postprocedure recovery (Fanti et al, 2004;
Chapter 29) was first described in 1968 (McCune et al, 1968) Wehrman et al, 1999). In some centers, general anesthesia may
and rapidly became accepted as an important diagnostic modal- be used if the endoscopic procedure is expected to be difficult,
ity for patients with hepatobiliary and pancreatic diseases the patient has significant comorbid medical conditions, or if
(Cotton, 1977; Cotton et al, 1972; Freeney, 1988; Oi et al, there are any signs of functional or mechanical intestinal
1970). Over the last 4 decades, multiple advances in technology obstruction. Oxygen is administered by nasal cannula to avoid
and training have enhanced and expanded the scope of ERCP. hypoxemia, which has been described in 40% of patients under-
The improvement of the side-viewing duodenoscope with an going ERCP (Woods et al, 1989). The patient’s electrocardio-
elevator helped facilitate cannulation of the papilla of Vater gram, blood pressure, oxygen saturation, and overall condition
(Kasugai et al, 1971; Ogoshi et al, 1970; Takagi et al, 1970). are continually monitored throughout the procedure by a dedi-
In addition, the development of therapeutic applications cated nurse. Antibiotics are not given routinely for diagnostic
through ERCP, including sphincterotomy (Classen et al, 1975; procedures. All endoscopic equipment used for this procedure,
Cotton et al, 1976; Kawai et al 1974) and stenting (Laurence including the endoscopes, is either chemically disinfected or gas
et al, 1980; Soehendra et al, 1980), has evolved ERCP from a sterilized.
diagnostic into a therapeutic procedure. ERCP is considered to The patient is placed in a semiprone position with special
be an advanced procedure, requiring skills more difficult to positioning of the arms, to help optimize access to the ampulla
learn than routine endoscopic procedures, and is offered as an of Vater. A side-viewing duodenoscope is used to afford excel-
additional year of training beyond the standard gastroenterol- lent visualization of the ampulla of Vater. In patients with a
ogy fellowship. Billroth II type gastrojejunostomy, a standard forward-viewing
upper endoscope or pediatric colonoscope may be needed to
Indications successfully approach the ampulla. An initial endoscopic evalu-
With the exception of a few scenarios, diagnostic ERCP has ation of the stomach and duodenum is performed before can-
largely been replaced by noninvasive imaging techniques. In nulation of the ampulla. The ampulla is usually located in the
2002, the National Institutes of Health sponsored a consensus second portion of the duodenum but, in rare instances, may be
conference on ERCP and issued a statement proposing the found more proximal or distal. It is usually easily identified,
indications for ERCP (Cohen et al, 2002). They concluded although, in some cases, it may be distorted due to malignancy
A B
C D
FIGURE 20.5. Pancreatic duct stent placement. A, Wire placement into the pancreatic duct. B, Contrast injection to confirm position in pancreatic
duct. C, Placement of a 5-Fr × 5-cm pancreatic duct stent with a full external pigtail and a single internal flap. D, Endoscopic view of 5-Fr pancreatic
duct stent.
or edema from pancreatitis, hidden behind a fold, or within a motility can be controlled by bolus injections of 0.25 to 1 mg
diverticulum. The orifice of the CBD is usually located on the of intravenous glucagon. If the pancreatic duct is inadvertently
left upper corner of the ampulla. In most patients, the CBD cannulated, placement of a pancreatic duct stent may help can-
and main pancreatic duct share a common channel. In a minor- nulation by both blocking the pancreatic duct orifice as well as
ity of patients, the orifices are separate. The minor papilla is by providing more information to the endoscopist about the
located about 1 to 2 cm above the major papilla. The CBD is angle of the bile duct, especially in cases of distorted anatomy
selectively cannulated with either a cannula or sphincterotome. (Fig. 20.5) (Goldberg et al, 2005). Another technique involves
Studies have shown that access to the biliary tree is easier and placing a guidewire into the pancreatic duct, which adds more
faster with a sphincterotome compared with a cannula (Kara- information about the optimal angle to approach cannulation
manolis et al, 2005; Laasch et al, 2003; Rossos et al, 1993). (Fig. 20.6). The studies are limited, but this technique
More recently, some endoscopists have been using guidewire- does not appear to increase the risk of post-ERCP pancreatitis
assisted cannulation. It remains controversial whether insertion (Draganov et al, 2005; Saad, 2004). Other more invasive
of a guidewire, as opposed to the more conventional technique maneuvers, including precut sphincterotomy with or without
of contrast injection, should be the preferred technique to pancreatic duct stent placement, may also improve success rate
access the bile ducts. A guidewire does not produce the hydro- but is associated with an increased risk of complications, includ-
static pressure associated with contrast injection and decreases ing bleeding, perforation, and pancreatitis, even in experienced
the risk of trauma to the pancreatic duct, thereby theoretically hands (Harewood et al, 2002).
decreasing the risk of post-ERCP pancreatitis. The data, Duodenal diverticula are common and almost always are
however, are mixed. A recent meta-analysis looking at 12 ran- located near the papilla in the descending duodenum. Although
domized trials with 3450 patients concluded that wire-guided usually asymptomatic, diverticula have been shown to be asso-
technique had a higher cannulation success rate (84% vs. 77%) ciated with choledocholithiasis. Periampullary diverticula
and a lower risk of post-ERCP pancreatitis (3.5% vs. 6.7%) may make cannulation more difficult, but the data are mixed
(Tse et al, 2013). However, a prospective trial with 1249 (Rajnakova et al, 2003; Tham & Kelly, 2004).
patients showed no significant difference in post-ERCP pancre- In patients with surgically altered anatomy, ERCP may be
atitis in the guidewire group (5.2%) compared with the contrast quite challenging. In patients with a Billroth II gastrojejunos-
injection group (4.4%) (Mariani et al, 2012). tomy, success rates for bile duct cannulation are much lower
In the event of a difficult cannulation, there are techniques (Forbes et al, 1984). The papilla is found in the afferent limb,
to help facilitate access to the bile duct. Excess duodenal which may be difficult to traverse. In addition, the orientation
A B C
FIGURE 20.6. Guidewire placement in pancreatic duct to facilitate bile duct cannulation. A, Wire placement into pancreatic duct. B, This provided
more information to the endoscopist about the angle of the bile duct, and the cannulatome was adjusted under fluoroscopic guidance. C, Contrast
injection confirmed position in common bile duct.
of the ampulla is upside down compared with standard ERCP.

In patients with Roux-en-Y gastrectomies, it is often difficult
even to reach the papilla, given the long Roux limb. If the
papilla is identified, cannulation may still be extremely difficult
due to the location and position of the ampulla (Byron et al,
2002). Recently, double-balloon enteroscopy has been reported
to be successful in enabling access to the afferent limb (Sato
et al, 2005). This is mostly only offered in tertiary care centers.
ERCP after bariatric surgery is extremely difficult, and most
consider it too complex to attempt.
Once the bile duct is selectively cannulated, a contrast agent
is injected under fluoroscopic control, with subsequent radio-
graphic images obtained of the duct anatomy. Choice of con-
trast agent differs among endoscopists. Many choose to use half
strength contrast when looking for stones. Material for patho-
logic and cytologic evaluation can be obtained from either the
biliary or pancreatic duct system with a variety of dedicated
endoscopic biopsy forceps and cytology brushes. Pathologic
and cytologic material may be obtained from the ampulla of
Vater, duodenum, and stomach for diagnostic purposes during
the procedure as well.
Pancreatography
Imaging the pancreatic duct can be an important adjunct to
cholangiography during ERCP. Strictures, stones, and other FIGURE 20.7. Pancreas divisum. The cannula is in the major papillary
obstructing lesions can be identified. Most recommend guide- orifice, and contrast injection opacifies the proximal portion of the major
wire cannulation of the pancreatic duct over contrast injection pancreatic duct (Wirsung). Injection through the minor papillae allows
as this produces less hydrostatic pressure in the duct, possibly opacification of the duct of Santorini and the distal duct of Wirsung.
decreasing the risk of post-ERCP pancreatitis. In general, the
pancreatic duct is about 20 cm in length and variable in caliber.
With increasing age comes progressive atrophy and fibrosis of
the pancreas. The diameter of the main pancreatic duct also Cholangioscopy and Pancreatoscopy
increases with age, although one study found no difference in Cholangioscopy and pancreatoscopy involve using miniature
pancreatic duct length among patients younger than 40 years endoscopes through the channel of the duodenoscope, allowing
compared with older patients. Duct diameter throughout the direct visualization of the bile and pancreatic ducts, respec-
pancreas was significantly greater, however, in patients older tively. A new skill set is necessary to perform these procedures,
than 40 years (Anand et al, 1989). given that this technique uses two different endoscopes. Diag-
Anatomic variations, such as pancreas divisum (Fig. 20.7), nostic cholangioscopy may be used to evaluate indeterminate
also may be identified on a pancreatogram (see Chapter 2). This biliary strictures and filling defects. Similarly, diagnostic pan-
abnormality has been described in 7.5% of ERCP procedures creatoscopy may be used to evaluate pancreatic strictures and
and can be confirmed by cannulation of the main pancreatic intraductal papillary mucinous neoplasms. Studies have shown
duct through the orifice in the minor papilla (Bernard et al, that it is an accurate diagnostic tool for patients with pancreati-
1990). cobiliary disorders (Fukuda et al, 2005; Itoi et al, 2010; Yamao
et al, 2003). One prospective multicenter study of 87 patients et al, 1985), and a meta-analysis showed no significant differ-
reported that endoscopists were able to distinguish benign from ence in post-ERCP pancreatitis among different contrast media
malignant indeterminate biliary lesions 92.1% of the time with (George et al, 2004).
cholangioscopy with visualization alone (Osanai et al, 2013). Although it remains unclear whether these mechanisms
Complications of cholangiopancreatoscopy include bacteremia, work independently or in conjunction, recent data have helped
bleeding, and pancreatitis. One retrospective study reported to elucidate both patient- and procedure-related risk factors
that complications of cholangiopancreatoscopy are increased that are independently associated with post-ERCP pancreatitis.
compared with ERCP alone and are associated with a much These risk factors are additive (Freeman et al, 2001). Patient-
higher risk of cholangitis (Sethi et al, 2011). related factors include younger age, female sex, normal serum
bilirubin, recurrent pancreatitis, history of post-ERCP pancre-
Complications atitis, and sphincter of Oddi dysfunction. Procedure-related
Complications of ERCP include those associated with endo- factors include difficult cannulation, pancreatic duct injection,
scopic procedures in general as well as those specific to ERCP. precut sphincterotomy, pancreatic sphincterotomy, minor
Incidence rates of complications vary in the literature. A sys- papilla sphincterotomy, balloon sphincteroplasty, ampullec-
tematic survey of prospective studies reviewed 21 studies with tomy, and sphincter of Oddi manometry. One study showed
16,855 patients and reported a specific complication rate of that trainee participation was an independent risk factor (Cheng
6.9% and a mortality rate of 0.33% (Andriulli et al, 2007). The et al, 2006). Most studies, however, have not shown a correla-
experience of the endoscopist and case volume also has an tion between case volume and rates of pancreatitis (Freeman
impact on the complication rate. In a study conducted in et al, 2001; Loperfido et al, 1998; Testoni et al, 2010).
Austria, endoscopists performing more that 50 ERCP proce- Specific techniques and measures to decrease the risk of
dures per year were compared with those performing fewer than pancreatitis have been studied. The risk is reduced by minimiz-
50 per year. Those in the higher case-volume group had a ing the number of attempts of cannulation, avoiding pancreatic
significantly higher success rate (86.9% vs. 80.3%; P < .001) duct cannulation if not necessary, and by avoiding overdisten-
and a lower overall complication rate (10.2% vs. 13.6%; P = sion or “acinarization” of the pancreatic duct by minimizing
.007) (Kapral et al, 2008). the volume of contrast medium into the pancreatic duct.
Placement of a temporary pancreatic duct stent may also
Pancreatitis reduce the risk. One meta-analysis demonstrated that the use
The most common complication of ERCP is pancreatitis, with of pancreatic duct stents in high-risk patients decreased the rate
reported incidences ranging from 1% to 40%, but most fre- of post-ERCP pancreatitis by about two thirds (Singh et al,
quently reported around 3% to 5% (Andriulli et al, 2007; 2004). Their use is not routinely recommended but is reserved
Cheng et al, 2006; Christensen et al, 2004; Freeman et al for high-risk patients. Studies have shown a benefit of pancre-
2001; Loperfido et al, 1998) (see Chapter 54). The incidence atic duct stents in biliary sphincterotomy for sphincter of Oddi
of post-ERCP pancreatitis in children is quite low, about 2.5% dysfunction, precut sphincterotomy, balloon sphincteroplasty,
in one study (Iqbal et al, 2008). The consensus classification endoscopic ampullectomy, and difficult cannulation (Fazel
defined post-ERCP pancreatitis as the clinical picture of new et al, 2003; Freeman et al, 2004; Singh et al, 2004).
or worsened abdominal pain with amylase at least three times Many pharmacologic agents have been studied. A recent
normal at 24 hours after the procedure and requiring hospital- systematic review included 85 randomized clinical trials and 28
ization (Cotton et al, 1991). They further defined it as mild if meta-analyses evaluating pharmacologic prevention of post-
hospitalization is 2 to 3 days, moderate if hospitalization is 4 ERCP pancreatitis. They concluded that rectal nonsteroidal
to 10 days, and severe if hospitalization is more than 10 days, antiinflammatory drugs (NSAIDs) were beneficial, especially in
if there is a pseudocyst or hemorrhagic pancreatitis, or if an high-risk patients. Data on bolus-administered somatostatin,
intervention is required. Post-ERCP pancreatitis should be dis- sublingual nitroglycerin, and some protease inhibitors were
tinguished from transient asymptomatic hyperamylasemia, considered promising, but confirmatory studies are necessary
which occurs in 40% to 75% of cases and disappears within 1 (Kubiliun et al, 2015).
to 2 days (Classen & Demling, 1975; Okuno et al, 1985). Most NSAIDs inhibit prostaglandin synthesis, phospholipase A2
cases of post-ERCP pancreatitis are mild and usually resolve in activity, and neutrophil/endothelial cell attachment, which are
a few days with conservative measures of bowel rest and intra- all thought to play a major role in the pathogenesis of pancre-
venous fluids. The management is the same as for pancreatitis atitis, and thus may have a role in post-ERCP pancreatitis
from other causes. prevention. A meta-analysis that included 10 randomized con-
There have been multiple potential mechanisms proposed trolled trials with a total of 2269 patients concluded that
to explain the pathogenesis of post-ERCP pancreatitis. Mechan- NSAID use decreased the risk of post-ERCP pancreatitis (risk
ical injury to the pancreatic duct from manipulation of the ratio, 0.57; 95% confidence interval [CI], 0.38 to 0.86; P =
papilla, instrumentation of the pancreatic duct, or injection of .007) (Ding et al, 2012). However, these studies were extremely
the pancreatic duct likely plays a role (Johnson et al, 1997). heterogeneous in regard to type of NSAID as well as to route
Similarly, thermal injury from electrocautery leading to edema and timing of administration. A randomized, placebo-
and possible obstruction of the duct has been invoked (Ratani controlled, double-blind clinical trial of high-risk patients
et al, 1999). Hydrostatic pressure from contrast injection showed that patients who received rectal indomethacin imme-
leading to injury is also likely a component. The contrast itself diately after the procedure were less likely to develop post-
theoretically may cause a chemical or allergic injury. However, ERCP pancreatitis than the control group (9.2% vs. 16.9%;
the use of nonionic contrast medium of low osmolarity has P = .0005) and were less likely to develop moderate to severe
shown no advantage over the less expensive ionic contrast pancreatitis (4.4% vs. 8.8%; P = .03) (Elmunzer et al, 2012).
medium in preventing ERCP-related pancreatitis (Hannigan A meta-analysis that specifically looked at studies of rectal
indomethacin included four studies with 1470 patients showed (25%; P = .012) (Hao et al, 2009). In both studies, however,
that the rate of pancreatitis was significantly lower using indo- the consensus definition for pancreatitis was not used, which
methacin compared with placebo (odds ratio, 0.49; CI, 0.34 to may account for the high rates of pancreatitis in the control
.71; P = .0002) (Yaghoobi et al, 2013). A network meta-analysis groups. One recent randomized controlled trial of 300 patients
compared rectal indomethacin to pancreatic duct stenting and showed that the combination of rectal indomethacin and sub-
concluded that rectal indomethacin alone was superior to pan- lingual nitroglycerin was superior to rectal indomethacin alone
creatic duct stenting in post-ERCP prevention (odds ratio, at prevention of post-ERCP pancreatitis (6.7% vs. 15.3%; P =
0.48; 95% CI, 0.26 to 0.87) (Akbar et al, 2013). The European .016) (Sotoudehmanesh et al, 2014). More studies are needed
Society of Gastrointestinal Endoscopy guidelines recommend to further clarify the benefit of nitroglycerin in post-ERCP
routine prophylactic use of rectal NSAIDs immediately before pancreatitis.
or after ERCP in all patients without a contraindication Protease inhibitors, such as gabexate mesylate, nafamostat
(Dumonceau et al, 2014). mesylate, and ulinastatin, have been investigated, given that
Somatostatin and its analogue octreotide inhibit pancreatic activation of proteolytic enzymes likely contributes to the
secretions, decrease sphincter of Oddi pressure, modulate pathogenesis of pancreatitis. A meta-analysis of 18 studies with
cytokines, and lead to apoptosis of pancreatic acinar cells, and a total of 4966 patients showed a significant, yet small, risk
therefore may be protective against post-ERCP pancreatitis. reduction in post-ERCP pancreatitis with the protease inhibi-
They have been studied extensively with conflicting results. tors (Seta & Noguchi, 2011). They stated there was no solid
One meta-analysis included seven homogeneous high-quality evidence to support their use at this time. A recent pilot study
studies involving 3130 patients and concluded that somatosta- investigated whether aggressive periprocedural hydration
tin administered as a bolus was effective in prevention of post- reduced the risk of post-ERCP pancreatitis and found that
ERCP pancreatitis (Rudin et al, 2007). A more recent none of the patients in the aggressive hydration group devel-
meta-analysis looked at 17 studies with a total of 3818 patients oped pancreatitis compared with 17% of patients in the stan-
and found that somatostatin and high-dose octreotide pre- dard hydration group (Buxbaum et al, 2014). Larger studies
vented post-ERCP pancreatitis if given over 12 hours or in are needed to corroborate these findings. Many other agents
bolus form (Omata et al, 2010). Subsequent randomized trials have been investigated, including secretin, corticosteroids, allo-
have yielded mixed results. One double-blinded, placebo- purinol, and topical epinephrine, with conflicting results, and
controlled, randomized study involved 391 patients in three are not recommended at this time.
hospitals. Patients were randomized to receive 3 mg of soma-
tostatin in 500 mL normal saline infused for 12 hours, starting Infection
30 minutes before the ERCP or 500 mL normal saline infused A serious complication of ERCP is the development of post-
for 12 hours, starting 30 minutes before the ERCP. They procedure infectious complications, most commonly cholangi-
found a significantly lower risk of pancreatitis in the soma- tis and cholecystitis. In a systematic survey of 21 prospective
tostatin group (3.6% vs. 9.6% in the placebo group; P = .02) studies with 16,855 patients, the incidence of infectious com-
(Lee et al, 2008). Another study involved 133 patients who plications was 1.4% (Andriulli et al, 2007).
were randomized to a bolus of somatostatin infusion before Cholangitis most commonly occurs when there is failed or
ERCP, followed by continuous infusion for 12 hours, a bolus incomplete biliary drainage. The risk is increased in patients
of somatostatin before ERCP only, and placebo alone; no sig- with hilar obstruction and sclerosing cholangitis, given the
nificant differences were found among the three groups (Chan increased risk of incomplete drainage (Bangarulingam et al,
et al, 2008). A recent randomized trial of 510 patients ran- 2009; Rerknimitr et al, 2004). Other risk factors include jaun-
domized to an intravenous bolus of 250 µg of somatostatin dice, small endoscopy center, and delay in performing ERCP
prior to cannulation, followed by a 4-hour continuous infusion (Loperfido et al, 1998; Navaneethan et al, 2013). Treatment
of the drug at 250 µg/hr, or placebo with normal saline showed involves supportive care with antibiotics and decompression of
no significant difference in rates of post-ERCP pancreatitis the obstruction. Studies have shown that prophylactic antibiot-
(Concepcion et al, 2014). In addition, when pancreatitis has ics significantly reduce the frequency of procedure-related bac-
developed, the use of somatostatin does not improve its clini- teremia but have not shown a difference in rates of cholangitis
cal course (Phillip et al, 1985). Future research is necessary to (Sauter et al, 1990). Two meta-analyses failed to demonstrate
elucidate the role of somatostatin in prevention of post-ERCP a benefit of giving routine prophylactic antibiotics (Bai et al,
pancreatitis. 2009; Harris et al, 1999). Antibiotics added to the injected
Nitroglycerin acts as a smooth muscle relaxant and subse- radiographic contrast medium are also of no benefit (Jendrze-
quently may decrease sphincter of Oddi pressures, thereby jewski et al, 1980).
decreasing the risk of post-ERCP pancreatitis. Only three out It has also been found that endoscopic instruments used in
of seven randomized controlled studies showed that nitroglyc- ERCP can be the source of serious infections (Earnshaw et al,
erin was effective, but two of the three positive studies used 1985). Salmonella species are the most common organisms
sublingual nitroglycerin (Kubiliun et al, 2015). One compared transmitted by contaminated endoscopes (Axon, 1991), and for
2 mg sublingual nitroglycerin given 5 minutes prior to endos- this reason, it is imperative to ensure that the endoscopic equip-
copy with placebo in 186 patients and found a lower incidence ment used for ERCP is adequately disinfected or sterilized.
of pancreatitis in the nitroglycerin group (7/90 vs. 17/96; P <
.05) (Sudhindran et al, 2001). The second one enrolled 74 Bleeding
patients and randomly assigned them to 5 mg sublingual glyc- Bleeding is a rare complication of diagnostic ERCP and is
eryl trinitrate versus 100 mg vitamin C, 5 minutes prior to the most commonly seen with sphincterotomy. Risk factors for
ERCP. They found a significant difference in post-ERCP pan- postsphincterotomy bleeding include bleeding during the
creatitis in the study group (7.9%) compared with placebo procedure, concomitant thrombocytopenia or coagulopathy,
anticoagulation started within 3 days of the sphincterotomy, confluence (Figs. 20.10 and 20.11B). A right sectoral duct
and low case volume of the endoscopist (Freeman et al, 1996). crosses to the left to join the left hepatic duct in 28% of cases,
Rarely, there may be a Mallory-Weiss tear from scope trauma according to Healey and Schroy (1953); in 22%, this is the
or submucosal hemorrhage from manipulation of the papilla posterior sectoral duct (see Figs. 20.9B and 20.11B), and in
(Loperfido et al, 1998). There have been case reports of intra- 6%, this is the anterior duct (see Figs. 20.9C and 20.11D).
peritoneal hemorrhage from injury to abdominal vessels, liver, Occasionally, a right sectoral or segmental duct, posterior more
and spleen (Lo et al, 1994; Wu & Katon, 1993). often than anterior, courses inferiorly and enters the common
hepatic duct directly (Fig. 20.12). The confluence of the right
Perforation and left ducts takes the form of a trifurcation rather than a
The most common type of perforation associated with ERCP bifurcation in 12% of cases according to Couinaud (1957; see
is retroperitoneal perforation and is usually associated with Fig. 20.9A).
sphincterotomy, with an incidence ranging from 0.5% to 2.1% In the left lobe, the superior and inferior lateral segment
(Cotton et al, 1991). In their systematic survey of 21 prospec- ducts, segments II and III, unite in the line of, or to the right
tive studies with 16,855 patients, Loperfido and colleagues
(1998) reported 101 perforations (0.6%) and 10 deaths from
perforation (0.06%). Free bowel wall perforation is quite rare
and is usually associated with a structural abnormality such as TABLE 20.1 Segmental Nomenclature
a stricture or Billroth II gastrectomy (Wilkinson et al, 1994). I Caudate lobe
The management of the perforation depends on the size, II Left lateral superior segment
location, and clinical picture. Free bowel wall perforations III Left lateral inferior segment
usually require surgery. Use of endoscopic clips for the treat- IV Left medial segment or quadrate lobe
ment of duodenal perforations has also been reported V Right anterior inferior segment
(Katsinelos et al, 2005; Solomon et al, 2012). VI Right posterior inferior segment
VII Right posterior superior segment
DIRECT CHOLANGIOGRAPHY AND VIII Right anterior superior segment
PANCREATOGRAPHY BY PERCUTANEOUS
TRANSHEPATIC CHOLANGIOGRAPHY OR
ENDOSCOPIC RETROGRADE RPSD
CHOLANGIOPANCREATOGRAPHY RHD
LHD
The anatomy of the bile ducts is discussed in Chapter 2. LHD
Knowledge of the common variations of ductal branching is
essential for accurate interpretations of cholangiograms, and CHD
these are shown in Figures 20.8 and 20.9; segmental nomen-
RASD Sectoral
clature is summarized in Table 20.1. In the right lobe, the
duct
posterior segments lie more laterally than the anterior seg- A D
ments, so that the most lateral ducts on a cholangiogram are
usually segments VI inferiorly and VII superiorly. The posterior RPSD
sectoral duct is often recognizable by an arched course near the
RHD
II
RPSD LHD
III
VII VIII
II RASD
IV
I B E
VI
V IV III RHD
RPSD
II
RASD LHD
IV III
RHD
LHD
CHD
RASD
FIGURE 20.8. Standard intrahepatic ductal anatomy. The seg-
ments are numbered according to Couinaud’s description (see Table C F
20.1). CHD, Common hepatic duct; LHD, left hepatic duct; RASD, right FIGURE 20.9. Variations of perihilar ductal anatomy. CHD, Com-
anterior sectoral duct; RHD, right hepatic duct; RPSD, right posterior mon hepatic duct; LHD, left hepatic duct; RASD, right anterior sectoral
sectoral duct. duct; RHD, right hepatic duct; RPSD, right posterior sectoral duct.
The upper limits of normal for the diameter of the extrahe-

patic bile ducts as measured by ERCP vary between 9 and
14 mm (Niederau et al, 1984). Combined radiologic and
manometric studies (Poralla et al, 1985) have shown that even
in the absence of extrahepatic cholestasis, the diameter of the
bile ducts and the pressure difference therein increases with
advancing age. The diameter of the bile ducts as measured by
ultrasonography is less than those measurements obtained
during ERCP (Niederau et al, 1984). Anatomic abnormalities
of the hepatobiliary system include cystic dilations (Fig. 20.13)
of the bile duct or of the intrahepatic bile ducts (Caroli disease;
see Chapter 46). There is a wide variation in where the cystic
duct joins the common hepatic duct. A low junction with a
correspondingly long cystic duct (Fig. 20.14) may result in
difficulties if not recognized. This is especially true when a
cholecystojejunostomy is performed as a palliative biliary bypass
for carcinoma of the head of the pancreas, and the jaundice
either is not relieved or recurs rapidly in the postoperative
period.
Interpretations
FIGURE 20.10. Hilar cholangiocarcinoma involving first-order right Because of its inherent weakness of only allowing visualization
hepatic duct, proximal common hepatic duct, and faintly opacified left of the bile duct lumen, the main problem with the use of direct
hepatic duct (arrowhead). Note the characteristic arched course of the cholangiography is its lack of specificity. There are few, if any,
right posterior sectoral duct (arrow). pathognomonic radiologic findings. Many disease entities, from
benign to malignant, overlap greatly in their cholangiographic
appearances. The combination of history, blood markers, asso-
ciated radiologic findings, and clinical scenario can often sig-
TABLE 20.2 Average Duct Diameters Measured Directly from
nificantly narrow the differential diagnosis. However, it must
50 Normal Percutaneous Transhepatic Cholangiography be stressed strongly that because the cholangiographic appear-
Examinations ance of many biliary diseases may be indistinguishable, biopsy
Duct Diameter (mm) is often required to rule out malignancy or to confirm suspected
Right hepatic = 4.7 diagnosis.
Left hepatic = 5.2
Bile Leaks
Common hepatic = 6.5
Bile leaks are seen as sites of free extravasation of contrast
Common bile = 7.6
agent at a site of bile duct injury. Injury may be secondary to
From Okuda K, et al: Frequency of intrahepatic arteriovenous fistula as a sequela to percutaneous trauma, but most commonly it is iatrogenic in nature. Associ-
needle puncture of the liver, Gastroenterology 74:1204–1207, 1978.
ated bilomas may be seen at the point of bile leak. Diagnosis
of bile leak can usually be made by noninvasive imaging
techniques such as ultrasound, CT or MRI, but because the
of, the umbilical fissure in 92% of cases. In the latter instance, intrahepatic bile ducts are usually not dilated, a bile leak may
the quadrate lobe, segment IV, may drain wholly or partially not be evident on imaging. Cholescintography with techne-
into the segment II duct. Rarely, segment II and III ducts join tium 99m–hepatic iminodiacetic acid (HIDA scan) may be
at or close to the confluence (see Figs. 20.9E and 20.11C), and useful for diagnosis when the other imaging modalities are
segment IV drains directly into the common hepatic duct in inconclusive. ERCP can diagnose bile duct leaks effectively
1% of cases (see Fig. 20.9F) (Healey & Schroy, 1953). but is usually reserved for cases when a therapeutic interven-
The caudate ducts are often difficult to identify. Usually two tion is anticipated.
or three ducts drain most commonly into the right posterior
sectoral duct, right hepatic duct, or left hepatic duct (Healey Filling Defects
& Schroy, 1953). The recognizable caudate ducts are usually a Air Bubbles, Blood Clots, Calculi, Primary and Secondary Bile
few centimeters long and drain downward or to the right. Duct Cancers, and Parasitic Diseases
The left hepatic duct (average length, 17 mm) is consider- Air bubbles, although confusing, most commonly declare
ably longer than the right hepatic duct (average length, 9 mm) themselves by their perfectly circular shape and their distribu-
and has a longer extrahepatic course. The normal diameters of tion to nondependent structures. Blood clots in the bile duct
the main bile ducts as measured at PTC are shown in Table are seen more frequently with PTC than with ERCP. Hemobi-
20.2. These figures are greater than the true duct dimensions lia can sometimes take more than 48 hours to resolve, and when
because of some distension produced by direct cholangiogra- it is more severe, it can be castlike and may mask other filling
phy, together with considerable magnification occurring on any defects (see Chapter 125).
fluoroscopic “spot film.” The magnification is of the order of Calculous disease (see Chapters 36 and 39) remains the
40% (Nichols & Burhenne, 1984) and affects all structures in most common filling defect in the biliary system. Distinguishing
the image, including calculi, tubes, and strictures. characteristics of gallstones and primary bile duct calculi
A B
C D
FIGURE 20.11. Postcholecystectomy strictures graded according to Bismuth. A, Grade I (>2 cm from the confluence of the right and left hepatic
ducts; arrowheads): Calculi lie above and below the stricture (arrow). B, Grade II (<2 cm from the confluence): There has been a previous hepato-
jejunostomy; the right posterior sectoral duct (arrowhead) has an exaggerated arched course and enters the left hepatic duct as a normal variant.
C, Grade III (the confluence is involved by stricture, but the right and left hepatic ducts are not completely separated): Ducts of segment II (white
arrow) and segment III (black arrow) join the confluence independently as a normal variant. D, Grade IV (the right and left ducts are separated by
the stricture): The right anterior sectoral duct (A) is draining into the left hepatic duct (L), which is separated from the right posterior sectoral duct (P)
by the stricture (arrows).
include a faceted appearance and disposition to move to gravity- more unusual causes of cholangiographic filling defects (Riopel,
dependent positions. Calculi may be seen as discrete filling 1997).
defects (Figs. 20.15 and 20.16) or castlike structures filling Parasitic infections, such as hydatid disease (see Chapter 74)
entire ducts, as occurs in recurrent pyogenic cholangitis, cystic and infections with Ascaris lumbricoides or Clonorchis sinensis (see
diseases of the bile ducts, or even proximal to strictures of any Chapter 45), are diseases with worldwide distribution that can
etiology. Impacted calculi may be difficult to differentiate from also present cholangiographically as intraductal filling defects.
strictures or tumors. A proportion of hepatic hydatid cysts (5% to 10%) rupture into
Primary bile duct cancer (see Chapter 51), specifically papil- the bile ducts and may simulate choledocholithiasis. Calcified
lary cholangiocarcinomas, can also present as cholangiographic cysts are easy to recognize on radiographs, and daughter cysts
filling defects (Fig. 20.17). T-shaped filling defects may also be can cause biliary obstruction. When the calcified cyst is not
detected (Fig. 20.18 ) (Torok & Gores, 2001), and papillary obvious, biliary strictures with obstruction may be noted on
bile duct cancers may cause filling defects as a result of mucin cholangiogram. The biliary ducts can show considerable irregu-
production (Fig. 20.19). Other malignancies, including mela- larities in caliber and extensive displacement of the intrahepatic
noma and intraductal metastases, such as colon cancer, are also branches secondary to the mass effect of a large hydatid cyst
FIGURE 20.12. Pancreatitis producing a typical, incomplete long stric-

ture of the common bile duct. The right posterior sectoral duct has a
low entrance into the common hepatic duct, an uncommon but impor-
tant normal variant. FIGURE 20.13. Choledochal cyst.
FIGURE 20.14. Long cystic duct. FIGURE 20.15. Choledocholithiasis.

FIGURE 20.16. Benign stricture of the right hepatic duct (arrow) with
multiple ductal calculi proximal to it. The hepatojejunostomy is partially
strictured.
FIGURE 20.17. Papillary hilar cholangiocarcinoma (arrowheads). Only

the right ducts are opacified.
FIGURE 20.18. “Golf tee” appearance of a papillary bile duct cancer FIGURE 20.19. Nasobiliary cholangiogram opacifying left hepatic
(arrow) involving the common hepatic duct. ducts. Main left hepatic duct contains a small mucin-secreting papillary
cholangiocarcinoma (arrow). The mucin results in expansion of the
common bile duct below the tumor and appears as strandlike filling
defects. (Courtesy Dr. A. Speer.)
FIGURE 20.21. Extracted liver flukes.
biliary sludge and stones, and intrahepatic duct strictures,

which are predominantly found in the branches of the left
hepatic duct (Yellin & Donovan, 1981). The eggs of C. sinensis
act as a nucleus for the development of the bile duct stones
(Lam et al, 1978). Naval and colleagues (1984) reported suc-
FIGURE 20.20. Liver flukes in the distal bile duct presenting as “biliary cessful endoscopic biliary lavage to eliminate the eggs.
sludge.” Invasion of Fasciola hepatica into the biliary tract also may
cause serious lesions (see Chapter 45). In the chronic stage, F.
hepatica infection can resemble sclerosing cholangitis (Hauser
(Cottone et al, 1978; Dyrszka & Sanghavi, 1983; Ibrahim & & Bynum, 1984). The appearance on ERCP is that of dilated
Kawanishi, 1981). bile ducts with unexplained sludge in the distal bile duct (Figs.
Ascaris lumbricoides is a commonly seen helminth with a 20.20 and 20.21).
prevalence of 90% in some parts of Africa and Asia (see Chapter
45). If the worm passes through the sphincter of Oddi, it may CONCLUSION
cause acute pancreatitis or a cholestasis syndrome (Winters
et al, 1984). In the acute stage, the worm occasionally may be The role of direct cholangiography in the diagnosis of biliary
found and extracted from the ampulla, and it can be detected disease has for the most part been supplanted by less invasive
in the biliary tract by cholangiography. imaging modalities, such as MR cholangiography and CT with
Eating raw meat has been associated with Clonorchis sinensis contrast. The ability to visualize the entire biliary tree, the bile
infestation. The prevalence of this disease has been estimated duct wall, and other structures other than the bile duct lumen
to be 60% of the general population of Hong Kong, based on increases the diagnostic accuracy and clinical utility of these
stool ova examinations. This worm can penetrate through the noninvasive techniques, which have become the gold standard
papilla into the bile ducts. In an ERCP study of 31 consecutive for cholangiography. Direct cholangiography is generally
patients, the typical filamentous, wavy, or elliptic appearance reserved for clinical scenarios involving concomitantly planned
of the worm in the bile ducts was believed to be pathogno- therapeutic interventions.
monic. Other common cholangiographic findings include
widely dilated extrahepatic bile ducts, which are filled with References are available at expertconsult.com.
Chapter 20 Direct cholangiography: approaches, techniques, and current role 391.e1
REFERENCES Draganov P, et al: A new technique to assist in difficult bile duct can-
nulation at the time of endoscopic retrograde cholangiopancreatog-
Adler DG, et al: ASGE guideline: the role of ERCP in diseases of raphy, JSLS 9(2):218–221, 2005.
the biliary tract and the pancreas, Gastrointest Endosc 62(1):1–8, Dumonceau JM, et al: Prophylaxis of post-ERCP pancreatitis: Euro-
2005. pean Society of Gastrointestinal Endoscopy (ESGE) Guideline—
Akbar A, et al: Rectal nonsteroidal anti-inflammatory drugs are supe- updated June 2014, Endoscopy 46(9):799–815, 2014.
rior to pancreatic duct stents in preventing pancreatitis after endo- Dyrszka H, Sanghavi B: Hepatic hydatid disease: findings on endo-
scopic retrograde cholangiopancreatography: a network meta-analysis, scopic retrograde cholangiography, Gastrointest Endosc 29:248–249,
Clin Gastroenterol Hepatol 11:778–783, 2013. 1983.
Anand BS, et al: Effect of aging on the pancreatic ducts: a study based Earnshaw JJ, et al: Outbreak of Pseudomonas aeruginosa following endo-
on endoscopic retrograde pancreatography, Gastrointest Endosc 35: scopic retrograde cholangiopancreatography, J Hosp Infect 6:95–97,
210–231, 1989. 1985.
Andriulli A, et al: Incidence rates of post-ERCP complications: a sys- Elmunzer BJ, et al: A randomized trial of rectal indomethacin to prevent
tematic survey of prospective studies, Am J Gastrotenterol 102(8): post-ERCP pancreatitis, N Engl J Med 366:1414–1422, 2012.
1781–1788, 2007. Fanti L, et al: Target-controlled propofol infusion during monitored
Axon AT: Disinfection and endoscopy: summary and recommenda- anesthesia in patients undergoing ERCP, Gastrointest Endosc 60:361–
tions. Working party report to the World Congress of Gastroenterol- 366, 2004.
ogy, Sydney, J Gastroenterol Hepatol 6:23–24, 1991. Fazel A, et al: Does a pancreatic duct stent prevent post-ERCP pan-
Badalov N, et al: The prevention, recognition and treatment of post- creatitis? A prospective randomized study, Gastrointest Endosc 57:
ERCP pancreatitis, JOP 10(2):88–97, 2009. 291–294, 2003.
Bai Y, et al: Prophylactic antibiotics cannot prevent endoscopic Ferrucci JT, et al: Fine needle transhepatic cholangiography: a new
retrograde cholangiopancreatography-induced cholangitis: a meta- approach to obstructive jaundice, AJR Am J Roentgenol 127:403–
analysis, Pancreas 38:126–130, 2009. 407, 1976.
Bakhal MS, et al: Multimodality imaging of biliary malignancies, Surg Flemma RJ, Shingleton WW: Clinical experience with percutaneous
Oncol Clin N Am 2:225–239, 2009. transhepatic cholangiography: experience from 107 cases, Am J Surg
Bangarulingam SY, et al: Complications of endoscopic retrograde chol- 111:13–22, 1966.
angiopancreatography in primary sclerosing cholangitis, Am J Gas- Forbes A, Cotton PB: ERCP and sphincterotomy after Billroth II
troenterol 104(4):855–860, 2009. gastrectomy, Gut 25:971–974, 1984.
Benjamin IS, et al: The early use of fine needle percutaneous transhe- Fraser GM, et al: Percutaneous transhepatic cholangiography with the
patic cholangiography in an approach to the diagnosis of jaundice in Chiba needle, Clin Radiol 29:101–112, 1978.
a surgical unit, Br J Surg 65:92–98, 1978. Freeman ML, et al: Complications of endoscopic biliary sphincterot-
Bernard JP, et al: Pancreas divisum is a probable cause of acute pan- omy, N Engl J Med 335:909–918, 1996.
creatitis: a report of 137 cases, Pancreas 5:248–254, 1990. Freeman ML, et al: Risk factors for post-ERCP pancreatitis: a prospec-
Burckhardt H, Müller W: Verusche über die Puncktion der Gallenblase tive, multicenter study, Gastrointest Endosc 54:425–434, 2001.
und ihre Röntgendarstellung, Dtsch Chir 162:168, 1921. Freeman ML, et al: Prevention of post-ERCP pancreatitis: a compre-
Buxbaum J, et al: Aggressive hydration with lactated ringer’s solution hensive review, Gastrointest Endosc 59:830–834, 2004.
reduces pancreatitis after endoscopic retrograde cholangiopancre- Freeney PC: Radiology of the pancreas: two decades of progress in
otography, Clin Gastroenterol Hepatol 12(2):303–307, 2014. imaging and intervention, AJR Am J Roentgenol 150:975–981, 1988.
Byron E, et al: ERCP in patients with long-limb Roux-en-Y gastroje- Fukuda Y, et al: Diagnostic utility of peroral cholangioscopy for various
junostomy and intact papilla, Gastrointest Endosc 56:225–232, bile-duct lesions, Gastrointest Endosc 62(3):374–382, 2005.
2002. George S, et al: Role of osmolarity of contrast media in the develop-
Carter RF, Saypol GM: Transabdominal cholangiography, JAMA ment of post-ERCP pancreatitis: a meta-analysis, Dig Dis Sci 49:
148:253–255, 1952. 503–508, 2004.
Chan HH, et al: Effect of somatostatin in the prevention of pancreatic Glenn F, et al: Percutaneous transhepatic cholangiography, Ann Surg
complications after endoscopic retrograde cholangiopancreatogra- 156:451, 1962.
phy, J Chin Med Assoc 71(12):605–609, 2008. Gold RP, Price JB: Thin needle cholangiography as the primary method
Cheng C, et al: Risk factors for post-ERCP pancreatitis: a prospective for the evaluation of the biliary-enteric anastamosis, Radiology 136:
multicenter study, Am J Gastroenterol 101(1):139–147, 2006. 309–316, 1980.
Christensen M, et al: Complications of ERCP: a prospective study, Goldberg E, et al: Pancreatic-duct stent placement facilitates difficult
Gasotrointest Endosc 60:721–731, 2004. common bile duct cannulation, Gastrointest Endosc 62(4):592–596,
Classen M, Demling L: Hazards of endoscopic retrograde cholangio- 2005.
pancreatography, Acta Hepatogastroenterol 22:1–3, 1975. Hadjis NS, et al: Malignant masquerade at the hilum of the liver, Br J
Cohen S, et al: National Institutes of Health State-of-the-Science Con- Surg 72:659–661, 1985.
ference Statement: ERCP for diagnosis and therapy, January 14-16, Hannigan BF, et al: Hyperamylasemia after ERCP with ionic and non-
2002, Gastrointest Endosc 56(6):803–809, 2002. ionic contrast media, Gastrointest Endosc 31:109–110, 1985.
Concepcion MM, et al: Somatostatin for prevention of post-ERCP Hao JY, et al: Prophylactic effect of glyceryl trinitrate on post-
pancreatitis: a randomized, double-blind trial, Endoscopy 46(10):851– endoscopic retrograde cholangiopancreatography pancreatitis: a ran-
856, 2014. domized placebo-controlled trial, World J Gastroenterol 15(3):366–368,
Cotton PB: Progress report ERCP, Gut 18:316–341, 1977. 2009.
Cotton PB, et al: Cannulation of the papilla of Vater via fiber- Harbin WP, et al: Transhepatic cholangiography: complications and
duodenoscope, Lancet 1:53–58, 1972. use patterns of the fine-needle technique: a multi-institution survey,
Cotton PB, et al: Duodenoscopic papillotomy and gallstone removal, Radiology 135:15–22, 1980.
Br J Surg 63:709–714, 1976. Harewood GC, et al: An assessment of the learning curve for precut
Cotton PB, et al: Endoscopic sphincterotomy complications and their biliary sphincterotomy, Am J Gastroenterol 97:1708–1712, 2002.
management: an attempt at consensus, Gastrointest Endosc 37:383– Harris A, et al: Meta-analysis of antibiotic prophylaxis in endoscopic
391, 1991. retrograde cholangiopancreatography (ERCP), Endoscopy 31:718–
Cottone M, et al: Endoscopic retrograde cholangiography in hepatic 724, 1999.
hydatid disease, Br J Surg 66:107, 1978. Hauser SC, Bynum TE: Abnormalities on ERCP in a case of human
Couinaud C: Le foie: études anatomiques et chirurgicales, Paris, 1957, fascioliasis, Gastrointest Endosc 30:80–81, 1984.
Masson. Healey JE, Schroy PC: Anatomy of the biliary ducts within the human
Covey AM, Brown KB: Percutaneous transhepatic biliary drainage, liver, AMA Arch Surg 66:599–616, 1953.
Tech Vasc Interv Radiol 11:16–20, 2008. Hinde GDB, et al: Percutaneous cholangiography with the Okuda
Ding X, et al: Nonsteroidal anti-inflammatory drugs for prevention of needle, Gut 18:610–614, 1977.
post-ERCP pancreatitis: a meta-analysis, Gastrointest Endosc 76: Huard P, Do-Xuan-Hop ••: La ponction transhepatique des canaux
1152–1159, 2012. biliares, Bull Soc Med Chir Indochine 15:1090, 1937.
Ibrahim MAH, Kawanishi H: Endoscopic retrograde cholangiography Nichols DM, Burhenne HJ: Magnification in cholangiography, AJR
in the evaluation of complicated echinococcus of the liver, Gastroin- Am J Roentgenol 141:947–949, 1984.
test Endosc 27:20–22, 1981. Niederau C, et al: Comparison of the extrahepatic bile duct size mea-
Iqbal CW, et al: Pediatric endoscopic injuries: incidence, management sured by ultrasound and by different radiographic methods, Gastro-
and outcomes, J Pediatr Surg 43(5):911–915, 2008. enterology 87:615–621, 1984.
Itoi T, et al: Diagnostic peroral video cholangioscopy is an accurate Ogoshi K, et al: Endoscopic observation of the duodenum and pancre-
diagnostic tool for patients with bile duct lesions, Clin Gastroenterol atocholedochography using duodenofiberscope under direct vision,
Hepatol 8(11):934–938, 2010. Gastrointest Endosc 12:83–96, 1970.
Jain S, et al: Percutaneous transhepatic cholangiography using the Ohto M, Tsuchiya Y: Non-surgically available percutaneous transhe-
“Chiba” needle: 80 cases, Br J Radiol 50:175–180, 1977. patic cholangiography: technique and cases, Medicina (Tokyo) 6:
Jaques PF, et al: The failed transhepatic cholangiogram, Radiology 735–739, 1969.
134:33–35, 1980. Oi I, et al: Endoscopic pancreato-cholangiography, Endoscopy 2:103–
Jendrzejewski JW, et al: Antibiotics and ERCP: in vitro activity of 106, 1970.
amino-glycosides when added to iodinated contrast agents, Gastro- Okuno M, et al: Changes in serum levels of pancreatic isoamylase,
enterology 78:745–748, 1980. lipase, trypsin, and elastase after endoscopic retrograde pancreatog-
Johnson GK, et al: Evaluation of post-ERCP pancreatitis: potential raphy, Hepatogastroenterology 32:87–90, 1985.
causes noted during controlled study of differing contrast media. Omata F, et al: Meta-analysis: somatostatin or its long-acting analogue,
Midwest Pancreatidcobiliary Study Group, Gastrointest Endosc 46: octreotide, for prophylaxis against post-ERCP pancreatitis, J Gastro-
217–222, 1997. enterol 45(8):885–895, 2010.
Kapral C, et al: Case volume and outcome of endoscopic retrograde Osanai M, et al: Peroral video cholangioscopy to evaluate indetermi-
cholangiopancreatography: results of a nationwide Austrian bench- nate bile duct lesions and preoperative mucosal cancerous extension:
marking project, Endoscopy 40(8):625–630, 2008. a prospective multicenter study, Endoscopy 45(8):635–642, 2013.
Kadir S: Diagnostic angiography, Philadelphia, 1985, Saunders. Park DH, et al: Accuracy of magnetic resonance cholangiopancreatog-
Karamanolis G, et al: A prospective cross-over study using a sphinc- raphy for locating hepatolithiasis and detecting accompanying biliary
terotome and a guidewire to increase the success rate of common strictures, Endoscopy 36:987–992, 2004.
bile duct cannulation, World J Gastroenterol 11(11):1649–1652, Pereiras R Jr, et al: Percutaneous transhepatic cholangiography with
2005. the “skinny” needle: a rapid, simple, and accurate method in the
Kasugai T, et al: Fiberduodenoscopy: analysis of 353 examinations, diagnosis of cholestasis, Ann Intern Med 86:562–568, 1977.
Gastrointest Endosc 18:9–16, 1971. Phillip J, et al: Effekt von Somatostatin auf pankreatitische Komplika-
Katsinelos P, et al: Treatment of a duodenal perforation secondary to tionen nach ERCP und Eingriffen an der Papilla Vateri. In Henning
an endoscopic sphincterotomy with clips, World J Gastroenterol 11: H, editor: Fortschritte der gastroenterologischen Endoskopie, Grafelfing,
6232–6234, 2005. 1985, Demeter Verlag, pp 124–129.
Kawai K, et al: Endoscopic sphincterotomy of the ampulla of Vater, Poralla T, et al: Age and sex dependency of bile duct diameter and bile
Gastrointest Endosc 20:148–151, 1974. duct pressure: an ERC manometry study, Z Gastroenterol 23:235–
Kittredge RD, Baer JW: Percutaneous transhepatic cholangiogra- 239, 1985.
phy: problems in interpretation, AJR Am J Roentgenol 125:35–46, Rajnakova A, et al: ERCP in patients with periampullary diverticulum,
1975. Hepaticogastroenterology 50(51):625–628, 2003.
Kubiliun NM, et al: Evaluation of pharmacologic prevention of pan- Ratani RS, et al: Electrophysical factors influencing endoscopic sphinc-
creatitis following endoscopic retrograde cholangiopancreatography: terotomy, Gastrointest Endosc 49:43–52, 1999.
a systematic review, Clin Gastroenterol Hepatol 13(7):1231–1239, Rerknimitr R, et al: Result of endoscopic biliary drainage in hilar chol-
2015. angiocarcinoma, J Clin Gastroenterol 38(6):518–523, 2004.
Laasch HU, et al: Comparison of standard and steerable catheters for Riopel MA: Intrabiliary growth of metastatic colon adenocarcinoma: a
bile duct cannulation in ERCP, Endoscopy 35(8):669–674, 2003. pattern of intrahepatic spread easily confused with primary neoplasia
Lam SK, et al: Recurrent pyogenic cholangitis: a study by endoscopic of the biliary tract, Am J Surg Pathol 21(9):1030–1036, 1997.
retrograde cholangiography, Gastroenterology 74:1196–1203, 1978. Romagnuolo J, et al: Magnetic resonance cholangiopancreatography: a
Laurence BH, Cotton PB: Decompression of malignant biliary meta-analysis of test performance in suspected biliary disease, Ann
obstruction by duodenoscope intubation of the bile duct, Br Med J Intern Med 139:547–557, 2003.
280(6213):522–523, 1980. Rossos PG, et al: Selective common bile duct cannulation can be sim-
Lee KT, et al: The prophylactic effect of somatostatin on post- plified by the use of a standard papillotome, Gastrointest Endosc
therapeutic endoscopic retrograde cholangiopancreatography pan- 39(1):67–69, 1993.
creatitis: a randomized, multicenter controlled trial, Pancreas 37(4): Rudin D, et al: Somatostatin and gabexate for post-endoscopic retro-
445-448,2008. grade cholangiopancreatography pancreatitis prevention: meta-
Lo AY, et al: Splenic trauma following endoscopic retrograde cholan- analysis of randomized placebo-controlled trials, J Gastroenterol
giopancreatography (ERCP), Surg Endosc 8(6):692–693, 1994. Hepatol 22(7):977–983, 2007.
Loperfido S, et al: Major early complications from diagnostic and Saad M: Biliary cannulation and pancreatic guide-wire placement,
therapeutic ERCP: a prospective multicenter study, Gastrointest Endoscopy 36(8):743, 2004.
Endosc 48:1–10, 1998. Sato H, et al: Application of double-balloon enteroscopy for afferent
Mariani A, et al: Guidewire biliary cannulation does not reduce post- limb lesions of Roux-en-Y surgical anastomoses, Gastrointest Endosc
ERCP pancreatitis compared with the contrast injection technique 61:AB238, 2005.
in low-risk and high-risk patients, Gastrointest Endosc 75:339–346, Sauter G, et al: Antibiotic prophylaxis of infectious complications with
2012. endoscopic retrograde cholangiopancreatography: a randomized
McCune WS, et al: Endoscopic annulation of the ampulla of Vater: a controlled study, Endoscopy 22:164–167, 1990.
preliminary report, Ann Surg 167:752–756, 1968. Seldinger SI: Percutaneous transhepatic cholangiography, Acta Radiol
Miller G, et al: The use of imaging in the diagnosis and staging of 253(Suppl):1–134, 1966.
hepatobiliary malignancies, Surg Oncol Clin N Am 16(2):343–368, Seta T, Noguchi Y: Protease inhibitors for preventing complications
2007. associated with ERCP: an updated meta-analysis, Gastrointest Endosc
Mueller PR, et al: Fine needle cholangiography: reflections after 450 73(4):700–706, 2011.
cases, AJR Am J Roentgenol 136:85–90, 1981. Sethi A, et al: ERCP with cholangiopancreatoscopy may be associated
Mujahed Z, Evans JA: Percutaneous transhepatic cholangiography, with higher rates of complications than ERCP alone: a single-center
Radiol Clin North Am 4:535–545, 1966. experience, Gastrointest Endosc 73:251–256, 2011.
Naval F, et al: Endoscopic biliary lavage in a case of Clonorchis sinensis, Shaldon S, et al: Percutaneous transhepatic cholangiography: a modi-
Gastrointest Endosc 30:292–294, 1984. fied technique, Gastroenterology 42:371–379, 1962.
Navaneethan U, et al: Delay in performing ERCP and adverse events Simone M, et al: Three-dimensional virtual cholangioscopy: a reliable
increase the 30-day readmission risk in patients with acute cholan- tool for the diagnosis of common bile duct stones, Ann Surg 240:82–
gitis, Gastrointest Endosc 78(1):81–90, 2013. 88, 2004.
Chapter 20 Direct cholangiography: approaches, techniques, and current role 391.e3
Singh P, et al: Does prophylactic pancreatic stent placement reduce the Tsuchiya Y: A new safe method of percutaneous transhepatic cholan-
risk of post-ERCP acute pancreatitis? A meta-analysis of controlled giography, Jpn J Gastroenterol 63:438, 1969.
trials, Gastrointest Endosc 60:544–550, 2004. Vaishali MD, et al: Magnetic resonance cholangiography in obstructive
Soehendra N, Reijnders-Frederix V: Palliative bile duct drainage. A jaundice, J Clin Gastroenterol 38:887–890, 2004.
new endoscopic method of introducing a transpapillary drain, Endos- Wehrman T, et al: Efficacy and safety of intravenous propofol sedation
copy 12:8–11, 1980. during routine ERCP: a prospective, controlled study, Gastrointest
Solomon M, et al: Iatrogenic duodenal perforation treated with endo- Endosc 49:677–683, 1999.
scopic placement of metallic clips: a case report, Case Rep Med Weicher KL: Percutaneous transhepatic cholangiography: technique
2012:609750, 2012. and application, Acta Chir Scand 11(Suppl):330, 1964.
Sotoudehmanesh R, et al: A randomized trial of rectal indomethacin Wetter LA, et al: Differential diagnosis of sclerosing cholangitis of the
and sublingual nitrates to prevent post-ERCP pancreatitis, Am J common hepatic duct (Klatskin tumors), Am J Surg 161:57–63,
Gastroenterol 109(6):903–909, 2014. 1991.
Stroszczynski C, Hünerbein M: Malignant biliary obstruction: value of Wilkinson ML, et al: Intestinal perforation after ERCP in Billroth II
imaging findings, Abdom Imaging 30(3):314–323, 2005. partial gastrectomy, Gastrointest Endosc 40(3):389–390, 1994.
Sudhindran S, et al: Prospective randomized double-blind placebo- Winters C, et al: Endoscopic documentation of Ascaris-induced acute
controlled trial of glyceryl trinitrate in endoscopic retrograde pancreatitis, Gastrointest Endosc 30:83–84, 1984.
cholangiopancreatography-induced pancreatitis, Br J Surg 88(9): Woods SD, et al: Hypoxia and tachycardia during endoscopic retro-
1178–1182, 2001. grade cholangiopancreatography: detection by pulse oximetry, Gas-
Takagi K, et al: Retrograde pancreatography and cholangiography by trointest Endosc 35:523–525, 1989.
fiber-duodenoscope, Gastroenterology 59:445–452, 1970. Wu WC, Katon RM: Injury to the liver and spleen after diagnostic
Testoni PA, et al: Risk factors for post-ERCP pancreatitis in high- and ERCP, Gastrointest Endosc 39(6):824–827, 1993.
low-volume centers and among experts and non-expert operators, a Yaghoobi M, et al: Meta-analysis: rectal indomethacin for the preven-
prospective multicenter study, Am J Gastroenterol 105:1753–1761, tion of post-ERCP pancreatitis, Aliment Pharmacol Ther 38:995–
2010. 1001, 2013.
Tham TCK, Kelly M: Periampullary duodenal diverticula, bile duct Yamao K, et al: Efficacy of peroral pancreatoscopy in the diagnosis of
stones, and ERCP, Endoscopy 36:1050–1053, 2004. pancreatic diseases, Gastrointest Endosc 57(2):205–209, 2003.
Torok N, Gores GJ: Cholangiocarcinoma, Semin Gastrointest Dis Yellin AE, Donovan AJ: Biliary lithiasis and helminthiasis, Am J Surg
12(2):125–132, 2001. 142:128–135, 1981.
Tse F, et al: Guide wire-assisted cannulation for the prevention of Zhong L, et al: Imaging diagnosis of pancreatico-biliary diseases: a
post-ERCP pancreatitis: a systematic review and meta-analysis, control study, World J Gastroenterol 9:2824–2827, 2003.
Endoscopy 45(8):605–618, 2013.
CHAPTER 21
Diagnostic angiography in hepatobiliary and
pancreatic disease: indications
Hooman Yarmohammadi
OVERVIEW ANGIOGRAPHY TECHNIQUE

Indications for catheter angiography have substantially During the past few years, the angiography suites have been
changed. This is mainly due to the development of noninvasive updated with state-of-the-art imaging technology. Cut-film
imaging modalities such as multidetector computed tomogra- angiography has been replaced with digital flat-panel detectors
phy (MDCT) (see Chapter 18) and magnetic resonance and biplane angiography units. Biplane angiography is capable
imaging (MRI) (see Chapter 19). These imaging techniques of producing high-quality images in 3D views. These advances
can accurately demonstrate the visceral arterial and venous in technology allow for less contrast and minimum radiation
structures and are not associated with the morbidities of cath- exposure to the patients and the interventionalists.
eter angiography. Depending on the procedure, catheter angiography may be
Catheter angiography was commonly used for evaluation of performed under conscious sedation or general anesthesia.
the hepatic arterial anatomy, preoperative planning before liver Most procedures are performed as an outpatient procedure;
resection, assessment of vascular invasion by pancreatic cancer however, some patients may require overnight stay predomi-
or cholangiocarcinoma and to ascertain the organ of origin of nantly for pain control. All patients are seen in the clinic before
an abdominal mass. However, CT angiography (see Chapter the diagnostic or interventional catheter angiography proce-
18) and magnetic resonance angiography (MRA) (see Chapter dure. During this clinic visit, indications for performing the
19) have completely replaced catheter angiography for these procedure are reviewed. Additionally, patients are assessed for
indications. Currently, the main indications for catheter angi- any history of heart disease, lung disease, or renal problems.
ography are for treatment and intervention purposes, including Prior history of angiography or other surgical interventions are
embolization of gastrointestinal bleeding (see Chapters 27, 124, also assessed. Thorough physical examination, which includes
and 125), hepatic artery embolization and chemoembolization detailed pulse examination and assessment of airways, lungs,
(see Chapter 96A), pretreatment mapping and arterial anatomy and heart, is performed. Finally, patients’ performance status is
assessment prior to radioembolization (see Chapter 96B), and evaluated. Most institutions either use Eastern Cooperation
chemoperfusion (see Chapters 99 and 102). Oncology Group performance status or the Karnofsky perfor-
Recent progress in imaging technology, including advances mance status. The procedure is explained to the patient in
in C-arm cone-beam technology, has enabled the visualization detail, and after reviewing the risks and benefits of the proce-
of three-dimensional (3D) vascular anatomy with a single dure, consent is obtained.
acquisition by using nonselective C-arm computed tomography The relevant laboratory parameters include serum creatinine
(CBCT). Such capability allows visualization of the vascular or estimated glomerular filtration rate, hematocrit level, platelet
tree associated with the target tumor from multiple projections count, and international normalized ratio (INR), and these
during an embolization procedure. In addition to these should be reviewed. Additionally, at our practice a baseline 12
advances, another technologic advance is the development of lead electrocardiogram is obtained.
hybrid imaging. Hybrid imaging, or fusion of images, couples In patients with significant morbidities, a cardiology or
two imaging modalities to create unique images that can provide geriatric consult may be requested before the procedure to
information that would not be available with either technique ensure safety of conscious sedation or general anesthesia.
alone. Catheter angiography can be coupled with a cross- If the patient has impairment of renal function, a number of
sectional imaging modality, usually MDCT and, occasionally, prophylactic measures have been recommended to diminish the
MRI. The images can be postprocessed and rendered in a 3D risk of contrast-induced renal failure. These measures include
format to provide an anatomic depiction that would not be intravenous hydration with a sodium bicarbonate solution as
available by either technique alone. well as the administration of N-acetyl cysteine. The effective-
In this chapter, we will discuss angiographic anatomy relevant ness of these prophylactic measures is controversial; however,
to hepatobiliary and pancreatic surgery and discuss the main they carry a minimal risk to the patient and have been adopted
current indications for performing catheter angiography relevant by the majority of practitioners.
to hepatobiliary and pancreas. Additionally, we will discuss The risk of bleeding from the puncture site is low, and
localization of occult neuroendocrine tumors of the pancreas. hematomas complicate 1% of femoral punctures and 3% of
Our discussion on splanchnic veins will include venographic upper extremity punctures. An abnormal bleeding profile related
anatomy, venous sampling, techniques of catheter-based venous to thrombocytopenia or prolongation of the INR increases
imaging, and venous imaging before surgical or percutaneous the risk of hemorrhage. The necessity to correct an underlying
venous interventions. coagulopathy is dependent upon the specific procedure to be
392
Chapter 21 Diagnostic angiography in hepatobiliary and pancreatic disease: indications 393
performed and the preference of the angiographer. Based on the

recommendations from the Society of Interventional Radiology
(SIR) consensus guideline, catheter angiography (arterial inter- RHA
vention with access size up to 7 Fr) is classified as a category II
CyA LHA
procedure with moderate risk of bleeding. SIR recommends
platelet count higher than 50,000 × 106 per liter and an INR of
LGA
1.5 or less (Patel et al, 2013). Patients are advised to stop eating
6 hours prior to the procedure. Intravenous hydration is recom- CHA
RGEA GDA
mended before, during, and after the arteriogram to diminish
adverse effects of contrast media on renal function. The patient RGA SA
should also be encouraged to drink plenty of fluids following CA
the procedure.
The right common femoral artery is the most common
puncture site. The left common femoral artery, axillary artery,
brachial artery, or radial artery may be used as alternatives
when clinically appropriate. In the past few years, the radial
artery puncture site is becoming more favorable, particularly
with interventional cardiology procedures. A transradial FIGURE 21.1 Conventional celiac artery anatomy. CA, Celiac axis;
approach with regard to bleeding complications allows immedi- CHA, common hepatic artery; CyA, cystic artery; GDA, gastroduodenal
ate postprocedure ambulation, which is usually appreciated by artery; LGA, left gastric artery; LHA, left hepatic artery; RGA, right gastric
the patients. Additionally, recent studies demonstrate similar artery; RGEA, right gastroepiploic artery; RHA, right hepatic artery; SA,
procedural and clinical outcomes compared with a transfemoral splenic artery.
approach (Kiemeneij et al, 1997). The desired puncture area is
cleansed, and the patient is draped in sterile fashion. In most
centers, arterial entry is performed under real-time ultrasound
guidance by using a 21 gauge micropuncture set. The use of
ultrasonography allows assessment of the quality of the common
femoral artery, depicts the position of the profunda femoris,
and detects the presence of aberrant veins extending ventral to II
the puncture site. Following entry into the vessel, the appropri-
ate catheter is inserted for catheterization of the desired vessel. VIII IV
III
After diagnostically adequate images are obtained, the catheter
V
is removed, and one of a variety of closure devices is deployed
VII LHA
to seal the arteriotomy. Patients are observed in a postproce- RHA
dural area until they have recovered from sedation, and most
VI PHA
can be discharged home 2 to 4 hours later.
HEPATOBILIARY AND PANCREATIC

ARTERIAL ANATOMY
Arterial Anatomy
Arterial anatomy has been discussed elsewhere (see Chapters
1 and 2) and will be only briefly reviewed here. The most fre-
quently encountered anatomy (Fig. 21.1) is the left gastric
FIGURE 21.2 Arterial anatomy of the liver. LHA, Left hepatic artery;
artery, splenic artery, and common hepatic artery (CHA), PHA, proper hepatic artery; RHA, right hepatic artery. Additional vessels
taking origin from the celiac axis. The CHA divides into the are labeled in accordance to the Couinaud segment they supply.
gastroduodenal artery (GDA) and proper hepatic artery, with
the latter dividing into the right and left hepatic arteries. The
right gastric artery most often originates from the base of the variant origins. When the entire vessel has a variant origin, it is
left hepatic artery, and the cystic artery most often originates termed replaced. If the entire trunk does not take a variant
from the right hepatic artery (RHA), but considerable varia- origin, the vessel is termed accessory. For example, if the right
tions in the origins of these arteries exist (Kobayashi et al, lobe is supplied by a RHA originating from the CHA, as well
2014). Moreover, accessory duodenal arteries, either represent- as a RHA taking origin from the SMA, the latter would be
ing a supraduodenal or a retroduodenal artery, are frequently termed an accessory RHA. If the entire right lobe was supplied
encountered; this is critical to recognize when embolization, by an artery taking origin from the SMA, it would be called a
chemoembolization, and radioembolization are planned. replaced RHA.
The normal arterial supply to the liver is shown in Fig. 21.2, In most patients, the arteries to Couinaud segments I, II,
which shows the commonly recognized variations of the left III, and IV are branches of the LHA, and arteries to segments
hepatic artery (LHA), taking origin from the left gastric artery, V, VI, VII, and VIII are branches of the RHA. The most variable
and the RHA, taking origin from the superior mesenteric artery segmental branch is to segment IV. Although most frequently
(SMA). It is important to recognize that either a part or the the artery to segment IV takes origin from the LHA, it may also
entirety of the right and left hepatic arteries may have these take origin from the RHA, assuming the misnomer of a “middle
hepatic” artery in older works. Separate origins of segment IVA, confluence, but it may also enter the SMV either at or just
usually from the LHA, and segment IVB from the RHA are caudal to the confluence. The coronary vein most often drains
frequently identified. Also, a branch from the segment IV artery into the cephalic aspect of the main portal vein just beyond the
is often seen extending outside of the liver, supplying the confluence of the SMV and splenic vein. The number and
falciform ligament. Recognition of this vessel is important when location of veins draining the pancreas is variable. Multiple
embolization, chemoembolization, or radioembolization are small, unnamed veins drain directly into the splenic vein. Typi-
conducted to avoid nontarget embolization. cally, the anterior-superior pancreaticoduodenal vein drains
The RHA conventionally divides into an anterior (ventral) directly into the portal vein, and the posterior-superior pancre-
and a posterior (dorsal) branch. The anterior branch usually is aticoduodenal vein drains into the SMV. The inferior pancre-
more vertically oriented and supplies segments V and VIII. The aticoduodenal veins drain into the SMV at the caudal margin
posterior branch is usually more horizontally oriented and of the pancreas, and the portal vein courses obliquely cephalad
supplies segments VI and VII. More than one projection is from near the midline toward the liver, where it divides to
usually required to ascertain with certainty which is the anterior supply the right and left lobes. This division, as well as the
branch and which the posterior branch. In the right anterior division into segmental branches, is variable and must be
oblique projection, the anterior branch moves medially and the delineated when clinically relevant for each individual patient.
posterior branch moves laterally when compared with the
posteroanterior projection. The entire segmental arterial supply ANGIOGRAPHY INDICATIONS
to the liver should be accounted for before hepatic arterial
therapy (see Chapters 96 and 99), major hepatic resection, As mentioned earlier in this chapter, historic indications
partial hepatectomy (see Chapter 103), or living-donor liver for performing diagnostic catheter angiography have been
transplantation (see Chapters 104 and 117). replaced by multidetector computed tomography angiography
The arterial supply to the pancreas is somewhat variable. (MDCTA), including visceral angiography to assess the resect-
The most consistent supply is to the pancreatic head, through ability of pancreatic and biliary tract tumors by demonstrating
an arcade formed by the inferior pancreaticoduodenal branch presence or absence of vascular invasion; angiography for
of the SMA to the anterior- and posterior-superior pancreati- diagnosis of a hepatic mass, such as differentiating a cavernous
coduodenal branches of the GDA (Fig. 21.3). The transverse hemangioma from a hepatocellular carcinoma; and preopera-
pancreatic artery runs along the middle portion of the long axis tive arterial mapping of the arterial anatomy of the liver prior
of the pancreas and may take origin from the arterial arcade in to major hepatic resection. MDCTA provides higher sensitivity
the head of the pancreas, directly from the GDA, or as a branch and diagnostic accuracy for these indications. On rare occa-
of the dorsal pancreatic artery, which variably originates from sions, there is a specific piece of critical anatomic information
the CHA or the splenic artery (Fig. 21.4). A number of small that cannot be ascertained with certainty by MDCTA, such as
branches from the splenic artery supply the pancreatic body the origin and course of the artery to segment IV prior to a
and tail, but the number and location of these arteries vary and living-donor partial hepatectomy. In this circumstance, catheter
must be identified in each individual patient when clinically angiography can be useful to answer the question.
relevant. On extremely rare occasions, large tumors are identified on
cross-sectional imaging, but the organ of origin cannot be
Venous Anatomy determined. The majority of these are large sarcomas of the
The splenic vein and superior mesenteric vein (SMV) join retroperitoneum, but excluding a pancreatic source may be
to form the main portal vein (see Chapter 2). The inferior difficult. A similar situation can occur with large right adrenal
mesenteric vein usually enters the splenic adjacent to the or renal tumors blending with the hepatic parenchyma. In these
GDA
RGA
PSPD
DP
ASPD
TP
FIGURE 21.3 Arterial anatomy of the pancreas. ASPD, Anterior-

superior pancreaticoduodenal artery; GDA, gastroduodenal artery;
PSPD, posterior-superior pancreaticoduodenal artery; RGA, right gastric FIGURE 21.4 Arterial anatomy of the pancreas. DP, Dorsal pancreatic
artery. artery; TP, transverse pancreatic artery.
highly selected cases, catheter angiography can be useful in 5. Diagnosis of vasculitis

delineating the organ of origin by demonstrating the primary 6. Diagnosis of other visceral vascular disease
arterial supply. 7. Localization of functional pancreatic neuroendocrine tumors
In recent years, the most common indication for arteri
ography is planning an arterial-based intervention, such as Treatment of Bleeding/Hemorrhage
embolization, chemoembolization, radioembolization, or che- Bleeding from the liver, spleen or pancreas is most often sec-
moperfusion, to treat a primary or metastatic hepatic malig- ondary to iatrogenic or noniatrogenic trauma (see Chapters 122
nancy. These specific interventions will be discussed elsewhere to 125), but it may occur spontaneously in patients with mycotic
in this book (see Chapters 96 and 99). Accurate delineation of aneurysms or pancreatitis. Angiography is usually not used to
arterial anatomy is remarkably important in procedures such as ascertain whether arterial hemorrhage is present, but rather to
embolization, chemoembolization, and radioembolization. precisely localize and treat the bleeding vessel. Embolization of
Inadvertent administration of embolic particles, radiation par- arterial bleeding will be discussed elsewhere in this book, but
ticles, or chemotherapeutic agents into arteries supplying the salient features will be reviewed in this chapter.
stomach or duodenum can lead to significant adverse outcomes
(Riaz et al, 2009). Communications from intrahepatic branches Splenic Bleeding
to the lower esophagus, stomach, and diaphragm are of equal The most common cause of bleeding from the spleen is blunt
importance (Miyayama et al, 2009). trauma, and nonoperative management is currently the stan-
As explained earlier in this chapter, most modern angiogra- dard of practice. Splenic artery embolization has been estab-
phy suites are equipped with CBCT imaging technology. This lished as a method to increase the success rate of nonoperative
technology uses a fixed C-arm system equipped with a flat- management of traumatic splenic injuries (Schnuriger et al,
panel detector and requires 3D CT volumetric images (Tacher 2011). A comparative study between two cohorts consisting of
et al, 2015). CBCT has improved the feasibility, effectiveness, 625 patients during a 15 year period revealed an improved
and safety of many image-guided procedures by allowing the success rate of nonoperative management from 77% to 96%
interventionalist to identify extrahepatic perfusion from aber- with the advent of splenic embolization (Rajani et al, 2006).
rant hepatic arterial branches (Fig. 21.5). The indications for splenic arteriography and splenic arterial
Other current indications of performing catheter angiogra- embolization are based upon CT findings and include active
phy in the hepatobiliary and pancreatic system are as follows: contrast blush beyond or within the splenic parenchyma,
1. Treatment of bleeding/hemorrhage from liver, spleen, and pseudoaneurysm, an associated large hemoperitoneum, and a
pancreas high-grade splenic injury (Schnuriger et al, 2011). Moreover,
2. Diagnosis of arterial occlusive diseases the American Association for the Surgery of Trauma recom-
3. Diagnosis and treatment of arterial stenosis mends angiography for grade III, IV, and V splenic injuries
4. Treatment of visceral arterial aneurysms (Raikhlin et al, 2008).
A B
FIGURE 21.5 Extrahepatic perfusion detected with C-arm computed tomography (CCT). A, Selective cystic arteriogram shows the gallbladder and
several branches extending medially (arrow). B, CCT performed during contrast injection of the cystic artery confirms aberrant perfusion of the
duodenum (arrowheads). (Courtesy Daniel Sze, Stanford University.)
Two techniques are used to perform splenic embolization: Obviously, the contribution of the embolization to these com-
the first is occlusion of the proximal splenic artery with coils or plications may be difficult to distinguish from sequelae of the
Amplatzer plugs (AGA Medical, Plymouth, MN), and the underlying traumatic injury.
second is selective small intrasplenic arterial embolization with Iatrogenic injury to the hepatic artery is suspected when a
a gelatin sponge or coils. Collaterals through the short gastric biliary drainage tube puts out blood following placement, or
arteries and the gastroepiploic arcade usually maintain splenic melena secondary to hemobilia develops in the patient follow-
viability following occlusion of the proximal splenic artery. ing an intervention through the hepatic parenchyma. CT in
Intrasplenic embolization generally results in a variable degree these patients may or may not reveal an abnormality, and
of splenic infarction, depending upon the size of the artery patients should undergo hepatic arteriography when a clinical
occluded, as intrasplenic arteries have no significant collateral suspicion of a hepatic arterial injury exists. Patients with iatro-
routes. Both techniques have equivalent rates of major infarc- genic injuries usually have a single, discrete bleeding source that
tions and infections requiring splenectomy (Ekeh et al, 2013). can be addressed with superselective embolization techniques.
Distal splenic embolization is associated with higher rates of Complications related to the embolization procedure tend to
infarction; however, these infarctions are limited to the seg- be lower when compared with patients who have had blunt
ments just distal to the site of embolization with questionable trauma, as the traumatic injury to the liver is less extensive.
clinical relevance. In summary, the current literature is incon-
clusive regarding whether the proximal or distal embolization Pancreas Bleeding
should be used. Minor complications including fever, pleural Bleeding is more commonly encountered in patients with
effusion, and partial splenic infarction have been reported in as pancreatitis (see Chapter 56) or pancreatic surgery (see Chap-
many as 34% of patients. Major complications including ters 66 and 67). Posttrauma bleeding from the pancreas is not
splenic abscesses, splenic infarction, splenic atrophy, and post- common. Hemorrhage can occur in the retroperitoneum, or it
procedure bleeding have been observed in 14% of patients may extend from the retroperitoneum into the peritoneal cavity
(Ekeh et al, 2013). or into the gastrointestinal tract via communication with the
pancreatic duct (hemosuccus pancreaticus). Vascular complica-
Hepatic Bleeding tions are seen in 25% of patients suffering from pancreatitis and
Hemorrhage from the liver may be encountered from blunt or are usually arterial in origin. Proteolytic enzymes coupled with
penetrating trauma (see Chapter 122) but is most commonly intense inflammation can erode small arterial branches and
due to an iatrogenic injury related to a biopsy (see Chapter 22) create foci of extravasation or create small pseudoaneurysms.
or percutaneous transhepatic biliary drainage (PTBD) (see Although this complication is encountered in only 2% to 5%
Chapters 30 and 52). Arteriographic findings indicating a of cases, it may be life threatening. Pseudocysts may erode small
source of hemorrhage include extravasation, pseudoaneurysm arterial branches, leading to hemorrhage within the pseudocyst,
formation, and/or arteriovenous fistula (see Chapter 124). or it may erode into larger arterial branches and create a large
Superselective catheterization by using coaxial microcatheters pseudoaneurysm (Fig. 21.6). These pseudoaneurysms are most
to deposit coil-spring emboli both distal and proximal to the frequently identified in the splenic artery or its branches (60%
area of injury is the preferred technique for embolization when to 65%), followed by the GDA (20% to 25%), pancreaticoduo-
a discrete bleeding site can be identified. In contrast to the denal arteries (10% to 15%), hepatic artery (5% to 10%), and
spleen, a rich collateral network exists in the hepatic arterial left gastric artery (2% to 5%) (Aswani et al, 2015; Kirby et al,
bed, making proximal occlusion of the offending artery ineffec- 2008).
tive in many cases. For more diffuse injuries with multiple MDCT has been reported to have 90% sensitivity in detect-
bleeding sites secondary to blunt trauma, using particulate ing pseudoaneurysms in patients with acute pancreatitis;
embolization with a gelatin sponge may be a useful adjunct. however, tiny aneurysms in the intrapancreatic arteries may
As with blunt splenic injuries, nonoperative management not be visible (Hyare et al, 2006). When a patient has clinical
has become the preferred method of management of hemody- evidence of significant hemorrhage, or a CT scan reveals a
namically stable patients. The success rate of this management pseudoaneurysm, arteriography followed by embolization is
exceeds 90% (Christmas et al, 2005). Two main indications for indicated. When the bleeding site is identified angiographically,
hepatic angioembolization are primary hemostatic control in a it can be controlled with embolotherapy in as many as 88% of
hemodynamically stable patient that has radiographic evidence patients (Mansueto et al, 2007).
of active arterial hemorrhage and adjunctive hemostatic control Angiography and embolization are feasible and safe in treat-
following surgical exploration and packing of a hepatic paren- ment of hemorrhagic complications following pancreatic surgery
chymal injury with evidence of continued bleeding or hemody- (Casadei et al, 2014; Yekebas et al, 2007). Yekebas and col-
namic instability. Additionally, patients who are seen with leagues (2007) reported significant hemorrhage in 5.7% of 1669
hemodynamic instability who can be successfully resuscitated consecutive patients following partial or total pancreatectomy.
can be treated effectively with transcatheter embolization (Mis- In a more recent study, Casadei and colleagues (2014) reported
selbeck et al, 2009). The yield of arteriography in identifying significant hemorrhage in 9.8% of 182 patients following pan-
an arterial injury amenable to embolization is higher when the creatic resection for pancreatic and periampullary diseases.
CT scan suggests a vascular injury. Bleeding may manifest clinically as gastrointestinal hemorrhage,
Complications following embolization of a hepatic arterial retroperitoneal or intraperitoneal bleeding, or bleeding through
injury secondary to blunt trauma are relatively frequent percutaneous or surgically placed drains. When a pancreatico-
(Letoublon et al, 2011). In a retrospective study, Letoublon jejunal anastomosis has been created, disruption of the anasto-
and colleagues reported a 70% liver complication rate. These mosis may lead to false localization of the bleeding site.
complications include hepatic ischemia, infarction, hepatic Specifically, an extraluminal bleeding site may drain into the
failure, gallbladder ischemia, bile leak, and abscess formation. bowel through the dehisced anastomosis, or bleeding from the
A B
C
FIGURE 21.6 Pseudoaneurysm secondary to pancreatitis treated with coil embolization. A, Contrast-enhanced computed tomography reveals a
pseudoaneurysm in the pancreatic head (arrow). B, Selective gastroduodenal arteriogram reveals the pseudoaneurysm (arrow) taking origin from a
branch of the anterior-superior pancreaticoduodenal artery. C, The pseudoaneurysm is occluded with coils (arrowheads).
bowel at the anastomosis may extend outside the lumen to cause but may be suggested by CT. If the underlying etiology is
an intraperitoneal hematoma or hemorrhage through a drain. compression of the splenic vein, percutaneous transhepatic
Angiography can be extremely useful in the diagnosis and stenting of the splenic vein may be potentially useful.
treatment of these patients. In 25 of 43 patients undergoing
angiography to diagnose and treat postpancreatectomy hemor- Diagnosis of Arterial Occlusive Disease
rhage, a bleeding site was located and embolized with an 80% Many arterial disorders that involve the primary branches of
success rate in controlling the hemorrhage (Yekebas et al, the celiac trunk, as well as the SMA, can be assessed adequately
2007). The relatively low rate of positive angiograms may be with MDCT or MRA. However, delineation of pathology in
explained by the high incidence of venous bleeding encountered smaller branches, such as with vasculitis, or subtle pathologic
in the postpancreatectomy patient. Regional portal hyperten- changes may require the increased morphologic detail afforded
sion develops in many of these patients as a result of splenic or by catheter angiography.
superior mesenteric venous compression or occlusion second- Arterial occlusive disease as a result of atherosclerosis and
ary either to the underlying pathology or a surgical complica- compression of the celiac axis by a median arcuate ligament are
tion. Venous extravasation is rarely identified by arteriography common diseases that do not generally influence the conduct
of hepatobiliary or pancreatic surgery. The principle exception Treatment of Visceral Arterial Aneurysms
is in the performance of orthotopic liver transplantation (OLT), Visceral artery aneurysms are rare entities that involve the
when preservation of brisk hepatic arterial flow is essential. In celiac, splenic, superior mesenteric, or inferior mesenteric
that circumstance, the inflow must be corrected either by surgi- arteries and their branches (see Chapter 124). The prevalence
cal or endovascular interventions. Another potential area of of visceral artery aneurysm is 0.1% to 2% (Hemp & Sabri,
concern is pancreaticoduodenectomy, particularly in jaundiced 2015; Tulsyan et al, 2007). True aneurysms involve all three
patients. In this situation, sacrifice of the GDA is required, vessel walls and are usually atherosclerotic or developmental
which interrupts the retrograde arterial flow from the SMA to in origin and differ from those encountered in pancreatic
the liver and could result in hepatic ischemia and necrosis. inflammatory disease, which are typically pseudoaneurysms.
Likewise, fibromuscular dysplasia may rarely involve the SMA, Depending on the size and location of the aneurysm, mortal-
but associated clinical sequelae are extremely uncommon. The ity from rupture ranges from 25% to 100% (Cordova &
SMA may also be compromised in very rare circumstances by Sumpio, 2013).
a median arcuate ligament. The splenic artery is the most common affected artery,
followed by the hepatic artery. Splenic artery aneurysms in
Diagnosis and Treatment of Arterial Stenosis women of childbearing age are of particular concern because
Hepatic arterial anastomotic stenoses are observed in as many of their propensity to rupture during childbirth. Most splenic
as 11% to 12% of patients following OLT (see Chapters 116 aneurysms are saccular and located in the mid to distal segment
and 120) (Duffy et al, 2009; Kim et al, 2012). Although these of the artery (Nosher et al, 2006). The rate of rupture ranges
stenoses are usually detected by surveillance duplex ultrasound from 3% to 20% (Lagana et al, 2006). Endovascular treatment
and confirmed with MDCTA, catheter angiography is often of splenic artery aneurysms depends on the tortuosity and
performed to improve morphologic delineation and assess the location of the aneurysm. Stent placement is used for more
degree of stenosis in preparation for endovascular treatment proximal disease and distal disease in a tortuous vessel is usually
(Vaidya et al, 2007). Stenosis of the hepatic artery anastomosis treated with coil embolization (Hemp & Sabri, 2015).
usually leads to allograft dysfunction, and stenoses of the Traditionally, surgical resection or ligation of the visceral
hepatic artery may also be more diffuse (Fig. 21.7). Severe aneurysms was the gold standard of therapy; however, endovas-
stenosis may lead to hepatic artery thrombosis. The hepatic cular treatments have largely replaced open resection. Large
artery supplies the bile duct, and the bile duct has been ren- studies have reported technical success rates ranging from 89%
dered devoid of collateral arterial supply during donor hepatec- to 98% (Hemp & Sabri, 2015; Sachdev et al, 2006; Tulsyan
tomy. Arterial insufficiency typically leads to biliary strictures et al, 2007). Endovascular treatments are associated with
and to potentially diffuse biliary ductal infarction. Hepatic shorter hospital stay, when compared with surgical repair: 3.8
artery thrombosis may lead to irretrievable allograft damage versus 12 days, respectively.
and may necessitate retransplantation. When a hepatic arterial
stenosis is identified before advanced biliary injury, endovascu- Diagnosis of Vasculitis
lar therapy with either balloon angioplasty and/or stent place- Vasculitis
ment is warranted. Vasculitides that involve the hepatic arterial system may require
the spatial resolution provided by catheter angiography for
definitive diagnosis, because the characteristic microaneurysms
and areas of arterial narrowing may not be apparent by MDCT.
The most common vasculitis with hepatic involvement is
polyarteritis nodosa, which may not lead to symptoms despite
involvement of the visceral arteries, although pancreatitis,
cholecystitis, and hepatic dysfunction may be observed.
Arterial abnormalities in the liver have also been identified
in patients with systemic lupus erythematosus and Wegener
granulomatosis. In most patients, the underlying diagnosis is
apparent, and the angiographic findings do not represent a
diagnostic dilemma.
Diagnosis of Other Visceral Vascular Disease

Segmental Arterial Mediolysis
A rare vascular disorder, segmental arterial mediolysis, may also
require catheter angiography for definitive diagnosis. This dis-
order is characterized pathologically by noninflammatory
destruction of smooth muscle cytoplasm, leading to arterial
dissection and aneurysm formation. Multifocal lesions with
skip areas involving the superior mesenteric, hepatic, renal, and
middle colic arteries in patients in their fourth to sixth decades
are typical (Fig. 21.8). Clinically, these lesions are prone to both
occlusion by dissection as well as rupture, leading to cata-
FIGURE 21.7 Hepatic artery stenosis following orthotopic liver trans- strophic abdominal hemorrhage. Treatment with coil emboliza-
plantation. Celiac arteriogram shows mild diffuse narrowing of the tion within noncritical arterial beds has been reported (Davran
hepatic artery with two areas of critical stenosis (arrowheads). et al, 2010).
Hereditary Hemorrhagic Telangiectasia depicting the malformations shunting hepatic arterial blood
Hereditary hemorrhagic telangiectasia (HHT) also known as into the hepatic venous system (Fig. 21.9) (Memeo et al, 2008).
Osler-Weber-Rendu syndrome, is an autosomal dominant vascular Although this disorder has been treated with transcatheter
disease, characterized by telangiectasias of the skin and mucous techniques, embolotherapy has currently fallen out of vogue
membranes. HHT is also associated with arteriovenous malfor- because of the risk of precipitating hepatic failure. OLT is
mations in the pulmonary and hepatic circulation that may be curative if high-output cardiac failure develops in the patient.
life-threatening. Diagnosis is usually based on family history
and clinical criteria. Cross-sectional imaging, including ultra- Peliosis Hepatitis
sound, CT or MRI, plays a fundamental role in detecting vis- Peliosis is an abnormality of the reticuloendothelial system that
ceral involvement in HHT. The hepatic arterial malformations is most commonly encountered in the liver (peliosis hepatis)
that shunt blood into the hepatic venous system may be initially (see Chapter 89). Pathologically, it is characterized by blood-
noted on a cross-sectional imaging study, but the findings may filled cystic spaces that range in size from a few millimeters to
be nonspecific. Catheter angiography can be diagnostic by almost a centimeter. This abnormality has been associated with
human immunodeficiency virus infection as well as with the use
of certain drugs, including immunosuppressives, antimetabo-
lites, and oral contraceptives. Although usually benign, it has
been associated with spontaneous massive hemorrhage and
therefore may be encountered angiographically during the
investigation of hepatic bleeding (Choi et al, 2009). Angio-
graphically (Fig. 21.10), the lesions are easily visible as a dis-
organized collection of amorphous channels not dissimilar to
those observed in HHT or hemangioma; however, the lack of
shunting to the hepatic venous system distinguishes it from
HHT, and the absence of sharp definition with persistence into
the late venous phase distinguishes it from hemangioma.
Localization of Functional Pancreatic Neuroendocrine
Tumors (See Chapter 65)
Calcium stimulation arteriography was developed and described
in 1991 (Doppman et al, 1991). Although unnecessary when
imaging can confidently detect the offending pancreatic lesion,
it is an extremely useful adjunct when the location of the lesion
cannot be defined with confidence (Fig. 21.11). When 1 mL of
10% calcium gluconate is injected into an artery supplying the
FIGURE 21.8 Segmental arterial mediolysis (SAM). Abdominal aorto- pancreas, tumor cells degranulate and release insulin into the
gram reveals a dissection with a small aneurysm in the celiac artery portal venous circulation. Sequential blood samples obtained
(arrow) as well as undulating irregularity of the common hepatic artery from the hepatic vein will demonstrate a significant rise in the
(arrowhead) typical of SAM. concentration of insulin when the portion of the pancreas
A B
FIGURE 21.9 Hereditary hemorrhagic telangiectasia (HHT). A, Selective hepatic arteriogram shows disorganized and dilated intrahepatic vessels
typical of HHT. Coils are being placed preoperatively for impending liver transplantation. B, Slightly later in the sequence, early opacification of the
hepatic vein is visible (arrowheads).
containing the tumor is injected. By sequentially injecting the determining the approximate areas of pancreatic arterial supply.
arterial supplies to different regions of the pancreas, it is pos- In the article by Guettier and colleagues (2009), calcium stimu-
sible to ascertain the location of the tumor. The GDA supplies lation arteriography was the most sensitive technique for local-
the head of the pancreas, the SMA supplies the head and izing surgically proven insulinomas, with an accuracy of 84%,
uncinate process, and the splenic artery supplies the body a false-negative rate of 11%, and a false-positive rate of 4%.
and tail. The origin of small branches supplying the pancreas Percutaneous transhepatic sampling of the splenic, superior
from the splenic artery and CHA are important to note in mesenteric, and portal venous system can be performed to
diagnose occult neuroendocrine tumors of the pancreas. This
may be done in conjunction with calcium stimulation as
described earlier, or it may be performed without stimulation
because of the higher concentration of the hormone when
obtained directly or adjacent to the venous tributary.
Insulinomas
Greater than 90% of insulinomas are single, benign tumors for
which surgical resection is curative. These tumors are the most
common tumors originating from the islets of Langerhans (see
Chapter 65). The most effective method of diagnosing these
tumors is a combination of dual-phase thin-section CT scan and
RIGHT endoscopic ultrasonography (EUS) (Jackson, 2005). Advances
in imaging technology have improved the sensitivity of CT and
MRI to greater than 80% and 70%, respectively (Nikfarjam
et al, 2008). The sensitivity of EUS in detecting and localizing
insulinomas ranges from 82% to 94% (McLean, 2004).
Gastrinomas
Gastrin-secreting tumors are the cause of Zollinger-Ellison
syndrome and in approximately one-third of the cases occur in
association with multiple endocrine neoplasia (MEN) type 1.
Fifty percent of gastrinomas occur in the pancreas, with the
FIGURE 21.10 Peliosis hepatitis. Multiple small contrast collections duodenum being the most common extrapancreatic location.
can be seen within the hepatic parenchyma. This patient had spontane- Approximately 60% to 90% of gastrinomas are malignant.
ous hemorrhage that created displacement of the hepatic parenchyma When a sporadic gastrinoma is identified, surgical resection is
(arrowheads). indicated. The role of surgery in patients with MEN type 1 is
A B
FIGURE 21.11 Insulinoma. A, Selective gastroduodenal arteriogram shows no definite abnormality during the arterial phase. B, A subtle area of
hypervascularity is identified during the capillary phase, possibly representing an insulinoma (arrowheads). This was confirmed as the region of tumor
by calcium stimulation with venous sampling.
A B
FIGURE 21.12 Arterial portography in a patient with cavernous transformation of the main portal vein. A, Venous phase of superior mesenteric
arterial injection shows the superior mesenteric vein (arrow) and cavernous transformation of the main portal vein. B, Venous phase of splenic arterial
injection from the same patient shows the splenic vein (arrow).
more controversial because of multiplicity of tumors and lack trunks and their primary branches in the vast majority of
of an established survival benefit. patients. Cross-sectional imaging can accurately depict the
The diagnosis and location of a gastrinoma can be estab- relationship of a mass in the pancreas to both the splenic and
lished with a combination of somatostatin receptor scintigraphy superior mesenteric veins. It also has the advantage of simulta-
and EUS in approximately 90% of patients. When an occult neous opacification of all of the venous structures. However, in
gastrinoma is encountered, angiography has been used for certain clinical situations, it is desirable to visualize the venous
localization. The principles are identical to the localization of structures with a higher level of clarity. These situations include
insulinomas; however, secretin has been used in addition to planning of percutaneous venous intervention in situations
calcium gluconate as the stimulating agent. Sensitivities of where occlusions are suspected or occasionally to plan a surgi-
arterial stimulation venous sampling (ASVS) in the detection cal portosystemic shunt.
of gastrinomas have ranged from 70% to 100% (Cohen et al, The most common technique to visualize the splanchnic
1997; Turner et al, 2002). vein is transarterial portography, in which selective splenic and
Angiographically, gastrinomas are less hypervascular and superior mesenteric arteriograms are performed with delayed
more difficult to detect in comparison to insulinomas. Sensitiv- imaging into the venous phase, depicting the splenic and supe-
ity of angiography alone without ASVS is less than 50%. rior mesenteric veins, respectively (Fig. 21.12). When detailed
Moreover, the 50% extrapancreatic location makes detection visualization of the venous anatomy is required, a higher dose
more difficult, often requiring superselective catheterization to of contrast media can be used for the arterial injection, increas-
evaluate the duodenum. ing the clarity of venous opacification.
When an extremely detailed evaluation is required, a com-
Glucagonoma bination of transarterial portography during MDCT can be
Glucagonomas may occur sporadically or may be associated performed (Fig. 21.13). The increased contrast sensitivity of the
with MEN type 1. These tumors are usually larger than gastri- MDCT coupled with multiplanar and 3D reconstruction allows
nomas or insulinomas at presentation; therefore localization can visualization of the veins that cannot be achieved by any other
generally be achieved with cross-sectional imaging or EUS. single technique. This is particularly useful in the presence of
Equchi and colleagues (2012) reported eight patients with portal vein occlusion, when a complex venous reconstruction
pancreatic hypervascular tumor and elevated serum glucagon or bypass is being considered.
level that glucagonomas were successfully localized using ASVS. Direct venography of the splanchnic veins can be achieved
by three routes: transjugular, transhepatic, and transsplenic. In the
Venographic Technique transjugular approach, a catheter is placed into the jugular vein
MDCT, magnetic resonance venography, and ultrasound are and advanced into a hepatic vein. Free and wedged hepatic
usually sufficient for depiction of the major visceral venous pressures can be obtained through this catheter. Using these
undergoing a transjugular intrahepatic portosystemic shunt, but

it is also used to manage selected patients with portal and SMV
thrombosis (Liu et al, 2009).
The portal venous system may also be accessed by a percu-
taneous transhepatic approach (Semiz-Oysu et al, 2007). After
sterile preparation of the right upper quadrant, an intrahepatic
portal venous radicle is punctured under real-time ultrasono-
graphic guidance, permitting a vascular access sheath to be
placed by using a Seldinger technique. Standard guidewire and
catheter manipulations are then used through the sheath for
selectively catheterizing the splenic vein and the SMV or its
branches. Once the diagnostic images have been taken or
venous interventions have been performed, the catheter is
removed. The transhepatic tract is then occluded by a variety
of techniques, including insertion of Gelfoam pledgets, deploy-
ment of coils, or injection of fibrin glue. The transhepatic
venographic procedure is usually performed as part of a direct
pancreatic venous sampling, lobar portal venous embolization
to stimulate contralateral hypertrophy before major hepatic
resection, assessment and management of an anastomotic
portal venous stenosis following OLT, pharmacomechanical
lysis of a splanchnic vein thrombosis, or, rarely, to control
bleeding from a splanchnic vein.
Percutaneous injection of the splenic parenchyma can also
be used to delineate the anatomy of the splenic and portal veins
as well as the draining tributaries. This examination is somewhat
FIGURE 21.13 Multidetector computed tomography with three- antiquated and has been generally replaced by MDCT or
dimensional reconstruction following superior mesenteric arterial injec- magnetic resonance venography. In infants, cross-sectional
tion. The superior mesenteric vein and cavernous transformation are
imaging may be inconclusive, and transarterial portography is
easily visualized.
risky because of the diminutive size of the femoral arteries;
however, it remains an alternative in planning a portosystemic
shunt procedure.
two pressures, the corrected sinusoidal pressure can be calcu- Percutaneous catheterization of the splanchnic veins is also
lated and can be used to assess portal hypertension. A biopsy possible from a transsplenic approach using a technique identi-
of the hepatic parenchyma (transjugular liver biopsy) may be cal to that described for transhepatic catheterization (Tuite
performed when desirable. Injection of carbon dioxide through et al, 2007). This procedure is usually done in conjunction with
a catheter wedged in a hepatic vein or through a balloon cath- a percutaneous procedure to assess a portal vein stenosis or
eter will often yield an image of the portal vein. If direct por- occlusion following OLT, but it may also be used in conjunction
tography is warranted, one of several specially designed needles with the assessment and treatment of a splenorenal venous
can be inserted to puncture the portal vein through the inter- bypass, when access cannot be achieved from the systemic
vening hepatic parenchyma and through catheters advanced venous circulation.
into the portal, splenic, or superior mesenteric venous system.
This procedure is performed almost exclusively in patients References are available at expertconsult.com.
Chapter 21 Diagnostic angiography in hepatobiliary and pancreatic disease: indications 402.e1
REFERENCES Letoublon C, et al: Hepatic arterial embolization in the management

of blunt hepatic trauma: indications and complications, J Trauma
Aswani Y, Hira P: Venous complications of pancreatitis: a review, JOP 70(5):1032–1036, discussion 1036–1037, 2011.
16(1):20–24, 2015. Liu FY, et al: Interventional treatment for symptomatic acute-subacute
Casadei R, et al: Interventional radiology procedures after pancreatic portal and superior mesenteric vein thrombosis, World J Gastroenterol
resections for pancreatic and periampullary diseases, JOP 15(4): 15(40):5028–5034, 2009.
378–382, 2014. Mansueto G, et al: Endovascular treatment of arterial bleeding in
Choi SK, et al: Spontaneous liver rupture in a patient with peliosis patients with pancreatitis, Pancreatology 7(4):360–369, 2007.
hepatis: a case report, World J Gastroenterol 15(43):5493–5497, 2009. McLean A: Endoscopic ultrasound in the detection of pancreatic islet
Christmas AB, et al: Selective management of blunt hepatic injuries cell tumours, Cancer Imaging 4(2):84–91, 2004.
including nonoperative management is a safe and effective strategy, Memeo M, et al: Diagnostic imaging in the study of visceral involve-
Surgery 138(4):606–610, discussion 610–611, 2005. ment of hereditary haemorrhagic telangiectasia, Radiol Med 113(4):
Cohen MS, et al: Prospective study of provocative angiograms to local- 547–566, 2008.
ize functional islet cell tumors of the pancreas, Surgery 122(6):1091– Misselbeck TS, et al: Hepatic angioembolization in trauma patients:
1100, 1997. indications and complications, J Trauma 67(4):769–773, 2009.
Cordova AC, Sumpio BE: Visceral artery aneurysms and Miyayama S, et al: Anastomosis between the hepatic artery and the
pseudoaneurysms—should they all be managed by endovascular extrahepatic collateral or between extrahepatic collaterals: observa-
techniques?, Ann Vasc Dis 6(4):687–693, 2013. tion on angiography, J Med Imaging Radiat Oncol 53(3):271–282,
Davran R, et al: Radiological findings and endovascular management 2009.
of three cases with segmental arterial mediolysis, Cardiovasc Intervent Nikfarjam M, et al: Improved contemporary surgical management of
Radiol 33(3):601–606, 2010. insulinomas: a 25-year experience at the Massachusetts General
Doppman JL, et al: Insulinomas: localization with selective intraarterial Hospital, Ann Surg 247(1):165–172, 2008.
injection of calcium, Radiology 178(1):237–241, 1991. Nosher JL, et al: Visceral and renal artery aneurysms: a pictorial essay
Duffy JP, et al: Vascular complications of orthotopic liver transplanta- on endovascular therapy, Radiographics 26(6):1687–1704, 2006.
tion: experience in more than 4,200 patients, J Am Coll Surg Patel IJ, et al: Addendum of newer anticoagulants to the SIR consensus
208(5):896–903, discussion 903–905, 2009. guideline, J Vasc Interv Radiol 24(5):641–645, 2013.
Eguchi H, et al: Arterial stimulation and venous sampling for gluca- Raikhlin A, et al: Imaging and transcatheter arterial embolization for
gonomas of the pancreas, Hepatogastroenterology 59(113):276–279, traumatic splenic injuries: review of the literature, Can J Surg 51(6):
2012. 464–472, 2008.
Ekeh AP, et al: Complications arising from splenic artery embolization: Rajani RR, et al: Improved outcome of adult blunt splenic injury: a
a review of an 11-year experience, Am J Surg 205(3):250–254, dis- cohort analysis, Surgery 140(4):625–631, discussion 631–632, 2006.
cussion 254, 2013. Riaz A, et al: Complications following radioembolization with yttrium-
Guettier JM, et al: Localization of insulinomas to regions of the pan- 90 microspheres: a comprehensive literature review, J Vasc Interv
creas by intraarterial calcium stimulation: the NIH experience, J Clin Radiol 20(9):1121–1130, quiz 1131, 2009.
Endocrinol Metab 94(4):1074–1080, 2009. Sachdev U, et al: Management of aneurysms involving branches of the
Hemp JH, Sabri SS: Endovascular management of visceral arterial celiac and superior mesenteric arteries: a comparison of surgical and
aneurysms, Tech Vasc Interv Radiol 18(1):14–23, 2015. endovascular therapy, J Vasc Surg 44(4):718–724, 2006.
Hyare H, et al: Multi-section CT angiography compared with digital Schnuriger B, et al: Outcomes of proximal versus distal splenic artery
subtraction angiography in diagnosing major arterial hemorrhage in embolization after trauma: a systematic review and meta-analysis,
inflammatory pancreatic disease, Eur J Radiol 59(2):295–300, 2006. J Trauma 70(1):252–260, 2011.
Jackson JE: Angiography and arterial stimulation venous sampling in Semiz-Oysu A, et al: Interventional radiological management of prehe-
the localization of pancreatic neuroendocrine tumours, Best Pract Res patic obstruction of [corrected] the splanchnic venous system,
Clin Endocrinol Metab 19(2):229–239, 2005. Cardiovasc Intervent Radiol 30(4):688–695, 2007.
Kiemeneij F, et al: A randomized comparison of percutaneous translu- Tacher V, et al: How I do it: Cone-beam CT during transarterial che-
minal coronary angioplasty by the radial, brachial and femoral moembolization for liver cancer, Radiology 274(2):320–334, 2015.
approaches: the access study, J Am Coll Cardiol 29(6):1269–1275, Tuite DJ, et al: Percutaneous transsplenic access in the management of
1997. bleeding varices from chronic portal vein thrombosis, J Vasc Interv
Kim SJ, et al: Surgical outcome of right liver donors in living donor Radiol 18(12):1571–1575, 2007.
liver transplantation: single-center experience with 500 cases, Tulsyan N, et al: The endovascular management of visceral artery
J Gastrointest Surg 16(6):1160–1170, 2012. aneurysms and pseudoaneurysms, J Vasc Surg 45(2):276–283, dis-
Kirby JM, et al: Vascular complications of pancreatitis: imaging and cussion 283, 2007.
intervention, Cardiovasc Intervent Radiol 31(5):957–970, 2008. Turner JJ, et al: Localization of gastrinomas by selective intra-arterial
Kobayashi S, et al: Anatomic variations of hepatic artery and new calcium injection, Clin Endocrinol (Oxf) 57(6):821–825, 2002.
clinical classification based on abdominal angiographic images of Vaidya S, et al: Liver transplantation: vascular complications, Ultrasound
1200 cases, Hepatogastroenterology 61(136):2345–2348, 2014. Q 23(4):239–253, 2007.
Lagana D, et al: Multimodal approach to endovascular treatment of Yekebas EF, et al: Postpancreatectomy hemorrhage: diagnosis and
visceral artery aneurysms and pseudoaneurysms, Eur J Radiol 59(1): treatment: an analysis in 1669 consecutive pancreatic resections,
104–111, 2006. Ann Surg 246(2):269–280, 2007.
CHAPTER 22
Percutaneous biopsy
Anne M. Covey and Lynn A. Brody
OVERVIEW equipment and technique. Optimal results may be achieved

when both fine needle and core biopsy are performed and mate-
Image-guided needle biopsy is the mainstay for diagnosis of rial is considered jointly (Sigel et al, 2011; Stewart et al, 2002).
nonpalpable masses almost anywhere in the body. The indica- To increase the likelihood of a diagnostic biopsy and to
tions for biopsy continue to evolve, and although percutaneous minimize complications, a high-quality imaging study should
biopsy of a given mass is possible, biopsy is not always indi- be available for review both when the biopsy is being planned
cated. As with any invasive procedure, percutaneous biopsy has and when it is performed. In some cases, careful review of
associated risks that should be weighed before the procedure. imaging studies may provide a definitive diagnosis, obviating
Needle biopsy is warranted when the result of the biopsy will the need for biopsy. Review of preprocedure imaging also influ-
influence patient management. These indications include ences selection of the most appropriate modality for guidance
establishing a diagnosis before surgery or in patients who are and patient positioning during the procedure so that potential
not surgical candidates, evaluation of organ dysfunction, and obstacles, such as interposed lung, bowel, or blood vessels, may
obtaining material for receptor or gene mutation evaluation. be anticipated and, optimally, avoided. In addition, knowledge
While we enter the era of personalized medicine, this last indi- of imaging findings allows appropriate discussion of relevant
cation is becoming increasingly important (see Chapter 9). risks when informed consent is obtained. With proper prepro-
Several methods are available to obtain tissue, including cedure imaging, the biopsy can be planned to avoid unusual or
biopsy by percutaneous needle or by endoluminal and transve- unsustainable positions, complex needle angulation, and chal-
nous techniques. Needles ranging in size from 25 to 14 gauge lenging breathing instructions. Further, when performing biop-
are typically used, and outside the central nervous system, a sies under computed tomography (CT) guidance, careful
biopsy can be obtained from almost any abnormality that can planning may help reduce the radiation dose to the patient; in
be imaged. Most needles contain an inner stylet, which is many cases, especially those performed using CT fluoroscopy,
removed when the needle is appropriately positioned immedi- the majority of the patient dose is administered during the
ately before obtaining the specimen; the stylet prevents the localizing/planning stage of the procedure (Sarti et al, 2012).
needle from coring interposed structures and accumulating With good-quality reference imaging, fewer localizing images
nontarget material before optimal placement. Smaller “fine” may be required, and the localizing images may be acquired
needles (25 to 20 gauge) are used to obtain cytologic specimens with lower dose parameters.
or to obtain samples for culture or flow cytometry. Larger Review of imaging studies before biopsy also facilitates tar-
needles (20 to 14 gauge) are used to obtain tissue cores when geting of the most viable area of a mass. While large masses
histologic material is required to assess tissue architecture. outgrow their blood supply, they often become necrotic, and
Common malignancies in which core specimens may be pre- diagnostic material cannot be reliably obtained in a background
ferred for diagnosis include lymphoma, sarcoma, thymoma, of necrosis. Preprocedure contrast-enhanced CT, magnetic
and mesothelioma. Core biopsy is necessary for parenchymal resonance imaging (MRI), and positron-emission tomography
organ biopsy in the setting of organ dysfunction/failure. Core (PET) can allow optimal targeting of areas most likely to yield
material may also be required to diagnose well-differentiated a diagnostic specimen.
hepatocellular carcinoma (HCC) or benign hepatic tumors.
Tissue cores also may be preferred for immunohistochemical BIOPSY TECHNIQUE
staining (IHC) and/or mutational analysis in patients for whom
targeted therapies are being considered; most pathology labo- Fine Needle Aspiration
ratories have IHC validated for formalin-fixed material. In Limited scanning is performed to localize the lesion and select
many laboratories however, fine needle biopsy specimens are an approach; imaging is performed using one or more modali-
sufficient for these purposes. Immunochemistry may be per- ties that will, or may, be used to guide needle placement. The
formed using a cell block prepared from a fine needle biopsy, needle can either be advanced alone or coaxially, after place-
provided there are an adequate number of cells, the appropriate ment of a guiding needle. The guiding needle is typically one
solution is selected for the cells, and appropriate panels are gauge larger. The choice of “bare” needle placement versus use
selected (Fowler & Lachar, 2008); mutational analysis can also of a guiding needle is at the discretion of the operator. Advan-
be performed from cytologic material (Boldrini et al, 2007). A tages of a single-needle technique largely relate to ease of taking
number of different strategies and technologies are being devel- the specimen without having to work through another needle,
oped and used to allow mutational analysis of very small as well as keeping the tract as small as possible. Advantages of
amounts of DNA, further enhancing the utility of fine needle a guiding needle include ease of taking multiple specimens from
biopsy (Kanagal-Shamanna et al, 2014). Different operators the same site and having access to embolize the track if so
and laboratories may also have individual preference based on desired.
403
Either alone, or through a guiding needle, a 25 to 20 gauge are commonly placed in formalin. Occasionally, specimens may
needle is advanced into the target lesion, and real-time (ultra- be sent “fresh” to pathology in saline or on saline-soaked gauze
sound [US], fluoroscopy, MR or CT fluoroscopy) or inter- for special studies. Because cells placed in saline eventually
rupted imaging (conventional CT or MRI) is performed to undergo cell lysis related to osmotic shifts of saline into the cell,
guide and assess needle position (Billich et al, 2008). a specimen in saline needs to be fixed or frozen within a few
Fine needles come in a variety of tip designs. Many physi- hours to avoid deterioration of the tissue sample. Tissue also
cians use a Chiba-style needle, in which the needle tip and inner may be snap frozen for future studies. The preferred method
stylet are beveled. Our preference is a Westcott-style needle: In for processing tissue may vary from institution to institution,
addition to a beveled tip, it has a notch in the sidewall of the and the preference of the pathologists reviewing the material
needle just proximal to the tip; this facilitates the operator’s should be determined before initiating a biopsy. For core biop-
ability to choose more precisely the location from which the sies obtained to evaluate organ parenchyma, compared with
cells will be aspirated. side-notch needles, end-cut needles may yield more diagnostic
When needle position has been confirmed, the needle is samples in terms of number of portal triads or glomeruli (Con-
attached to a disposable syringe. The plunger of the syringe is stantin et al, 2010). In the setting of organ dysfunction or
retracted to apply suction while the needle is moved back and failure, no touch preparation is required; specimens are sent in
forth within the lesion to obtain a sample. The needle is with- formalin or saline, depending on the indication and the prefer-
drawn after suction is released, and the specimen is deposited ence of the pathologist.
on a glass slide. Smears are made from the biopsy specimen, As mentioned earlier, many authors advocate performing
and these may be used for immediate analysis (e.g., Diff-Quik) both fine needle and core biopsy to maximize the diagnostic
or fixed in ethanol for immunohistochemical stains. The resid- yield of every given biopsy (Sigel et al, 2011; Stewart et al,
ual material within the syringe and needle is rinsed in a cell- 2002). Core biopsy may be particularly useful to distinguish a
preserving solution for preparation of a cell block. The solution well-differentiated HCC from nodular regeneration in the
used for the cell block should be selected based on the sus- setting of cirrhosis (see Chapter 76) and in confirmation of
pected diagnosis, the need for special studies, and preference benign diagnoses, including hemangioma, adenoma, and focal
of the cytopathologist. For cases in which non-Hodgkin’s lym- nodular hyperplasia (Kulesza et al, 2004; Kuo et al, 2004) (see
phoma is suspected, a specimen for flow cytometry may be Chapter 90A), although recent literature describes findings that
prepared by rinsing the needle and syringe in a nutrient-rich may increase the diagnostic accuracy of fine needle biopsy for
culture medium (e.g., Roswell Park Memorial Institute HCC versus cirrhotic nodules (Geramizadeh et al, 2012). Yang
medium). When infection is suspected, material can be set and colleagues (2004) described an interesting method of dis-
aside for culture. tinguishing well-differentiated HCC from benign hepatic
In the ideal situation, an on-site cytopathologist or cytotech- lesions by using the physical features of the smear from the fine
nologist can provide an immediate interpretation of the sample; needle aspirates; namely, the smear in malignant lesions will
this has been shown to increase the sensitivity of the biopsy, appear finely granular, whereas a benign lesion would produce
shorten the procedure time, and minimize the number of passes fragments of rigid fine needle cores. The authors attribute this
required to obtain a diagnostic specimen (Nasuti et al, 2002; to the reticulin status of the tissue.
Tsou et al, 2009).
IMAGING GUIDANCE
Core Biopsy
The technique for localizing a lesion is identical for fine needle Ultrasound
and core biopsies. To obtain a good core sample, a target lesion Ultrasound (see Chapter 15) is commonly used to guide per-
should ideally be at least 1 cm in maximum dimension, although cutaneous biopsies of the liver because it is widely available,
we have found that useful core biopsy material can be obtained inexpensive, and portable. US may be used at the bedside or
from even smaller lesions. Because of the typically larger needle in patients in whom CT guidance is impractical. Lesions larger
size (14 to 20 gauge) used to perform a core biopsy, care should than 1 cm are often visible, and US allows real-time visualiza-
be taken to minimize the possibility of traversing medium-sized tion of the needle while it courses from the skin into the lesion.
arteries, which lack the muscular wall of larger arteries and have This is especially helpful in small lesions that move with respira-
an increased tendency to bleed; traversing the colon can result tion and are difficult to target with interrupted imaging modali-
in peritonitis or abscess. Core biopsy needles come in a variety ties. Visualization of smaller-caliber needles may be difficult,
of sizes and styles, including biopsy “guns” and cutting needles although many manufacturers make needles specially designed
that obtain histologic samples manually as the needle is passed to enhance visibility with US.
back and forth within the lesion. Another advantage of US guidance is the capability of mul-
To confirm adequacy of the specimen, a touch preparation tiplanar imaging and the absence of ionizing radiation—a
may be prepared on a glass slide for immediate evaluation by benefit to both the patient and the operator. This is particularly
a cytopathologist or cytotechnologist. Before placement in for- useful in finding a route to lesions at the hepatic dome that
malin or saline, the core of tissue is placed on a glass slide and avoids aerated lung and eliminates the risk of causing pneumo-
gently moved over the slide to allow some cells to collect on thorax. Because CT provides a two-dimensional image,
the surface. Care should be taken to avoid excessive vigorous complex triangulation often is required to accomplish the same
touch preparations, because this has been shown to deplete the result. Doppler imaging is occasionally helpful in performing a
cellularity and DNA content of the specimen (Rekhtman et al, biopsy by enhancing the localization of small lesions. US con-
2015). In addition to the tissue sample, material that is suitable trast agents may help identify the most viable regions of a
for cytology is often collected at the same time, and this tumor. Limitations of US include interoperator variability and
increases the diagnostic yield of the procedure. Core samples poor transmission through air, fat, and bone, which limits the
Chapter 22 Percutaneous biopsy 405
visualization of some lesions. Additionally, patients who are for endoluminal biopsy, when satisfactory decompression of the
sedated for biopsy may not be able to cooperate with breathing biliary tree has been achieved. Biopsy forceps or brush-biopsy
instructions necessary to facilitate visualization of some masses, catheters can be used through the existing tract to obtain tissue
including small lesions or those high in the hepatic dome. samples of suspicious areas (Fig. 22.2). The sensitivity of
forceps biopsy is in the range of 40% to 80%, higher than that
Computed Tomography of brush biopsy, which is in the range of 30% to 60%. Specific-
Computed tomography (see Chapter 18) is a common modality ity for each approaches 98% (Govil et al, 2002; Stewart et al,
for guiding percutaneous biopsies because it provides superb 2001; Weber et al, 2008), and sensitivity is highest for intra-
anatomic detail, which gives the operator the ability to plan a ductal lesions and when biopsy is done in conjunction with
path from skin to lesion using the safest approach, clearly visu- choledochoscopy to provide direct visualization of the lesion
alizing interposed structures. CT is the imaging modality of (Ponchon et al, 1996). Combining forceps and brush biopsy of
choice for biopsies of the lung, pancreas, adrenal glands, the bile duct may provide superior results to either alone.
abdominal and retroperitoneal lymph nodes, and bone and liver Kulaksiz and colleagues (2011), noted sensitivity of brushing
lesions that are not well visualized with US or when biopsy is alone of 49%, forceps alone of 69%, and combined of 80%;
performed by operators who are not skilled with US. specificity for malignancy was 100%. However, a recent report
of a new percutaneous forceps biopsy technique cites sensitivity
Computed Tomographic Fluoroscopy of 93.3%, specificity of 100%, positive predictive value of
Many manufacturers now offer CT fluoroscopy as an option 100%, and negative predictive value of 70%; overall accuracy
on diagnostic scanners. This produces CT images in near real was 94.2% (Patel et al, 2015).
time. It does expose the operator to ionizing radiation and Alternatively, after the biliary tree is opacified, a direct per-
requires that the operator wear lead, as with conventional fluo- cutaneous needle biopsy of a bile duct lesion may be targeted
roscopy, but it is preferred by some physicians, owing to with fluoroscopy using a transhepatic approach (Chawla et al,
decreased time between needle manipulation and image 1989) (see Chapter 30). This technique is most useful for
availability. intrinsic bile duct lesions but may also be used to diagnose
lesions adjacent to the bile duct. With this technique, contrast
Cone-Beam Computed Tomography is injected into an indwelling biliary drainage catheter to delin-
Cone-beam computed tomography (CBCT), also sometimes eate the targeted bile duct abnormality. A needle is advanced
referred to as C-arm CT, uses a flat-panel X-ray detector that through the anterior abdomen to the lesion, and a specimen is
rotates around the patient; the X-rays are divergent, forming a obtained. Confirmation of accurate needle position is made by
cone. Images can be reconstructed in multiple planes, and obtaining oblique fluoroscopic images and by real-time fluoros-
three-dimensional reconstruction can be performed and rotated copy, when the needle is seen to move the duct or the indwell-
in multiple planes. The soft tissue resolution is not comparable ing catheter or both (Fig. 22.3). Fluoroscopy may also be useful
to conventional CT for abdominal and pelvic soft tissue, but to guide percutaneous biopsy of lung nodules and for nontar-
the resolution for lung and bone is adequate for biopsy guid- geted transvenous biopsies of the liver or kidney.
ance. Further, available biopsy path–planning and needle-
navigation software may assist the operator with needle Positron-Emission Tomography
placement (Floridi et al, 2014). Occasionally, a lesion is only well demonstrated by 18F-
fluorodeoxyglucose (FDG) PET imaging. In these cases, it is
Magnetic Resonance Imaging possible to use PET and CT to guide accurate needle place-
Biopsies guided by MRI (see Chapter 19) have been made ment. Certain lesions may also have varying FDG avidity; PET
possible by the advent of open-bore MRI systems that provide guidance allows the most hypermetabolic portion of a lesion to
access to patients during imaging and the availability of nonfer- be targeted. Operators should be mindful of the patient as a
rous biopsy needles and monitoring equipment. The superior source of radiation; the major source of radiation to the opera-
contrast resolution of MR allows targeting lesions that are dif- tor during PET-guided interventions was found to be the time
ficult to visualize with US and noncontrast CT, and the ability spent in close proximity to the patient (Ryan et al, 2013) (see
of MRI to image in any plane enhances the targeting of lesions Chapter 17).
that are not safe to approach or easy to access in the axial plane
(Stattaus et al, 2008) (Fig. 22.1). Owing to high cost and Other Guidance
limited availability, as well as a relative paucity of MRI- New technology is under development that allows fusion of
compatible biopsy needles, MR-guided biopsies are generally multiple modalities. For example, CT and MRI scans can be
reserved for patients whose lesion cannot be seen or targeted overlaid with real-time US images to achieve the clarity of the
with US or CT. In addition, it is necessary to ensure that non- one imaging modality and the real-time visualization capabili-
ferromagnetic, MRI-compatible equipment is on hand and that ties of the other. PET images (see Chapter 17) also can be fused
the patient is able to undergo MRI fluoroscopy. to demonstrate sites of highest metabolic activity. Additionally,
Fluoroscopy is useful for guiding bile duct biopsies. Benign robotic guidance systems have been piloted in an effort to opti-
and malignant biliary strictures (see Chapters 42 and 51) often mize speed and accuracy of needle placement.
have similar cholangiographic appearances and rarely can be
distinguished based on imaging alone (Corvera et al, 2005; BIOPSY OF SPECIFIC SITES
Hadjis et al, 1985). Lesions originating within the duct may be
sampled by either an endoluminal (see Chapter 29) or a direct Liver Biopsy
percutaneous approach (see Chapter 30). Percutaneous tran- Percutaneous liver biopsy is performed to determine the nature
shepatic biliary drainage allows direct access to the biliary tract of focal liver masses, the cause and extent of cirrhosis, or the
A B
C D
FIGURE 22.1. Magnetic resonance (MR)-guided percutaneous liver biopsy. A, Axial T1 MR after contrast administration shows a focal lesion in
segment IV in a patient with hepatitis B. B, Noncontrast computed tomography in the same patient at the same level failed to show the lesion. C,
The lesion was hypermetabolic on fluorodeoxyglucose positron-emission tomography. D, biopsy was successfully performed using MR guidance in
the sagittal plane to avoid the risk of pneumothorax. The biopsy specimen confirmed the diagnosis of hepatocellular carcinoma.
cause of liver dysfunction (see Chapters 76 and 89). Modern diagnoses should be considered in all solitary liver lesions,
cross-sectional imaging techniques, including US, CT, and unless they are known to be new in the setting of a known
MRI, allow percutaneous biopsy of both small and deep lesions cancer or in patients at risk for primary HCC.
to be performed on an outpatient basis with minimal morbidity. HCC may also be diagnosed based on imaging and clinical
With current techniques, almost any liver lesion can be targeted criteria (see Chapter 91). According to the American Associa-
percutaneously; many liver lesions larger than 5 mm are suit- tion for the Study of Liver Diseases (AASLD) guidelines, in
able for biopsy (Yu et al, 2001). Fine needle biopsy of liver patients with cirrhosis, liver lesions identified on screening US
lesions has a sensitivity of 69% to 97% (Ohlsson et al, 2002). that are larger than 1 cm may be diagnosed as HCC based on
a single CT or MR imaging study demonstrating classic find-
Focal Liver Lesions ings of arterial enhancement and portal venous or delayed-
Focal liver lesions may be solitary or multiple. For a solitary phase washout (Bruix & Sherman, 2011); the National
lesion, several benign conditions—cyst, hemangioma, focal Comprehensive Cancer Network guidelines are the same for
nodular hyperplasia, and adenoma—often can be diagnosed lesions 2 cm and larger, but for lesions 1 to 2 cm in size, they
confidently by high-quality cross-sectional imaging, obviating advise that two concordant imaging studies are necessary to
the need for biopsy (Gibbs et al, 2004; Hussain et al, 2004; confirm a diagnosis of HCC (Benson et al, 2009). When the
Kim et al, 2004) (see Chapters 15, 18, and 19). These diagnosis of HCC is considered, and these criteria are not met,
Child B or C cirrhotic patients with tumor-nodes-metastasis

classification stage I to III tumors greater than 3 cm and
α-fetoprotein greater than 200 ng/mL. They also note, however,
that accurate diagnosis of HCC is extremely important because
confirmation of diagnosis alters the priority for liver transplan-
tation (Rockey et al, 2009).
Multifocal solid liver lesions most commonly represent met-
astatic disease. In such cases, biopsy may be requested to (1)
confirm the presence of metastatic disease in a patient with a
known primary, (2) establish tumor type and stage simultane-
ously at initial presentation, (3) acquire tissue for genetic analy-
sis, and (4) obtain required samples for patients undergoing
experimental therapies.
Liver Parenchyma Biopsy

Core liver biopsy is used to grade and stage liver disease in
patients with abnormal liver function studies, chronic hepatitis,
and known or suspected cirrhosis (see Chapters 9D and 76).
Gastroenterologists have historically performed most liver biop-
sies without imaging guidance; however, unusual anatomy,
A obesity, and other exigencies occasionally make imaging-guided
biopsy advisable. Adequate tissue cores can be obtained with
needles 20 gauge and larger, although needles 18 gauge and
larger provide a more generous specimen for analysis (Guido
& Ruggae, 2004; Maharaj et al, 1986). Schiano and colleagues
(2005) further suggest that specimens of at least 1 cm in length
are preferred. Biopsy may also be performed in transplanted
livers and in living donors before transplant. Indications for this
last group are controversial. Herrero and colleagues (2014)
recommend following the recommendations of the Vancouver
Forum (Barr et al, 2006) in performing biopsy in potential
donors who have a clinical or imaging-based reason to do so,
but not as a matter of routine.
Transvenous Biopsy
Transjugular liver biopsy is a useful alternative to percutaneous
biopsy in patients with coagulopathy or when hepatic venous
pressure measurements are required (Behrens & Ferral, 2012;
Garcia-Compean & Cortes, 2004). Although transvenous
biopsy is seemingly more invasive than percutaneous biopsy,
the risk of significant bleeding in patients with coagulopathy is
minimized using a transvenous approach, because bleeding
B from the biopsy site tracks back into the venous circulation and
FIGURE 22.2. Biliary brush biopsy. A, Transhepatic cholangiogram
not into the abdominal cavity as can happen with percutaneous
shows a high bile duct stricture (arrow) Prior brush biopsy was nondi- biopsies. Proper technique to avoid puncture of the liver capsule
agnostic. B, A brush (curved arrow) was advanced to the stricture and is crucial to optimize safety.
used to obtain a specimen for cytology. The specimen was diagnostic In this technique, a venous sheath is introduced into a
of cholangiocarcinoma. hepatic vein, most commonly the right hepatic vein, typically
from a right internal jugular approach. Pressure measure-
ments may be obtained to evaluate the source (presinusoidal,
biopsy may be required. Smaller, encapsulated tumors are more sinusoidal, or postsinusoidal) and degree of portal hyperten-
likely to be well differentiated, and tissue cores may be required sion. A biopsy needle is introduced through the sheath into
to distinguish a well-differentiated tumor from normal or cir- the liver parenchyma to obtain histologic samples (Fig. 22.4).
rhotic liver. Because of the risk of tumor seeding (Chaudhry Care must be taken to avoid performing the biopsy in too
et al, 2004; Durand et al, 2007; Perkins, 2007), when a cura- peripheral an area of liver, because this increases the risk that
tive treatment is possible, biopsy for HCC should be performed the liver capsule will be punctured and that peritoneal hemor-
only after consultation with a hepatobiliary surgeon and after rhage will result.
referencing the most current imaging and clinical criteria. The Transvenous biopsy is only used for nontargeted parenchy-
AASLD position paper on liver biopsy reinforces caution based mal biopsy; it cannot be used for biopsy of a focal lesion.
on concern for needle-track seeding, sampling error, and, based Transvenous biopsy of the kidney also has been reported and
on a paper by Saborido and colleagues (2005), possible may be useful for evaluation of renal dysfunction in coagulo-
increased rate of recurrence posttransplant, although only in pathic patients (Misra et al, 2008). This may be particularly
A B
C
FIGURE 22.3. Percutaneous bile duct biopsy. A, Transhepatic cholangiogram shows a high bile duct stricture (marked by the clamp) of uncertain
etiology. B, Under fluoroscopic guidance, a 22 gauge needle was used to target the stricture. C, Oblique imaging confirmed the position of the
needle at the site of the bile duct stricture.
convenient in coagulopathic patients who require both paren- adenomas occur in 5% of individuals (Abecassis et al, 1985;
chymal hepatic and renal biopsies. Libe et al, 2002). Adrenal biopsy can often be avoided,
because MRI or adrenal protocol CT can sometimes distin-
Biopsy of Other Organs guish between an adenoma and a metastasis (Davarpanah &
Adrenal Biopsy Israel, 2014; National Institutes of Health, 2002). In some
The adrenal gland is a common site of metastatic disease. cases, the mass remains indeterminate, and biopsy is indi-
This gland also is the site of many benign neoplasms, and cated. Because of the posterior approach used for all left and
some right adrenal biopsies, it is common to traverse the lung,

introducing the possibility of a pneumothorax. The right
adrenal gland can alternatively be targeted from a lateral,
transhepatic approach to eliminate the risk of pneumothorax
(Fig. 22.5). Some authors advocate placing patients in the
ipsilateral decubitus position for adrenal biopsy, to take
advantage of diminished aeration of the lower lobe of the lung
in this position and decrease the likelihood of either putting
the needle through lung, or using an out-of-plane trajectory to
avoid lung (Odisio et al, 2012) (see Fig. 22.5C).
Pancreas Biopsy
Despite the best efforts of clinicians and advances in imaging,
most patients with pancreatic cancer still initially present with
unresectable disease (see Chapter 59). Needle biopsy may be
requested for tissue diagnosis so that treatment can be initiated,
and it is essential to evaluate the preprocedure imaging carefully
to determine the safest approach. Increasingly, endoscopic
ultrasound (EUS) has become an excellent alternative to per-
cutaneous biopsy (see Chapter 16). In cases where in which
EUS-guided biopsy is not feasible or adequate material is not
obtained, percutaneous biopsy may still be necessary. An ante-
rior approach to the pancreas may be difficult because of inter-
posed colon, spleen, or mesenteric vessels, which we would
prefer not to traverse. We often prefer a posterior, transcaval
approach (Sofocleous et al, 2004) or a transhepatic approach,
particularly for lesions in the head or uncinate process
(Fig. 22.6).
Retroperitoneal and Pelvic Biopsy

A Most retroperitoneal and pelvic biopsies are performed to
determine the cause of lymphadenopathy. When the patient has
a known malignancy with the proclivity to spread via retroperi-
toneal lymphatics, needle biopsy is indicated, and this is gener-
ally easily and simply performed. When lymphoma is a
consideration, biopsy specimens should be obtained either for
flow cytometry and/or surgical pathology (cores), depending on
local practice and expertise.
Occasionally, a biopsy is performed to diagnose a soft tissue
mass suspected to be a sarcoma. In patients who are candidates
for resection, the approach to the lesion should be discussed
with the surgeon, as they may wish to resect the needle path to
minimize the risk of local recurrence. For primary diagnosis of
sarcoma, histologic material is useful for classifying the sarcoma,
and core biopsy should be performed. If recurrence is the issue,
the diagnosis usually can be established on the basis of cytology
alone. In women with pelvic masses that might be adnexal,
needle biopsy should be performed only after the diagnosis of
ovarian cancer has been excluded or after an oncologic gyne-
cologist has been consulted. The biopsy procedure itself
may result in peritoneal contamination, relegating the patient
to intraperitoneal chemotherapy, when a simpler treatment
regimen might have been possible. In addition to previously
described guidance modalities, deep pelvic pathology is some-
times best approached using transrectal or transvaginal US
guidance.
Lung and Mediastinum Biopsy

B Lung biopsies can be performed with CT, CT fluoroscopy,
FIGURE 22.4. Transjugular liver biopsy. A, Right hepatic venogram CBCT, or conventional fluoroscopy. The size, location, and
through a sheath placed via the right internal jugular vein. B, A 19-gauge nature of the target lesion, as well as operator experience, and
automated flexible biopsy needle with a 20-mm throw (arrow) is available equipment determine which modality is used.
advanced through a metal cannula and directed anteriorly as the biopsy
specimen is obtained.
A B
C D
FIGURE 22.5. Adrenal biopsy. A, Transhepatic right adrenal mass (arrows) biopsy eliminates the risk of pneumothorax. B, Coronal reformat of a
contrast enhanced computed tomography shows a right adrenal mass (arrows). C, With the patient positioned prone, there is interposition of lung
between the skin and adrenal mass (arrows). D, With the patient in a right-side-down position, there is no need to traverse aerated lung for biopsy
of the adrenal mass (arrows).
Mediastinal biopsies are usually performed using CT guidance. a tyrosine kinase inhibitor compared with chemotherapy in
The size of the needle used and type of specimen acquired vary nonsmokers or former light smokers (Mok et al, 2009). Patients
depending on the clinical indication. Diagnostic rates for needle with adenocarcinoma and an EGFR mutation were also shown
biopsy of malignant lung lesions have been reported as greater to do better when a tyrosine kinase inhibitor was used as main-
than 90% (Swischuk et al, 1998). Lung biopsy may be useful tenance after first-line treatment rather than as second-line
to provide material in the setting of metastatic disease, primary therapy (Wang et al, 2011). The KRAS and ALK mutations
malignancy, and/or infection. Personalized medicine is already have also been shown to influence response to treatment (Eber-
making its mark in primary lung cancer. Not long ago, all cases hard et al, 2005; Kwak et al, 2010).
of non–small-cell lung cancers were treated in a similar fashion.
However, recent studies have shown that histologic subtype and Bone Biopsy
certain molecular alterations influence the response to various Bone biopsies are often done using CT guidance. MRI and
chemotherapies and targeted agents (Moreira et al, 2012; fluoroscopy may also be used. More recently, operators have
Travis et al, 2010). For example, bevacizumab, a humanized described excellent results using CBCT with needle-guiding
monoclonal antibody targeted at endothelial growth factor, is software (Tselikas et al, 2015). Bone biopsies are typically per-
contraindicated in patients with squamous cell lung carcinoma formed using larger needles (11 to 15 gauge) than those used
due to increased risk of pulmonary hemorrhage (Johnson et al, for organ or soft tissue. Purely lytic lesions are the exception,
2004). Patients with adenocarcinoma also respond better to the because small-caliber needles will often suffice to document
antifolate agent pemetrexed than those with squamous carci- metastases and are able to penetrate bone in the setting of
noma (Scagliotti et al, 2008), and patients with adenocarci- cortical destruction. For surgical candidates, the biopsy path
noma and an epidermal growth factor receptor gene (EGFR) should be discussed with the surgeon to ensure that violating
mutation have been shown to have a better response to an uninvolved compartment does not compromise a planned
carefully. For patients with an elevated INR, transfusion of

plasma or supplementation with vitamin K is required. Ele-
vated partial thromboplastin time (PTT) in patients not on
heparin usually is due to circulating antiphospholipid cardio-
lipins and is not typically clinically significant. A test of bleeding
time may be performed to evaluate the significance of an ele-
vated PTT in select patients. It is advisable to have patients
stop antiplatelet medications, if possible, to minimize the risk
of bleeding; however, the risk of stopping antiplatelet therapy
must be weighed against the risk of bleeding from the biopsy;
occasionally, the balance favors performing the biopsy while the
patient remains on his or her regular medication(s).
The risk of significant bleeding after liver biopsy is less than
1% (Piccinino et al, 1986). In many cases, the bleeding is self-
limited, and conservative management comprising observation
and hydration will suffice. In patients with more significant
A bleeding and/or symptomatic hemorrhage, hepatic angiogra-
phy and potential embolization of an injured artery may be
required (Fig. 22.7). Some authors advocate placing absorb-
able gelatin sponge (Gelfoam) pledgets in the biopsy tract
through the needle after core biopsy, but this has not been
shown definitively to decrease the risk of major bleeding (Hat-
field, 2008).
Pain out of proportion to imaging findings after liver biopsy
may be due to bile peritonitis (Ruben & Chopra, 1987). Care
should be taken to minimize needle passes through the gallblad-
der, cystic duct, or dilated bile ducts. If the gallbladder is
inadvertently punctured, it should be aspirated as completely
as possible before removing the needle. Bile leaks resulting in
discernible collections are rare after liver biopsy in the absence
of downstream biliary obstruction.
Adrenal masses and lesions in the dome of the liver some-
times require an approach for biopsy that crosses the lung base,
putting patients at risk for pneumothorax. The two most
B common complications after lung biopsy include hemoptysis
FIGURE 22.6. Transcaval pancreatic biopsy. A, Contrast-enhanced and pneumothorax (Covey et al, 2004). Hemoptysis occurs in
computed tomography scan shows a cystic mass (arrow) in the head approximately 10% of patients who undergo lung biopsy and
of the pancreas in a patient with a history of non–small-cell lung cancer. is usually self-limited, but it can be frightening to the patient.
B, Transcaval biopsy performed in the prone position with a 22 gauge Pneumothorax occurs in 20% to greater than 40% of patients
needle confirmed the diagnosis of metastasis. The inferior vena cava is after biopsy and requires placement of a chest tube in approxi-
indicated by the arrows. mately 6% to 12% of cases. The risk of pneumothorax is typi-
cally related more to patient than technical factors, although
depth of the target lesion, number of pleural surfaces trans-
procedure. Liu and colleagues (2007) published useful ana- gressed, and patient positioning (prone positioning decreases
tomic guidelines for biopsy of bone tumors and further stress the risk of pneumothorax) have been shown to affect the likeli-
the need to consult with the surgeon to maximize each patient’s hood. Elderly patients and patients with underlying chronic
chance of getting optimal surgical treatment. obstructive pulmonary disease are more prone to pneumotho-
rax requiring treatment (Covey et al, 2004; Hiraki et al, 2010;
COMPLICATIONS OF PERCUTANEOUS BIOPSY Takeshita et al, 2015). A symptomatic or enlarging pneumo-
thorax is treated with a small-bore chest tube (generally 8 to
Any invasive procedure has risks. Risks of biopsy include bleed- 12 Fr) and often necessitates hospital admission. A number of
ing, pneumothorax, infection, bile leak, and needle-tract methods have been described in an attempt to decrease the
seeding of tumor. The risks vary depending on the organs incidence of pneumothorax requiring chest tube placement
involved. To minimize the risk of bleeding, we advocate that after percutaneous lung biopsy. These include autologous
all patients should have appropriate laboratory work before blood patch injection into the needle track (Herman & Weis-
biopsy, including a complete blood count and coagulation brod, 1990), embolization of the needle track using gelatin
profile. Although criteria differ from institution to institution sponge slurry (Tran et al, 2014), rapidly placing a patient in a
and from physician to physician, a platelet count of greater than “puncture-side-down” position following removal of the biopsy
50,000 µL and an international normalized ratio (INR) less guiding needle (Kim et al, 2015), and use of commercial
than 1.5 are acceptable in most cases. Biopsy in thrombocyto- track plugs, among others. Although some studies showed a
penic patients can be performed with platelet coverage, although lower rate of pneumothorax requiring chest tubes in the “inter-
the decision to proceed with biopsy should be considered vention group,” others did not; there is no consensus regarding
which, if any, technique may be most useful in decreasing the

rate of pneumothorax related to percutaneous lung biopsy.
Hemorrhagic pericardial tamponade is a rare, potentially life-
threatening complication after mediastinal biopsy (Kucharczyk
et al, 1982). Although hypoxemia may be a feature, this com-
plication can be distinguished from iatrogenic pneumothorax
clinically by the development of hypotension with narrowing of
the pulse pressure and diminished amplitude of the electrocar-
diogram complex on the monitor. The diagnosis can be con-
firmed immediately by scanning the heart and pericardium, and
it can be treated by directly placing a drainage catheter into the
pericardial space.
Needle-tract seeding after percutaneous biopsy is a worri-
some complication. The interval between biopsy and appear-
ance of a tract metastasis is 6 to 24 months (Kosugi et al, 2004;
A Schotman et al, 1999). The risk overall is likely greater than
reported, but current understanding is that the risk in HCC is
in the range of 2% to 3% (Perkins, 2007; Stigliano et al, 2007).
Although the incidence is relatively small, the possibility of
rendering a patient ultimately incurable because of tract or
peritoneal seeding should be considered in the risk/benefit
analysis for each patient. For this reason, some surgeons do not
advocate percutaneous biopsy for patients with potentially
resectable lesions highly suspicious for malignancy (Al-Leswas
et al, 2008; Cha et al, 2002); this sentiment is particularly
strong among transplant surgeons in regard to patients with
HCC or other malignant disease undergoing evaluation for a
new liver graft (see Chapter 115).
CONCLUSION
B Percutaneous needle biopsy is a well-established and safe diag-
nostic tool, and it is an important instrument in the diagnosis
of tumors, for determining the cause of organ dysfunction
staging, and for documenting recurrent or metastatic disease.
Increasingly, needle biopsy is essential to provide material for
genetic analysis. Complications occur infrequently, and most
are easily treated or self-limiting. Tract seeding has been
reported but occurs infrequently. There is no such thing as a
“negative” biopsy (Phillips et al, 1998). If a diagnosis of malig-
nancy is not made, a specific benign diagnosis needs to be
confirmed. If nonspecific findings are evident on cytology,
including inflammatory or reactive changes, fibrous tissue, or
normal site tissue, or, if atypical cells are present, another
biopsy should be performed, or the lesion should be closely
followed up, depending on the pretest probability of disease.
The role of biopsy in patient management is evolving in
tandem with the development of associated fields, including
functional and molecular imaging. Until biopsies are no longer
necessary, every effort should be made to keep morbidity low
and diagnostic rates high.
C
FIGURE 22.7. Hemorrhage complicating liver biopsy. A, Contrast-
enhanced computed tomographic scan 5 days after liver biopsy shows
a large subcapsular hematoma (curved arrow) with active extravasation
of contrast from the liver mass (curved arrow). B, Angiography shows
active extravasation (curved arrow) from peripheral segment VIII artery
that was successfully embolized with stainless steel coils (C).
Chapter 22 Percutaneous biopsy 412.e1
REFERENCES Hiraki T, et al: Incidence and risk factors for pneumothorax and chest
tube placement after CT fluoroscopy-guided percutaneous lung
Abecassis M, et al: Serendipitous adrenal masses: prevalence, signifi- biopsy: retrospective analysis of the procedures conducted over a
cance and management, Am J Surg 149:783–788, 1985. 9-year period, AJR Am J Roentgenol 194(3):809–814, 2010.
Al-Leswas D, et al: Biopsy of solid liver tumors: adverse consequences, Hussain SM, et al: Focal nodular hyperplasia: findings at state-of-the-
Hepatobiliary Pancreat Dis Int 7(3):325–327, 2008. art MR imaging, US, CT and pathologic analysis, Radiographics
Barr ML, et al: A report of the Vancouver Forum on the Care of the 24:3–17, 2004.
Live Organ Donor: lung, liver, pancreas, and intestine data and Johnson DH, et al: Randomized phase II trial comparing bevacizumab
medical guidelines, Transplantation 81(10):1373–1385, 2006. plus carboplatin and paclitaxel with carboplatin and paclitaxel alone
Behrens G, Ferral H: Transjugular liver biopsy, Semin Intervent Radiol in previously untreated locally advanced or metastatic non-small-cell
29(2):111–117, 2012. lung cancer, J Clin Oncol 22(11):2184–2191, 2004.
Benson AB, et al: NCCN clinical practice guidelines in oncology: hepa- Kanagal-Shamanna R, et al: Mod Pathol 27(2):314–327, 2014.
tobiliary cancers, J Natl Compr Canc Netw 7(4):350–391, 2009. Kim J, et al: An algorithm for the accurate identification of benign liver
Billich C, et al: CT-guided lung biopsy: incidence of pneumothorax lesions, Am J Surg 187:274–279, 2004.
after instillation of NaCl into the biopsy track, Eur Radiol 18(6):1146– Kim JI, et al: Rapid needle-out patient-rollover approach after cone
1152, 2008. beam CT-guided lung biopsy: effect on pneumothorax rate in 1,191
Boldrini L, et al: Mutational analysis in cytolological specimens of consecutive patients, Eur Radiol 25(7):1845–1853, 2015.
advanced lung adenocarcinoma: a sensitive method for molecular Kosugi C, et al: Needle-tract implantation of hepatocellular carcinoma
diagnosis, J Thorac Oncol 2(12):1086–1090, 2007. and pancreatic carcinoma after ultrasound-guided percutaneous
Bruix J, Sherman M: Management of hepatocellular carcinoma: an puncture: clinical and pathologic characteristics and the treatment
update, Hepatology 53(3):1020–1022, 2011. of needle-tract implantation, World J Surg 28:29–32, 2004.
Cha C, et al: Surgery and ablative therapy for hepatocellular carci- Kucharczyk W, et al: Cardiac tamponade as a complication of thin-
noma, J Clin Gastroenterol 35(Suppl 2):S130–S137, 2002. needle aspiration lung biopsy, Chest 82:120–121, 1982.
Chaudhry SI, et al: Preoperative biopsy of hepatocellular carcinoma Kulaksiz H, et al: A novel method of forceps biopsy improves the
renders 1 in 5 operable cases incurable [abstract], Br J Surg 91(Suppl diagnosis of proximal biliary malignancies, Dig Dis Sci 56:596–601,
1):103, 2004. 2011.
Chawla S, et al: Cholangiographically guided aspiration cytology in the Kulesza P, et al: Cytopathologic grading of hepatocellular carcinoma
management of malignant biliary obstruction, Indian J Gastroenterol on fine-needle aspiration, Cancer 102:247–258, 2004.
8:95–96, 1989. Kuo FY, et al: Fine needle aspiration cytodiagnosis of liver tumors,
Constantin A, et al: Percutaneous US guided renal biopsy: a retrospec- Acta Cytol 48:142–148, 2004.
tive study comparing and 16 gauge end-cut and 14 gauge side notch Kwak EL, et al: Anaplastic lymphoma kinase inhibition in non-small-
needles, J Vasc Interv Radiol 21(3):357–361, 2010. cell lung cancer, N Engl J Med 363(18):1693–1703, 2010.
Corvera CU, et al: Clinical and pathologic features of proximal biliary Libe R, et al: Long-term follow up study of patients with adrenal inci-
strictures masquerading as hilar cholangiocarcinoma, J Am Coll Surg dentalomas, Eur J Endocrinol 147:489–494, 2002.
201:862–869, 2005. Liu PT, et al: Anatomically based guidelines for core needle biopsy of
Covey AM, et al: Factors associated with pneumothorax following bone tumors:implications for limb-sparing surgery, Radiographics
percutaneous lung biopsy in 443 consecutive patients, J Vasc Interv 27:189–206, 2007.
Radiol 15:479–483, 2004. Maharaj V, et al: Sampling variability and its influence on the diagnos-
Davarpanah AH, Israel GM: MR imaging of the kidneys and adrenal tic yield of percutaneous needle biopsy of the liver, Lancet 8:523–
glands, Radiol Clin North Am 52(4):779–798, 2014. 535, 1986.
Durand F, et al: Liver transplantation for hepatocellular carcinoma: Misra S, et al: Safety and diagnostic yield of transjugular renal biopsy,
role of biopsy, Liver Transpl 13:S17–S23, 2007. J Vasc Interv Radiol 19(4):546–551, 2008.
Eberhard DA, et al: Mutations in the epidermal growth factor receptor Mok TS, et al: Gefitinib or carboplatin-paclitaxel in pulmonary adeno-
and in KRAS are predictive and prognostic indicators in patients carcinoma, N Engl J Med 361(10):947–957, 2009.
with non-small-cell lung cancer treated with chemotherapy alone Moreira AL, et al: Personalized medicine for non-small-cell lung
and in combination with erlotinib, J Clin Oncol 23(25):5900–5909, cancer: implications of recent advances in tissue acquisition for
2005. molecular and histologic testing, Clin Lung Cancer 13(5):334–339,
Floridi C, et al: C-arm cone-beam computed tomography needle path 2012.
overlay for percutaneous biopsy of pulmonary nodules, Radiol Med Nasuti JF, et al: Diagnostic value and cost-effectiveness of on-site
199:820–827, 2014. evaluation of fine-needle aspiration specimens: review of 5688 cases,
Fowler LJ, Lachar WA: Application of immunohistochemistry to cytol- Diagn Cytopathol 27:1–4, 2002.
ogy, Arch Pathol Lab Med 132:373–383, 2008. National Institutes of Health: NIH state-of-the-science statement on
Garcia-Compean D, Cortes C: Transjugular liver biopsy: an update, management of the clinically inapparent adrenal mass (“inciden-
Ann Hepatol 3:100–103, 2004. taloma”), NIH Consens State Sci Statements 19:1–25, 2002.
Geramizadeh B, et al: Cytologic comparison between malignant and Odisio BC, et al: CT-guided adrenal biopsy;comparison of ipsilateral
regenerative nodules in the background of cirrhosis, Hepat Mon decubitus versus prone positioning for biopsy approach, Eur Radiol
12(7):448–452, 2012. 22(6):1233–1239, 2012.
Gibbs JF, et al: Contemporary management of benign liver tumors, Ohlsson B, et al: Percutaneous fine-needle aspiration cytology in the
Surg Clin North Am 84:463–480, 2004. diagnosis and management of liver tumours, Br J Surg 89:757–762,
Govil H, et al: Brush cytology of the biliary tract: retrospective study 2002.
of 278 cases with histopathologic correlation, Diagn Cytopathol 26: Patel P, et al: Improved accuracy of percutaneous biopsy using
273–277, 2002. “cross and push” technique for patients with suspected malignant
Guido M, Ruggae M: Liver biopsy sampling in chronic viral hepatitis, biliary stricutures, Cardiovasc Intervent Radiol 38(4):1005–1010,
Semin Liver Dis 24:89–97, 2004. 2015.
Hadjis NS, et al: Malignant masquerade at the hilum of the liver, Br J Perkins JD: Seeding risk following percutaneous approach to hepatocel-
Surg 72:659–661, 1985. lular carcinoma, Liver Transpl 13(11):1603, 2007.
Hatfield MK: Percutaneous imaging-guided solid organ core needle Phillips MS, et al: Negative predictive value of imaging-guided abdom-
biopsy: coaxial versus noncoaxial method, AJR Am J Roentgenol inal biopsy results: cytologic classification and implications for
190(2):413–417, 2008. patient management, AJR Am J Roentgenol 171:693–696, 1998.
Herman SJ, Weisbrod GL: Usefulness of the blood patch technique Piccinino F, et al: Complications following percutaneous liver biopsy:
after transthoracic needle aspiration biopsy, Radiology 176(2):395– a multicentre retrospective study on 68,276 biopsies, J Hepatol
397, 1990. 2:165–173, 1986.
Herrero JI, et al: Is liver biopsy necessary in the evaluation of a living Ponchon T, et al: Methods, indications and results of percutaneous
donor for liver transplantation?, Transplant Proc 46(9):3082–3083, choledochoscopy: a series of 161 procedures, Ann Surg 223:26–36,
2014. 1996.
Rekhtman N, et al: Depletion of core needle biopsy cellularity and Stewart CJ, et al: Comparison of fine needle aspiration cytology and
DNA content as a result of vigorous touch preparations, Arch Pathol needle core biopsy in the diagnosis of radiologically detected abdom-
Lab Med 139(7):907–912, 2015. inal lesions, J Clin Pathol 55(2):93–97, 2002.
Rockey DC, et al; American Association for the Study of Liver Disease: Stigliano R, et al: Seeding following percutaneous diagnostic and thera-
Liver biopsy, Hepatology 49(3):1017–1044, 2009. peutic approaches for hepatocellular carcinoma. What is the risk and
Ruben RA, Chopra S: Bile peritonitis after liver biopsy: nonsurgical the outcome? Seeding risk for percutaneous approach to HCC,
management of a patient with an acute abdomen: a case report with Cancer Treat Rev 33(5):437–447, 2007.
review of the literature, Am J Gastroenterol 82:265–268, 1987. Swischuk JL, et al: Percutaneous transthoracic needle biopsy of the
Ryan ER, et al: PET/CT-guided interventions: personnel radiation lung: review of 612 lesions, J Vasc Interv Radiol 9:347–352, 1998.
dose, Cardiovasc Intervent Radiol 36(4):1063–1067, 2013. Takeshita J, et al: CT-guided fine-needle aspiration and core needle
Saborido BP, et al: Does preoperative fine needle aspiration-biopsy biopsies of pulmonary lesions: a single-center experience with 750
produce tumor recurrence in patients following liver transplantiation biopsies in Japan, AJR Am J Roentgenol 204(1):29–34, 2015.
for hepatocellular carcinoma?, Transplant Proc 37:3874–3877, 2005. Tran AA, et al: Tract embolization with gelatin sponge slurry for pre-
Sarti M, et al: Low-dose techniques in CT-guided interventions, Radio- vention of pneumothorax after percutaneous computed tomography-
graphics 32:1109–1119, 2012. guided lung biopsy, Cardiovasc Intervent Radiol 37(6):1546–1553,
Scagliotti GV, et al: Phase III study comparing cisplatin plus gem- 2014.
citabine with cisplatin plus pemetrexed in chemotherapy-naive Travis WD, et al: Pathologic diagnosis of advanced lung cancer based
patients with advanced-stage non-small-cell lung cancer, J Clin Oncol on small biopsies and cytology: a paradigm shift, J Thorac Oncol
26(21):3543–3551, 2008. 5(4):411–414, 2010.
Schiano TD, et al: Importance of specimen size in accurate needle liver Tselikas L, et al: Percutaneous bone biopsies: comparison between
biopsy evaluation of patients with chronic hepatitis C, Clin Gastro- flat-panel cone beam CT and CT-scan guidance, Cardiovasc Inter-
enterol Hepatol 3:930–935, 2005. vent Radiol 38:167–176, 2015.
Schotman SN, et al: Subcutaneous seeding of hepatocellular carci- Tsou M-H, et al: CT-guided needle biopsy:Value of on-site cytopatho-
noma after percutaneous needle biopsy, Gut 45:626–627, 1999. logic evaluation of core specimen touch preparations, J Vasc Interv
Sigel CS, et al: Subtyping of non-small cell lung carcinoma: a compari- Radiol 20:71–76, 2009.
son of small biopsy and cytology specimens, J Thorac Oncol Wang ZJ, et al: Immediate versus delayed treatment with EGFR tyro-
6(11):1849–1856, 2011. sine kinase inhibitors after first-line therapy in advanced non-small-
Sofocleous CT, et al: CT-guided transvenous or transcaval needle cell lung cancer, Chin J Cancer Res 23(2):112–117, 2011.
biopsy of pancreatic and peripancreatic lesions, J Vasc Interv Radiol Weber A, et al: Diagnostic approaches for cholangiocarcinoma, World
15:1099–1104, 2004. J Gastroenterol 14(26):4131–4136, 2008.
Stattaus J, et al: MR-guided core biopsy with MR fluoroscopy using Yang GC, et al: Distinguishing well-differentiated hepatocellular carci-
a short, wide-bore 1.5 Tesla scanner: feasibility and initial results, noma from benign liver by the physical features of fine-needle aspi-
J Magn Reson Imaging 27(5):1181–1187, 2008. rates, Mod Pathol 17:798–802, 2004.
Stewart CJ, et al: Brush cytology in the assessment of pancreatico- Yu SC, et al: US-guided percutaneous biopsy of small (<1 cm) hepatic
biliary strictures: a review of 406 cases, J Clin Pathol 54:449–455, lesions, Radiology 218:195–199, 2001.
2001.
CHAPTER 23
Intraoperative diagnostic techniques
Kelly M. Collins and M. B. Majella Doyle
OVERVIEW 35% more hepatic lesions than preoperative imaging studies.

D’Hondt and colleagues (2011), in a study of 418 patients with
The management of hepatobiliary and pancreatic disease has liver malignancy evaluated with MRI and/or CT and IOUS,
evolved considerably since the 1990s. Today’s operative surgeon found that IOUS was responsible for a change in operative
has a number of valuable adjuncts to aid in accurate preopera- strategy in 69 patients (16.5%), with additional lesions
tive evaluation and planning for benign and malignant disease. being responsible for the change 11.2% of the time. The cor-
Advances in contrast-enhanced computed tomography (CT) relation rate with pathology for each modality is as follows:
(see Chapter 18), endoscopic ultrasound (EUS) (see Chapter 8-slice multidetector (MDCT), 0.627; 64-slice MDCT, 0.785;
16), magnetic resonance imaging (MRI) (see Chapter 19), and contrast-enhanced MRI, 0.657; and IOUS, 0.913 (D’Hondt
endoscopic retrograde cholangiopancreatography (ERCP) (see et al, 2011). The yield of IOUS is dependent on the type and
Chapters 20 and 29) have revolutionized the treatment of quality of preoperative imaging obtained. Even in cases where
benign and malignant hepatobiliary and pancreatic disease. additional intraoperative findings are made, it may not affect
Despite these techniques, present-day management of these the planned resection (Jarnagin et al, 2001). Regardless, it is
conditions relies on the appropriate use of intraoperative diag- an integral adjunct to parenchymal assessment and operative
nostic modalities, particularly when confronted with challeng- planning.
ing and unanticipated findings during surgical exploration. Several reports have documented increased resectability
This chapter will review the use of the following intraoperative rates in patients with hepatobiliary malignancy evaluated with
diagnostic modalities in hepatobiliary and pancreatic surgery: IOUS (Jarnagin et al, 1999; Lo et al, 1998; Soyer et al, 1993).
ultrasound, cholangiography, and laparoscopy. More recent studies estimate that the use of IOUS will lead to
changes in 15% to 25% of procedures (Jarnagin et al, 2001;
INTRAOPERATIVE ULTRASONOGRAPHY Kruskal & Kane, 2006; Zacherl, 2002).
When evaluating the liver parenchyma, the sonographer
Intraoperative ultrasound (IOUS) is commonly used during must first evaluate the extent of intrahepatic disease and then
both open and laparoscopic procedures. It can be a valuable assess for vascular occlusion or invasion. Sonographic features
tool in the assessment of the hepatic and biliary anatomy, local- of liver tumors vary. The surgeon should anticipate the need
ization of tumors, determination of the extent of and/or resect- for assistance in the case of ambiguous or challenging lesions
ability of disease, and determination of the extent of mesenteric and communicate with radiologists preoperatively so that per-
vasculature involvement of pancreatic tumors (Jakimowicz, sonnel and equipment are readily available.
1993). It may also be used in the evaluation of biliary calculi. The sonographer should be familiar with the characteristics
of common parenchymal lesions (see Chapters 15 and 48).
Hepatic Disease Hemangiomas are typically soft when palpated and either lack
Evaluation of the Liver any visible flow or demonstrate minimal flow compared with
IOUS of the liver was first introduced into clinical practice in the adjacent liver parenchyma. They have uniform hyperecho-
the early 1980s and rapidly became routine practice for man- genicity and well-defined margins (Moody & Wilson, 1993).
agement of malignant liver disease (Belghiti et al, 1986; There may be posterior echo enhancement. Colorectal liver
Bismuth & Castaing, 1985; Bloed et al, 2000; Gouillat et al, metastases are usually hyperechoic or isoechoic with adjacent
1992; Gozzetti et al, 1986; Gunven et al, 1985; Makuuchi liver parenchyma and are frequently surrounded by an ill-
et al, 1981; Nagasue et al, 1984; Olsen, 1990; Solomon et al, defined hypoechoic rim that may give the lesion a bull’s-eye or
1992). Early reports demonstrated that IOUS frequently altered target appearance (see Chapters 15 and 92). Mucinous variant
the operative management of malignant hepatic tumors by colorectal cancer metastases may contain calcification that pro-
more precisely delineating the relationship between tumors and duces acoustic shadowing. Hepatocellular carcinoma (HCC)
major vascular and biliary structures and by detecting hepatic frequently invades major vascular structures and may be associ-
tumors not revealed on preoperative imaging studies (Fig. ated with portal lymphadenopathy (Jeffers et al, 1988) (see
23.1). In an analysis of 210 patients with primary and second- Chapters 15 and 91).
ary hepatic malignancies, Bismuth and colleagues (1987)
reported that IOUS provided additional information in 35% of Technical Considerations
cases and changed the operative plan in 20%. Parker and asso- IOUS is applied with increasing frequency during hepatic seg-
ciates (1989) reported similar findings, with US influencing mental or nonanatomic resection, and partial liver transplanta-
49% of cases. A more recent analysis by Kruskal and Kane tion. It provides for dynamic and precise delineation of the
(2006) demonstrated that IOUS routinely identifies 25% to spatial relationships of hepatic tumors to major vascular and
413
The planned surgical margin of the lesion may be outlined

using electrocauterization of the liver capsule, as it will
produce acoustic shadowing. The depth of the lesion can be
Tumor estimated using its relationship to other structures or direct
measurement.
Biliary Disease
LHD Evaluation of the Biliary Tree
As laparoscopic cholecystectomy has become the standard of
care for gallbladder disease, the use of laparoscopic US during
this procedure has emerged as an alternative to cholangiogra-
phy (see Chapters 35 and 36). Laparoscopic US is recom-
mended as the primary screening modality for evaluating bile
duct calculi because of its safety, efficiency, and overall cost-
effectiveness (Machi et al, 1999). IOUS is at least as good as
PV cholangiography, if not better, at detecting stones during lapa-
roscopic cholecystectomy (Catheline et al, 1999; Siperstein
et al, 1999; Tranter & Thompson, 2003). Laparoscopic US, in
particular, has the added advantage of reducing the risk of
biliary and vascular injury. Perry and associates (2008) reported
on the use of laparoscopic US in a prospective study performed
FIGURE 23.1. Intraoperative ultrasound view of a tumor (arrows) at between 1995 and 2005. Laparoscopic US was performed in
the base of segment IV. The confluence of the right and left portal venous 236 (60%) of 396 laparoscopic cholecystectomies performed
branches is shown (PV). Dilation of the left hepatic duct (LHD) is indi- for cholelithiasis. US had a positive predictive value of 100%,
cated and suggests possible involvement by tumor. a negative predictive value of 99.6%, a sensitivity of 92.3%, and
a specificity of 100% for detecting CBD stones (Perry et al,
2008). However, proficiency with IOUS requires some time to
biliary structures. Intraoperative transducers appropriate for develop, which may explain why it is not used more liberally
liver surgery include high-frequency (6 to 10 MHz) T-shaped for biliary disease.
linear- or curvilinear-array transducers. These probes provide Laparoscopic US is a useful adjunct for staging gallbladder
excellent high-resolution images and are capable of identifying carcinoma and cholangiocarcinoma (see Chapters 49 to 51).
lesions as small as approximately 1 to 2 mm in size at depths Sonographic findings not only reveal subtle liver metastases but
of penetration of approximately 10 to 12 cm. Transducers with may help to define the local extent of these tumors and their
color flow Doppler imaging enable further discrimination of relationship to the ductal system. Doppler and color flow
tumors and normal hepatic vasculature. images may help distinguish biliary sludge from polyps and
Techniques for IOUS continue to evolve. Particular atten- other intraluminal tumors. Other applications of laparoscopic
tion must be given to port placement when IOUS will be US include evaluating biliary strictures and malignant biliary
performed laparoscopically (see Chapter 105). Typically, no obstructions in the planning of surgical reconstructions, such
coupling gel is required because the natural surface moisture as biliary bypass procedures. These approaches and others are
of the liver provides adequate acoustic coupling. The trans- summarized in a review by Berber and Siperstein (2004).
ducer is initially held in a transverse position, and the survey
of the liver begins by first identifying the confluence of the right Technical Considerations
and left portal pedicles. Next, all segmental pedicles on the The biliary tree and gallbladder are imaged best from a right
right are visualized, followed by identifying those on the left. subcostal port or from the periumbilical port. Sonographic
Each hepatic vein is visualized by scanning peripherally and examination of the gallbladder is approached through the liver
then by traversing toward the vena cava. It is easiest to identify by using a 5- or 7-MHz transducer. A nasogastric or orogastric
the hepatic veins by placing the transducer cranially, in a tube should be placed for decompression. The bile ducts can
midline position and angled toward the heart. Only light pres- be visualized through a compressed duodenum or gastric
sure should be applied to the liver. The confluence of the left antrum using a 7-MHz transducer, but placing the transducer
and middle vein has a characteristic appearance and should be directly on the ducts should be avoided because reverberation
observed in all sonographic examinations. If the entire liver artifact limits its sensitivity. Color flow images are helpful for
is to be scanned—for example, to search for metastases— distinguishing the common bile duct (CBD) from the portal
sequential overlapping sagittal strokes are made sweeping from vein, for identifying the insertion of cystic duct into the CBD,
superior to inferior, beginning at the most lateral margin of and for identifying any aberrant biliary anatomy.
segment II and traveling toward the right. More focal scanning
can be used to localize impalpable lesions situated deep within Pancreatic Disease
the liver parenchyma. Special care must be taken when imaging Evaluation of the Pancreas
the superior portion of the right liver, the posterior subdia- The principal application of IOUS for pancreatic disease is
phragmatic bare area, and surface lesions, such as hamartomas, during staging laparoscopy (SL) in patients with pancreatic
as these areas are particularly challenging to completely visual- malignancy. Its utility lies in demonstrating vascular invasion,
ize. Once the lesion has been delineated and measured, photo- liver metastases, and the degree of resectability. Approxi-
graphic documentation should be obtained and saved. mately 20% to 35% of patients deemed resectable based on
Chapter 23 Intraoperative diagnostic techniques 415
preoperative radiologic assessment harbor occult metastases

and/or locally invasive disease, which preclude curative resec- Biliary Injuries
tion. These concepts will be discussed further for SL (see In the late 1980s, IOC emerged with the hope that the tech-
Chapter 62). nique could potentially help avoid major injury to the biliary
Other applications for US in pancreatic surgery include system. Today, laparoscopic cholecystectomy is the most com-
localizing islet cell tumors (see Chapter 65). These particularly monly performed elective abdominal surgical procedure in the
small lesions may be difficult to visualize or palpate in the body United States (see Chapter 35), with more than 750,000 cases
or tail, and sonographic images may help delineate these and performed annually. Injury to the bile duct is rare but is a major
other malignant or premalignant tumors in the distal portions contributor to patient morbidity (see Chapters 38 and 42).
of the pancreas (Hayashibe et al, 2005). Finally, US can help These injuries are typically due to either technical error or
identify pancreatic ductal abnormalities among patients with misidentification of the duct: The technical error is that the
pancreatitis and may be useful during drainage procedures, CBD is inadvertently transected, and a surgical clip is some-
such as cystogastrostomy (see Chapter 58). times applied to a misidentified CBD, with or without subse-
quent transection.
Technical Considerations Four mechanisms of injury have been described (Way et al,
IOUS can be a useful adjunct in the staging of pancreatic and 2003). Class I injuries involve an incision (i.e., an incomplete
periampullary tumors. When performed laparoscopically, a transection) of the CBD and typically occur when the CBD is
right upper or left upper quadrant port is used to orient the mistaken for a cystic duct. Class II injuries represent lateral
probe transversely along the long axis of the pancreas. Fre- damage to the common hepatic duct. Class III injuries are the
quently, a periumbilical port is required to image the head, most common and represent a CBD mistaken for a cystic duct;
neck, and uncinate process. A 7-MHz transducer is used and these are not immediately recognized intraoperatively. Finally,
can be placed directly on the surface of the pancreas or through class IV injuries occur when a right hepatic duct is mistaken for
the overlying omentum. a cystic duct.
Flum and colleagues (2003) analyzed nationwide Medicare
INTRAOPERATIVE CHOLANGIOGRAPHY data from 1992-1999. Among 1.6 million cholecystectomies
performed, roughly 8000 bile duct injuries occurred overall
Intraoperative cholangiography (IOC) is most commonly used (0.5%); these occurred in 0.39% of patients undergoing
during elective cholecystectomy to assess for choledocholithia- cholecystectomy with IOC and in 0.58% of patients under
sis and to provide clarification of the biliary anatomy; rarely is going cholecystectomy without cholangiography (P < .001). The
it necessary or helpful in assessing extent of biliary tumors or adjusted relative risk of bile duct injury when cholangiography
during hepatic resection (Sotiropoulos et al, 2004). Interest- was not used was 1.71 (95% confidence interval, 1.38 to 2.28).
ingly, Mirizzi first described the procedure in 1937, as a method Inexperienced surgeons were twice as likely to cause bile duct
to help delineate the anatomy of the biliary tree in advanced injuries when cholangiography was not used, compared with
cases of biliary disease (see Chapter 35). their more experienced colleagues. These estimates had been
observed in an earlier study (Fletcher et al, 1999).
Choledocholithiasis Another potential benefit of IOC, with respect to biliary
Two factors must be considered when performing IOC for the injuries, is the earlier recognition and treatment of biliary injury
evaluation of bile duct stones (Hyser et al, 1999; Sigel et al, (see Chapter 42). Ideally, IOC should be performed prior to
1983; Tokumura et al, 2002). First, most stones in the biliary dividing the presumed cystic duct. If cannulation of the CBD,
tree are suspected on clinical grounds or after serologic testing rather than the cystic duct, occurs, the cholangiocatheter can
for liver enzyme elevation and are often visualized during be removed, and the ductotomy can be addressed by primary
preoperative US (see Chapter 15), ERCP (see Chapters 20 repair or by placing a T-tube, without the need for formal
and 29), or MRCP (see Chapter 19). Common presentations biliary reconstruction. A T-tube for large defects would allow
include obstructive jaundice, cholangitis, biliary colic with the CBD to heal without stricture formation, and the tube can
abnormal liver function test results, and acute pancreatitis. The be removed nonoperatively several weeks after the cholecystec-
incidence of clinically silent choledocholithiasis is low, occur- tomy following cholangiography. The alternative scenario is
ring in roughly 1 in 25 cases of biliary colic (Metcalfe et al, complete transection of the CBD, a class III injury. In the
2004; Nugent et al, 2005). Even when missed during cholecys- absence of cholangiography, this transection is seldom recog-
tectomy, these residual stones rarely, if ever, cause symptoms nized intraoperatively, typically presents postoperatively, and
or become clinically relevant (Sarli et al, 2003a) (see Chapters results in significant patient morbidity. Management requires
36 and 37). Charfare and Cheslyn-Curtis (2003) investigated an open bilioenteric anastomosis for repair; thus IOC has the
the incidence of retained biliary stones after 600 laparoscopic potential to not only prevent bile duct injury, but also to miti-
cholecystectomies. With a median follow-up of 3 years among gate its impact if injury does occur (Giurgiu et al, 1999).
438 patients (73%), findings from this study indicate that resid-
ual bile duct stones occurred in only 7 cases (1.2%). Controversies
The second factor to consider when performing IOC is that The routine application of IOC during elective cholecystec-
false-positive results are not uncommon, occurring up to 4% tomy has been called into question in recent years and is still
of the time (Metcalfe et al, 2004). As a result, routine applica- a matter of considerable debate (Massarweh & Flum, 2007;
tion of IOC for the evaluation of unsuspected choledocholithia- Metcalfe et al, 2004; Mills et al, 1985; Nickkholgh et al, 2006;
sis may subject patients to unnecessary bile duct explorations Soper & Brunt, 1994 ). Critics of routine cholangiography
or ERCP studies, with considerable cost to both patients and suggest that such an approach increases the risk of ductal com-
providers (Cuschieri & Berci, 1984; Sarli et al, 2003b). plications (i.e., stricture) and pancreatitis, wastes time and
money, and is seldom indicated if the critical view of safety is

BOX 23.1 Benefits of Laparoscopic Staging
achieved (see Chapter 35). Proponents argue that routine chol-
angiography is a safe, accurate, quick, and cost-effective method Shorter recovery
for evaluating the bile duct (Amott et al, 2005; Wenner et al, Shorter hospital stay
Decreased procedure-related morbidity
2005). Although we favor a selective approach, we recognize
Improved quality of life
that the success of either approach is highly variable and likely
Shorter time to initiation of adjuvant therapy
to depend on each surgeon’s familiarity with cholecystectomy Reduced hospital cost
and cholangiography. Whether selective or routine cholangiog-
raphy is adopted, it is critical that surgeons develop a familiarity
with the interpretation of the cholangiogram. Way and col-
leagues (2003) reported that although 43 cholangiograms dem-
onstrated a bile duct injury, only 9 were correctly interpreted
BOX 23.2 Potential Disadvantages of Staging Laparoscopy
at the time of surgery. Potential for morbidity
Risk of port-site recurrence
Technical Considerations Increased time of surgery and increased cost
The first step of IOC consists of identification of the cystic duct
at its junction with the gallbladder neck (Yamakawa, 1976).
The gallbladder is retracted laterally, and the cystic duct and
artery are dissected free and cleared of the fat and overlying et al, 2003; Vollmer et al, 2002). These findings demonstrate
peritoneum in the area of the triangle of Calot. A small incision that SL can clarify resectability, identify metastatic disease, and
(less than 50% of the duct circumference) is made in the cystic help specialty surgeons tailor curative and palliative operations
duct adjacent to the gallbladder neck. Laparoscopically, the for patients with hepatobiliary and pancreatic malignancies
cystic duct is best approached from a right subcostal port or (Box 23.1). Controversy still exists, however, regarding the
from the periumbilical port (we prefer the former). A 60-cm, added benefit of the procedure, particularly in an era of more
tapered 5-Fr cholangiocatheter is then advanced directly into effective preoperative imaging. Potential advantages in patients
the cystic duct through the ductotomy. A specialized cholan- who avoid nontherapeutic laparotomy include the following:
giogram clamp, often termed an Olsen cholangiogram clamp, shorter recovery, decreased hospital stay, decreased morbidity,
secures the catheter in place (Decker et al, 2003). In the open improved quality of life, shorter time to initiation of adjuvant
setting, the cholangiocatheter can be secured using a silk therapy, and reduced hospital costs. Disadvantages include
ligature. increased potential morbidity, increased operating room times
Alternatively, a percutaneous method may be used, where and costs, and potential port-site seeding (Box 23.2).
access to the cystic duct is achieved via a separate puncture in
the abdominal wall by using at least a 2-inch, 14-gauge needle. Surgical Technique
A 5-Fr cholangiocatheter is guided through this needle and Technical Considerations
directly into the cystic duct. Regardless of approach, the cath- SL is typically performed under general anesthesia. The patient
eter is first flushed with saline to confirm its patency, then is placed supine on the operating table and fully prepared as
radiographic contrast is infused, and fluoroscopic images are for laparotomy. The abdomen is marked for the intended open
obtained. A complete study demonstrates flow of contrast into incision; at our institution, we favor a bilateral subcostal inci-
the duodenum and shows opacification of both the right and sion two to three fingerbreadths below the costal margin. In the
left hepatic ducts. case of pancreatic disease, access to the peritoneum is made
through a 1-cm subumbilical incision; for hepatobiliary cases,
STAGING LAPAROSCOPY a 1-cm incision is made along the midclavicular line. For both,
the fascia and peritoneum are incised under direct vision, a
The detection of unresectable disease and the avoidance of blunt 5- or 10-mm port is placed, and insufflation is begun with
futile surgery are critical in the management of hepatobiliary the initial flow rate set at 2 L/min and gradually increased to
and pancreatic malignancies. Despite refinements in several 15 L/min to achieve an intraabdominal pressure of 10 to
diagnostic modalities that include, but are not limited to, EUS, 14 mm Hg.
ERCP, CT, and MRCP, in up to 20% to 25% of patients, We use a 5- or 10-mm, 30-degree angled telescope to inspect
occult disease missed on imaging is discovered at the time of the abdomen. Although we prefer a multiport technique to
operative intervention (de Rooij et al, 1991; Espat et al, 1999; facilitate obtaining biopsy specimens and laparoscopic US,
Lillemoe et al, 1999). Oncologic staging seeks to determine others may use a single-port approach. Two more ports are
tumor size and characteristics, the extent of regional spread to placed under direct vision, in the left and right upper quad-
adjacent structures, and the magnitude of any distant spread. rants. For pancreatic cases, another 5-mm port is inserted along
Since its introduction in the 1980s, laparoscopy has become the incision line to the right of the midline (Fig. 23.2). Entry
increasingly used in the staging of hepatobiliary and pancreas of these ports is followed by a systematic examination of the
cancers. Jarnagin and associates (2000) demonstrated that peritoneal cavity; adhesions may need to be divided to facilitate
among a cohort of 186 patients with primary and secondary adequate inspection of the abdominal viscera and peritoneum.
hepatic malignancies, 25% were noted to harbor unresectable All four quadrants are examined for evidence of peritoneal
disease at the time of laparoscopic staging, and 65% of this deposits; biopsies are taken of any suspicious lesions using
group were spared an unnecessary laparotomy. Callery and standard biopsy forceps, and samples are sent for histologic
colleagues (1997) similarly revealed unresectable pancreatic examination (Fig. 23.3). If clinically indicated, peritoneal cytol-
and periampullary disease among 11 of 50 patients (D’Angelica ogy can be sampled at this time. The steps for obtaining
5 mm
5 mm
10 mm
Hasson
FIGURE 23.4. Examination of the liver. A blunt 10-mm instrument is
used in conjunction with a 5-mm grasper.
FIGURE 23.2. Port placement for laparoscopic staging. For patients

undergoing staging for hepatobiliary diseases, the camera may be
placed along the line of a subcostal incision.
FIGURE 23.5. The hepatoduodenal ligament can be examined from

the left (shown) or right side. Periportal adenopathy (arrow) is shown.
the 10-degree Trendelenburg position, and the omentum is

retracted into the left upper quadrant. The ligament of Treitz
and the inferior surface of the transverse mesocolon are
inspected for metastatic deposits and for lymphadenopathy
(Fig. 23.6).
FIGURE 23.3. Peritoneal nodule. Arrow indicates a metastatic deposit,
The patient is repositioned in a leveled position, and the left
which can be sampled easily using a cup forceps to obtain a biopsy. lateral segment of the liver is elevated superiorly with a retractor
via the left upper quadrant port. The lesser omentum is opened
to visualize the caudate lobe of the liver and the vena cava. Care
peritoneal cytology are as follows. After instillation of 200 mL should be taken to evaluate for, and avoid damage to, aberrant
of warm normal saline and gentle agitation, the irrigant is aspi- left hepatic arterial anatomy. From this vantage point, the ante-
rated from the right and left subhepatic spaces and from the rior aspect of the head of the pancreas, posterior wall of the
pelvis, and a sample is sent for cytologic examination. stomach, and “gastric pillar,” which contains the left gastric
Following inspection of the peritoneum, attention is turned artery and vein, can be seen (Fig. 23.7). The left gastric artery
to the anterior and posterior surfaces of the left lateral segment can be followed to its origin to allow inspection of the celiac
of the liver and the anterior and inferior surfaces of the right axis (Box 23.3).
lobe (Fig. 23.4). Next, the hepatoduodenal ligament, the
foramen of Winslow, and hilum of the liver are examined for Laparoscopic Ultrasound
lymphadenopathy (Fig. 23.5). Any suspicious nodes are excised US can be used as an adjunct in diagnostic laparoscopy. A
and sent for pathology analysis. The patient is then placed in systematic examination of the liver, biliary tree, and pancreas
FIGURE 23.8. Laparoscopic ultrasound of the hepatoduodenal liga-

ment. Doppler capability facilitates identification of the hepatic artery.
FIGURE 23.6. Examination of the ligament of Treitz. Proximal jejunum
(solid arrow) and inferior mesenteric vein (dashed arrow) are shown.
FIGURE 23.7. Lesser sac. The “gastric pillar” (dashed arrow) and
hepatic artery (solid arrow) are shown.
BOX 23.3 Steps for Staging Laparoscopy

Examination of peritoneal cavity
Placement of laparoscopic trocars
Instillation of 200 mL of normal saline and aspiration of cytologic
specimens FIGURE 23.9. Hepatic metastasis (arrow) in a patient with pancreatic
Assessment of primary tumor adenocarcinoma.
Examination of the liver and porta hepatis
Division of gastrohepatic omentum and examination of caudate Placement of the probe in the transverse direction over the
lobe, vena cava, celiac axis, and lesser sac hepatoduodenal ligament allows the common hepatic duct,
Identification of the ligament of Treitz and inspection of the CBD, hepatic arteries, and portal vein to be visualized (Fig.
mesocolon, duodenum, and jejunum 23.8). Similarly, at the confluence of the portal vein, the splenic
Laparoscopic ultrasound
and superior mesenteric veins can be identified. Finally, the
From Conlon KC, Brennan MF: Laparoscopy for staging abdominal malignancies, St Louis, 2000, superior mesenteric artery is delineated, and its relationship to
Mosby. any pancreatic tumor is assessed. The probe is placed on the
gastrocolic omentum and advanced first caudally and then
through the window in the gastrohepatic omentum. Laparo-
takes place as previously described. A 6- to 10-MHz T-shaped scopic US can help facilitate pathology biopsies and needle
linear- or curvilinear-array transducer is placed over the left aspirations of any suspicious lesions (Fig. 23.9, Box 23.4).
lateral segment to assess segments I, II, and III. The probe is
moved to the right liver and placed on the dome of the liver. Complications
The vena cava is visualized posteriorly, and the hepatic and Morbidity associated with SL is uncommon, occurring in only
portal veins are seen as the probe is moved anteriorly. 1% to 2% of cases. Major complications are similar to those
BOX 23.4 Steps for Laparoscopic Ultrasound BOX 23.5 Criteria for Clinical Risk Score to Determine Risks for
Insertion of laparoscopic ultrasound probe Extrahepatic Disease
Examination of liver: left lateral segment, right lobe Lymph node–positive tumor
Transverse scan of hepatoduodenal ligament Disease-free interval between primary and detection of metastasis:
Identification of superior mesenteric artery, portal vein, splenic vein <12 mo
Examination of pancreas Number of hepatic tumors: >1 (based on preoperative imaging)
Assessment of tumor Carcinoembryonic antigen: >200 ng/mL within 1 mo of surgery
Size of the largest hepatic tumor: >5 cm
From Minnard EA, et al: Laparoscopic ultrasound enhances standard laparoscopy in the staging
of pancreatic cancer, Ann Surg 228:182–187, 1998. From Jarnagin WR, et al: A clinical scoring system predicts the yield of diagnostic laparoscopy in
patients with potentially resectable hepatic colorectal metastases, Cancer 91:1121–1128, 2001b.
for other laparoscopic procedures and include hemorrhage,
visceral perforation, and infection. A misconception is that
suspected malignancy precludes staging laparoscopic examina- criteria) hepatobiliary malignancy. Laparoscopy was performed
tion. An early report raised concern about “port-site” tumor in 401 patients between 1997 and 2001; 291 (73%) cases were
implants 2 weeks after laparoscopy in a patient with malignant considered complete, 88 (22%) incomplete, and 22 (5.5%)
ascites (Dobronte et al, 1978); this concern was also expressed were considered failures. Unresectable disease was discovered
by later studies in the 1990s, which reported rates of port-site at the time of laparoscopy in 84 cases, for an overall incidence
recurrence ranging from 0.8% to 2% (Curet, 2004; Nieveen of 21%. Sixty-nine patients had unresectable disease identified
van Dijkum et al, 1999; Pearlstone et al, 1999). Shoup and during open exploration, for an overall false-negative rate of
colleagues (2002) examined this question in a series of 1650 22%. This high false-negative rate was attributed primarily to
diagnostic laparoscopic procedures, in which 4299 trocars were failure to identify lymph node metastases or to detect vascular
inserted. The overall rate of port-site recurrence was 0.8% (13 invasion.
cases); 5 of those occurred in patients with local disease; the Laparoscopy is most accurate for identifying peritoneal
remaining 8 occurred in patients with documented metastatic deposits (80% accuracy) and hepatic disease (63% accuracy);
disease. The findings of this study suggest that recurrence is a it is least accurate for identifying nodal metastases (7% accu-
marker of more advanced disease rather than being an event racy) and vascular invasion (18% accuracy) (D’Angelica et al,
resulting from the laparoscopy. Overall, Shoup and associates 2003). Factors that can improve the yield of laparoscopy
concluded that SL can be performed safely in the setting of include the surgeon’s preoperative judgment regarding the like-
presumed malignancy (Shoup et al, 2002; Shoup et al, 2004). lihood of resectability, the completeness of laparoscopic staging
This concept has been confirmed by others, including Velanov- examinations, the addition of laparoscopic ultrasound, and the
ich (2004), who demonstrated that the incidence of port-site primary diagnosis. The yield was highest with gallbladder ade-
recurrence (3%) is equivalent to that of wound recurrence in nocarcinoma and cholangiocarcinoma and lowest with colorec-
patients who had an exploratory laparotomy alone (3.9%). tal metastases (Callery et al, 1997; D’Angelica et al, 2003; John
et al, 1994).
Staging Laparoscopy for Potentially Resectable Disease Laparoscopic staging can be used in the evaluation of
Hepatobiliary Malignancy colorectal metastases (Rahusen et al, 1999) (see Chapter 92).
Determination of resectability of hepatobiliary malignancy Approximately one half of all patients with new diagnoses of
begins with CT or MRI; however, laparoscopy plays an essen- colorectal cancer will subsequently develop liver metastases, yet
tial role in detecting occult disease (Jarnagin et al, 2001a; only 20% are candidates for curative hepatic resection. Most
Jeffers et al, 1988; Lightdale, 1982; Torzilli et al, 2002). authors agree that hepatic cirrhosis, extrahepatic tumor spread,
Roughly 20% to 25% of patients initially believed to have and significant bilobar disease are relative contraindications for
resectable disease will harbor occult disease detected by lapa- hepatic resection. In a large series of patients with colorectal
roscopy that precludes curative resection and thereby can be metastases identified as resection candidates who underwent
spared laparotomy (Conlon et al, 2003; Weber et al, 2002). SL with IOUS, Thaler (2005) reported that a change in the
In addition to the quality of the preoperative imaging, the preoperatively determined plan occurred in 67% (46/69) of
yield of laparoscopy depends on the underlying disease and the patients. Of those in whom resection was considered impossi-
quality of preoperative imaging. For example, hilar cholangio- ble, 18 were untreatable altogether, and 28 patients had alterna-
carcinoma rarely gives rise to peritoneal disease; however, it is tive treatments (i.e., regional therapy or a palliative hepatic
often unresectable due to local tumor extension and/or vascular arterial infusion pump) (Thaler et al, 2005). Of 36 patients in
involvement, which is difficult to determine laparoscopically. whom radiofrequency ablation was planned, 14 had procedures
The yield of laparoscopy is therefore lower than for other diag- altered by SL/IOUS (Thaler et al, 2005).
noses, such as gallbladder cancer, which frequently results in In an effort to improve the yield of SL for hepatic colorectal
peritoneal disease. metastases, a study used a clinical risk score (CRS) to try to
In HCCs, laparoscopy can be used to evaluate for cirrhosis determine which patients are more likely to have disease that
and aid in determining resectability. Weitz and colleagues (2004), is occult on preoperative imaging. The CRS uses five clinical
in a study of patients with HCC who underwent SL, found that parameters: Each is shown as an independent predictor of
22% of patients were spared nontherapeutic laparotomy. The outcome after resection, and each criterion is assigned 1 point.
reported yield of SL over the past 10 years ranges from 16% to Forty-two percent of patients with a CRS greater than 2 had
57% (Gaujoux & Allen, 2010) (see Chapter 91). unresectable disease detected at laparoscopy versus none of the
In a prospective analysis, D’Angelica and colleagues (2003) patients with CRSs of zero to 1 (Box 23.5) (Grobmyer et al,
studied the use of SL in patients with resectable (by radiologic 2004; Jarnagin et al, 2001b). The score was validated by Mann
Pancreatic
mass on CT
Resectable Unresectable
Laparoscopic
staging ± LUS Localized Metastatic
disease disease
Resectable Unresectable Laparoscopy Chemotherapy
Open Appropriate
exploration therapy No mets Mets
Chemotherapy Chemotherapy
/radiotherapy
FIGURE 23.10. Use of laparoscopic staging in the management of adenocarcinoma of the pancreas. CT, Computed tomography; mets, metastases;
LUS, laparoscopic ultrasound.
and colleagues (2007) in a study of 200 patients with colorectal 2. For locally advanced unresectable pancreatic cancer without
metastases who underwent SL. The CRS predicted the likeli- radiographic evidence of distant metastasis, SL may be used
hood of finding incurable disease (P < .001) and predicted to rule out subclinical metastatic disease to optimize treat-
curability determined by laparoscopy (P < .001). Laparoscopy ment selection (Callery et al, 2009).
did not change the management of any patient with a CRS of Cuschieri and colleagues first reported their experience
zero or 1; however, it did alter the course of patients with a using SL for pancreatic malignancy in 1988. In this report,
score of 2 to 3 (18/129 patients affected) and a score of 4 to 5 laparoscopic examinations demonstrated that 42 (57%) of 73
(21/40 patients affected) (Mann et al, 2007). Using such a risk patients had unresectable disease, mainly because of peritoneal
scoring system to help guide the addition of laparoscopy to or omental metastases. Since these early papers, laparoscopy
higher-risk patients should improve the overall yield from has been refined and is now comparable to open exploration in
laparoscopy. defining resectability. Other groups have reported similar find-
ings. Warshaw and colleagues (1990) compared CT, MRI,
Pancreatic and Periampullary Malignancy angiography, and laparoscopy in 88 patients and found that CT
Our approach for the staging of pancreatic adenocarcinoma is failed to show peritoneal disease in 22 cases discovered at lapa-
illustrated in Figure 23.10. A minority of patients (20% to roscopy. MRI conferred no further benefit, and laparoscopy
35%) with pancreatic and periampullary malignancies (see had a sensitivity of 96%, having understaged 1 patient. This
Chapters 62 and 63) are suitable candidates for curative resec- same group reported that 39 patients (44%) were spared a
tion (Conlon & Brennan, 2000; Conlon & Minnard, 1997; laparotomy because of metastases discovered at laparoscopy
Jimenez et al, 2000). Similar to the management of hepa- not seen on spiral CT. SL was performed as a day-case proce-
tobiliary malignancies, initial diagnostic evaluation begins with dure, further reducing inpatient hospital time. Reddy and asso-
high-resolution contrast-enhanced CT. Imaging studies can ciates (1999) confirmed these findings in a later report; 29%
predict resectability in 57% to 88% of cases (Contreras et al, were spared further intervention. A more recent prospective
2009; Freeny, 2001; Mayo et al, 2009; Pisters et al, 2001). study (Doran et al, 2004) reported outcomes of 239 patients
Other useful imaging modalities that may augment CT eva with pancreatic cancer who were submitted to SL; of these, 190
luations include MRI, ERCP, EUS, and positron-emission were considered to have resectable disease based on CT find-
tomography. ings. After SL, roughly 15% of patients were spared a lapa-
The Society of Surgical Oncology guidelines from the 2009 rotomy. This estimate likely reflects advancements in
consensus statement regarding the pretreatment assessment of high-resolution imaging, but it also demonstrates that, despite
pancreatic cancer state the following regarding the use of refinements in imaging techniques, peritoneal disease is not
laparoscopy: easily revealed on CT.
1. For apparent resectable pancreatic cancer, SL should be As imaging has improved, the yield of laparoscopy has been
used selectively on the basis of clinical parameters that opti- challenged. In 1996, Conlon reported the Memorial Sloan Ket-
mize yield. These include pancreas head tumors of greater tering experience. In a review of 115 patients between the years
than 3 cm, tumors of the pancreas body and tail, equivocal of 1992 and 1994 deemed resectable based on imaging, 38%
findings on a CT scan, and high cancer antigen 19-9 levels had findings at the time of SL that precluded resection (Box
(>100 U/mL). 23.6) (Conlon, 1996). However, in an updated review of their
TABLE 23.1 Reviews on Staging Laparoscopy for Pancreatic Cancer

No. Resectable on CT but Laparoscopy Changed
Reference N No. Resectable on CT Unresectable at Laparoscopy Operative Plan (%)
Reddy et al, 1999 109 99 29 29

Jimenez et al, 2000 125 70 39 31
Conlon & Brennan, 2000 577 577 211 36
Vollmer et al, 2002* 72 72 22 31
Doran et al, 2004 305 190 28 15
Contreras et al, 2009† 58 58 18 31
Mayo et al, 2009† 86 86 24 28
White et al, 2008 1045 1045 145 14
*Excluding ampullary malignancies.
†
Pancreatic adenocarcinoma only (including head and body/tail).
CT, Computed tomography.
similarly reported 96% specificity and 92% sensitivity when

BOX 23.6 Criteria for Unresectability in Pancreatic Cancer
laparoscopic US was combined with SL. Vollmer (2002) and
Histologically confirmed hepatic, serosal, peritoneal, or omental others studied 72 patients with pancreatic head cancers and
metastases found that 22 had metastatic disease that precluded resection;
Tumor extension outside pancreas laparoscopy alone identified 14 of these 22 patients; the remain-
Celiac or high portal node involvement confirmed by frozen section
ing 8 patients had major vessel encasement or liver metastases
Extensive portal vein involvement by tumor or invasion/encasement
of celiac axis, hepatic artery, or superior mesenteric artery
revealed by laparoscopic US. Minnard and colleagues (1998)
similarly reported that laparoscopic US altered the planned
Data from Conlon KC, Brennan MF: Laparoscopy for staging abdominal malignancies, St Louis, surgical treatment in 14% of patients, whose staging laparo-
2000, Mosby; and Conlon KC, Minnard EA: The value of laparoscopic staging in upper scopic procedures were equivocal. Sonography was particularly
gastrointestinal malignancy, Oncologist 2:10–17, 1997.
useful in identifying venous (42%) and arterial (38%) involve-
ment, which precludes curative resection.
experience, the yield of the procedure has decreased. In a
review of 1045 patients with imaging-determined pancreatic Controversies
and peripancreatic tumors, 12% of patients with pancreatic Some critics of routine SL believe that unresectable disease
tumors had findings on diagnostic laparoscopy that precluded due to vascular involvement or local extension can only
resection. The yield was greater for patients diagnosed with be confirmed during open explorations (Friess et al, 1998).
pancreatic adenocarcinoma and less for patients with tumors Others suggest that the role for laparoscopy is limited only to
of the ampulla, distal bile duct, duodenum, and neuroendo- those patients who do not require some form of palliation,
crine tumors. The decrease in sensitivity between the time either biliary or gastric bypass (Andren-Sandberg et al, 1998;
periods is likely in part due to the improvement in preoperative Holzman et al, 1997). This is determined in part by the type
imaging with the addition of MDCT scanners and thin-slice of malignancy (Brooks et al, 2002; Vollmer et al, 2002). SL has
imaging (White et al, 2008). not been shown to be useful among patients with ampullary or
In addition to the benefits related to recovery from laparos- duodenal tumors, as these cancers tend to present earlier and
copy, several authors have found that the laparoscopic approach are less likely to have metastatic disease at the time of presenta-
to staging may be superior to the open exploration for the tion. In contrast, SL may have a role in patients with nonfunc-
detection of occult metastases (Shoup et al, 2004). Contreras tioning islet cell tumors (Hochwald et al, 2001), which manifest
and colleagues (2009) discovered further that within a 4-year with a large tumor mass and often in the presence of metastatic
period, occult metastases among 52 patients with potentially disease. The relative yield of diagnostic laparoscopy may be
resectable pancreatic tumors were more likely to be detected lower at centers that have an aggressive palliative surgical
during SL when compared with open explorations (32% vs. approach, whereas even if a disease is unresectable, laparotomy
18%). Other centers have reported similar results (Table 23.1). is performed regardless for palliative intervention with bypass.
Laparoscopic US may further increase the sensitivity of SL; Finally, some argue that SL is costly, time consuming, and of
however, it can be technically demanding, and it is unclear if decreasing diagnostic yield because of improvements in radio-
many centers routinely use this adjunct. John and colleagues logic imaging techniques (Spitz et al, 1997).
(1995) demonstrated that the addition of US improved the
sensitivity from 65% to 80%. Callery and associates (1997) References are available at expertconsult.com.
Chapter 23 Intraoperative diagnostic techniques 421.e1
Flum DR, et al: Bile duct injury during cholecystectomy and survival
REFERENCES in medicare beneficiaries, JAMA 290:2168–2173, 2003.
Freeny PC: Pancreatic carcinoma: imaging update 2001, Dig Dis
Amott D, et al: Prospective comparison of routine and selective opera- 19:37–46, 2001.
tive cholangiography, Aust N Z J Surg 75:378–382, 2005. Friess H, et al: The role of diagnostic laparoscopy in pancreatic and
Andren-Sandberg A, et al: Computed tomography and laparoscopy in periampullary malignancies, J Am Coll Surg 186:675–682, 1998.
the assessment of the patient with pancreatic cancer, J Am Coll Surg Gaujoux S, Allen PJ: Role of staging laparoscopy in peri-pancreatic and
186:35–40, 1998. hepatobiliary malignancy, World J Gastrointest Surg 2:283–290, 2010.
Belghiti J, et al: Surgical treatment of hepatocellular carcinoma in cir- Giurgiu DI, et al: Laparoscopic common bile duct exploration: long-
rhosis: value of preoperative ultrasonography, Gastroenterol Clin Biol term outcome, Arch Surg 134:839–844, 1999.
10:244–247, 1986. Gouillat C, et al: Value of intraoperative ultrasonography in the surgical
Berber E, Siperstein AE: Laparoscopic ultrasound, Surg Clin North Am treatment of malignant tumors, Ann Chir 45:534–539, 1992.
84:1061–1084, 2004. Gozzetti G, et al: Intraoperative ultrasonography in surgery for liver
Bismuth H, Castaing D: Échographie per-opératoire du foie et des voies tumors, Surgery 99:523–529, 1986.
biliaires, Paris, 1985, Blammarion. Grobmyer SR, et al: Diagnostic laparoscopy prior to planned hepatic
Bismuth H, et al: The use of operative ultrasound in surgery of primary resection for colorectal metastases, Arch Surg 139:1326–1330, 2004.
liver tumors, World J Surg 11:610–614, 1987. Gunven P, et al: Preoperative imaging of liver metastases: comparison
Bloed W, et al: Role of intraoperative ultrasound of the liver with of angiography, CT scan and ultrasonography, Ann Surg 202:573–
improved preoperative hepatic imaging, Eur J Surg 166:691–695, 579, 1985.
2000. Hayashibe A, et al: A resected case of multiple intraductal papillary
Brooks AD, et al: The value of laparoscopy in the management of mucinous tumor of the pancreas with US-guided ductal branch-
ampullary, duodenal, and distal bile duct tumors, J Gastrointest Surg oriented partial pancreatectomy, Pancreatology 5:462–465, 2005.
6:139–146, 2002. Hochwald SN, et al: Laparoscopy predicts metastatic disease and
Callery MP, et al: Staging laparoscopy with laparoscopic ultrasonogra- spares laparotomy in selected patients with pancreatic nonfunction-
phy: optimizing resectability in hepatobiliary and pancreatic malig- ing islet cell tumors, Ann Surg Oncol 8:249–253, 2001.
nancy, J Am Coll Surg 185:33–39, 1997. Holzman MD, et al: The role of laparoscopy in the management of
Callery MP, et al: Pretreatment assessment of resectable and borderline suspected pancreatic and periampullary malignancies, J Gastrointest
resectable pancreatic cancer: expert consensus statement, Ann Surg Surg 1:236–244, 1997.
Oncol 16:1727–1733, 2009. Hyser MJ, et al: Laparoscopic transcystic management of choledocho-
Catheline J, et al: A comparison of laparoscopic ultrasound versus lithiasis, Am Surg 65:606–609, 1999.
cholangiography in the evaluation of the biliary tree during laparo- Jakimowicz J: Review: intraoperative ultrasonography during minimal
scopic cholecystectomy, Eur J Ultrasound 10:1–9, 1999. access surgery, J R Coll Surg Edinb 38:231–238, 1993.
Charfare H, Cheslyn-Curtis S: Selective cholangiography in 600 Jarnagin WR, et al: Liver resection for metastatic colorectal cancer:
patients undergoing cholecystectomy with 5-year follow-up for resid- assessing the risk of occult irresectable disease, J Am Coll Surg
ual bile duct stones, Ann R Coll Surg Engl 85:167–173, 2003. 188:33–42, 1999.
Conlon KC, Brennan MF: Laparoscopy for staging abdominal malignan- Jarnagin WR, et al: A prospective analysis of staging laparoscopy in
cies, St Louis, 2000, Mosby. patients with primary and secondary hepatobiliary malignancies,
Conlon KC, Minnard EA: The value of laparoscopic staging in upper J Gastrointest Surg 4:34–43, 2000.
gastrointestinal malignancy, Oncologist 2:10–17, 1997. Jarnagin WR, et al: What is the yield of intraoperative ultrasonography
Conlon KC, et al: The value of minimal access surgery in the staging during partial hepatectomy for malignant disease? J Am Coll Surg
of patients with potentially resectable peripancreatic malignancy, 192:577–583, 2001.
Ann Surg 223:134–140, 1996. Jarnagin WR, et al: What is the yield of intraoperative ultrasound
Conlon R, et al: The value of intraoperative ultrasound during hepatic during partial hepatectomy for malignant disease? J Am Coll Surg
resection compared with improved preoperative magnetic resonance 192:577–583, 2001a.
imaging, Eur J Ultrasound 16:211–216, 2003. Jarnagin WR, et al: A clinical scoring system predicts the yield of diag-
Contreras CM, et al: Staging laparoscopy enhances detection of occult nostic laparoscopy in patients with potentially resectable hepatic
metastases in patients with pancreatic adenocarcinoma, J Surg Oncol colorectal metastases, Cancer 91:1121–1128, 2001b.
100:663–669, 2009. Jeffers L, et al: Laparoscopically directed fine needle aspiration for the
Curet MJ: Port site metastases, Am J Surg 187:705–712, 2004. diagnosis of hepatocellular carcinoma: a safe and accurate technique,
Cuschieri A: Laparoscopy for pancreatic cancer: does it benefit the Gastrointest Endosc 34:235–237, 1988.
patient? Eur J Surg Oncol 14:41–44, 1988. Jimenez RE, et al: Impact of laparoscopic staging in the treatment of
Cuschieri A, Berci G: Operative biliary endoscopy. In Cuschieri A, pancreatic cancer, Arch Surg 135:409–415, 2000.
Berci G, editors: Common bile duct exploration, Boston, 1984, Marti- John TG, et al: Superior staging of liver tumors with laparoscopy and
nus Nijhoff, pp 55–59. laparoscopic ultrasound, Ann Surg 220:711–719, 1994.
D’Angelica M, et al: The role of staging laparoscopy in hepatobiliary John TG, et al: Carcinoma of the pancreatic head and periampullary
malignancy: prospective analysis of 401 cases, Ann Surg Oncol 10: region: tumor staging with laparoscopy and laparoscopic ultrasonog-
183–189, 2003. raphy, Ann Surg 221:156–164, 1995.
Decker G, et al: One hundred laparoscopic choledochotomies with Kruskal JB, Kane RA: Intraoperative US of the liver: techniques and
primary closure of the common bile duct, Surg Endosc 17:12–18, clinical applications, Radiographics 26:1067–1084, 2006.
2003. Lightdale CJ: Laparoscopy and biopsy in malignant liver disease,
de Rooij PD, et al: Evaluation of palliative surgical procedures in unre- Cancer 50(Suppl 11):2672–2675, 1982.
sectable pancreatic cancer, Br J Surg 78:1053–1058, 1991. Lillemoe KD, et al: Is prophylactic gastrojejunostomy indicated for
D’Hondt M, et al: Is intra-operative ultrasound still useful for the unresectable periampullary cancer? A prospective randomized trial,
detection of a hepatic tumour in the era of modern pre-operative Ann Surg 230:322–330, 1999.
imaging? HPB (Oxford) 13:665–669, 2011. Lo CM, et al: Laparoscopy and laparoscopic ultrasonography avoid
Dobronte Z, et al: Rapid development of malignant metastases in the exploratory laparotomy in patients with hepatocellular carcinoma,
abdominal wall after laparoscopy, Endoscopy 10:127–130, 1978. Ann Surg 227:527–532, 1998.
Doran HE, et al: Laparoscopy and laparoscopic ultrasound in the Machi J, et al: Laparoscopic ultrasonography versus operative
evaluation of pancreatic and periampullary tumours, Dig Surg 21: cholangiography during laparoscopic cholecystectomy: review of the
305–313, 2004. literature and a comparison with open intraoperative ultrasonogra-
Espat NJ, et al: Patients with laparoscopically staged unresectable pan- phy, J Am Coll Surg 1999:360–367, 1999.
creatic adenocarcinoma do not require subsequent surgical biliary or Makuuchi M, et al: Intraoperative ultrasonic examination for hepatec-
gastric bypass, J Am Coll Surg 188:649–657, 1999. tomy, Jpn J Clin Oncol 11:367–372, 1981.
Fletcher DR, et al: Complications of cholecystectomy: risks of the lapa- Mann CD, et al: Clinical Risk Score predicts yield of staging laparos-
roscopic approach and protective effects of operative cholangiogra- copy in patients with colorectal liver metastases, Br J Surg 94:855–
phy: a population-based study, Ann Surg 229:449–457, 1999. 859, 2007.
Massarweh NN, Flum DR: Role of intraoperative cholangiography in Sigel B, et al: Comparative accuracy of operative ultrasonography and
avoiding bile duct injury, J Am Coll Surg 204:656–664, 2007. cholangiography in detecting common duct calculi, Surgery 94:715–
Mayo SC, et al: Evolving preoperative evaluation of patients with pan- 720, 1983.
creatic cancer: does laparoscopy have a role in the current era? J Am Siperstein A, et al: Comparison of laparoscopic ultrasonography and
Coll Surg 208:87–95, 2009. fluo-rocholangiography in 300 patients undergoing laparoscopic
Metcalfe MS, et al: Is laparoscopic intraoperative cholangiogram a cholecystectomy, Surg Endosc 13:113–117, 1999.
matter of routine? Am J Surg 187:475–481, 2004. Solomon MJ, et al: A new classification of hepatic territories using
Mills JL, et al: Routine operative cholangiography, Surg Gynecol Obstet intraoperative ultrasound, Am J Surg 163:336–338, 1992.
161:343–345, 1985. Soper NJ, Brunt LM: The case for routine operative cholangiography
Minnard EA, et al: Laparoscopic ultrasound enhances standard lapa- during laparoscopic cholecystectomy, Surg Clin North Am 74:953–
roscopy in the staging of pancreatic cancer, Ann Surg 228:182–187, 959, 1994.
1998. Sotiropoulos GC, et al: Intraoperative cholangioscopy in the manage-
Mirizzi PL: Operative cholangioraphy, Surg Gynecol Obstet 65:702–710, ment of biliary adenomas, J Surg Oncol 85:91–92, 2004.
1937. Soyer P, et al: Surgical treatment of hepatic metastases: impact
Moody AR, Wilson SR: Atypical hepatic hemangioma: a suggestive of intraoperative sonography, AJR Am J Roentgenol 160:511–514,
sonographic morphology, Radiology 188:413–417, 1993. 1993.
Nagasue N, et al: Intraoperative ultrasonography for the surgical treat- Spitz FR, et al: Preoperative and postoperative chemoradiation strate-
ment of hepatic tumors, Acta Chir Scand 150:311–316, 1984. gies in patients treated with pancreaticoduodenectomy for adenocar-
Nickkholgh A, et al: Routine versus selective intraoperative cholangi- cinoma of the pancreas, J Clin Oncol 15:928–937, 1997.
ography during laparoscopic cholecystectomy: a survey of 2,130 Thaler K, et al: The evolving role of staging laparoscopy in the treat-
patients during laparoscopic, Surg Endosc 20:868–874, 2006. ment of colorectal hepatic metastasis, Arch Surg 140:727–734,
Nieveen van Dijkum EJ, et al: Staging laparoscopy and laparoscopic 2005.
ultrasonography in more than 400 patients with upper gastrointesti- Tokumura H, et al: Laparoscopic management of common bile duct
nal carcinoma, J Am Coll Surg 189:459–465, 1999. stones: transcystic approach and choledochotomy, J Hepatobiliary
Nugent N, et al: Low incidence of retained common bile duct stones Pancreat Surg 9:206–212, 2002.
using a selective policy of biliary imaging, Surgeon 3:352–356, 2005. Torzilli G, et al: Ultrasound-guided liver resections for hepatocellular
Olsen AK: Intraoperative ultrasonography and the detection of liver carcinoma, Hepatogastroenterology 49:21–27, 2002.
metastases in patients with colorectal cancer, Br J Surg 77:998–999, Tranter SE, Thompson MH: A prospective single-blinded controlled
1990. study comparing laparoscopic ultrasound of the common bile duct
Parker GA, et al: Intraoperative ultrasound of the liver affects operative with operative cholangiography, Surg Endosc 17:216–219, 2003.
decision making, Ann Surg 209:569–577, 1989. Velanovich V: The effects of staging laparoscopy on trocar site and
Pearlstone DB, et al: Laparoscopy in 533 patients with abdominal peritoneal recurrence of pancreatic cancer, Surg Endosc 18:310–313,
malignancy, Surgery 125:67–72, 1999. 2004.
Perry KA, et al: Laparoscopic ultrasound as the primary method for Vollmer CM, et al: Utility of staging laparoscopy in subsets of peripan-
bile duct imaging during cholecystectomy, Surg Endosc 22:208–213, creatic and biliary malignancies, Ann Surg 235:1–7, 2002.
2008. Warshaw AL, et al: Preoperative staging and assessment of resectability
Pisters PW, et al: Laparoscopy in the staging of pancreatic cancer, of pancreatic cancer, Arch Surg 125:230–233, 1990.
Br J Surg 88:325–337, 2001. Way LW, et al: Cause and prevention of laparoscopic bile duct injuries:
Rahusen FD, et al: Selection of patients for resection of colorectal analysis of 242 cases from a human factors and cognitive psychology
metastases to the liver using diagnostic laparoscopy and laparoscopic perspective, Ann Surg 237:460–469, 2003.
ultrasonography, Ann Surg 230:31–37, 1999. Weber SM, et al: Staging laparoscopy in patients with extrahepatic
Reddy KR, et al: Experience with staging laparoscopy in pancreatic biliary carcinoma: analysis of 100 patients, Ann Surg 235:392–399,
malignancy, Gastrointest Endosc 49(4 Pt 1):498–503, 1999. 2002.
Sarli L, et al: Scoring system to predict asymptomatic choledocholi- Weitz J, et al: Selective use of diagnostic laparoscopy prior to planned
thiasis before laparoscopic cholecystectomy: a matched case-control hepatectomy for patients with hepatocellular carcinoma, Surgery
study, Surg Endosc 17:1396–1403, 2003a. 135:273–281, 2004.
Sarli L, et al: Preoperative endoscopic sphincterotomy and laparo- Wenner DE, et al: Actual time required for dynamic fluoroscopic intra-
scopic cholecystectomy for the management of cholecystocholedo- operative cholangiography, JSLS 9:174–177, 2005.
cholithiasis: 10-year experience, World J Surg 27:180–186, 2003b. White R, et al: Current utility of staging laparoscopy for pancreatic and
Shoup M, et al: Port-site metastasis after diagnostic laparoscopy for peripancreatic neoplasms, J Am Coll Surg 206:445–450, 2008.
upper gastrointestinal tract malignancies: an uncommon entity, Ann Yamakawa T: An improved choledocho-fiberscope and non-surgical
Surg Oncol 9:632–636, 2002. removal of retained biliary calculi under direct visual control, Gas-
Shoup M, et al: Is there a role for staging laparoscopy in patients with trointest Endosc 22:160–165, 1976.
locally advanced, unresectable pancreatic adenocarcinoma? J Gas- Zacherl J, et al: Current value of intraoperative sonography during
trointest Surg 8:1068–1071, 2004. surgery for hepatic neoplasms, World J Surg 26:550–554, 2002.
CHAPTER 24
Intraoperative and immediate
postoperative management
Mary Fischer, Vittoria Arslan-Carlon, and Jose Melendez
OVERVIEW complications; however, this is not specific to liver surgery

(Cuthbertson et al, 2007).
Improvements in patient selection criteria, advances in hepatic Anesthetic management can be a Sophie’s choice: modifying
surgical techniques, and perioperative care have enabled one component to prevent a complication that leads to increas-
increasing numbers of older and previously deemed inoperable ing risk by a different pathway. Anesthetic care, like all of life,
patients to undergo hepatobiliary surgery. Despite this increase has risk, and this risk is determined by the preoperative pres-
in comorbidities, recent studies have documented the progres- ence of one or more comorbidities that significantly augment
sive safety of liver resection (Agrawal et al, 2011; Cescon et al, the incidence of postoperative adverse events. A key component
2009; Kingham et al, 2014; Mullen et al, 2006) in noncirrhotic of complications after surgery is the failure to recognize the
patients; nonetheless, morbidity remains significant even at patients at risk so that appropriate assessments occur prior to
tertiary care centers (Aloia et al, 2009, Kamiyama et al, 2010; surgery (Jhanji et al, 2008). Because anesthesia care is facilita-
Kingham et al, 2014; Kneuertz et al, 2012). This chapter tive rather than therapeutic, the main outcome of anesthesia
addresses the unique perioperative anesthetic considerations of care has been traditionally measured in terms of absence of
noncirrhotic patients undergoing hepatobiliary operations, with “complications.” Today it is rare for a patient to develop
emphasis on the role of the anesthesia care provider (ACP) to complications due directly to the act of anesthesia, yet anesthe-
improve morbidity. sia clinical decisions may impact the perioperative outcome.
Given that anesthetic drugs are fairly short acting, it is not
PREOPERATIVE EVALUATION obvious that consequences of anesthetic management could last
more than hours or days after surgery. There is arguably a shift
Risk and Outcome Improvement toward avoiding anesthesia-related harm after surgery, such as
Components to consider for postoperative morbidity include, the maintenance of normothermia and β-blockade or prescrib-
but are not limited to, comorbid medical conditions, periopera- ing target-controlled fluid management. The choices we make
tive care to modify risk, the volume of the future liver remnant, in the operating room may have an impact not only during the
and the ability to rescue should an adverse event occur. case, or in the immediate postoperative time, but long after the
Identifying patients at increased risk of a poor outcome patient is discharged from the postanesthesia care unit (PACU)
prior to surgery remains challenging, and what the anesthesi- (Grocott et al, 2012).
ologist can do to improve perioperative outcome is not always The occurrence of a 30-day postoperative complication is
obvious. more important than the preoperative patient risk and intraop-
There are a number of scoring systems to assist clinical risk erative factors in determining the survival after surgery (Khuri
assessment. By entering data into a multivariable prediction et al, 2005). Postoperative morbidity has been shown to adversely
model, an individualized patient risk score for certain morbidi- affect long-term outcome after hepatic resection; therefore
ties and mortality can be included to guide surgical planning efforts aimed at reducing perioperative morbidity will not only
and informed consent (Moonesinghe et al, 2013). Surgical risk reduce usage of resources but will likely further enhance the
prediction models, such as the Physiological and Severity Score therapeutic benefit of resection (Correa-Gallego et al, 2015;
for the Enumeration of Mortality and Morbidity (POSSUM), Ito et al, 2008). Given the impact of anesthesia intervention on
are not always based solely on preoperative data, nor are they long-term postoperative outcomes and costs, a natural evolution
procedure specific, and they are complex to use (Jones et al, of the Michigan Surgical Quality Collaborative has been to
1992). Anesthesiologists prefer a less complicated risk score— expand the data collection and collaboration efforts to include
the American Society of Anesthesiologists’ Physical Status the anesthesiology provider and process. Preoperative and
Score (Table 24.1) (Saklad, 1941)—which scores a patient on intraoperative data are collected and used to identify variations
a scale on which other patients may be compared, not the in anesthetic care and optimal care patterns (Kheterpal, 2011,
individualized risk prediction of an adverse outcome (Adams 2012). The data collected are used as a foundation of collabora-
et al, 2012). However, even if the anesthesiologist knows the tion between surgeons and anesthesiologists to establish process
risk prediction or risk score, it may not be apparent which of care and outcome measures to recommend best practice
anesthesia management factors are independent risk factors for clinical standards where prospective effectiveness trials may be
perioperative morbidity and amenable to preventive measures absent.
to improve outcome. Another option is to use a biomarker, such Although it is not possible to alter all risk factors, such as
as plasma B-type natriuretic peptide concentrations, to predict age, or avoid every potential consequence, such as pain, there
423
424 PART 3 ANESTHETIC MANAGEMENT: PREOPERATIVE AND POSTOPERATIVE CARE
TABLE 24.1 American Society of Anesthesiologists’ Physical Status Score (ASA-PS)

ASA Classification Examples
ASA I A normal healthy patient Healthy; no smoking; no or very minimal drinking

ASA II A patient with mild systemic disease Smoker; more than minimal drinking; pregnancy; obesity; well-controlled
diabetes; well-controlled hypertension; mild lung disease
ASA III A patient with severe systemic disease, not Diabetes, poorly controlled hypertension; distant history of MI, CVA, TIA,
incapacitating cardiac stent; COPD; ESRD; dialysis; active hepatitis; implanted
pacemaker; ejection fraction below 40%; congenital metabolic
abnormalities
ASA IV A patient with severe systemic disease that Recent history of MI, CVA, TIA, cardiac stent; ongoing cardiac ischemia or
is a constant threat to life severe valve dysfunction; implanted ICD; ejection fraction below 25%
ASA V A moribund patient who is not expected to Ruptured abdominal or thoracic aneurysm; intracranial bleed with mass
survive without the operation effect; ischemic bowel in the face of significant cardiac pathology
ASA VI A patient who has already been declared
brain-dead and whose organs are being
removed for transplant
The addition of an “E” indicates emergency surgery.
COPD, Chronic obstructive pulmonary disease; CVA, cerebrovascular accident; ESRD, end-stage renal disease; ICD, internal cardiac debifrillator; MI, myocardial infarction; TIA, transient ischemic attack.
are modifiable risk factors. Detailed evaluation and correction this noninvasive test to the LRCRI (Hwang et al, 2015). The
of all modifiable risk factors combined with best practice 2014 American College of Cardiology/American Heart Associa-
guidelines are our best choices to help avoid the most prevent- tion (ACC/AHA) guidelines on perioperative cardiovascular
able complications. Medical guidelines have rapid turnover, evaluation and care for noncardiac surgery are an excellent
with medical reversal a reality, and staying current to apply framework for evaluating cardiac risk in the perioperative
evidence-based practice is recognizing that the correct “scien- period for patients with clinical risk factors (Fleischer et al,
tific answer” may shift over time (Neuman et al, 2014; Prasad 2014).
et al, 2013). Not only is the patient population becoming older Cardiac functional status can be expressed in metabolic
with ever expanding comorbidities, but improvements in the equivalents or simply the inability to perform various activities
medical management of some chronic illnesses mean that the such as climbing two flights of stairs or walking four blocks
implications of such illnesses may be quite different today than (Eagle et al, 2002; Girish et al, 2001). Over the years, periop-
years past. Select patients might be better served having hepatic erative management has shifted from treating coronary obstruc-
surgery at a major medical center where the multidisciplinary tion with coronary revascularization toward medical therapy
care team provides optimum preparation of patients as well as aiming at prevention of myocardial oxygen supply and demand-
the ability and availability to minimize the impact on patients ing mismatch and coronary plaque stabilization. Today
when adverse events occur (Fong et al, 2005; Nathan et al, preoperative cardiac testing, cardiac stenting, and coronary
2009). revascularization are only performed for the same indications
as the nonoperative setting (Neuman et al, 2014).
Cardiac Evaluation The risks and benefits of continuing perioperative medical
Although the perioperative event rate has declined as a conse- management are complex. Perioperative β-blockade (PBB) has
quence of better anesthetic and surgical techniques, periopera- been shown to reduce the incidence of perioperative ischemic
tive cardiac complications remain a significant problem. The events and MI in patients with coronary artery disease but
first step in preoperative care is an adequate identification of confers questionable benefit and possible harm in patients
patients at risk for perioperative cardiac events. Clinical history, without coronary artery disease (Goldman, 2001; Lindenauer
physical examination, and review of a baseline electrocardio- et al, 2005). In 2008, the PeriOperative Ischemic Study Evalu-
gram usually provide enough data to estimate cardiac risk. ation (POISE) trial completely transformed the premise of
Estimation of a stable patient’s cardiac risk can be derived from PBB, finding PBB stroke morbidity outweighed any PMCE
the Lee Revised Cardiac Risk Index (LRCRI), a simple index prevention (Devereaux et al, 2006). POISE confirmed what
that identifies six independent risk factors to provide the risk clinical anesthesiologists had experienced increased intraopera-
of a cardiac complication in percent (Lee et al, 1999). The risk tive hypotension and bradycardia in low-risk patients with
of a perioperative major cardiac event (PMCE), defined as newly initiated PBB. PBB guidelines revisions followed rapidly
cardiac death, myocardial infarction (MI), or pulmonary edema in 2009. The guidelines recommend continuation of β-blockers
within 30 days postoperative, is the summation of an individual for patients who are already on them, the initiation of PBB for
patient’s risk and functional capacity and the cardiac stress high-risk patients, but not to initiate PBB in low-risk patients
related to the surgery. There are active cardiac conditions that in the perioperative period. β-Blockers should still be used to
may lead to cancellation of the procedure, unless the surgery is manage acute hypertension and tachycardia perioperatively in
emergent, but most PMCE risk is silent, and the LRCRI has patients at risk for myocardial supply and demand imbalance.
shown only moderate predictive performance (Ford et al, Similarly, balancing the risk-to-benefit ratio of dual antiplatelet
2010). The absence of coronary computed tomographic angi- therapy or aspirin interruption and the risk of stent thrombosis
ography findings of coronary artery disease confers low PMCE versus continuation and the risk of bleeding is challenging
risk regardless of clinical risk, and some have suggested adding (Newsome et al, 2008). The POISE-2 trial showed that
Chapter 24 Intraoperative and immediate postoperative management 425
perioperative aspirin did not reduce PCME at 30 days but did or dypsnea may benefit from this test of cardiopulmonary
increase perioperative bleeding (Devereaux et al, 2014). Only reserve to determine complication risk (Snowden et al, 2010).
a small number of patients in this trial had coronary stenting, Although it seems reasonable to assume that fitter patients will
and patients with a bare-metal stent for less than 6 weeks or a have better outcome, a recent study suggested that cardiopul-
drug-eluting stent for less than 1 year were excluded. monary exercise testing should not be used as a barrier to
Communication among the patient’s cardiologist, surgeon, patients undergoing liver surgery (Declan et al, 2014). Potential
and anesthesiologist is essential for the management of patients interventions to reduce PPCs include smoking cessation, pre-
with active or quiescent coronary artery disease. At our institu- operative exercise training, early mobilization, postoperative
tion, the medical consult takes into account both the ACC/ parental nutrition, and optimal treatment of pain (Smetana
AHA patient’s number of clinical factors (does the patient have et al, 2006).
active cardiac condition?; planned surgery, low or high risk?; Although obesity presents the anesthesiologist with signifi-
good functional capacity?; further testing required?) and the cant challenges, obesity per se is not a significant risk factor for
LRCRI (high-risk type of surgery?; history of ischemic heart PPCs and should not be used to deny a patient hepatic surgery.
disease?; history of congestive heart failure?; history of cerebro- There are two subsets of obese patients: One group is “the
vascular disease?; diabetes mellitus [insulin dependent]?; renal metabolically healthy but obese”; the other group is the “meta-
insufficiency [creatinine > 2]?), to determine if the patient is at bolically unhealthy but obese” (Glance et al, 2010). When an
acceptable risk to proceed to the planned surgery or if there is obese patient has three or more of the following criteria—
the need for further testing prior to surgery or pharmacologic abdominal obesity, increased trigycerides, decreased high-
intervention perioperatively (Neuman et al, 2014). density lipoprotein, elevated cholesterol, hypertension, and
The elderly patient will also have a geriatric assessment to glucose intolerance—the patient has a 2.5 increased incidence
test cognitive function, fall risk, and risk of postoperative of PPCs (Neligan & Fleischer, 2006). Obese patients are at risk
delirium. This assessment also includes advice on medications of suffering from a number of respiratory derangements,
to be used with caution and nonpharmacologic strategies to including obstructive sleep apnea (OSA), obesity-hypoventilation
reduce delirium and analgesics while the patient is in the syndrome, and restrictive impairment. The increase in body
hospital. mass also results in increased oxygen consumption and carbon
In hepatobiliary patients with a history of alcohol abuse, dioxide production. With these issues in mind, it is not surpris-
cardiac assessment needs to stress the evaluation of myocon- ing that acute PPCs are twice as likely in OSA patients (Dindo
tractile function. Two basic patterns of alcohol-induced cardio- et al, 2003). Many patients with OSA are undiagnosed, but
myopathy have been shown: left ventricular dilation with there is a strong relationship between obesity and OSA. The
impaired systolic function and left ventricular hypertrophy with ASA addressed this issue with practice guidelines, including
diminished compliance and normal or increased contractile assessment of patients for possible OSA before surgery and
performance. careful postoperative monitoring for those suspected to be at
risk (Gross, et al, 2006). It is unclear whether screening for
Pulmonary Evaluation OSA will affect surgical morbidity, but it is reasonable to ques-
Despite steady advances in care, patients with respiratory tion obese patients about symptoms that may suggest sleep
disease are still at increased risk for postoperative pulmonary apnea prior to liver surgery. At our institution, all obese patients
complications (PPCs). PPCs continue to rival cardiovascular are given the STOP (Snoring, Tiredness, Observed apnea, high
complications in frequency and severity after hepatic surgery blood Pressure)-Bang (Body mass index, Age, Neck circumfer-
(Kingham et al, 2014). ence, and gender) questionnaire (Chung et al, 2012). Given the
There are many limitations of studies that examine risk association of obesity and OSA with multiple medical
factors for PPCs, but there are some consistent patterns. Impor- conditions—increased risk of venous stasis, pulmonary embo-
tant risk factors for PPCs are the presence of pulmonary disease, lism, hypertension, cerebral vascular accidents, cardiomyopathy,
cigarette smoking, low preoperative arterial oxygen saturation, arrythmias, and ischemic heart disease—the anesthesiologist is
acute respiratory infection during the previous month, age, in a position to have an informed discussion with the patient
preoperative anemia, site of surgery (with upper abdominal, about the increased risk of morbidity and mortality and work
especially near the diaphragm, or intrathoracic surgery being the with other members of the patient’s care team to determine
highest risk), surgery duration of at least 2 hours, and emergency whether any interventions should be initiated before surgery in
surgery (Canet et al, 2010; Warner, 2000). Despite the increased an effort to minimize the risk of complications (Gupta et al,
risk of PPCs in patients with preexisting pulmonary disease, no 2001). Polysomonography is the gold standard for diagnosis of
prohibitive level of pulmonary function has been established for OSA, but it is expensive and a limited resource. The most
which surgery is contraindicated. Neither abnormal pulmonary reasonable approach is to check room air-pulse oximetry. If the
function testing or arterial blood gas analysis are useful in patient has an oxygen saturation level less than 96%, further
predicting risk. Thus these tests are only justified as part of an evaluation is warranted. A 2-week period of continuous positive
effort to optimize preoperative pulmonary status, either with an airway pressure (CPAP) therapy has been shown to be effective
immediate perioperative course of systemic corticosteroids or in correcting abnormal ventilatory drive and improving cardiac
antibiotics, or to advise if surgery should be delayed (McAlister function (Loadsman et al, 2001).
et al, 2003). Submaximal cardiopulmonary exercise testing is a After liver resection, the risk of venous thromboembolism
noninvasive objective test that measures a patient’s anaerobic (deep vein thrombosis, pulmonary embolus) is not insignificant
threshold. Poor functional capacity and especially low anaerobic and is higher in the obese patient. Prophylaxis is not routine
threshold has been associated with a high risk of postoperative for these patients after major hepatectomy due to the concern
complications and death (Older et al, 1999). This test is expen- that the risk of bleeding may outweigh the benefit of pharma-
sive; however, patients with low subjective functional capacity cologic prophylaxis, as well as the misconception that after
major hepatectomy the abnormal coagulation profile will

prevent this adverse event. The ACP and the surgeon should Alcohol Use Disorder
have a discussion about risk versus benefit. Patients with unhealthy alcohol use face increased perioperative
risks from the medical consequences associated with alcohol
Hepatic Evaluation (See Chapter 3) consumption as well as from physiologic dependence and
Risk factors and symptoms of liver disease are not as well withdrawal. Up to 50% of patients presenting for gastrointes-
defined as in other organ systems. There is no single biomarker tinal cancer surgery have alcohol use disorder. Studies have
for liver dysfunction; instead, the diagnosis of liver disease found that many surgical patients have not had been appropri-
requires a high degree of suspicion, with a careful probing of ately assessed for alcohol use in the preoperative evaluation
the clinical history to identify specific risk factors for liver (Chang & Steinberg, 2001). Abstinence from drinking, as
disease, such as previous blood transfusions, jaundice, travel, imposed by a hospital admission, places patients at risk for
tattoos, high-risk sexual behavior, illicit drug use, excessive alcohol withdrawal syndrome (AWS). The preoperative evalua-
alcohol intake, or chemotherapy (Suman et al, 2006). tion of patients with unhealthy alcohol use should include
The operative outcome for hepatobiliary surgery is linked to effective screening strategies to identify the presence of heavy
severity of hepatic parenchymal disease and the extent of hepa- alcohol use, detect end-organ damage secondary to alcohol
tectomy. Most hepatic resections are performed for metastatic consumption, and prompt intervention to address alcohol use
cancer. In these patients with mild or well-compensated hepatic prior to surgery. There are several screening tools to identify
disease, operative outcome is likely to be indistinguishable alcohol use disorder. The authors’ preference is to use the
from the outcome in the general population, although heavy CAGE (cut down, annoyed, guilty, eye opener) questionnaire.
pretreatment with chemotherapy can negatively impact the AWS prophylaxis should begin on admission to the hospital.
perioperative outcome (see Chapter 100). On the other hand, Evidence supports the use of benzodiazepines as first-line treat-
patients with hepatocellular carcinoma (HCC) have poorer ment (Schuckit, 2014). Two strategies are recommended: either
outcome than those with metastatic disease, given the propen- a fixed-dosage or an as-needed regimen triggered by symptoms.
sity of HCC to develop in the cirrhotic liver (Fattovich et al, These patients may have increased or decreased anesthesia
2003). requirements during induction and maintenance. The most
The goal of preoperative screening is to determine the pres- common 90-day postoperative complications are infections,
ence of preexisting liver disease without the need for extensive bleeding, and cardiopulmonary insufficiency; however, these
or invasive monitoring. Liver function tests can measure complications are only increased if the patient has alcohol abuse
different aspects of hepatic function, but as a group of tests, at the time of surgery (Rubinsky et al, 2013). For those patients
they lack specificity and are often affected by nonhepatic func- with alcohol abuse at the time of surgery, the development of
tion. These biochemical markers can not quantify hepatic cel- AWS is associated with a longer hospital stay and increased
lular dysfunction. In contrast, anesthesiologists are often mortality (Chang & Steinberg, 2001).
confronted with abnormal hepatic function tests in asymptom-
atic patients. In general, for asymptomatic patients with mildly Jaundice
elevated alanine and asparatate aminotransferase levels and a The etiology of hyperbilirubinemia may have an obstructive or
normal bilirubin concentration, cancellation of surgery is rarely nonobstructive cause (see Chapters 3, 8, and 51B). No matter
indicated. In patients with significant abnormalities, additional the root cause, jaundice adversely affects outcome. Unlike
investigation is warranted, given the higher risk to patients elevated bilirubin associated with hepatocyte disturbance,
having hepatic surgery to evaluate whether there is underlying obstructive jaundice is typically seen in patients with bile duct
cirrhosis or steatosis. Serologic testing to exclude viral hepatitis obstruction and is not a contraindication for hepatic resection
and human immunodeficiency virus should always be if the resection is being performed to remove the cause of the
performed. obstruction. In situations where there is clinical concern for the
The shift from extensive major resections to parenchymal- development of acute cholangitis, rapid biliary decompression
sparing operations does not eliminate the most feared and IV antibiotics should be administered preoperatively, and
complication—postoperative liver failure (PLF), the inability of surgery should be delayed until the infection resolves (see
the remaining liver to carry on its metabolic functions (see Chapter 43). Preoperative biliary drainage (either by percuta-
Chapters 100 and 108). The risk of PLF is closely related to neous or endoscopic approaches) does not improve the periop-
the volume and function of the remaining liver. Although the erative morbidity of patients in whom there is no evidence of
standard of care is CT, combined with three-dimensional CT infection and is therefore not routinely recommended. Elevated
volumetry to measure remnant liver volume (RLV) (Denys bilirubin is a risk factor for postoperative liver failure. Bile duct
et al, 2002) and to recommend if portal vein embolization is obstruction affects hepatic hemodynamics. Although acute
indicated prior to hepatic resection (Vauthey et al, 2000), biliary obstruction is associated with an increase in liver blood
preoperative virtual planning (Pathfinder software; Pathfinder flow, chronic obstruction is not, and its relief is not associated
Therapeutics, Nashville, TN) of the future RLV has been with a return to normal pressures. The exact cause of hemody-
shown to accurately predict RLV better than CT scans (Simpson namic derangement in the face of biliary obstruction is unknown,
et al, 2014). The quantity of hepatocyte dysfunction, whether but increased portal resistance is suggested to play a role.
induced by preoperative alcohol, chemotherapy, or metabolic Biliary sepsis may contribute to the exacerbation of the hemo-
syndrome, is more elusive to identify. The development of dynamic instability. Patients with hyperbilirubinemia are a
functional imaging techniques and virtual planning to deter- subset of patients with an increased risk of renal compromise
mine the quantity and the health of the liver left behind and after low central venous pressure (LCVP)-assisted hepatectomy
follow the liver as it regenerates function is promising (Simpson (Melendez et al, 1998). The surgeon and anesthesiologist must
et al, 2014; Wakamatsu et al, 2010). have a detailed discussion of risk versus benefit.
management. The liver is unique in that it receives a dual

Bloodless Surgery: Blood Conservation afferent blood supply equivalent to about 20% of the cardiac
In the early years of hepatic resection, major intraoperative blood output (see Chapter 8). The majority of HPF (70%) is via the
loss was common, transfusion of blood products was routine, portal vein and flows through the hepatic sinusoids, hepatic
and morbidity and mortality were unacceptably high (Foster & veins, vena cava, and back to the right atrium, and the remain-
Berman, 1977) (see Introduction chapter and Chapters 117 to der is derived from the hepatic artery. It is a low-resistance
120). General improvements in hepatic resection techniques reservoir that can store blood during hypervolemia and a source
have led to substantial reductions in intraoperative blood loss of blood during times of hypovolemia. Although hepatic outflow
and transfusion rates; nevertheless, perioperative transfusion may vary, hepatic inflow is constant by the reciprocal relation-
remains a potent predictor of increased perioperative morbidity ship of portal vein flow and hepatic artery flow (Lautt, 2007).
and mortality (Cescon et al, 2009; Jarnagin et al, 2002; Kingham When portal venous flow increases, hepatic arterial flow
et al, 2014). In recent years, liver resection without the need for decreases, and when portal vein flow decreases, hepatic arterial
blood transfusion has become increasingly possible. flow increases. This is the hepatic arterial buffer response
Patient blood management is based on the three pillars: because the hepatic artery adjusts its tone to keep hepatic blood
detecting and treating preoperative anemia, reducing the loss flow at a steady state. The reverse is not true; hepatic arterial
of red blood cells (RBCs) perioperatively, and optimizing tone does not affect portal venous flow.
the treatment of anemia (Spahn & Goodnough, 2013). Thor- Veins, particularly splanchnic veins, are much more com-
ough preoperative planning is essential to avoid perioperative pliant than arteries (Gelman, 2008), and with a higher density
allogeneic transfusion. Any history of bleeding disorders and of α-adrenergic receptors, are more sensitive to sympathetic
management of anticoagulation must be evaluated, including activation than arteries (Birch et  al, 2008). HBF is directly
discontinuation of drugs that adversely affect clotting (e.g., proportional to perfusion pressure (mean arterial or portal
acetylsalicylic acid, nonsteroidal antiinflammatory drugs, and vein pressure minus hepatic vein pressure) across the liver
anticoagulants). In patients with anemia, iron therapy may help and inversely related to splanchnic vascular resistance. Auto-
optimize the starting operative hemoglobin. Preoperative regulation of HPB is not prominent; therefore total HBF
autologous donation (PAD) also has been used to reduce the (arterial plus portal) can be modified by “surgery-related”
need for allogeneic red blood cell products (Brecker & Good- factors—stimulation, retraction, or manipulation—and several
nough, 2001). However, PAD may not avoid allogeneic blood “anesthesia-related” factors—positive pressure ventilation,
because almost half of the patients who donate blood before anesthetic technique, or drug effect on perfusion pressure or
surgery are anemic on the day of surgery, and preoperative splanchnic vascular resistance.
strategies to augment the RBC mass require more time than is Oxygen delivery to the liver may already be marginal because
generally reasonable for optimal efficacy. Patients with low most of the blood flow is with desaturated hemoglobin delivered
hemoglobin levels at the start of surgery are at an increased risk via the portal vein. A well-planned anesthetic maximizes the
of receiving allogeneic blood (Armas-Loughran et al, 2003). In relationship between oxygen supply and demand, with the
addition, PAD is costly, it can be associated with clerical errors, premise that reductions in systemic pressure will reduce HBF.
and for every two units donated, usually only one unit gets A good rule of thumb is anything that could result in significant
transfused (Goldman et al, 2002). If the patient is optimized reductions of systemic pressure and/or blood flow (cardiac
and surgery is bloodless, the autologous units are discarded. output–induced hypotension, hypovolemia, anesthetic over-
Other blood conservation strategies, such as intraoperative doses) should be avoided.
blood salvage (cell saver) and acute normovolemic hemodilu-
tion (ANH), have been used successfully for patients having Volatile Anesthetics
major liver resections and for Jehovah’s Witnesses (Jabbour Anesthetic Hepatotoxicity
et al, 2005). ANH has been shown in two prospective studies Fulminant hepatic necrosis and jaundice after halothane
to reduce the amount of RBCs transfused per patient in major (“halothane hepatitis”) is rare (1 in 6000 to 35,000 deliveries
liver resections (Jarnagin et al, 2008; Matot et al, 2002). Risk of anesthetics) but often fatal. Halothane hepatitis is an
factors for transfusion are cirrhosis, the extent of liver resection, immunologic phenomenon initiated by halothane metabolism
and portal hypertension. A score with good discriminatory and the binding of its metabolite to liver proteins, forming
ability to predict the necessity of RBC transfusion during liver trifluoroacetylated proteins, which stimulate the formation of
resection was developed (Sima et al, 2009), and subsequently, antibodies in susceptible individuals. Upon subsequent halo-
this score was incorporated into a nomogram to predict which thane reexposure, these antibodies mediate massive hepatic
patients would benefit from ANH to decrease transfusion necrosis. Because the extent of metabolism of enflurane,
during hepatic resection (Frankel et al, 2013). isoflurane, and desflurane is so much less than that of halo-
thane, fulminant hepatitis from enflurane, isoflurane, and
INTRAOPERATIVE MANAGEMENT desflurane is far less common than with halothane (Elliot et  al,
1993). In 1987, the U.S. Food and Drug Administration
The main goals of anesthesia for liver resection are to minimize concluded that there was no conclusive association between
blood loss, maintain intraoperative hemodynamic stability isoflurane exposure and postoperative hepatitis. Sevoflurane
during blood loss and vascular clamping, and follow an appro- metabolism is different from that of the other volatile anesthet-
priate transfusion protocol. ics because it does not result in trifluoroacetyled liver proteins,
and immune-based hepatitis after sevoflurane has not been
Hepatic Blood Flow reported. With the disappearance of halothane and enflurane
The impact of anesthesia and surgery on hepatic blood from clinical practice in developed countries, and lack of
flow (HBF) has important implications for intraoperative hepatotoxicity from either isoflurane, desflurane or sevoflurane,
anesthetic volatile hepatotoxicity is not a significant concern concentration is little changed after passage through the liver
(Elliot et  al, 1993). and depend on the intrinsic clearance (liver size, total enzyme
capacity) of the liver. The elimination of drugs with a low-
Hemodynamics extraction rate (benzodiazepines) depends more on the meta-
Volatile anesthetics reduce HBF in a dose-dependent fashion bolic capacity of the liver and less on the hepatic blood flow.
by affecting cardiac output and systemic pressure. Isoflurane In patients with impaired liver function, such drugs experience
has been considered the agent of choice in cases in which a prolonged length of activity with no increase in peak levels.
preservation of splanchnic blood flow is required. Liver blood The safety of IV anesthestic agents and muscle relaxants is
flow and the hepatic artery buffer response are maintained uncontested, yet, increasingly, anesthesiologists prefer agents
better in the presence of isoflurane than with any other volatile that are not influenced by liver function or using multiple drugs
anesthetic agent (Berendes et al, 1996a). In addition, isoflurane for the same effect despite liver dysfunction. Although the use
is shown to attenuate the increases in hepatic oxygen consump- of opioids is appropriate during liver surgery and the manage-
tion associated with surgery and liver manipulation. Desflurane ment is similar to other abdominal surgery, remifentanil, a short
is shown to have no deleterious effects on liver function and rapidly acting opioid, given by continual infusion and metabo-
hepatocyte integrity. Desflurane anesthesia is associated with lized by plasma esterases, is gaining in popularity over fentanyl.
significantly greater gut blood flow than equipotent isoflurane. The muscle relaxants atracurium and cisatracurium both
This difference cannot be explained by systemic hemodynamics undergo Hoffman degradation and ester hydrolysis, of which
alone. There is no difference in total hepatic flow between neither is dependent on liver function. Dexmedetomidine, an
isoflurane and desflurane groups, however, implying that an α2-agonist, and ketamine do depend on hepatic function;
intact hepatic arterial supply buffers response with desflurane however, perioperatively, their weak analgesic effects decrease
(O’Riordan et al, 1997). Sevoflurane seems to be similar to the minimum alveolar concentration of volatile vapors and the
isoflurane and desflurane with a few exceptions. Indocyanine postoperative opioid requirements (De Kock et al, 2001; Lin
green clearance is better preserved during sevoflurane anesthe- et al, 2014).
sia. Sevoflurane seems similar to isoflurane in its effect on
regional hepatic blood flow (Ebert et al, 1995). Epidural Anesthesia (EDA)
Nitrous oxide is used extensively in patients with hepatic Hepatic resection is performed under general anesthesia with
disease. It is not shown to contribute to hepatic disease exac- or without thoracic epidural block. The epidural block may be
erbation (Lampe et al, 1990). The sympathomimetic effects of used intraoperatively or solely to provide postoperative analge-
nitrous oxide decreases HBF. sia. Its benefit-to-risk ratio is controversial due to the role the
liver plays in the postoperative coagulation cascade (Matot
Intravenous Anesthetics and Muscle Relaxants et al, 2002) and the possibility that EDA may drive increased
Inhaled anesthetics supply all the aspects needed for anesthesia fluid therapy and transfusion (Page et al, 2008). Studies do not
in one package, but today most anesthesiologists choose mul- support the routine use of EDA and epidural analgesia (EPA),
tiple drugs to reach their goals: immobility, amnesia, suppres- rather than general anesthesia and parenteral analgesia, to
sion of autonomic reflexes, muscle relaxation, and analgesia. prevent postoperative mortality (Park et al, 2001; Rigg et al,
Over the last several decades, dramatic advances have been 2002; Wijeysundera et al, 2008). Nonetheless, enthusiasts
made in IV anesthesia, with the result that total IV anesthesia point out that there may be several other advantages to regional
is now a workable alternative to the traditional inhalation anesthesia, especially when combined with postoperative EPA:
anesthetic. Anesthesiologists using multiple drugs take advan- better pain control, attenuation of the stress response, decreased
tage of the interactions of drugs with different mechanisms of requirement of volatile anesthetics, and a reduction in the need
action but similar therapeutic effects. The therapeutic goal of for perioperative opioids.
the anesthetic can often be achieved with less toxicity and faster Perhaps the strongest arguments for the use of thoracic EDA
recovery than when the individual drugs are used alone in for liver surgery is its possible modulation of the immune
higher doses. response and its possible effect on tissue microperfusion (Sinis-
The liver plays a major role in biotransformation, the process calchi et al, 2015). Preclinical data, animal studies, and retro-
through which drugs are broken down into metabolites that can spective reviews demonstrate the potential for a decreased
be more easily eliminated. The main mechanisms that affect recurrence rate in some cancer types. Animal studies have
hepatic elimination of a drug are changes in HBF and changes shown that EDA could have an important role in modifying
in the ability of the liver cells to biotransform a drug for excre- tissue microperfusion and protecting tissue from ischemic
tion. These two mechanisms, hepatocyte function and HBF, damage, regardless of the effects on hemodynamics. The notion
have an important role in the choice of anesthetics for patients that anesthesiologists may be able to impact the short-term and
undergoing hepatobiliary surgery because even small changes long-term outcome for a cancer patient simply by incorporating
in liver function or blood flow can change the concentrations regional anesthesia is appealing, although unproven, and more
of drugs and their metabolites. High-extraction drugs (ket- prospective randomized research is needed.
amine, flumazenil, morphine, fentanyl, sufentanil, lidocaine)
are directly related to liver blood flow and essentially cleared SPECIAL ANESTHETIC CONSIDERATIONS
as they pass through the liver. Protein binding, enzymatic
induction, intrahepatic shunting, and the effect of anesthetics Anesthesia for Hepatectomy
on liver blood flow may affect the elimination of drugs with a For hepatic resection, a relationship between extent of intraop-
high-extraction rate. Reductions in metabolic clearance result erative blood loss and mortality and morbidity has been con-
in increases of peak drug level with minimal change in the sistently shown (see Chapter 103). To minimize blood loss, it
elimination half-life. Low-extraction drugs are those whose is common anesthesia practice to perform liver resections with
was required (Melendez, et al, 1998). However, in the modern

era of bloodless hepatic resection, to avoid the morbidity associ-
ated with central vein cannulation, our clinical practice has
abandoned the routine use of central venous lines. The authors
CVP have not adopted surrogate measures of CVP: external jugular
venous pressure, peripheral venous pressure, inferior vena cava
diameter using transesophageal echocardiography, inferior
A vena cava collapsibility using hand-carried ultrasound devices,
or stroke-volume variation (Dunki-Jacobs et al, 2013); instead,
the success of our LCVP is currently based on surgical visual
inspection of the vena cava or the amount of venous back
bleeding as reported by the surgeon. Patients should still be
prepared for large-volume transfusion, although this is infre-
quently needed. Close cooperation between the anesthesiologist
and the surgeon continues so that likely difficulties can be
CVP anticipated and appropriate measures can be taken. Fluid
B management is an important aspect of the LCVP anesthesia.
FIGURE 24.1 Vena caval injury profile under various central venous
Intraoperative fluid management is divided into two phases.
pressure (CVP) conditions. A, High CVP. B, Low CVP. Increased CVP
Phase 1: Prehepatic
leads to distention of vena cava, with ensuing enlargement in diameter
of injury and increase in the bleeding driving pressure. Prehepatic resection starts at anesthesia induction and ends at
the completion of parenchymal transection and hemostasis.
During this phase, inflow control of the portal vein and hepatic
the CVP less than 5 mm Hg. The blood loss resulting from a artery are achieved, and the vena cava and hepatic veins are
vascular injury is directly proportional to the pressure gradient dissected. Sixty percent of the time, hepatic parenchymal
across the vessel wall and the fourth power of the radius of the transection is performed with intermittent inflow occlusion
injury. If the CVP is lowered from 15 mm Hg to 3 mm Hg, the (Pringle technique) applied (Kingham et al, 2014).
blood loss through a vena caval injury consequently falls by a This phase avoids fluid excess and takes advantage of the
factor greater than 5. LCVP not only lessens the pressure vasodilatory effects of anesthetic drugs. Preoperative overnight
component of the equation but also minimizes the radial fluid replacement is withheld, and maintenance fluid require-
component of flow by reducing vessel distention (Fig. 24.1). ment at 1 mL/kg/hr of balanced crystalloid solution is infused
LCVP anesthesia is designed to preclude vena caval and hepatic until the liver resection is completed. Intermittently, small fluid
venous distention, facilitate mobilization of the liver and dis- boluses or vasoactive drugs may be given to maintain hemody-
section of retrohepatic vena cava, minimize hepatic venous namic stability. Some extent of permissive oliguria due to
back bleeding during parenchymal transection, and facilitate decreases in antidiuretic hormone or permissive relative hypo-
control of inadvertent venous injury (Melendez et al, 1998). tension while peripheral tone is decreased will allow for contin-
LCVP anesthesia is often performed in combination with ued minimal fluid infusion until the specimen is delivered.
surgical inflow and outflow vascular control (Pringle technique) Anesthesia is maintained with a combination of isoflurane in
before parenchymal transection (Cunningham et al, 1994). oxygen and fentanyl. Isoflurane provides vasodilation with
Twenty years ago, the authors developed and reported a minimal myocardial depression (Schwinn et al, 1990). Consis-
simple, effective, and reproducible technique for decreasing the tent with its minimal effect on cardiac output and systemic
intraoperative blood loss in patients undergoing liver resection pressure, fentanyl has no effect on liver blood flow and oxygen
based on fluid restriction and the vasodilatory effects of anes- delivery and, given its lack of toxic metabolites, can be admin-
thestic agents (Melendez et al, 1998). Around the same time, istered similar to any abdominal surgery without any dosing
a complex LCVP management technique was described (Rees reduction (Trescot et al, 2008). Shortly before transection of the
et al, 1996). This used epidural blockade and IV nitroglycerin. liver, sublingual nitroglycerine is applied. This combination of
These patients often required intraoperative dopamine for isoflurane, fentanyl, and sublingual nitroglycerine readily pro-
systemic pressure support. The technique seemed cumber- vides the favorable LCVP environment for hepatic resection.
some; adding an unnecessary level of complexity to an already
challenging situation. Despite this, both approaches contributed Postresection: Phase 2
to improved outcomes and continue to be practiced at major Posthepatic resection, the second phase, begins once the speci-
institutions (Correa-Gallego et al, 2015; Dunki-Jacobs et al, men has been delivered and hemostasis secured. During this
2013; Jones et al, 2013; Lin et al, 2014). phase, the goal of fluid prescription is to leave the operating
room with a normovolemic patient.
Low Central Venous Pressure Technique:
General Anesthesia Goal-Directed Individual Fluid Therapy
LCVP-assisted hepatic resection at the authors’ institution has Goal-directed individual fluid therapy (GDT) based on the
evolved over the years but is still true to the two original pillars: optimization of cardiac output or using dynamic indices of
fluid restriction and pharmacologic vasodilation. This anes- blood flow and arterial pressure to guide fluid requirement has
thetic technique was historically dependent on the presence of appeal and has become more common (Cannesson, 2010). The
a central venous catheter to provide hemodynamic information availability of minimally invasive hemodynamic monitoring
and expeditious and reliable access in case rapid resuscitation techniques (esophageal Doppler, arterial waveform analysis)
and the use of dynamic parameters of fluid responsiveness allow blockade, systemic hemodynamics, or even the circulating
the use of protocolized GDT strategies. Functional hemody- effects of local anesthetics. Splanchnic veins, with their higher
namic parameters provide a numeric representation of the density of α-adrenergic receptors, play the main role in main-
patient’s fluid responsiveness and are more reliable than using taining a ratio between stressed (Vs) and unstressed blood
standard static parameters, such as blood pressure, heart rate, volume (Vu) (Gelman, 2008). Vu is hemodynamically inactive,
urine output, or even CVP (Auler et al, 2008). A recent meta- but when venoconstriction changes it, this is equivalent to a
analysis supports GDT for patients having abdominal surgery, transfusion of a significant amount of blood. Controlled ventila-
to improve postoperative recovery and decrease complication tion and EDA both decrease venous return (VR) and must be
rates (Pearse et al, 2014). Liver surgery exposes patients to associated with an increase in Vs to maintain hemodynamics.
periods of cardiovascular insufficiency, either because of Decreased Vs and VR can be restored by fluid infusion to fill
anesthesia-induced loss of vasomotor tone and baroreceptor up the increased venous capacity or by an α-agonist to increase
responsiveness or because of blood loss and mechanical sympathetic tone of the compliant veins, which robs from Vu
obstruction to blood flow. In all cases, stroke volume will fall to give to Vs. The clinical advantage of using a vasopressor is
as well as global oxygen delivery to the tissues. GDT is targeted that it maintains tissue blood flow but avoids fluid infusion.
to detect hypovolemia and hypoperfusion early (at or before However, a clinician who practices LCVP should realize that
anesthetic induction) to be proactive to avoid hypoperfusion. this approach might decrease the margin of safety. Low Vu per
Early proactive GDT may compromise the effectiveness of se is not harmful. Up to 1000 mL of blood may be lost without
LCVP on decreased blood loss. Although most liver surgery change in standard hemodynamic parameters. However, beyond
does not result in profound tissue hypoperfusion, some degree this point, when the mobilization of blood from Vu to Vs is
of hypoperfusion does occur with the Pringle technique, adding approximately complete, even minor reduced VR, whether by
tissue perfusion injury before resuscitation. There is presently the Pringle technique, vena caval compression, or blood loss,
no approved treatment for ischemia reperfusion injury (IRI), can quickly lead to hemodynamic deterioration. Vasopressors
the inflammation that occurs when blood flow is returned to during LCVP-assisted liver surgery are certainly justified;
healthy liver tissue after diseased tissue has been surgically however, it may delay recognizing dangerous hypovolemia.
removed. IRI can harm the vascular barrier, in particular the On the other hand, if fluid is infused to counteract the
endothelial glycocalyx (Chappell et al, 2014), and results in hemodynamic effects of EPA this may lead to excessive hydra-
part from the deposition of complement, a protein that kills tion and increased packed RBC transfusion after hepatectomy
liver cells and impairs regeneration (Chappell et al, 2014). (Page et al, 2009). The liver is the only organ to have regional
Because surgery also creates a cytokine storm, the combination blood flow monitored by the autonomic nervous system, and
of relative hypoperfusion and immune modulation will alter the whether the effects of an epidural on perfusion pressure may
microcirculation, causing subclinical damage. Whether GDT be reversed by vasopressors is unclear, as well as the influence
for liver surgery can rescue patients from this insult is unknown. of EDA on fluid balance.
GDT has been shown to be effective when combined in an
enhanced recovery after liver surgery (ERAS) program (Hughes Air Embolus
et al, 2014; Jones et al 2013). A recent randomized trial for The goal of keeping a low central pressure to minimize back
hepatic surgery at our institution showed intraoperative GDT bleeding from the liver sinusoids during transection must be
was a safe technique, allowed for less intraoperative fluid, but counterbalanced by a central pressure that minimizes the risk
did not influence overall 30-day morbidity. The optimal peri- of air entrainment. The risk of intraoperative air emboli is likely
operative fluid resuscitation strategy for liver resections remains to increase under LCVP anesthesia. Elimination of nitrogen
undefined. from the anesthetic gas mixture is necessary to permit expira-
tory nitrogen monitoring for air emboli. Restriction of nitrous
Low Central Venous Pressure: Epidural Anesthesia oxide in the gas mixture prevents the diffusion-mediated
EDA may be used to provide LCVP conditions during hepatic increase in the size of circulating air. Transesophageal echocar-
surgery (Feltracco et al, 2008; Jones et al, 2013). Despite many diography can be used to monitor air emboli, but this technol-
animal and human studies, the effects of thoracic EDA on HBF ogy is sensitive and overdiagnoses clinically insignificant events.
are not entirely clear. An early study using the indocyanine At our institution, during open hepatectomy, surgical and
green plasma disappearance rate revealed that EDA resulted in anesthesia vigilance and communication are the keys to detect
a decrease in mean arterial pressure and HBF (Kennedy et al, and treat air emboli. With surgical watchfulness and rapid
1971). Another study using indocyanine green and EDA con- occlusion of open venous channels, and our monitoring of
cluded that maintaining normotension with volume infusion end-tidal carbon dioxide and hemodynamics, LCVP anesthesia
does not reverse its hemodynamic effect; however, the decrease results in a low incidence of clinically significant air emboli.
in flow is reversed by dopamine (Tanaka et al, 1997). Patients
treated with norepinephrine to compensate for the EDA- Bloodless Surgery: Blood Conservation, Transfusion
induced decrease in arterial blood pressure had an additional The anesthesiologist not only has a pivotal role in reducing
marked decrease in HBF (Meierhenrich et al, 2009). blood loss during hepatic surgery but also contributes to intra-
Subsequently, these studies have been shown to oversimplify operative blood conservation and controls transfusion protocol.
the numerous factors affecting HBF: hemodynamics, auto- Transfusion-free surgery, better known to the public as blood-
nomic nervous system, circulating neurohumoral agents, and less surgery, can only be achieved by the application of blood
local metabolites (adenosine). Today it is accepted that this management techniques to decrease allogeneic transfusion.
multiplicity of regulatory mechanisms provide overlapping The three pillars of blood conservation are (1) build up the
controls to maintain tissue perfusion (Siniscalchi et al, 2015). patient’s own blood, (2) reduce blood loss, and (3) recycle the
EDA can interfere with all these factors either by sympathetic patient’s own blood (Spahn & Goodnough, 2013). The liver
anesthesiologist should be familiar with recycling techniques, (Kooby et al, 2003). Patients receiving only autologous blood
intraoperative RBC salvage and hemodilution, maneuvers that or spared-blood transfusion from ANH also did not have better
may contribute to reduced allogeneic RBC transfusion in disease-free survival (Correa-Gallego et al, 2015; Kooby et al,
major liver surgery. Autologous RBC salvage (intraoperative 2003). Postoperative complications, especially infectious and
autotransfusion) involves recovery of the patient’s shed blood other tumor-related factors, such as tumor-infiltrating lympho-
from a surgical wound, washing or filtering, and reinfusion of cytes, are more dominant determinants of long-term cancer
the blood into the patient. Hepatectomies often are performed survival (Khan et al, 2014). The magnitude of operative blood
for cancer. Cell salvage had been excluded in oncologic surgery loss during resection of HCC was found to be a predictor of
because of the concern for potential dissemination of cancer recurrence and survival rates; however, the blood loss was
cells, but the availability of leukocyte-depleting filters allows its found to be related to tumor characteristics and extent of
use during cancer surgery. Another transfusion-sparing tech- surgery (Katz et al, 2009). RBC transfusions are indisputably
nique is acute normovolemic hemodilution (ANH). ANH is a associated with an increase in mortality and morbidity in liver
recycling technique that can be performed intraoperatively by surgery; however, transfusion may be a surrogate marker of one
the anesthesiologist. Blood is removed from the patient after or more other variables that is more directly related to the
induction and replaced with crystalloid or colloid fluid. The complication.
removed blood is stored at room temperature in the operating
room and is returned to the patient at the conclusion of the Minimally Invasive Liver Resection
operation. Two randomized studies of ANH in major hepatic Anesthesiologists continually adjust strategies as innovative
resection showed a significant reduction in the percentage of surgical techniques evolve. The benefits of laparoscopic liver
patients requiring allogeneic RBC transfusion (Jarnagin et al, resection (LLR) have been associated with less blood loss and
2008; Matot et al, 2002). Tolerance of normovolemic anemia earlier postoperative recovery (Ito et al, 2008), although all of
is important, and care must be taken not to confuse the the comparisons reported are retrospective and therefore asso-
momentary helpful effect of RBC transfusion on hypotension ciated with a huge selection bias. Pneumoperitoneum induces
and hypovolemia with an outcome benefit. A restrictive predictable pulmonary and renal responses as well as phasic
approach to blood transfusion with a threshold of 7 g/dL has hemodynamic changes. Intraperitoneal insufflation and head-up
been shown to reduce blood use and not cause harm in criti- tilt result in impairment of HBF secondary to decreases in
cally ill patients (Holst et al, 2014), as well as in liver resection cardiac output (Berendes et al, 1996b; Eleftheriadis et al,
patients (Wehry et al, 2015). 1996). In well-selected patients, the consequences of these
In a recent review, Kingham and colleagues reported that changes are not relevant. However, the challenges of the effect
transfusions have decreased by 50% over 2 decades, but peri- of pneumoperitoneum, positioning, and mechanical ventilation
operative blood loss and transfusion remain associated with on cardiopulmonary function in addition to longer surgical time
morbidity. Cannon and colleagues (2013) specifically showed may not be the correct choice for every patient. Before position-
that transfusion with packed RBCs was associated with post- ing, the patient should be prepared the same as for an open
operative complications in patients undergoing hepatectomy. case, including the ability to do large-volume transfusion in
The correlation of transfusion with complications should not case inadvertent major vessel bleeding occurs. A protocol
be interpreted as a direct cause and effect relationship; instead, should be in place as well as a fully open instrument tray and
what is important is the fact that patients having liver surgery equipment available in case the robot must be undocked emer-
receiving very few units of blood transfusion nevertheless gently to convert to open surgery. It is important that the
have higher complication rates (Kooby et al, 2003). There are anesthesiologist and surgeon have a discussion before the start
increased infectious complications in those transfused: The of the laparoscopic or robotic liver case because the risk versus
more transfusions that are done, the higher the rate of total benefit of LCVP-assisted hepatic resection is less clear.
complications and infectious complications (Page et al, 2009). The pneumoperitoneum compresses the portal vein, reduces
The cause is not known, but one possible mechanism proposed portal blood flow, and seems to reduce hepatic back bleeding
is “transfusion-related immunomodulation” (Vamvakas et al, during transection. Decreasing blood loss for open resection is
2014). Patients who have received a blood transfusion stored based on keeping the radius of an inadvertent venous injury
29 days or more have twice the rate of infections. A hypothesis small and the pressure head low so that less blood will be lost
is that as stored RBCs break down, they release cytokines that through the opening. However, retractions can tort vessels and
can lower immune function (Vamvakas, 2002). One could actually stent them open during LLR. Fluid management is
hypothesize that by reducing transfusion, complications would complicated by compromised hemodynamics resulting from
also be reduced. A prospective trial of ANH reduced RBC positioning and decreased lung compliance. Renal parenchyma
transfusion but did not reduce morbidity (Jarnagin et al, 2008). and venous compressions during pneumoperitoneum are the
Most liver surgery is performed for cancer, and studies have etiology of oliguria during LLR. The effects are reversible
examined the association of cancer progression and RBC and usually cause no harm. Yet, many anesthesiologists prefer
transfusion for primary and metastatic cancer. Because it was to optimize intravascular volume to minimize the effects of
reported in the journal Anesthesiology that blood transfusion in intraabdominal pressure (IAP) on renal and cardiac function.
rats promoted cancer progression (Atzil et al, 2008), there has The issue of gas embolism (GE) during LLR is still debated.
been a perception that an anesthesiologist’s choice to transfuse Some authors consider it little or no problem, and some con-
RBCs can influence long-term survival. Human studies have sider it a real threat to patient safety (Min et al, 2007). The
not supported animal studies. A retrospective reivew of 1300 debate centers on the theory that GE occurs when the IAP
hepatectomy patients with metastatic colorectal cancer reported exceeds CVP (Eriksson et al, 2011). The opposite situation,
the major effect on survival is in the immediate postoperative when CVP exceeds IAP, does not prevent GE because it can
period, but transfusion did not predict long-term survival occur irrespective of whether CVP is greater or less than IAP
(Fors et al, 2012). Positive pressure ventilation causes rhythmic

variations in venous return for both pressure and flow. With Cardiopulmonary
an open vein because of entrainment during the phase with Cardiopulmonary complications have decreased 50% over the
higher flow, gas from the abdomen might reach the venous past 3 decades (cardiovascular: 20% to 13%; pulmonary: 30%
circulation. Both carbon dioxide and the argon beam coagula- to 9%) according to a recent review of patients undergoing liver
tion (ABC) during liver surgery have been associated with GE resection for malignancy (Kingham et al, 2014). This improve-
(Vibert et al, 2006). ment parallels decreased blood loss with parenchymal sparing
surgery, decreased transfusion rate, and improved perioperative
Ablation fluid management.
Major hepatectomies have decreased, whereas hepatectomy
with concurrent surgery and repeat hepatectomies with or Cardiac Dysfunction
without simultaneous ablation have increased. The develop- The incidence of postoperative myocardial infarction may be
ment of ablative techniques for tumor ablation has been one of explained by better patient selection and optimization;
the major advances for liver cancer. Ablation therapy for benign however, the absence of benefit of preoperative revasculariza-
or malignant liver tumors is often used as an alternative to tion in the face of known coronary artery disease may be
surgery, the principal aim being to ablate the undesirable areas explained, first, by the overall skill with which ACPs manage
without damaging the surrounding healthy tissue. At our perioperative stressors, and second by the nature of periopera-
institution, ablation is mostly used for parenchymal-sparing tive MI. Ischemia can occur due to excess demand versus
procedures during concurrent major hepatectomy or when supply (type 2 MI) (Landesberg et  al, 2009) as triggered by
comorbid conditions preclude major liver surgery. The treat- hypertension, tachycardia, hypoxia, anemia, or hypotension.
ments currently available, such as low-temperature cryosurgery, Anesthesiologists can and do manage these conditions both
nonselective chemical ablation, focused ultrasound, radiofre- intraoperatively and postoperatively. Ischemia can also occur
quency ablation (RFA), microwave ablation (MWA), or elec- with acute plaque rupture and thrombosis formation (Lande-
troporation, have their own specific advantages, disadvantages, berg et  al, 2009). These events occur more often in the
and particular applications. At our institution, the ACP may operative setting due to increased hypertension, tachycardia
care for patients undergoing an ablative procedure during open (shear forces), hypercoagulability, and surgery-associated
or laparoscopic surgery or percutaneous ablation in the inter- inflammatory response. The POISE trials give credibility to a
ventional suite. Experience and rapidly changing technology calculated strategy of decreasing heart rate while avoiding
have overcome many of the issues that challenged the ACP in perioperative hypotension (avoid MI and stroke). An anesthe-
the early years of RFA or cryotherapy and replaced them with siologist should be able to manage these hemodynamic issues
newer concerns. RFA is by far the most frequently used proce- that can precipitate an MI, but the triggers to inflammation
dure; however, for technical reasons, MWA and irreversible and hypercoagulation are poorly understood and not yet
electroporation are becoming common. MWA is faster and can modifiable by an anesthetic method. The ACPs involved in
be used for larger tumors, creating greater cell lysis (Groeschl nonpreventable events may have their clinical vigilance and
et al, 2014). Electroporation requires the use of an electrocar- ability to rescue the patient questioned, or they may feel as
diogram synchronizer to protect the patient from arrhythmias personally responsible as those involved in preventable events
and dense muscle relaxation to prevent upper muscle body (Gazoni et  al, 2012).
contraction (Ball et al, 2010). Myocardial injury after noncardiac surgery is common and
not necessarily revealed by ischemic features (symptoms or
POSTOPERATIVE CARE electrocardiographic findings). Elevated troponin T measured
in a large international study reported that an elevated troponin
Hepatobiliary surgery is increasingly common at major T, irrespective of ischemic features, independently predicted
medical centers, and postoperative recovery is becoming indis- 30-day mortality (Botto et al, 2014).
tinguishable from that after other major abdominal surgery.
Some of the problems causing metabolic and functional Pulmonary Dysfunction
changes after hepatobiliary surgery are common to all major Postoperative pulmonary complications (PPCs) include pneu-
intraabdominal procedures, whereas others are unique and monia, respiratory failure, bronchospasm, pleural effusions,
require an in-depth understanding of liver physiology. Imme- atelectasis, hypoxemia, and exacerbation of underlying chronic
diate postoperative concerns are impaired cardiopulmonary lung disease. Atelectasis and pleural effusion are common
function and oxygenation, bleeding and intravascular volume, consequences of anesthesia and surgery after liver resection.
renal and liver dysfunction, delirium, analgesia, and infection With aggressive postoperative pulmonary toilet and early mobi-
(see Chapter 25). lization, these minor problems resolve without the need for
Important laboratory tests are hematocrit, electrolytes further intervention.
including magnesium and phosphorus, serum creatinine, blood At least three mechanisms contribute to impaired pulmonary
urea nitrogen, prothrombin time/partial thromboplastin time, function in the postoperative period. First, respiratory muscles
liver enzymes, and chest radiography. At our institution, most disrupted by surgical transection (abdominal muscles) will not
patients after major liver surgery recover in the postanesthesia function normally. Second, patients may limit motion of respira-
care unit day 1 and then are transferred to the floor. This allows tory muscles to minimize postoperative pain. Finally, stimula-
anesthesiologists to remain engaged in the postoperative man- tion of the visceral afferent nerves markedly changes the
agement of patients, to be confident that they will receive activation of respiratory muscles. For example, removal of the
appropriate care and not suffer morbidity from what was a gallbaldder activates vagal efferents, which produces a reflex
potentially avoidable anesthetic complication. inhibition of diaphramatic activity. Of note, laparoscopic surgery
may ameliorate the first two mechanisms but not the third, and Particular attention must be paid to anesthetic elimination
significant decrements in pulmonary function may still be and residual drug effect, especially sedative, analgesic, and
observed after laparoscopic surgery. neuromuscular blockers, in posthepatectomy patients who have
ACPs play a key role in the prevention of PPCs. Preoperative altered drug pharmacodynamics and pharmacokinetics. In a
preparation, intraoperative management, and immediate post- recent prospective cohort study, the use of intermediate-acting
operative care can have a major impact on the occurrence of neuromuscular blocking agents was associated with postopera-
this morbidity. It has long been known that the induction of tive desaturation (90%) and reintubation, regardless of train-
general anesthesia, both total IV anesthesia and inhaled anes- of-four monitoring and use of reversal agents (Grosse-Sundrup
thesia, decreases lung volume and promotes dependent zone et al, 2012). In a cohort of 33,769 surgical cases, planned
atelectasis (Putensen et al, 2009; Serpa Neto et al, 2012). To reintubation within the first 3 days after surgery was associated
improve oxygenation, ACPs apply a high tidal volume (TV) and with a 72-fold increased risk of in-hospital mortality (Brueck-
high fraction of inspired oxygen (FiO2). High FiO2 via absorp- mann et al, 2013).
tion atelectasis has also been linked to the development of The findings of a recent systematic review and meta-analysis
atelectasis during the postoperative period (Edmark et al, suggest that the use of early CPAP for the prevention of hypox-
2003). After liver surgery, like all abdominal surgery, this emia after abdominal surgery may reduce the incidence of
impaired oxygenation caused by intraoperative atelectasis per- PPCs compared with just supplemental oxygen (Ferreyra et al,
sists for days after the surgical procedure (Duggan et al, 2005). 2008).
An association, but not a causal effect, of atelectasis and PPC Anticipated respiratory compromise after liver surgery in
has been reported (Milic-Emili et al, 2007). There is experi- patients without an epidural may preclude early extubation,
mental work that links atelectatic areas of the lung to transloca- especially in combination with any baseline abnormality in gas
tion and increased bacterial growth, providing an optimal nidus exchange. If a patient must be placed on mechanical ventila-
for infection (Van Kaam et al, 2004). Regardless of the etiology, tion, diaphragmatic weakness, atrophy, and respiratory muscle
PPCs increase 30-day mortality (Canet et al, 2010). fatigue can occur within hours. Similar to all muscles, complete
The use of high TV (10 to 15 mL/kg) ventilation, encour- rest can lead to diaphragm atrophy, and modes that allow
aged to prevent atelectasis, results in high peak ventilatory patient triggering, such as assist-control ventilation are neces-
airway pressure, which has been shown to be associated with sary to maintain diaphragm muscle function (Powers et al,
acute lung injury (Fernandez-Perez, et al, 2009). Low TV 2008).
ventilation (TV 6 to 8 mL/kg ideal body weight) with positive
end-expiratory pressure (PEEP) (6 to 8 mm H2O) and recruit- Venous Thromboembolism
ment maneuvers every 30 minutes (lung protective ventilation) Patients having cancer surgery have a moderate risk for venous
has demonstrated reduced mortality in patients with acute thromboembolism (VTE) (Agnelli et al, 2006; Alcalay et al,
respiratory distress syndrome and decreased PPCs in patients 2006; Catheline et al, 2000). Guidelines recommend low-
at risk of PPCs after abdominal surgery (Futier et al, 2013). molecular-weight heparin for patients undergoing general
Yet, most anesthesiologists decrease TVs without the addition surgery procedures with at least moderate (3%) risk of VTE, if
of PEEP or respiratory maneuvers (likely because of their the risk of bleeding does not negate the risk of VTE (Gould
effects on hemodynamics), which in a large retrospective cohort et al, 2012). After major hepatectomy, the concern for postop-
may have lead to increased mortality (Levin et al, 2014). The erative bleeding, combined with an erroneous presumption of
authors speculated that this was increased atelectasis from protection due to the coagulopathy, often preclude the use of
lower TV ventilation. Lung protective ventilation is being routine prophylaxis despite evidence to the contrary (Tzeng
proposed as a standard of care to be bundled into ERAS or et al, 2012). In a recent retrospective review at our institution,
perioptimal care to further decrease the incidence of PPCs and postoperative VTE occurred in 2.6% of patients and was
serve as a measure of the quality and safety of care. The liver independently associated with higher postoperative interna-
ACP must weigh the evidence for lung protective ventilation tional normalized ratio (INR) and LWMH had no relationship
and the effect of high PEEP and respiratory maneuvers on to VTE incidence or bleeding complications (Nathan et al,
decreased venous return or liver congestion on a patient-by- 2013). Patients with more extensive liver resections and higher
patient basis. operative blood loss had a higher incidence of VTE. Despite an
Postoperative respiratory failure is most commonly defined elevated INR and lower platelet count, a patient having major
as the need for mechanical ventilation for more than 48 hours liver cancer surgery is in a normocoagulable or even a hyper-
or unplanned postoperative reintubation. The nature and coagulable state and at risk for venous thromboembolism; yet
magnitude of the preexisting respiratory conditions determine there is no consensus of opinion for pharamacologic prophylaxis
the effect of a given standard anesthetic on respiratory function. (Barton et al, 2013).
The ACP plays a key role in the prevention of respiratory
muscle-related PPCs. In patients with obstructive lung disease, Liver Dysfunction (See Chapters 79 to 81)
high airway resistance favors deep slow respiration, which may Jaundice
not be possible in patients with large abdominal incisions. The Liver regeneration after partial hepatectomy involves rapid cell
treatment of this pain with opioids will reduce minute ventila- division of 24 to 72 hours and can be characterized by signifi-
tion and respiratory drive. A meta-analysis of patients undergo- cant changes in cellular phosphorus metabolism (Zakian et al,
ing thoracic and abdominal surgery showed that the odds of 2005). Standard therapy for posthepatic resection should
developing postoperative pneumonia was reduced nearly 50% include the supplementation of IV fluids with potassium phos-
in patients receiving epidural versus IV analgesia, fueling the phate (30 to 40 mmol/day).
controversy of this choice in patients after liver resection Mild jaundice in the early postoperative period after partial
(Popping et al, 2008). hepatectomy is relatively common, particularly in patients who
undergo a major resection and who have a transient, self- surgery, by itself, is generally unimportant clinically. Even in
limited hyperbilirubinemia in the first few days after operation, the presence of severe hypoalbuminemia, albumin infusion is
which may be exacerbated by hematoma reabsorption and of no beneficial effect. The medical management of ascites
hemolysis of transfused blood. emphasizes bed rest and sodium restriction. If a spontaneous
The portal circulation is valveless: Changes in pressure are diuresis does not occur, therapy is begun with the aldosterone
transmitted across the entire portal circulation, with the large antagonist spironolactone. In patients with ascites, electrolyte
sinusiodal area of the normal liver offering very little resistance abnormalities (hyponatremia, hyperkalemia, hypochloremia,
to portal flow. After hepatic resection, there will be increased and hypokalemia) and contraction alkalosis are commonly
resistance to portal flow until the liver regenerates. Usually, this encountered as a result of the primary process and medical
increase in portal pressure is of little consequence. If the size therapy. If ascites is refractory to medical management, para-
or the health of the remaining liver is not adequate to carry out centesis may offer temporary palliation. If life-threatening
the metabolic functions of the liver, the patient may experience complications such as cardiac or respiratory compromise occur,
small-for-size syndrome, characterized by coagulopathy, cho- these patients require transfer to the intensive care unit.
lestasis, hyperbilirubinemia, and ascites. For these patients, it
is important to maintain HBF and perfusion pressure to avoid Renal Dysfunction
hepatic artery thrombosis, hepatic artery vasospasm, or portal After hepatobiliary surgery, one may encounter rapidly progres-
vein thrombosis. sive renal failure, although this is uncommon (Correa-Gallego
Progressive jaundice that does not resolve portends a more et al, 2015; Melendez et al, 1998) after routine resections. This
significant problem. One suggested definition of postoperative postoperative acute renal failure (ARF) generally resolves when
hepatic insufficiency is a peak bilirubin greater than 7 mg/dL. not correlated with liver failure or multiorgan dysfunction. The
In a cohort of 1059 noncirrhotic patients undergoing hepatic principal differential diagnoses include acute tubular necrosis,
resection, a postoperative bilirubin greater than 7 mg/dL was prerenal azotemia, and hepatorenal syndrome. Postoperative
associated with a 30% chance of dying of hepatic insufficiency renal clinical dysfunction is associated with perioperative trans-
(Mullen et al, 2006). However, in the same cohort, 70% sur- fusions, LOS, other major complications, 30- and 90-day
vived, highlighting the importance of identifying any correctable mortality, larger resections, and age (Correa-Gallego et al,
problem that may be contributing to the hyperbilirubinemia, 2015; Kim et al, 2014).
such as a bile duct injury at the time of surgery, a bile leak, Recently, using a large national database, procedure-specific
infection, or vascular embarrassment to the functioning liver. risk of perioperative acute kidney injury (AKI) for patients
Endoscopic retrograde cholangiopancreatography or imaging undergoing intraabdominal surgery was evaluated (Kim et al,
studies will help in the diagnosis. 2014). The results demonstrated that among patients undergo-
ing intraabdominal surgery, the risk of severe AKI varies con-
Coagulation siderably, depending on the specific procedure. Hepatobiliary
The half-life of coagulation factors is 6 hours, so coagulation procedures (liver and pancreas, elective and emergent) had a
factor deficiency is likely to evolve rapidly. After routine liver 1.8 % incidence of ARF with an adjusted risk of 1%. The
resections, a transient and self-limited prolongation of pro- authors hypothesized that AKI may be a spectrum of diseases
thrombin time or INR occurs commonly and usually resolves with different causes and consequences that depend on the
without the need for transfusion of fresh frozen plasma (FFP) clinical context in which it occurs. A retrospective review in a
or administration of other agents. Patients undergoing major similar time period, as in the report by Kim and colleagues
liver resection have higher peak postoperative INR (median, (2014), evaluated severe clinical AKI following LCVP-assisted
1.6) and lower nadir platelet count (median, 148,000) within liver resection. Clinically relevant AKI was rare (<1%) and
7 days than patients undergoing lesser procedures (Nathan resolved in half of these patients during a short follow-up period
et al, 2013). Routine parenteral administration of vitamin K (Correa-Gallego et al, 2015). These results are mirrored by
(10 mg/day) commonly corrects the coagulation abnormalities analysis of postoperative morbidity after liver resections (Virani
within 48 hours. More rapid correction may be accomplished et al, 2007). Whereas Kim only had severe end points (creati-
by FFP but is rarely needed unless the INR exceeds 2 or if nine >2, dialysis) to analyze, and may have underestimated the
there are otherwise concerns of ongoing postoperative blood incidence of clinically significant AKI defined by the RIFLE
loss (Martin et al, 2003). If coagulopathy persists, dysfibrino- (Risk, Injury, Failure, Loss of kidney function, and End-stage
genemia should be entertained. kidney disease) criteria, this review analyzed laboratory data
and was able to apply the RIFLE criteria to the incidence
Ascites of biochemical AKI (Correa-Gallego et al, 2015). Although
Ascites (see Chapter 81) manifests as a result of several factors: biochemically defined renal dysfunction was a relatively
(1) impaired albumin and protein synthesis; (2) increased common event (16%) in patients undergoing LCVP-assisted
hydrostatic pressure in hepatic sinusoids and splanchnic capil- liver resection, it was a transient phenomenon, and its clinical
laries; (3) overproduction of hepatic and splanchnic lymph, significance was limited. These results are mirrored by analysis
leading to a transudation of lymph into the peritoneal space; of postoperative morbidity after liver resections (Virani et al,
(4) limited or reduced reabsorption of water and protein by 2007).
peritoneal lymphatics; (5) sodium retention by the kidney There is no standard definition of ARF, but common criteria
secondary to hyperaldosteronism, increased sympathetic stimu- are either biochemical (a rise in serum creatinine of 50 μmol/l,
lation, and alterations in metabolism of prostaglandins and or of 50% from baseline, or beyond a set level, for example,
kinins; and (6) impaired renal water excretion, partially >500 μmol/l), or clinical (oliguria with urine output <400 ml/
owing to increased concentrations of antidiuretic hormone. The day), a combination of both, or the need for renal replacement.
decreased serum albumin 1 to 2 weeks after hepatobiliary A recent development is the RIFLE classification, which seeks
to standardize definitions of ARF based on a combination of group of anesthesiologists, surgeons, and surgical scientists that
biochemical and urine output criteria. recommend optimal postoperative pain management that is
The incidence of biochemical bARF is not in and of itself specific for different surgical procedures, as well as arguments
an argument against the use of LCVP anesthesia for liver resec- for and against the invasiveness of the analgesic technique and
tions. Some permissive, clinically relevant renal dysfunction is the consequences of pain on outcome (Neugebauer et al,
a very uncommon event in this patient population, and it is 2007). Hepatic surgery is not represented, and what operation
consistent with published literature regarding other types of is transferable is unclear.
major operations (Bihorac et al, 2013). The evidence does not The most commonly used incisions are subcostal or midline,
support LCVP-assisted liver surgery increasing the incidence with significant cephalic retraction of the chest wall. Therefore
of significant AKI. the patient’s pain is more similar to a postthoracotomy patient’s
pain. For postoperative pain after liver surgery, acute pain treat-
Glucose Metabolism ment services primarily offer thoracic EPA with dilute local
Glucose deprivation leads to hepatic glycogenolysis to maintain anesthetics, with the premise that neuraxial delivery of drug will
normoglycemia. After glycogen stores in the liver are exhausted result in equivalent or improved analgesia at doses much lower
(12 to 24 hours), the liver must synthesize glucose from than would be required by systemic delivery. Although thoracic
other substrates (gluconeogenesis). The hormones, insulin, EPA provides superior pain relief for thoracoabdominal opera-
glucagon, and epinephrine are important modulators of these tions (Joshi et al, 2008) and reduces PPCs (Popping et al,
processes. On balance, the postoperative patient would seem 2008), the risk versus benefit of an indwelling catheter with
to be prone to hypoglycemia secondary to diminished glycogen EDA for liver operations is debatable because of possible
reserves, increased insulin levels, and impaired gluconeogenesis, changes in postoperative coagulation. In addition, the number
but dangerous hypoglycemia is rare, and its presence must raise one complication following hepatectomy is abdominal infec-
the suspicion of acute hepatic failure. In most resections, iso- tions, with PPCs less common than thoracic surgery and result-
tonic saline solutions without dextrose can be used during the ing more from the act of surgery (pleural effusion or bile leak)
first 12 to 24 hours after resection, following which routine than the patient’s comorbidity. An elevated INR or decreased
dextrose solutions should be ordered and glucose levels platelet count can influence the timing of the removal of the
monitored. catheter because of a theoretical risk of spinal hematoma (Sinis-
calchi et al, 2015). In recent years, bloodless hepatic surgery,
Analgesia the known hypercoagulable state despite elevated INR and
After hepatic resection, most patients are extubated in the decreased platelet counts, and the safety of an indwelling cath-
operating room. Until recently, it was tacitly assumed that eter in other surgeries in patients who undergo anticoagulation
postoperative pain was essentially the pain of inflammation postoperatively allows reevaluating the modern age risk versus
plus, perhaps, some direct pain from cutting nerves. Subse- benefit in patients undergoing liver resection. There are no
quently, it has been demonstrated that there are important clinical guidelines as to the safety of epidural use in the patient
differences between pain from surgery and pain from inflam- for liver surgery. The guidelines for the placement and removal
mation (Brennan, 2005). More important from a clinical per- of neuraxial analgesia in patients with coagulation defects out-
spective, some drugs are effective to treat either surgery pain or lined by the American Society of Regional Anesthesia (ASRA)
inflammation, such as opioids, whereas others are unique to the apply to hepatic resection as well (Apfelbaum et al, 2012).
setting. A better understanding of the pain pathways and the Recently, basic science and retrospective reviews have sug-
concerted effort by ERAS to shorten and uncomplicate hospital gested that the anesthetic management during cancer surgery
stay, especially in the older patient, has drawn attention to the may influence the patient’s long-term survival. There are several
success of multimodality analgesia. Postoperative pain manage- multicenter prospective trials underway to examine whether the
ment has evolved considerably from the use of parental or use of regional anesthesia can truly decrease cancer recurrence.
neuraxial opioid monotherapy to multimodal opioid-sparing What is perhaps the most intriguing and an immediate hypoth-
analgesia. Today’s menu includes nonsteroidal antiinflammatory esis for liver surgery is whether EDA can improve immune
drugs/cyclooxygenase-2 inhibitors, local anesthetics, N-methyl- function and increase resistance to postoperative infection.
D-aspartate antagonists, α2-antagonists, thoracic epidural Surgical stress and pain may induce lymphocyte depletion,
catheters, IV patient-controlled analgesia, and catheters for the which may be associated with the risk of postoperative infec-
incisional wound (White & Kehlet, 2010). tious complications. Although circulating cytokines related to
Although opioids have stood the test of time to anchor monocyte activation and phenotype alterations are not influ-
postoperative pain, these drugs, whether infused epidurally or enced by postoperative pain reduction compared with systemic
parentally, are not without adverse effects; consequently, tar- opioid treatment, there has been evidence in animals that EPA
geted multimodality opioid-sparing analgesic algorithms are influences lymphocyte distribution, increases the postoperative
being proposed. These algorithms may not be safe for patients CD4/CD8 ratio and B cells, and decreases natural killer cells.
having hepatic surgery; instead, choices must be based on an This preservation of immunity has not been shown in human
individual basis, given the liver’s role in drug biotransformation studies (Cata et al, 2013) (see Chapter 11). EDA combined
and coagulation. with EPA has been shown to significantly reduce the amount
of postoperative inflammatory response of patients by altering
Epidural Analgesia the circulating leucocyte surface molecules CD11b and CD62L
The ASA postoperative evidenced-based pain management (L-selectin), which are known to be more sensitive markers of
guidelines are generalized and can be confusing when adapted the early detection of postoperative stress response inflamma-
to liver surgery. The web-based PROSPECT (Procedure- tory markers (Chloropoulou et al, 2013). If there is an attenu-
SPECific Postoperative Pain ManagemenT) is a collaborative ated neutrophil adhesive capability by EDA/EPA, and if this
improves patient resistance to infection, this may be more better pain relief, avoidance of polypharmacy, and good nutri-
important for liver surgery than for other types of surgery tion, but from a prevention or therapeutic point of view, there
because of the role that the liver plays in the immune and is no one target for decreasing the incidence after surgery.
inflammatory process. Ketamine is an anesthetic that has been used for 50 years, and
new research has shown that even in low doses intraoperatively,
Intraabdominal Infection in addition to regular anesthesia during surgery, it may reduce
Abdominal infections are the most common complication of the risk of delirium (Khan et al, 2015).
modern liver surgery (Kingham et al, 2014). Postoperative Encephalopathy is usually graded using the West Haven
complications in cancer patients with liver resections have been criteria for encephlopathy (Ferenci, 2013). This tries to semi-
shown to reduce disease-free survival and disease-specific sur- quantitate the grading of a patient’s mental status, as it pro-
vival (Ito et al, 2008). The mechanism is unknown, but peri- gresses from a trivial lack of awareness to lethargy, confusion,
operative inflammation and infection are a current theory. and coma. Clinically, a sudden change in a patient’s mental
Postoperative bilirubin, blood loss, transfusion, major hepatec- status or the onset of asterixis, a flapping tremor related to loss
tomy, and hepatectomy combined with a colon resection have of extensor tone, is a sign of hepatic insufficiency and should
all been shown to be associated with complications. Complica- precipitate a search for the cause and the start of therapy for
tions, especially infections, are powerful predictors and a pos- hepatic encephalopathy. Any number of perioperative factors,
sible cause of adverse disease-specific survival (Correa-Gallego including gastrointestinal bleeding, infection, drugs, diet, and
et al, 2015; Ito et al, 2008). A common condition causing a dehydration can precipitate hepatic encephalopathy (Sladen,
low-flow state to the liver is sepsis. Biliary manipulation and 2008). The exact pathogenesis of hepatic encephalopathy is
placement of biliary stents, culture-positive bile, retained gall- unknown. Accumulation of endogenous ammonia is associated
stones, ascites, blood loss, and transfusion all may predispose with the onset of encephalopathy and has been suggested as
to abdominal infection. Patients undergoing hepatic resection causative; however, it is only a marker of disordered protein
for hilar cholangiocarcinoma are at particular risk for postop- metabolism. In an acidic milieu, ammonia gains a hydrogen ion
erative infective complications. Preventing a hospital-acquired to become ionized ammonium (NH4+), which cannot cross lipid
infection is always better than treatment. membranes and enters the central nervous system (CNS). In
Initial treatment of abdominal infections falls under the an alkalotic milieu, ammonium loses the hydrogen ion to
domain of the interventional radiology. Ultrasound and CT are become ammonia, which, as it is nonionized, freely crosses lipid
used to locate collections that are then drained percutaneously, membranes and enters the CNS. Ammonia plays an important
often without sequelae (see Chapter 27). Radioisotopes that role in the development of cerebral edema because astrocytes
attach to white blood cells and glucose and can display high take up ammonia produced by bacteria in the bowel and convert
metabolic demand (biologic imaging) are on the horizon to be it into glutamine, which has considerable osmotic activity.
used in the future for identifying an abdominal collection. Ammonia also causes additional changes in neurotransmitter
synthesis and release, mitochondrial function, and neuronal
Delirium: Encephalopathy oxidative stress (Norenberg et al, 2014). Standard therapy is
Delirium (temporary inability to focus attention and think designed to reduce levels of ammonia and other potentially
clearly) occurs in one out of five older patients who undergo toxic metabolites. Metabolic alkalosis associated with potas-
major surgery. Delirium is associated with a slower recovery sium chloride should be corrected. Decreasing absorption of
and a poorer outcome, and a vicious circle may be initiated gut protein by using oral lactulose or rifaximin is preferred to
(delirium, physical restraint, and medication to treat delirium; dietary protein restriction, which may complicate wound
postoperative complications; then more delirium). Our hospital healing (Mas et al, 2003).
has a daily protection and intervention program based on early-
started and supporting treatment, with increased monitoring, References are available at expertconsult.com.
Chapter 24 Intraoperative and immediate postoperative management 436.e1
REFERENCES Chloropoulou P, et al: Epidural anesthesia followed by epidural anal-

gesia produces less inflammatory response than spinal anesthesia
Adams ST, Leveson SH: Methods of clinical prediction rules, BMJ followed by intravenous morphine analgesia in patients with total
344(1):d8312, 2012. knee arthroscopy, Med Sci Monit 19(1):73–80, 2013.
Agnelli G, et al: A clinical outcome-based prospective study on venous Chung F, et al: High STOP-Bang score indicates a high probability of
thromboembolism after cancer surgery: the @RISTOS project, Ann obstructive sleep apnoea, Br J Anaesth 108(5):768–775, 2012.
Surg 243(1):89–95, 2006. Correa-Gallego C, et al: Goal-directed fluid therapy using stroke volume
Agrawal S, Belghiti J: Oncologic resection for malignant tumors of the variation for resuscitation after low central venous pressure—assisted
liver, Ann Surg 253(4):656–665, 2011. liver resection: a randomized clinical trial, J Am Coll Surg 221(2):591–
Alcalay A, et al: Venous thrombembolism in patients with colorectal 601, 2015.
cancer incidence and effect on survival, J Clin Oncol 24(7):1112– Cunningham JD, et al: One hundred consecutive hepatic resections:
1118, 2006. blood loss, transfusion, and operative technique, Arch Surg 129:
Aloia TA, et al: Predicting poor outcome following hepatectomy:analysis 1050–1056, 1994.
of 2313 hepatectomies in the NSQIP database, HPB (Oxford) Cuthbertson BH, et al: Utility of B-type natriuretic peptide in predict-
11(6):510–515, 2009. ing perioperative cardiac events in patients undergoing non cardiac
Apfelbaum JK, et al: Practice guidelines for acute pain management in surgery, Br J Anaesth 99(2):170–176, 2007.
the perioperative setting: an updated report by the American Society De Kock M, et al: “Balanced analgesia” in the perioperative period: is
of Anesthesiologists Task Force on Acute Pain Management, Anes- there a place for ketamine? Pain 92(3):373–380, 2001.
thesiology 116(2):248–273, 2012. Denys A, et al: Indications for and limitations of portal vein emboliza-
Armas-Loughran B, et al: Evaluation and management of anemia and tion before major hepatic resection for hepatobiliary malignancy,
bleeding disorders in surgical patients, Med Clin North Am 87:229– Surg Oncol Clin N Am 11(4):955–968, 2002.
242, 2003. Devereaux PJ, et al: Aspirin in patients undergoing noncardiac surgery,
Atzil S, et al: Blood transfusion promotes cancer progression: A critical N Engl J Med 370(16):1494–1503, 2014.
role for aged erythrocytes, Anesthesiology 109(6):989–997, 2008. Devereaux PJ, et al: Rationale, design, and organization of the periop-
Auler JO, et al: Online monitoring of ppv to guide fluid therapy after erative ischemic evaluation (POISE) trial; a randomized controlled
cardiac surgery, Anesth Analg 106(4):1201–1206, 2008. trial of metoprolol versus placebo in patients undergoing noncardiac
Ball C, et al: Irreversible electroporation: a new challenge in out of surgery, Am Heart J 152(2):223–230, 2006.
operating theater anesthesia, Anesth Anal 110(5):1305–1309, 2010. Dindo D, et al: Obesity in general elective surgery, Lancet 361(9374):
Barton JS, et al: Coagulopathy after a liver resection: is it overdiagnosed 2032–2035, 2003.
and over treated? HPB (Oxford) 15(11):865–871, 2013. Duggan M, Kavanagh BP: Pulmonary atelectasis: a pathogenic periop-
Berendes E, et al: Effects of enflurane and isoflurane on splanchnic erative entity, Anesthesiology 102(4):838–854, 2005.
oxygenation in humans, J Clin Anesth 8:456–468, 1996a. Dunki-Jacobs EM, et al: Stroke volume variation in hepatic resection:
Berendes E, et al: Effects of positive end-expiratory pressure ventilation a replacement for standard central venous pressure monitoring, Ann
on splanchnic oxygenation in humans, J Cardiothorac Vasc Anesth Surg Oncol 21(3):473–478, 2013.
10:598–602, 1996b. Declan FJ, et al: Cardiopulmonary exercise testing before liver surgery,
Bihorac A, et al: National surgical Quality improvement program J Surg Oncol 110(6):439–444, 2014.
underestimates the risk associated with mild and moderate postop- Eagle KA, et al: ACC/AHA guideline update for perioperative cardio-
erative acute kidney injury, Crit Care Med 41(11):2570–2583, 2013. vascular evaluation for noncardiac surgery-executive summary:
Birch DJ, et al: Sympathetic innervation of human mesenteric artery a report of the ACC/AHA Task Force on Practice Guidelines
and vein, J Vasc Res 45(4):323–332, 2008. (Committee to update the 1996 Guidelines on Perioperative
Botto F, et al: Myocardial injury after noncardiac surgery: a large, Cardiovascular Evaluation for Noncardiac Surgery), Anesth Analg
international, prospective cohort study establishing diagnostic crite- 94(5):1052–1064, 2002.
ria, characteristics, predictors, and 30 day outcomes, Anesthesiology Ebert TJ, et al: Cardiovascular responses to sevoflurane: a review,
120(3):564–578, 2014. Anesth Analg 81(Suppl 6):S11–S22, 1995.
Brecker ME, Goodnough LT: The rise and fall of preoperative autolo- Edmark L, et al: Optimal oxygen concentration during induction of
gous blood transfusion, Transfusion 41:1459–1462, 2001. general anesthesia, Anesthesiology 98:28–33, 2003.
Brennan TJ: Incisional sensitivity and pain measurements: dissecting Eriksson KG, et al: High intra-abdominal pressure during experimental
mechanisms for postoperative pain after surgery, Anesthesiology laparoscopic liver resection reduces bleeding but increases the risk
103(1):3–4, 2005. of gas embolism, Br J Surg 98:845–852, 2011.
Brueckmann B, et al: Development and validation of a score for Eleftheriadis E, et al: Splanchnic ischemia during laparoscopic chole-
prediction of postoperative respiratory complications, Anesthesiology cystectomy, Surg Endosc 10:324–326, 1996.
118(6):1276–1285, 2013. Elliot RH, Strunin L: Hepatotoxicity of volatile anaesthetics: implica-
Canet J, et al: Prediction of postoperative pulmonary complications in tions for halothane, enflurane, isoflurane, sevoflurane, and desflurane,
a population-based surgical cohort, Anesthesiology 113(6):1338–1350, Br J Anaesth 70:339–348, 1993.
2010. Fattovich G, et al: Natural history and prognosis of hepatitis B seminar,
Cannesson M: Arterial pressure variability and goal directed therapy, Semin Liver Dis 23(1):47–58, 2003.
J Cardiothorac Vasc Anesth 24(3):487–497, 2010. Feltracco P, et al: Epidural anesthesia and analgesia in liver resection
Cannon RM, et al: Negative effects of transfused blood components and living donor hepatectomy, Transplant Proc 40(4):1165–1168,
after hepatectomy for metastatic colorectal cancer, Am Surg 79(1): 2008.
35–39, 2013. Ferenci P: Diagnosis of minimal hepatic encephalopathy: still a chal-
Cata P, et al: Effects of surgery, general anesthesia, perioperative epi- lenge, Gut 2(10):1394, 2013.
dural analgesia on the immune function of patients with non small Fernandez-Perez E, et al: Intraoperative ventilation for settings and
cell lung cancer, J Clin Anesth 25(6):255–262, 2013. acute lung injury after elective surgery: a nested case control study,
Catheline JM, et al: Thromboembolism prohylaxis and incidence of Thorax 64(2):121–127, 2009.
thromboembolic complications after laparoscopic surgery, Int J Surg Ferreyra GP, et al: Continuous positive airway pressure for treatment
Investig 2(1):41–47, 2000. of respiratory complications after abdominal surgery: a systematic
Chang PH, Steinberg MB: Alcohol withdrawal, Med Clin North Am review and meta-analysis, Ann Surg 247(4):617–626, 2008.
85(5):1191–1212, 2001. Fleischer LA, et al: ACC/AHA 2014 guidelines on perioperative car-
Cescon M, et al: Trends in perioperative outcome after hepatic resec- diovascular evaluation and management of patients undergoing
tion: analysis of 1500 consecutive unselected cases over 20 years, noncardiac surgery: a report of the ACC/AHA Task Force on
Ann Surg 249(6):995–1002, 2009. Practice Guidelines, J Am Coll Cardiol 64(22):e77–e137, 2014.
Chappell D, et al: protection of glycolyx decrease platelet adhesion Fong Y, et al: Long-term survival is superior after resection for cancer
after ischaemia/reperfusion: an animal study, Eur J Anaesthesiol in high volume centers, Ann Surg 242:540–544, discussion 544-547,
9:474–481, 2014. 2005.
436.e2 PART 3 ANESTHETIC MANAGEMENT: PREOPERATIVE AND POSTOPERATIVE CARE
Ford MK, et al: Systematic review: prediction of perioperative cardiac Jones C, et al: Randomized clinical trial on enhanced recovery versus
complications and mortality by the revised cardiac risk index, Ann standard care following open liver resection, Br J Surg 100(8):1015–
Intern Med 152(1):26–35, 2010. 1024, 2013.
Fors D, et al: Elevated PEEP without effect upon gas embolism fre- Joshi GP, et al: A systematic review of randomized trials evaluating
quency or severity in experimental laparoscopic liver resection, Br J regional techniques for postthoracotomy analgesia, Anesth Analg
Anaesth 109(2):272–278, 2012. 107(3):1026–1040, 2008.
Foster JH, Berman MM: Solid liver tumors, Major Probl Clin Surg Kamiyama T, et al: Perioperative management of hepatic resection
22:1–342, 1977. toward zero mortality and morbidity: analysis of 793 consecutive
Frankel TL, et al: Selecting patients for acute normovolemic hemodi- cases in a single institution, J Am Coll Surg 211(4):443–449,
luiton during hepatic resection: a prospective randomized evaluation 2010.
of nomogram-based allocation, J Am Coll Surg 217(2):210–220, Katz SC, et al: Operative blood loss independently predicts recurrence
2013. and survival after resection of hepatocellular carcinoma, Ann Surg
Futier E, et al: IMPROVE Study group: a trial of intraoperative low 249(4):617–623, 2009.
tidal volume ventilation in abdominal surgery, N Engl J Med Kennedy WF, et al: Simultaneous systemic and hepatic hemodynamic
369(5):428–437, 2013. measurements during high peridural and anesthesia in normal man,
Gazoni FM, et al: The impact of perioperative catastrophes on an Anesth Analg 50(6):1069–1077, 1971.
anesthesiologist: results of a national survey, Anesth Analg 114(3): Khan H, et al: The progostic value of liver tumor T cell infiltrates,
596–603, 2012. J Surg Res 191(1):189–195, 2014.
Gelman S: Venous function and central venous pressure: a physiologic Khan BA, et al: Update on pharmacotherapy for prevention and treat-
story, Anesthesiology 108(4):735–748, 2008. ment of post-operative delirium: a systematic evidence review, Curr
Girish M, et al: Symptom-limited stair climbing as a predictor of Anesthesiol Rep 5(1):57–64, 2015.
postoperative cardiopulmonary complications after high-risk surgery, Kheterpal S: Clinical research using an information system: the multi-
Chest 120(4):1147–1151, 2001. center perioperative outcomes group, Anesthesiol Clin 29(3):377–388,
Glance LG, et al: Perioperative outcomes among patients with the 2011.
modified metabolic syndrome who are undergoing noncardiac Kheterpal S: Random clinical decisions: identifying variation in peri-
surgery, Anesthesiology 113(4):859–872, 2010. operative care, Anesthesiology 116(1):3–5, 2012.
Goldman L: Assessing and reducing cardiac risks of noncardiac surgery, Khuri SF, et al: Determinant of long-term survival after major surgery
Am J Med 110(4):320–323, 2001. and the adverse effect of postoperative complications, Ann Surg
Goldman M, et al: Declining value of preoperative autologous dona- 242(3):326–341, 2005.
tion, Transfusion 42:819–823, 2002. Kim M, et al: Variations in acute renal failure in abdominal cases,
Gould MK, Cooke CR: Prevention of VTE in nonorthopedic surgical Anesth Analg 119(5):1121–1132, 2014.
patients : antithrombotic therapy and prevention of thrombosis, 9th Kingham TP, et al: Hepatic parenchymal preservation surgery: decreas-
ed: American College of Chest Physicians Evidence-Based Clinical ing morbidity and mortality rates in 4,152 resections for malignancy,
Practice guidelines, Chest 141(Suppl 2):e227S–e277, 2012. J Am Coll Surg 220(4):471–479, 2015.
Grocott MP, Pearse RM: Perioperative medicine: the future of anaes- Kneuertz PJ, et al: Risk of morbidity and mortality following hepato-
thesia? Br J Anaesth 108(5):723–726, 2012. pancreato-biliary surgery, J Gastrointest Surg 16(9):1727–1735,
Groeschl RT, et al: Microwave ablation for hepatic malignancies, Ann 2012.
Surg 259:1195–1200, 2014. Kooby DA, et al: Influence of transfusions on perioperative and long-
Gross JB, et al: Practice guidelines for the perioperative management term outcome in patients following hepatic resection for colorectal
of patients with sleep apnea: a report by the ASA Task Force on metastases, Ann Surg 237(6):860–889, 2003.
perioperative management of patients with sleep apnea, Anesthesiol- Lampe GH, et al: Nitrous oxide does not impair hepatic function in
ogy 104(5):1081–1093, 2006. young or old surgical patients, Anesth Analg 71:606–609, 1990.
Grosse-Sundrup M, et al: Intermediate acting non-depolarizing neuro- Landesberg G, et al: Perioperative myocardial infarction, Circulation
muscular blocking agents and risk of postoperative respiratory 119(22):2936–2944, 2009.
complications: prospective propensity score matched cohort study, Lautt WW: Regulatory processes to maintain hepatic blood flow con-
BMJ 345:e6329, 2012. stancy: vascular compliance, hepatic arterial buffer response, hepa-
Gupta RM, et al: Postoperative complications in patients with obstruc- torenal reflex liver regeneration, escape from vasoconstriction,
tive sleep apnea syndrome undergoing hip or knee replacement, Hepatol Res 37(11):891–903, 2007.
Mayo Clin Proc 76(9):897–905, 2001. Lee TH, et al: Derivation and prospective validation of a simple index
Holst LB, et al: Lower versus higher hemoglobin threshold for transfu- for prediction of cardiac risk of major noncardiac surgery, Circulation
sion in septic shock, N Engl J Med 371(15):1381–1391, 2014. 100(10):1043–1049, 1999.
Hughes MJ, et al: Enhanced recovery following liver surgery: a system- Levin MA, et al: Low intraoperative tidal volume ventilation with
atic surgery: a systematic review and meta-analysis, HPB (Oxford) minimal PEEP is assoociated with increased mortality, Br J Anaesth
16(8):699–706, 2014. 113(1):97–108, 2014.
Hwang JW, et al: Assessment of perioperative cardiac risk of Lin TF, et al: Effect of combining dexmedetomidine and morphine for
patients undergoing noncardiac surgery using coronary computed intravenous patient-controlled analgesia, Br J Anaesth 102(1):117–
tomographic angiography, Circ Cardiovasc Imaging 8(3):e002582, 122, 2014.
2015. Lindenauer PK, et al: Perioperative beta-blocker therapy and
Ito H, et al: Effect of postoperative morbidity on long-term survival mortality after noncardiac surgery, N Engl J Med 353(4):349–361,
after hepatic resection for metastatic colorectal cancer, Ann Surg 2005.
247(6):994–1002, 2008. Loadsman JA, Hillman DR: Anesthesia and sleep apnea, Br J Anaesth
Jabbour N, et al: Transfusion free surgery: Single institution experience 86(2):254–266, 2001.
of 27 consecutive liver transplants in Jehovah’s Witnesses, J Am Coll Martin RC, et al: The use of fresh frozen plasma after major hepatic
Surg 201(3):412–417, 2005. resection for colorectal metastasis: is there a standard for transfusion?
Jarnagin WR, et al: Improvement in perioperative outcome after J Am Coll Surg 196:402–409, 2003.
hepatic resection: analysis of 1803 consecutive cases over the past Mas A, et al: Comparison of rifaximin and lactitol in the treatment of
decade, Ann Surg 236:397–407, 2002. acute hepatic encephalopathy: results of a randomized, double-
Jarnagin WR, et al: A prospective randomized trial of acute normovole- blind, double-dummy, controlled clinical trial, J Hepatol 38(1):51–58,
mic hemodilution compared to standard intraoperative management 2003.
in patients undergoing major hepatic resection, Ann Surg 248(3): Matot I, et al: Epidural anesthesia and analgesia in liver resection,
360–369, 2008. Anesth Analg 95(5):1179–1181, 2002.
Jhanji S, et al: Mortality and utilisation of critical care resources Matot I, et al: Effectiveness of acute normovolemic hemodilution to
amongst high risk surgical patients in a large NHS trust, Anaesthesia minimize allogeneic blood transfusion in major liver resections,
63(7):695–700, 2008. Anesthesiology 97:794–800, 2002.
Jones DR, et al: Comparison of POSSUM with APACHE II for predic- McAlister FA, et al: Accuracy of preoperative assesssment in predicting
tion of outcome from a surgical high dependency unit, Br J Surg pulmonary risk after nonthoracic surgery, Am J Respir Crit Care Med
79(12):1293–1296, 1992. 167(5):741–744, 2003.
Chapter 24 Intraoperative and immediate postoperative management 436.e3
Meierhenrich R, et al: The effect of thoracic epidural anesthesia on Schuckit MA: Recognition and management of withdrawal delirium
hepatic blood flow in patients under general anesthesia, Anesth Analg (delirium tremens), N Engl J Med 371(22):2109–2113, 2014.
108(4):1331–1337, 2009. Schwinn DA, et al: Isoflurane-induced vasodilation: role of the adren-
Melendez J, et al: Perioperative outcome of major hepatic resections ergic nervous system, Anesth Analg 71:451–459, 1990.
under low central venous pressure anesthesia—blood loss, blood Serpa Neto A, et al: Association between use of lung protective ventila-
transfusion and the risk of post-operative renal dysfunction, J Am tion with lower tidal volumes and clinical outcomes among patients
Coll Surg 178:620–625, 1998. without acute respiratory distress syndrome: a meta-analysis, JAMA
Milic-Emili J, et al: Closing volume: a reappraisal (1967-2007), Eur J 308:1651–1659, 2012.
Appl Physiol 99(6):567–583, 2007. Sima CS, et al: Predicting risk of perioperative transfusion for patients
Min SK, et al: Carbon dioxide and argon gas emobolism during lapa- undergoing elective hepatectomy, Ann Surg 250(6):914–921, 2009.
roscopic hepatic resection, Acta Anaesthesiol Scand 51(7):949–953, Siniscalchi A, et al: Thoracic epidural anesthesia: effects on splanchnic
2007. circulation and implications in anesthesia and intensive care, World
Moonesinghe RM, et al: Risk stratification tools for predicting morbid- J Crit Care Med 4(1):89–104, 2015.
ity and mortality in adult patients undergoing major surgery, Sladen RN: Anesthetic concerns for the patient with liver disease,
Anesthesiology 119(4):959–981, 2013. Anesth Analg 106(4):1–7, 2008.
Mullen JT, et al: Hepatic insufficiency and mortality in 1,059 noncir- Simpson AL, et al: Liver planning software accurately predicts postop-
rhotic patients undergoing major hepatectomy, J Am Coll Surg erative liver volume and measures early regeneration, J Am Coll Surg
204(6):854–862, 2006. 219(2):199–207, 2014.
Nathan H, et al: The volume-outcomes effect in hepat-pancreato-biliary Smetana GW, et al: American College of Physicians: preoperative
surgery: hospital versus surgeon contributions and specialty relation- pulmonary risk stratification for noncardiothoracic surgery: system-
ship, J Am Coll Surg 208(4):528–538, 2009. atic review for the American College of Physicians, Ann Intern Med
Nathan H, et al: Pharmacologic prophylaxis , postoperative INR, and 144(8):581–595, 2006.
risk of venous thromboembolism after hepatectomy, J Gastrointest Snowden CP, et al: Submaximal cardiopulmonary exercise testing
Surg 218(2):295–302, 2013. predicts complications and hospital length of stay in patients under-
Neligan PJ, Fleischer LA: Obesity and diabetes: evidence of increased going major elective surgery, Ann Surg 251(3):535–541, 2010.
perioperative risk? Anesthesiology 104(3):398–400, 2006. Spahn DR, Goodnough LT: Alternatives to blood transfusion, Lancet
Neugebauer EA, et al: PROSPECT Working group: PROSPECT: a 381(9880):1855–1865, 2013.
practical method for formulating evidence-based expert recommen- Suman A, Carey WD: Assessing the risk of surgery in patients with
dations for the management of postoperative pain, Surg Endosc liver disease, Cleve Clin J Med 73:398–404, 2006.
21(7):1047–1053, 2007. Tanaka N, et al: The effect of dopamine on hepatic blood flow in
Neuman MD, et al: Learning from mistakes in clinical practice guide- patients undergoing epidural anesthesia, Anesth Analg 85(2):286–
lines: the case of perioperative B blockade, BMJ Qual Saf 23(11): 290, 1997.
957–964, 2014. Trescot AM, et al: Opioid pharmacology, Pain Physician 11(Suppl
Neuman MD, et al: Durability of class 1 American College of 2):S133–S153, 2008.
Cardiology/American Heart Association clinical practice guideline Tzeng CW, et al: Risk of venous thromboembolism outweighs post-
recommendations, JAMA 311(20):2092–2100, 2014. hepatectomy bleeding complications analysis of 5651 National
Newsome LT, et al: Coronary artery stents: II. Perioperative consider- Surgical Quality Improvement program patients, Ann Surg Oncol
ations and management, Anesth Analg 107(2):570–590, 2008. 20(2):482–490, 2012.
Norenberg MD, et al: Brain edema in acute liver failure: mechanisms Vamvakas E: Possible mechanisms of allogeneic blood transfusion-
and concepts, Metab Brain Dis 29(4):927–936, 2014. associated postoperative infection, Transfus Med Rev 16:144–160,
Older P, et al: Cardiopulmonary exercise testing as a screening test for 2002.
perioperative maangement of major surgery in the elderly, Chest Vamvakas EC: Allogeneic blood transfusion and cancer recurrence: 20
116:355–362, 1999. years later, Transfusion 54(9):2149–2153, 2014.
O’Riordan J, et al: Effects of desflurane and isoflurane on splanchnic Van Kaam AH, et al: Reducing atelectasis attenuates bacterial growth
microcirculation during major surgery, Br J Anaesth 78:95–96, 1997. and translocation in experimental pneumonia, Am J Respir Crit Care
Page A, et al: Epidural analgesia in hepatic resection, J Am Coll Surg Med 169(9):1046–1053, 2004.
206(3):1184–1192, 2008. Vauthey JN, et al: Standardized measurement of the future liver
Park WY, et al: Effect of epidural anesthesia and analgesia on periop- remnant prior to extended liver resection: methodology and clinical
erative outcome: a randomised controlled Veterans’ Administration associations, Surgery 127(5):512–519, 2000.
study, Ann Surg 234(4):560–569, 2001. Vibert E, et al: Laparoscopic liver resection, Br J Surg 93(1):67–72,
Pearse RM, et al: Effect of a perioperative, cardiacoutput-guided 2006.
hemodynamic therapy algorithm on outcomes following major gas- Virani S, et al: Morbidity and mortality after liver resection: results of
trointestinal surgery: a randomized clinical trial and systematic the patient safety in surgery study, J Am Coll Surg 204(6):1284–1292,
review, JAMA 311(21):2181–2190, 2014. 2007.
Popping DM, et al: Protective effects of epidural analgesia on pulmo- Wakamatsu H, et al: Predictive value of Tc-99m galactosyl human
nary complications after abdominal and thoracic surgery: a meta- serum albumin liver SPECT on the assessment of functional recov-
analysis, Arch Surg 143(10):990–999, 2008. ery after partial hepatectomy: a comparison with CT volumetry, Ann
Powers SK, et al: Pressure support ventilation attenuates ventilator- Nucl Med 24(10):729–734, 2010.
induced protein modifications in the diaphragm, Crit Care 12(6):191, Warner DO: Preventing postoperative pulmonary complications: the
2008. role of the anesthesiologist, Anesthesiology 92(5):1467–1472, 2000.
Prasad V, et al: A decade of reversal: an analysis of 146 contradicted Wehry J, et al: Restrictive blood transfusion protocol in liver resection
medical practices, Mayo Clin Proc 88(8):790–798, 2013. patients reduces blood transfusions with no increase in patient
Putensen C, et al: Meta-analysis: ventilation strategies and outcomes morbidity, Am J Surg 209(2):280–288, 2015.
of the acute respiratory distress syndrome and acute lung injury, Ann White PF, Kehlet H: Improving postoperative pain management: what
Intern Med 151(8):566–576, 2009. are the unresolved issues?, Anesthesiology 112(1):220–225, 2010.
Rees M, et al: One hundred and fifty hepatic resections: evolution of Wijeysundera D, et al: Epidural anaesthesia and survival after
technique toward bloodless surgery, Br J Surg 83:1526–1529, 1996. intermediate-to-high non-cardiac surgery: a population-based cohort
Rigg JR, et al: Epidural anaesthesia and anlagesia and outcome of major study, Lancet 372(9638):562–569, 2008.
surgery: a randomized trial, Lancet 359(9314):1276–1282, 2002. Zakian KL, et al: Liver regeneration in humans is characterized by
Rubinsky AD, et al: Postoperative risks associated with alcohol screen- significant changes in cellular phosporus metabolism: assessment
ing depend on documented drinking at the time of surgery, Drug using proton-decoupled 31P-magnetic resonance spectroscopic
Alcohol Depend 132(3):21–27, 2013. imaging, Magn Reson Med 54(2):264–271, 2005.
Saklad M: Grading of patients for surgical procedures, Anesthesiology
2:281–284, 1941.
CHAPTER 25
Perioperative critical care in
hepatopancreatobiliary patients
Louis P. Voigt, Stephen M. Pastores, and Neil A. Halpern
INTRODUCTION In 2000, the MELD score was developed (Malinchoc et al,

2000). The MELD score is based on the prothrombin time–
Patients undergoing hepatobiliary surgery or pancreaticoduo- international normalized ratio (PT-INR) and serum bilirubin
denectomy may require admission to the intensive care unit and creatinine levels and is calculated as follows:
(ICU), especially if the surgery is extensive or if complications
MELD score = 9.6 [ In Serum creatinine ( mg dL )]
occur. Compared with the general population, patients with
chronic liver diseases (e.g., cirrhosis due to alcohol or infections + 3.8 [ In Serum bilirubin ( mg dL )]
with the hepatitis B or C virus, alcoholic or autoimmune hepa- + 11.2 [ In INR ] + 6.4
titis, primary biliary cirrhosis, sclerosing cholangitis, and hemo-
chromatosis) are at increased risk for complications after major The MELD score ranges from 6 to 40. Moderate MELD
surgery (see Chapters 81 and 103D). Careful preoperative scores (>9) correspond to poor outcomes after hepatic resection
assessment, planning, monitoring, and optimization of the for hepatocellular carcinoma (HCC) (Delis et al, 2009). The
patient’s medical condition may mitigate the negative impact MELD score correlates well with the CTP score and is an
of chronic liver diseases and reduce the preoperative risks of accurate predictor of postoperative mortality in cirrhotic
morbidity and mortality of hepatopancreatobiliary surgery. patients undergoing elective and emergent surgery (Farnsworth
Although this chapter mainly focuses on the special periopera- et al, 2004; Fayad et al, 2015; Hoteit et al, 2008).
tive concerns of patients with cirrhosis and chronic liver disease, Imaging studies of the liver (i.e., magnetic resonance
a brief discussion of the preoperative assessment and periopera- imaging) and functional assays, such as the monoethylglycinex-
tive complications that can occur in patients undergoing pan- ylidide test and indocyanine green clearance (ICG), have been
creaticoduodenectomy is presented. used to assess the capacity of the liver to sustain stress and to
predict postoperative morbidity and mortality before liver
PREOPERATIVE ASSESSMENT AND CARE resection (Lorf et al, 2008; Scheingraber et al, 2008). Data on
the combination of MELD score and IGC (MELD-ICG) show
Patients with cirrhosis have higher morbidity and mortality promising results for use of transjugular intrahepatic portosys-
rates, not only after hepatic and pancreatobiliary surgical pro- temic shunts (TIPSs), but not for patients undergoing liver
cedures but also after major nonhepatic surgical interventions resection (Zipprich et al, 2010).
(Nguyen et al, 2009). When evaluating perioperative risks for
patients with chronic liver diseases, three main factors should Comorbid Conditions
be considered: hepatic reserve, comorbid conditions, and the Several intrinsic (e.g., active alcohol use, viral and autoimmune
complexity of the planned surgical procedure. hepatitis, and portal hypertension) and extrinsic (e.g., severe
obesity, diabetes mellitus, acute kidney injury [AKI]) comorbid
Hepatic Reserve with Child-Turcotte-Pugh and Model conditions may negatively impact the outcome of hepatic and
for End-Stage Liver Disease Scores nonhepatic surgical interventions in patients with cirrhosis
Several tools are available to assess hepatic reserve (see Chapter (Hickman & Macdonald, 2007) (see Chapter 24). Contrain
3). The most frequently used are the Child-Turcotte-Pugh dications to surgery include acute alcoholic hepatitis, decompen-
(CTP) and the Model for End-Stage Liver Disease (MELD) sated autoimmune hepatitis, acute hepatic failure, refractory
scores (Table 25.1) (Nicoll, 2012). The CTP score represents ascites, spontaneous bacterial peritonitis (SBP), variceal hemor-
the degree of hepatic synthetic dysfunction and is based on the rhage, uncontrolled hepatopulmonary syndrome, and portopul-
presence of hepatic encephalopathy, degree of ascites, pro- monary hypertension (Czaja, 2009). In patients with chronic
thrombin time, and serum levels of albumin and bilirubin. The hepatitis C infection and cirrhosis, diabetes increases the risk
score allows categorization of cirrhosis into three classes. Class for the development of ascites, renal dysfunction, and bacterial
A (5 to 6 points) refers to well-compensated cirrhosis; class B infections (Elkrief et al, 2014). In cases of autoimmune
(7 to 9 points) and class C (10 to 15 points) describe cases of and infectious hepatitis, immunomodulation and antiviral thera-
cirrhosis that are in a state of mild and severe decompensation, pies are beneficial in the preoperative period, and the doses of
respectively. The mortality rates for elective or emergent extra- steroids are often increased to mitigate adrenal insufficiency
hepatic surgery range from 0% to 7.1% for class A, 50% for (Chan, 2005; Hadziyannis, 2006; Keeffe et al, 2008; Lau et al,
class B, and 84% to 100% for class C cirrhosis (Franzetta et al, 2005; Marcellin et al, 2008). Ascites, upper gastrointestinal (GI)
2003; Hayashida et al, 2004). bleeding, poor nutritional state, anemia, thrombocytopenia,
437
438 PART 3 ANESTHETIC MANAGEMENT, PRE- AND POSTOPERATIVE CARE
TABLE 25.1 Scoring Systems to Assess Hepatic Reserve

Possible Scores Hepatic Encephalopathy Ascites Albumin Bilirubin PT-INR Creatinine
CTP 5-15 + + + + + −
MELD 6-40 − − − + + +
CTP, Child-Turcot-Pugh; MELD, Model for End-stage Liver Disease; PT-INR, prothrombin time–international normalized ratio.
electrolyte disorders with or without AKI, and preoperative

infection are additional risk factors for perioperative complica- Nutrition
tions and mortality among cirrhotic patients undergoing surgery Malnutrition is commonly underappreciated in cirrhotic
(Ziser et al, 1999). Preoperative cholestasis and hepatic steato- patients because the common parameters of nutritional status
sis are significant risk factors for major complications including (body weight, body mass index, and albumin level) are unreli-
mortality (McCormack et al, 2007). able (Juakiem et al, 2014). Up to 40% of cirrhotic patients
suffer from severe malnutrition, a comorbid condition that
Ascites, Portal Hypertension, and Variceal Bleeding confers a higher risk of perioperative morbidity and mortality
Ascites predisposes to bacterial peritonitis, delayed wound (Sam & Nguyen, 2009) (see Chapter 26). Combined with dia-
healing, and respiratory compromise as a result of reduced betes mellitus, malnutrition in cirrhotic patients increases the
functional residual capacity (FRC) (Gupta et al, 2000) (see risk of hepatic encephalopathy (Kalaitzakis et al, 2007).
Chapter 81). Adequate control of ascites may require several Enhancing the nutritional status of cirrhotic patients may be
months and can be achieved with a low-sodium diet, judicious associated with improved outcomes. Oral diet and enteral
use of diuretics, and/or large volume paracentesis. Intravenous feeding via a nasogastric tube are preferred and can safely be
(IV) administration of albumin solutions during paracentesis used in patients with esophageal varices (Hebuterne & Van-
may reduce the risk of perioperative hypovolemic shock and biervliet, 2011). Branched-chain amino acids are advocated
mortality (Kwok et al, 2013). because they may reduce the risk of hepatic encephalopathy,
Portal hypertension can be identified early by transient elas- but their real benefit is marginal (Harrison, 2006; Marchesini
tography, and this diagnosis confers a higher risk of hospital et al, 2003). The placement of a percutaneous endoscopic gas-
mortality (Berzigotti et al, 2013; Bureau et al, 2008; Nguyen trostomy (PEG) tube is contraindicated in patients with ascites
et al, 2009). Ascites and varices related to portal hypertension and should be avoided in patients with portal hypertension due
are associated with increased risks of variceal hemorrhage, bac- to the potential risk of puncturing a variceal vessel during PEG
terial peritonitis, and hospital mortality in cirrhotic patients insertion (Rahnemai-Azar et al, 2014). Short-term parenteral
undergoing various surgical procedures, such as cholecystec- nutrition is not advised because of increased risks of hypergly-
tomy, coronary artery bypass grafting, abdominal aortic aneu- cemia, infection, fluid overload, azotemia, and liver dysfunction
rysm repair, and colorectal surgery (Csikesz et al, 2009; Nguyen (Badia-Tahull et al, 2012).
et al, 2009) (see Chapter 77). Abnormal blood flow in the
portal venous system can be easily demonstrated by Doppler Anemia
ultrasound and correlates well, when present, with increased Anemia is a common complication of liver diseases that usually
mortality (Harrod-Kim & Waldman, 2005). Hepatic venous develops approximately 40 months after the onset of cirrhosis,
gradient, MELD score, and albumin can also predict clinical even in the absence of decompensation or esophageal varices
decompensation in patients with cirrhosis (Ripoll et al, 2007) (Qamar et al, 2009). Anemia is due to many factors, including
(see Chapters 81 and 82). bleeding, hemolysis, splenic sequestration, hepatic dysfunction,
Preoperative interventions should be initiated to reduce malnutrition, and iron and vitamin deficiency. Iron supple
hepatic venous pressure and the risks of variceal bleeding. The mentation should be administered only in the presence of
combination of octreotide and diuretics can lead to improve- documented deficiency. Autologous blood transfusion and
ment in portal and systemic hemodynamics (Kalambokis et al, administration of erythropoietin have been used successfully to
2006). Nonselective β-blockers, such as propranolol or nadolol, mitigate the need for red blood cell transfusions (Kato et al,
either alone or in combination with nitrates and esophageal 2009; Silver et al, 2006).
variceal band ligation are recommended, with TIPS as an alter-
native option (de Franchis & Baveno, 2010; Escorsell et al, Coagulopathy
2002; Garcia-Pagán et al, 2009; Lo et al, 2002). TIPS is indi- Similar to anemia, coagulation abnormalities are frequent in
cated for patients with ascites refractory or intolerant to diuret- cirrhotic patients and are due to many causes, including
ics, but patients who undergo TIPS should be closely monitored decreased hepatic synthetic function, abnormal synthesis of
for hepatic encephalopathy (Qi et al, 2015). Patients with mod- coagulation factors, malnutrition, vitamin K deficiency, throm-
erate to high MELD scores (>15) are at high risk for 30-day bocytopenia, and dysfibrinogenemia (Qamar et al, 2009). Pre-
mortality and should not undergo TIPS for refractory ascites operatively, a 3-day trial of IV administration of vitamin K is
(Ferral et al, 2004; Schepke et al, 2003). β-Blockers are con- appropriate in patients with prolonged PT-INR and suspected
traindicated in cirrhotic patients with refractory ascites, and vitamin K deficiency (Shah et al, 2014). In patients requiring
these patients should avoid abdominal surgery because of the interruption of oral anticoagulation, the administration of oral
high rates of postoperative morbidity and death (Serste et al, menadiol, the water-soluble formulation of vitamin K, on the
2010) (see Chapters 79 and 87). day before surgery can normalize the PT-INR by the day of
Chapter 25 Perioperative critical care in hepatopancreatobiliary patients 439
surgery and may not confer postoperative resistance to warfarin extrahepatic procedures in patients with HPS is associated with
(Woods et al, 2007). increased perioperative risk (Mazzeo et al, 2004). A preopera-
Recent studies and analyses have failed to establish a clear tive partial pressure of oxygen in arterial blood (PaO2) of
correlation between an increased risk of hemorrhage and a 50 mm Hg or a macroaggregated albumin scan quantification
prolonged PT or activated partial thromboplastin time (aPTT) shunt fraction of 20% or higher is associated with prohibitive
(Tripodi & Mannucci, 2011). In the absence of bleeding, mortality rates from cardiorespiratory complications (Arguedas
routine fresh frozen plasma (FFP) transfusions to correct pro- et al, 2003).
longed PT-INR and aPTT are no longer advocated because of Portopulmonary hypertension is an integral component of
better understanding of the coagulation paradigm in liver cir- HPS and predisposes to intraoperative cardiac arrhythmias and
rhosis (Monroe & Hoffman, 2009; Northup & Caldwell, 2013). cardiac arrest (Porres-Aguilar et al, 2012). Mortality of porto-
However, the final decision by the perioperative team is often pulmonary hypertension is high, with 1- and 5-year survival
individualized. Excessive FFP transfusions can lead to volume rates of 54% and 14%, respectively (Swanson et al, 2008).
overload and may exacerbate ascites. Human recombinant acti- Portopulmonary hypertension is often asymptomatic and dif-
vated factor VIIa (rFVIIa) or prothrombin complex concentrate ficult to diagnose. Right side of the heart catheterization helps
is an effective alternative to plasma in patients with acute hem- to determine the various causes of pulmonary hypertension and
orrhage (Pabinger et al, 2008; Romero-Castro et al, 2004), but is the best diagnostic tool for portopulmonary hypertension
their use has not been rigorously tested. Administration of (Krowka et al, 2006). Elective surgery under general anesthesia
cryoprecipitate is advocated when serum fibrinogen levels are in patients with severe HPS-associated pulmonary hyperten-
less than 100 mg/dL in the presence of bleeding (Anderson sion should be deferred until the pulmonary hypertension is
et al, 2013). Data on the benefits of fibrinogen concentrates are controlled with vasodilator agents.
limited to trauma and massive hemorrhage (Fenger-Eriksen
et al, 2008). Transfusion of platelets is recommended before Electrolytes
invasive procedures for moderate thrombocytopenia of less Electrolyte abnormalities are frequent in patients with active
than 50,000/µL (Johansson, 2009; Rebulla, 2001). Adjunctive alcohol use and chronic liver diseases. Hyponatremia is attrib-
desmopressin may be considered as rescue therapy for refrac- uted to an impaired ability to excrete free water and occurs
tory hemorrhage, particularly when AKI is present, but con- in the setting of severe ascites, hepatic encephalopathy, SBP,
flicting results persist about its benefits in chronic liver disease and hepatorenal syndrome (Angeli & Merkel, 2008). The
(Shah et al, 2014). causes of hyponatremia should be elucidated and addressed
through various strategies, such as correction of fluid deficit
Cardiovascular and Respiratory Parameters or fluid restriction, and cautious and transient use of
Repiratory alkalosis tends to develop in patients with chronic vasopressin-2 (V2)-receptor antagonists, such as conivaptan,
liver disease. Increased abdominal girth in the setting of tense satavaptan, and lixivaptan (Kwo, 2014). Other electrolyte
ascites can lead to restrictive ventilatory defect with reduced imbalances (hypokalemia, hypomagnesemia, and hypophos-
FRC, and hypoxia and paracentesis may lead to improvement phatemia) can result from the use of loop diuretics, chronic
in respiratory parameters (Gupta et al, 2000). Patients with respiratory alkalosis, or malnutrition and should be corrected
cirrhosis also are at risk for coronary artery disease because of before surgery to limit cardiac arrhythmias, peritonitis, and
a high prevalence of cigarette smoking and diabetes mellitus hepatic encephalopathy.
(Dam et al, 2013; Hickman & MacDonald, 2007). Cirrhotic
patients typically have a hyperdynamic hemodynamic profile Infection
similar to the systemic inflammatory response syndrome (SIRS) Patients with cirrhosis are at increased risk for infection, par-
with high cardiac output and vasodilatation of the pulmonary, ticularly SBP (Khan et al, 2009). In patients with cirrhosis,
splanchnic, and peripheral beds (Cazzaniga et al, 2009; Moller infections increase mortality fourfold with a median mortality
& Henriksen, 2008). Patients with alcoholic liver disease and rate of 38% (30% within 1 month after infection, whereas
iron overload are predisposed to cardiomyopathy and cardiac another 30% die within 1 year) (Arvaniti et al, 2010). Risk
arrhythmias (Moller & Henriksen, 2008). factors for SBP include gastrointestinal hemorrhage, metabolic
alkalosis, dehydration, hyponatremia, and a high MELD score
Hepatopulmonary Syndrome and (Obstein et al, 2007). SBP carries an increased mortality rate,
Portopulmonary Hypertension which can be worsened by age, renal dysfunction, bacteremia,
Hepatopulmonary syndrome (HPS) is relatively common in and elevated MELD score (Cho et al, 2007; Khan et al, 2009;
advanced cirrhosis and is often associated with portopulmonary Nobre et al, 2008). SBP can be confirmed by a diagnostic
hypertension as a result of alterations in the pulmonary vascu- paracentesis, especially in the setting of large-volume ascites
lature by various toxins that accumulate and bypass the cir- (Kuiper et al, 2007). If SBP is present, elective surgery should
rhotic liver (Ho, 2008) (see Chapter 79). Signs and symptoms be deferred until the infection has been adequately controlled.
of HPS include hypoxemia, platypnea, orthodeoxia, clubbing, Prophylactic intestinal decontamination with quinolones can
and spider angiomata (Singh & Sager, 2009). Transthoracic prevent SBP in high-risk patients, such as those in whom GI
echocardiography with agitated saline or bubble contrast allows hemorrhage develops, or who have a low protein level in the
identification of intrapulmonary and right-to-left intracardiac ascitic fluid (Loomba et al, 2009; Terg et al, 2008).
shunts indicative of HPS (Rodriquez-Roisin et al, 2005). The
severity of HPS can be quantified with a tagged, macroaggre- Surgical Procedures
gated albumin lung perfusion scan and arterial blood gas analy- The complexity of the surgical procedure influences outcomes
sis. HPS worsens the overall prognosis of liver cirrhosis (see Chapters 24, 77, and 103D). Timing of the surgery and
(Fauconnet et al, 2013). General anesthesia for hepatic and perioperative management should factor in the nature of the
surgical procedures, the administration of anesthetic agents, they are not metabolized through the hepatic or renal system
the severity of the cirrhosis, and occasionally the degree of (Hoetzel et al, 2012). Isoflurane and sevoflurane remain the
acuity and the causes of the liver disease. Emergent surgery preferred inhalational anesthetic agents in patients with liver
and unscheduled laparotomy for trauma carry a high mortality disease because they undergo less hepatic metabolism (Nishi-
(Demetriades et  al, 2004; Georgiou et  al, 2009). Cholecystec- yama et al, 2004). Of the two agents, isoflurane is more fre-
tomy and surgical repair of umbilical and inguinal hernia were quently selected because of its minimal effects on hepatic blood
associated with the lowest morbidity and mortality risk, flow, whereas enflurane and halothane are not used because of
whereas pancreatic surgery, cardiovascular, and trauma surgery the high risk of hepatotoxicity (Berendes et al, 1996).
correlated with the highest (De Goede et  al, 2012). The mor- Most narcotic, sedative, and analgesic agents are extensively
bidity and mortality rates of cardiovascular operations, open metabolized in the liver, and their half-life is often altered in
cholecystectomy, hysterectomy, nephrectomy, and transure- patients with chronic liver disease (Bosilkovska et al, 2012).
thral resection of the prostate are also higher in comparison Cirrhosis delays the elimination of alfentanil, whereas the
to laparoscopic and elective interventions (Csikesz et  al, metabolism of fentanyl, sufentanil, and remifentanil does not
2009; Filsouf et  al, 2007; Pavlidis et  al, 2009; Shaheen et  al, seem to be affected (Delco et al, 2005). Decreased protein
2009). binding and increased volumes of distribution hinder the clear-
ance of midazolam in cirrhotic patients (Verbeeck, 2008). IV
Assessment for Pancreatobiliary Surgery remifentanil is favored for induction because of rapid onset and
Surgery for gallbladder bile duct, and pancreatic cancer remains elimination. Propofol is also increasingly preferred versus mid-
very challenging and requires experienced surgical and anes- azolam for sedation during surgical or endoscopic procedures
thesiology teams. During the past four decades, the mortality because of shorter recovery time, better pharmacokinetics,
associated with standardized Kausch-Whipple pancreaticoduo- and less propensity to exacerbate subclinical encephalopathy
denectomy for adenocarcinoma of the pancreatic head has dra- (Khamaysi et al, 2011). In general, opioids are avoided or their
matically improved (Richter et al, 2003) (see Chapters 66 and use is limited intraoperatively.
67). However, postoperative complications are frequent and
include cardiopulmonary events, fistula, delayed gastric empty- Management of Mechanical Ventilation
ing, sepsis, and bleeding (Schmidt et al, 2004). From a ventilatory perspective, cirrhosis is a relative contrain-
Predictive models of postoperative morbidity based on the dication to the use of carbon dioxide during exploratory lapa-
estimation of physiologic ability and surgical stress score in roscopy. Low to moderate VT is advocated even in the absence
pancreatic surgery are associated with mixed results (Deyle of acute lung injury/acute respiratory distress syndrome because
et al, 2011; Hashimoto et al, 2010). Other scoring systems larger VT may cause abdominal hypertension, worsen respira-
allow preoperative risk stratification, prediction of major com- tory alkalosis, lead to barotrauma, reduce ventricular preload,
plications associated with pancreaticoduodenectomy, and and result in systemic hypotension and multiorgan injury
careful patient selection. The Preoperative Pancreatic Resec- (Determann et al, 2010). Patients with chronic liver disease are
tion (PREPARE) score is based on several variables, such as at increased risk of hypoxemia because of diffusion abnormali-
systolic blood pressure, heart rate, hemoglobin level, albumin ties, ventilation-perfusion mismatch, and decreased FRC
level, the American Society of Anesthesiologists category, elec- (Polverino et al, 2014). Muscle relaxants and inhaled anesthetic
tive versus emergent surgery, and disease of pancreatic origin agents may reduce the oxygen debt. Surgery is contraindicated
or not. The PREPARE score is quite accurate in identifying in cases of uncontrolled portopulmonary hypertension because
low-, intermediate-, and high-risk patients for pancreatic of high mortality rates (Kawut et al, 2005).
surgery (Uzunoglu et al, 2014). Other predictors of postopera-
tive outcome include texture of the pancreas, pancreatic duct Fluid/Blood Loss and Hemodynamic Parameters
diameter, and operative blood loss (Braga et al, 2011). Careful patient selection, attention to the volume of hepatic
tissue resected, and reduction in intraoperative blood loss con-
INTRAOPERATIVE MANAGEMENT tribute to improved perioperative morbidity and mortality.
Decreased effective circulatory volume is frequently present,
The core components of the intraoperative care of patients with despite the presence of total body fluid overload (Cazzaniga
cirrhosis include careful selection of anesthetic, narcotic, and et al, 2009). The initial periods of surgery may be associated
sedative agents; management of mechanical ventilation; main- with significant fluid shifts resulting from the removal of ascites
tenance of adequate intravascular volume in the setting of hem- when the peritoneal cavity is accessed. Surgery may also be
orrhage and fluid shift/loss; glucose control and correction of complicated by excessive intraoperative hemorrhage resulting
electrolyte disorders; improved surgical techniques; avoidance from the presence of portal hypertension and coagulopathy.
of hypothermia; and adjustment to blood losses in the setting Thromboelastography has been advocated as a guide for the
of relative anemia and coagulopathy (see Chapter 24). intraoperative monitoring and therapy of coagulopathy (Lud-
dington, 2005). The administration of IV fluids in the immedi-
Selection of Anesthetic, Narcotic, and Sedative Agents ate preoperative period may reduce the hemodynamic alterations
The choices and doses of anesthetics, muscle relaxants, anal- and prevent complications related to hypovolemia, such as
gesics, and sedatives should account for the degree of impair- hepatic encephalopathy and postoperative AKI. Large volumes
ment in hepatic synthetic function and clearance (Hoetzel et al, of blood products or IV fluids are sometimes necessary to
2012). Endotracheal intubation and induction are achieved maintain hemodynamic stability. Colloid solutions such as
in patients with liver disease similar to that in the general albumin (with vasoconstrictors) should be administered to
population. Atracurium and cisatracurium are the preferred minimize the risks of worsening portal hypertension and volume
neuromuscular blocking agents to facilitate intubation because overload (Appenrod et al, 2008).
Anesthetic techniques aimed at maintaining a low central Hepatic cirrhosis increases the risk for AKI, sepsis, ICU
venous pressure (CVP) during hepatic resection are associated admission, and postoperative mortality, particularly when asso-
with a reduction in blood loss, renal failure, and mortality ciated with alcohol dependence, hepatic encephalopathy, and
(Melendez et al, 1998). Alternatively, stroke volume variation GI hemorrhage (Lin et al, 2013). High MELD score, admis-
is an alternative strategy to CVP monitoring because it is a sion to the ICU, and need for mechanical ventilation and con-
useful indicator of intraoperative blood loss, and it can safely tinuous renal replacement therapy are predictors of increased
guide the administration of IV fluid during hepatic resection length of stay and both short- and long-term mortality
(Dunki-Jacobs et al, 2014; Harimoto et al, 2013). Intraopera- (Bahirwani, 2013; Cholongitas et al, 2009; Levesque et al,
tive blood transfusion is a harbinger of morbidity and mortality 2014). Among cirrhotic patients admitted to the ICU, sequen-
in patients undergoing hepatic resections (Neeff et al, 2014). tial organ failure assessment and MELD scores were more
Increased blood loss during surgical resection of hepatocellular accurate in predicting mortality, and those with failure of three
carcinoma is an independent predictor of cancer recurrence or more organs had a 90% mortality rate (Cholongitas et al,
and mortality (Katz et al, 2009). 2006). The three main risk factors for postoperative mortality
are older age, high MELD score of 20 or greater, and American
Correction of Electrolyte Abnormalities Society of Anesthesiologists class IV and V (Teh et al, 2007).
Correction of electrolyte abnormalities is important to prevent
arrhythmias and hemodynamic compromise. Compared with Postoperative Hepatic Failure
the general population, acute hypocalcemia during massive Hepatic failure, often manifested as encephalopathy, hyperbili-
blood transfusion is more frequent in patients with chronic liver rubinemia, and coagulopathy, can lead to acute respiratory
disease and is the result of high citrate load (Chung et al, failure, renal failure with/without hepatorenal syndrome, and
2012). Administration of IV calcium should be considered in bleeding in the setting of thrombocytopenia and acquired coag-
the setting of intraoperative hemorrhage and hypotension ulation defect. Several postoperative criteria have been used in
because it helps maintain vascular muscle tone and may predicting hepatic failure and mortality. The 50-50 criteria, a
enhance hemostasis. Intraoperative and postoperative serum combination of prothrombin time of less than 50% (or PT-INR
levels of potassium, calcium, magnesium, and phosphorus > 1.7) and serum bilirubin greater than 50 µmol/L (or > 2.9 mg/
should be monitored, and abnormalities should be promptly dL), has been validated as an excellent predictor of death on
corrected (Hayter et al, 2012). days 3 and 5 for patients admitted to the ICU for postoperative
liver failure following hepatic resection (Balzan et al, 2005;
Improved Surgical Techniques Paugam-Burtz et al, 2009). The Mullen criteria (bilirubin peak
During the past decade, advances in surgical techniques for > 7 mg/dL on postoperative days 1 to 7) were found to be more
hepatic surgery have resulted in significant improvement in accurate than the 50-50 criteria in predicting death from hepatic
postoperative outcomes (Rowe et al, 2009; Schiffman et al, failure after liver resection (Filicori et al, 2012). Encephalopa-
2015; Vavra et al, 2014) (see Chapter 103). Reduction in intra- thy and hyperbilirubinemia warrant avoidance of hepatotoxic
operative blood loss, ischemic preconditioning with intermit- medications, use of small incremental doses of analgesics and
tent portal triad clamping, and other techniques that allow other medications to achieve desired therapeutic effects, and
preservation of a larger volume of hepatic parenchyma are the administration of lactulose and stool softeners (Wright & Jalan,
primary contributing factors to these favorable outcomes, par- 2007). N-acetylcysteine does not offer any survival benefit in
ticularly at high-volume centers (Aragon & Solomon, 2012; the management of hepatic failure after liver resection (Robin-
Heizmann et al, 2008). However, the benefits of ischemic pre- son et al, 2013). Liver transplantation may be necessary but
conditioning could not be demonstrated in recent randomized transient supportive care with liver assist devices remains elusive
controlled trials and meta-analyses (O’Neill et al, 2013;Ye et al, (Schilsky, 2011; Sussman et al, 2009) (see Chapter 79).
2014; Zhu et al, 2014).
In regard to pancreatic surgery, pylorus-preserving pancre-
Fluid Management and Abdominal
aticoduodenectomy does not differ in intermediate- and long- Compartment Syndrome
term outcomes from the classic Whipple procedure as surgical Patients with cirrhosis are prone to postoperative hypotension,
treatment of pancreatic carcinoma (Diener et al, 2014; Tran vasodilation, and a hyperdynamic state of SIRS as well as portal
et al, 2004). However, a pylorus-preserving procedure offers hypertension (Cazzaniga et al, 2009). Surgical procedures can
significant advantages, such as shorter operative time and hos- cause significant fluid shifts, and ascites may occur even in
pital length of stay, reduced blood loss and blood transfusions, stable patients with cirrhosis without documented preoperative
decreased incidence of delayed gastric emptying, and more portal hypertension. Judicious administration of IV fluids may
enduring impact on long-term functional and nutritional status preserve renal function and prevent ascites (Cholongitas et al,
(Niedergethmann et al, 2006) (see Chapter 66). 2006). Oliguria is frequently caused by reaccumulation of
ascites at the expense of the intravascular space and is indicative
POSTOPERATIVE MANAGEMENT of the need for IV colloids and vasopressor agents (e.g.,
norepinephrine, terlipressin), rather than diuretics or α-
Despite careful planning, cautious patient selection, and judi- adrenoreceptor agonists such as midodrine (Appenrodt et al,
cious intraoperative monitoring and management, some patients 2008; Belcher et al, 2013).
experience hepatic failure, the most alarming postoperative The development of tense ascites predisposes to
complication. Patients with chronic liver diseases are also sus- surgical wound dehiscence, AKI, peritonitis, and abdominal
ceptible to other postoperative complications, such as abdomi- compartment syndrome. Abdominal compartment syndrome is
nal compartment syndrome, fluid derangements, and venous characterized by an acute and progressive elevation of the
thromboembolism (VTE). intraabdominal pressure to more than 20 mm Hg in association
with one or more organ dysfunction (Malbrain et al, 2007).

BOX 25.1 International Ascites Club Criteria for Diagnosing
Risk factors are obesity, shock/hypotension, sepsis, large-
Hepatorenal Syndrome
volume crystalloid resuscitation, and abdominal surgery (Holo-
dinsky et al, 2013). Abdominal compartment syndrome is quite 1. Presence of advanced hepatic failure and pulmonary
common and predisposes to a vicious cycle of hypotension from hypertension
decreased venous return, hypoxemia from decreased FRC and 2. Creatinine (>1.5 mg/dL)
3. Absence of shock, bacterial infection, nephrotoxic drugs,
ventilator dyssynchrony, and AKI and intestinal ischemia from
excessive diarrhea, or renal fluid losses
decreased perfusion of the kidneys and bowel, respectively
4. No sustained improvement in renal function following diuretic
(Daugherty et al, 2007). Measurement of intraabdominal pres- withdrawal and expansion of plasma volume with 1.5 L of
sure through a bladder catheter is the current mainstay of isotonic saline
diagnosis, and excessive fluid administration should be avoided 5. Absence of significant proteinuria (<500 mg/day)
by establishing appropriate endpoints of resuscitation (Vidal 6. No evidence of obstructive uropathy
et al, 2008). Administration of vasopressor agents along with 7. Low urine volume (<500 mL/day) and low urine sodium
restriction of a large volume of IV fluids can restore hemody- (<10 mEq/L)
namic stability, safeguard renal function, and prevent tense
ascites and abdominal compartment syndrome, whereas para-
centesis and surgical decompression may be necessary to
prevent organ injury and wound dehiscence (An & West, 2008; renal function with molecular adsorbent recirculation
Levesque et al, 2011). system seem promising (Lavayssiere et al, 2013) (see Chapter
80).
Acute Renal Failure and Hepatorenal Syndrome
AKI occurs in approximately 15% of patients who undergo liver Anemia and Hemorrhage
resection and correlates to preexisting cardiovascular disease, The risk of postoperative variceal rupture and bleeding is
preoperative serum alanine aminotransferase elevation, under- increased by excessive blood transfusion, large-volume ascites,
lying renal disease, and diabetes mellitus (Slankamenac et al, increased portal pressure, and infection, particularly when
2009). Postoperative AKI in patients with chronic liver disease coagulopathy and thrombocytopenia coexist (García-Pagán
may be caused by the administration of nephrotoxic drugs in et al, 2012). Hepatic venous pressure gradient does not cor-
the perioperative period, hypovolemia, parenchymal renal relate with the severity of portal hypertension (Bellis et al,
disease, and HRS (Gines & Schrier, 2009). Avoidance of 2007), but hepatic venous pressure gradient of greater than or
harmful medications and sensible use of IV fluids and vasopres- equal to 10 mm Hg was associated with older age, presence of
sor agents are useful management strategies for AKI. cirrhosis, and reduced hepatic reserve (Bureau et al, 2008). For
HRS, an alarming postoperative renal complication in patients with acute variceal hemorrhage, a MELD score of
patients with advanced-stage liver disease, represents a func- greater than or equal to 18, transfusion requirement of greater
tional renal failure caused by circulatory dysfunction (see than or equal to 4 U of red blood cells within the first 24 hours
Chapter 79). The underlying mechanism combines intrarenal of admission, and active bleeding at endoscopy are independent
vasoconstriction and marked arterial vasodilatation of the predictors of 6-week mortality and variceal rebleeding within
splanchnic circulation precipitated by infection or intravascular the first 5 days (Bambha et al, 2008).
volume depletion (i.e., hemorrhage, peritoneal fluid drainage, Variceal hemorrhage warrants transfusion of blood products
or diuretic administration) (Salerno et al, 2007). There are two and treatment with several agents, including terlipressin,
types of HRS: type 1 HRS is often associated with a rapid albumin, and octreotide (Abid et al, 2009; Martin-Llahi et al,
deterioration of renal function with increased levels of bilirubin 2008). Infected patients with cirrhosis are prone to alteration
and prothrombin time. Type 2 HRS is characterized by a more in anti–activated factor X assay and production of endogenous
gradual and less severe form of renal impairment related to heparinoids as demonstrated by thromboelastography (Mon-
refractory ascites (Angeli & Merkel, 2008). The updated diag- talto et al, 2002; Triantos et al, 2014; Zambruni et al, 2004).
nostic criteria of the International Ascites Club for HRS are Antibiotic prophylaxis can prevent infection and rebleeding and
summarized in Box 25.1 Arroyo et al, 1996; Salerno et al, reduce transfusion requirement in patients with acute GI hem-
2007). Acute tubular necrosis may be indistinguishable from orrhage (Hou et al, 2004; Pohl et al, 2004). In patients with
HRS, because both conditions share similar clinical presenta- ascites, paracentesis may reduce the hepatic venous pressure
tions and findings on urinalysis (Gines & Schrier, 2009). gradient and wall tension and prevent GI bleeding (Silkauskaite
Vasoconstrictor agents (terlipressin [not yet approved in the et al, 2009).
United States] or norepinephrine), either alone or in combina- Among patients with cirrhosis, postoperative anemia may
tion with albumin and TIPS, are the main therapies for HRS warrant transfusion of red blood cells, particularly in the setting
(Alessandria et al, 2007; Gluud et al, 2012; Martin-Llahi et al, of surgical or GI hemorrhage. Hematocrit should be kept
2008; Nassar Junior et al, 2014; Sanyal et al, 2008; Sharma between 21% and 24% because excessive blood transfusions
et al, 2008). Continuous or intermittent renal replacement can lead to increased variceal rebleeding (de Franchis & Baveno,
therapies should be offered to patients who do not respond to 2010). Erythropoietic growth factors appear to be beneficial for
vasoconstrictors or TIPS and to prospective candidates of the treatment of anemia in patients receiving combination anti-
orthotopic liver transplantation, and recovery of renal function viral agents for chronic hepatitis C infection (Del Rio et al,
can be achieved in as many as 50% of patients (Fabrizi et al, 2006). However, data are lacking on the potential reduction in
2013). Primary prophylaxis with norfloxacin can achieve sig- blood products use in the postoperative period.
nificant survival benefits and reduction in SBP and HRS The traditional strategy of FFP transfusion to correct coagu-
(Fernandez et al, 2007). New data on the improvement of lopathy in patients who are not bleeding has been abandoned
after recent studies and analyses have indicated no clear correla- Early enteral nutrition is advocated after liver resection
tion between an increased risk of hemorrhage and a prolonged because of a perceived decrease in postoperative complications
PT or aPTT (Tripodi & Mannucci, 2011). Moreover, FFP (Richter et al, 2006). Compared with enteral nutrition, paren-
transfusion may lead to volume overload and does not correct teral nutrition for 7 days before planned hepatectomy in patients
coagulation defect or decrease the risk of bleeding (Tinmouth, with liver cirrhosis and continued for 7 days afterward was
2012; Tripodi et al, 2012). The routine use of rFVIIa in acute associated with fewer septic complications and a lower mortal-
variceal hemorrhage is controversial (Bendtsen et al, 2014; ity rate (Fan et al, 1994). In reality, the benefits and recom-
Bosch et al, 2008). Data on the successful use of four-factor mended route of nutritional support after hepatopancreatobiliary
prothrombin and fibrinogen complex in patients with liver cir- surgery are debatable because of conflicting results from a few
rhosis and postoperative hemorrhage are lacking. randomized trials and systematic review (Fan et al, 1994;
Koretz et al, 2012). At present, no specific guidelines or recom-
Sedative and Pain Management and mendations are available, even during liver regeneration or in
Hepatic Encephalopathy instances of acute pancreatitis (Poropat et al, 2015).
In patients with cirrhosis, sedative and narcotic agents undergo
increased volume of distribution, slow hepatic metabolism, and Postoperative Care after Pancreaticoduodenectomy
delay elimination (Delco et al, 2005). Advanced cirrhosis is Postoperative complications of the Kausch-Whipple procedure,
often associated with renal dysfunction, which increases the such as anastomotic leak, fistula, and abdominal collections,
risks of adverse events, and the doses of sedative and analgesic are relatively frequent and vary with the type of pancreaticoen-
medications should be adjusted to prevent an exacerbation of teric anastomosis, pancreatic texture and intraoperative blood
hepatic encephalopathy and other complications (Verbeeck, transfusion (Fathy et al, 2008) (see Chapters 27 and 66).
2008). Inappropriate use of narcotics and sedatives can lead to Abdominal sepsis is relatively common, and pneumonia,
atelectasis, aspiration, and postoperative pneumonia by impair- abdominal abscess, and multiorgan dysfunction syndrome
ing respiratory drive and by delaying extubation or triggering (MODS) in the setting of hemorrhage are the main contribu-
reintubation. tors to the relatively low mortality rate of 2% to 5% (Satoi et al,
Among the narcotic agents available, remifentanil is pre- 2006). Early recognition is the most important component in
ferred because of its limited hepatic metabolism (Uchida the successful management of pancreatic leak, and in the
et  al, 2011). Continuous epidural analgesia is frequently absence of sepsis, hemorrhage, or MODS, most cases are
avoided because of concerns about coagulopathy, thrombocy- managed expectantly on the surgical wards (Ho et al, 2005).
topenia, and the increased risk of hematoma. In a small Imaging studies are sometimes indicated, and a few patients
series, patient-controlled analgesia has been found to be a may require IV antibiotics, blood transfusion, and close drain-
safe and effective method of managing pain after hepatic age. Surgical exploration and prognosis are often dictated by
resection (Roy et  al, 2006). Propofol is a better choice than the degree of destruction and inflammation of the retroperito-
benzodiazepines for sedation because of its shorter recovery neum (Ho et al, 2005).
time, which may allow more reliable serial neurologic assess- As many as 18% of patients may require ICU admission for
ment (Khamaysi et  al, 2011). Although data are limited, dex- postoperative complications of pancreaticoduodenectomy.
medetomidine is a potential alternative to IV propofol because Delayed admission to the ICU for septic shock with multiorgan
of rapid onset and offset of the drug (Wang et  al, 2014). failure and hemorrhage carries a mortality rate of close to 20%
Nonsteroidal antiinflammatory agents should be used spar- (Welsch et al, 2011). Prophylactic octreotide has no impact on
ingly because of the potential for peptic ulceration, fluid the rates of pancreatic fistula or total complications after pan-
retention from inhibition of renal prostaglandin synthesis, creaticoduodenectomy (Yeo et al, 2000). However, the periop-
precipitation of hepatorenal syndrome, and bleeding due to erative use of pasireotide was associated with significant
antiplatelet activity, gastrointestinal irritation, and renal failure decrease in postoperative pancreatic fistula, leak, or abscess
(Bosilkovska et  al, 2012; Dwyer et  al, 2014). (Allen et al, 2014).
Hepatic encephalopathy warrants assessment and treatment
of infections and administration of oral lactulose or lactitol Venous Thromboembolism Prophylaxis
suppository (Wright & Jalan, 2007). Second-line agents for the Compared with the general population, patients with cirrhosis
treatment of hepatic encephalopathy include oral antibiotics are at higher risk of bleeding because of anemia, thrombocyto-
such as neomycin, metronidazole, and rifaximin (Al Sibae & penia, and the inability of the liver to produce procoagulants in
McGuire, 2009). response to the stress of bleeding. Yet these patients manage to
strike a fine balance between prothrombotic and anticoagulat-
Liver Regeneration, Hyperglycemia, and Nutrition ing proteins and other endogenous products (Senzolo et al,
Hypophosphatemia is a frequent occurrence in the early days 2009). Severe thrombocytopenia rather than coagulopathy is
after hepatectomy and should be corrected, accordingly (George the most significant risk factor for bleeding after invasive pro-
& Shiu, 1992). Failure to develop hypophosphatemia is a cedures (Giannini et al, 2010; Rockey et al, 2009).
marker of postoperative hepatic insufficiency and mortality In patients with liver cirrhosis and with noncirrhotic chronic
(Squires et al, 2014). Liver regeneration is a protracted com- liver disease, the risks of VTE are relatively higher (Sogaard et al,
pensatory process of hepatocytes replication that can be assessed 2009). Postoperative VTE may occur in the setting of low serum
by magnetic resonance spectroscopic imaging (Fausto et al, albumin, despite thrombocytopenia and coagulopathy (Dabbagh
2006; Zakian et al, 2005) (see Chapter 6). Early postoperative et al, 2010; Northup et al, 2006). In postoperative patients with
hyperglycemia after pancreaticoduodenectomy increases the cirrhosis at moderate risk for VTE (i.e., Caprini score < 5 and
risk of infectious complications, delayed gastric emptying, and Rogers score < 10), early ambulation and mechanical thrombo-
reoperation (Eshuis et al, 2011). prophylaxis with intermittent pneumatic compression devices
and/or graduated compression stockings, are favored particu- (Gould et al, 2012). Although renal impairment must be con-
larly in the early perioperative period, when the consequences of sidered when choosing the type of heparin for VTE, a recent
bleeding are potentially severe (Gould et al, 2012). meta-analysis showed no meaningful differences in the inci-
Pharmacologic thromboprophylaxis with unfractionated dence of heparin-induced thrombocytopenia and/or thrombosis
heparin or low-molecular-weight heparins (LMWH) is appro- between LMWH and unfractionated heparin (Morris et al,
priate when the risk of bleeding subsides (Barclay et al, 2013). 2007).
Periodic surveillance with venous compression ultrasound and
routine placement of inferior vena cava filters are not advocated References are available at expertconsult.com.
Chapter 25 Perioperative critical care in hepatopancreatobiliary patients 444.e1
REFERENCES Cazzaniga M, et al: The systemic inflammatory response syndrome

in cirrhotic patients: relationship with their in-hospital outcome,
Abid S, et al: Terlipressin vs. octreotide in bleeding esophageal varices J Hepatol 51(3):475–482, 2009.
as an adjuvant therapy with endoscopic band ligation: a randomized Chan HL, et al: A randomized, controlled trial of combination therapy
double-blind placebo-controlled trial, Am J Gastroenterol 104(3):617– for chronic hepatitis B: comparing pegylated interferon-alpha2b and
623, 2009. lamivudine with lamivudine alone, Ann Intern Med 142(4):240–250,
Alessandria C, et al: Noradrenalin vs terlipressin in patients with hepa- 2005.
torenal syndrome: a prospective, randomized, unblinded, pilot study, Cho JH, et al: Bacteremia is a prognostic factor for poor outcome in
J Hepatol 47(4):499–505, 2007. spontaneous bacterial peritonitis, Scand J Infect Dis 39(8):697–702,
Allen PJ, et al: Pasireotide for postoperative pancreatic fistula, N Engl 2007.
J Med 370(21):2014–2022, 2014. Cholongitas E, et al: Risk factors, sequential organ failure assessment
Al Sibae MR, McGuire BM: Current trends in the treatment of hepatic and model for end-stage liver disease scores for predicting short term
encephalopathy, Ther Clin Risk Manag 5(3):617–626, 2009. mortality in cirrhotic patients admitted to intensive care unit, Aliment
An G, West MA: Abdominal compartment syndrome: a concise clinical Pharmacol Ther 23(7):883–893, 2006.
review, Crit Care Med 36(4):1304–1310, 2008. Cholongitas E, et al: RIFLE classification as predictive factor of mortal-
Anderson MA, et al: When do we transfuse cryoprecipitate?, Intern Med ity in patients with cirrhosis admitted to intensive care unit,
J 43(8):896–902, 2013. J Gastroenterol Hepatol 24(10):1639–1647, 2009.
Angeli P, Merkel C: Pathogenesis and management of hepatorenal Chung HS, et al: Effects of liver function on ionized hypocalcaemia
syndrome in patients with cirrhosis, J Hepatol 48(Suppl 1):S93– following rapid blood transfusion, J Int Med Res 40(2):572–582,
S103, 2008. 2012.
Appenrodt B, et al: Prevention of paracentesis-induced circulatory dys- Csikesz NG, et al: Nationwide volume and mortality after elective
function: midodrine vs albumin. A randomized pilot study, Liver Int surgery in cirrhotic patients, J Am Coll Surg 208(1):96–103, 2009.
28(7):1019–1025, 2008. Czaja AJ: Current and future treatments of autoimmune hepatitis,
Aragon RJ, Solomon NL: Techniques of hepatic resection, J Gastroin- Expert Rev Gastroenterol Hepatol 3(3):269–291, 2009.
test Oncol 3(1):28–40, 2012. Dabbagh O, et al: Coagulopathy does not protect against venous
Arguedas MR, et al: Prospective evaluation of outcomes and predictors thromboembolism in hospitalized patients with chronic liver disease,
of mortality in patients with hepatopulmonary syndrome undergoing Chest 137(5):1145–1149, 2010.
liver transplantation, Hepatology 37(1):192–197, 2003. Dam MK, et al: Smoking and risk of liver cirrhosis: a population-based
Arroyo V, et al: Definition and diagnostic criteria of refractory ascites cohort study, Scand J Gastroenterol 48(5):585–591, 2013.
and hepatorenal syndrome in cirrhosis. International Ascites Club, Daugherty EL, et al: Abdominal compartment syndrome is common
Hepatology 23(1):164–176, 1996. in medical intensive care unit patients receiving large-volume resus-
Arvaniti V, et al: Infections in patients with cirrhosis increase mortality citation, J Intensive Care Med 22(5):294–299, 2007.
four-fold and should be used in determining prognosis, Gastroenterol- de Franchis R, Baveno VF: Revising consensus in portal hypertension:
ogy 139(4):1246–1256.e1245, 2010. report of the Baveno V consensus workshop on methodology of
Badia-Tahull MB, et al: Liver function test alterations associated with diagnosis and therapy in portal hypertension, J Hepatol 53(4):762–
parenteral nutrition in hospitalized adult patients: incidence and risk 768, 2010.
factors, Nutr Hosp 27(4):1279–1285, 2012. De Goede B, et al: Morbidity and mortality related to non-hepatic
Bahirwani R, et al: Factors that predict short-term intensive care unit surgery in patients with liver cirrhosis: a systematic review, Best Pract
mortality in patients with cirrhosis, Clin Gastroenterol Hepatol Res Clin Gastroenterol 26(1):47–59, 2012.
11(9):1194–1200.e2, 2013. Delco F, et al: Dose adjustment in patients with liver disease, Drug Saf
Balzan S, et al: The “50-50 criteria” on postoperative day 5: an accu- 28(6):529–545, 2005.
rate predictor of liver failure and death after hepatectomy, Ann Surg Delis SG, et al: Perioperative risk assessment for hepatocellular carci-
242(6):824–828, discussion 828-829, 2005. noma by using the MELD score, J Gastrointest Surg 13(12):2268–
Bambha K, et al: Predictors of early re-bleeding and mortality after 2275, 2009.
acute variceal haemorrhage in patients with cirrhosis, Gut 57(6):814– Del Rio RA, et al: Cost-effectiveness of hematologic growth factors for
820, 2008. anemia occurring during hepatitis C combination therapy, Hepatol-
Barclay SM, et al: Evaluation of pharmacologic prophylaxis for venous ogy 44(6):1598–1606, 2006.
thromboembolism in patients with chronic liver disease, Pharmaco- Demetriades D, et al: Liver cirrhosis in patients undergoing laparotomy
therapy 33(4):375–382, 2013. for trauma: effect on outcomes, J Am Coll Surg 199(4):538–542,
Belcher JM, et al: Association of AKI with mortality and complications 2004.
in hospitalized patients with cirrhosis, Hepatology 57(2):753–762, Determann RM, et al: Ventilation with lower tidal volumes as com-
2013. pared with conventional tidal volumes for patients without acute
Bellis L, et al: Hepatic venous pressure gradient does not correlate with lung injury: a preventive randomized controlled trial, Crit Care
the presence and the severity of portal hypertensive gastropathy in 14(1):R1, 2010.
patients with liver cirrhosis, J Gastrointestin Liver Dis 16(3):273–277, Deyle S, et al: Estimation of physiologic ability and surgical stress score
2007. does not predict immediate outcome after pancreatic surgery, Pan-
Bendtsen F, et al: Effect of recombinant Factor VIIa on outcome of creas 40(5):723–729, 2011.
acute variceal bleeding: an individual patient based meta-analysis of Diener MK, et al: Pylorus-preserving pancreaticoduodenectomy (pp
two controlled trials, J Hepatol 61(2):252–259, 2014. Whipple) versus pancreaticoduodenectomy (classic Whipple) for
Berendes E, et al: Effects of enflurane and isoflurane on splanchnic surgical treatment of periampullary and pancreatic carcinoma,
oxygenation in humans, J Clin Anesth 8(6):456–468, 1996. Cochrane Database Syst Rev 11:CD006053, 2014.
Berzigotti A, et al: Elastography, spleen size, and platelet count identify Dunki-Jacobs EM, et al: Stroke volume variation in hepatic resection:
portal hypertension in patients with compensated cirrhosis, Gastro- a replacement for standard central venous pressure monitoring, Ann
enterology 144(1):102–111.e101, 2013. Surg Oncol 21(2):473–478, 2014.
Bosch J, et al: Recombinant factor VIIa for variceal bleeding in patients Dwyer JP, et al: Analgesia for the cirrhotic patient: a literature review
with advanced cirrhosis: A randomized, controlled trial, Hepatology and recommendations, J Gastroenterol Hepatol 29(7):1356–1360,
47(5):1604–1614, 2008. 2014.
Bosilkovska M, et al: Analgesics in patients with hepatic impairment: Elkrief L, et al: Diabetes mellitus is an independent prognostic factor
pharmacology and clinical implications, Drugs 72(12):1645–1669, for major liver-related outcomes in patients with cirrhosis and
2012. chronic hepatitis C, Hepatology 60(3):823–831, 2014.
Braga M, et al: A prognostic score to predict major complications after Escorsell A, et al: TIPS versus drug therapy in preventing variceal
pancreaticoduodenectomy, Ann Surg 254(5):702–707, discussion rebleeding in advanced cirrhosis: a randomized controlled trial,
707-708, 2011. Hepatology 35(2):385–392, 2002.
Bureau C, et al: Transient elastography accurately predicts presence of Eshuis WJ, et al: Early postoperative hyperglycemia is associated
significant portal hypertension in patients with chronic liver disease, with postoperative complications after pancreatoduodenectomy, Ann
Aliment Pharmacol Ther 27(12):1261–1268, 2008. Surg 253(4):739–744, 2011.
444.e2 PART 3 ANESTHETIC MANAGEMENT, PRE- AND POSTOPERATIVE CARE
Fabrizi F, et al: Hepatorenal syndrome and novel advances in its man- Hayter MA, et al: Massive transfusion in the trauma patient: Continu-
agement, Kidney Blood Press Res 37(6):588–601, 2013. ing Professional Development, Can J Anaesth 59(12):1130–1145,
Fan ST, et al: Perioperative nutritional support in patients undergoing 2012.
hepatectomy for hepatocellular carcinoma, N Engl J Med 331(23): Hebuterne X, Vanbiervliet G: Feeding the patients with upper gastro-
1547–1552, 1994. intestinal bleeding, Curr Opin Clin Nutr Metab Care 14(2):197–201,
Farnsworth N, et al: Child-Turcotte-Pugh versus MELD score as a 2011.
predictor of outcome after elective and emergent surgery in cirrhotic Heizmann O, et al: Ischemic preconditioning improves postoperative
patients, Am J Surg 188(5):580–583, 2004. outcome after liver resections: a randomized controlled study, Eur J
Fathy O, et al: 216 cases of pancreaticoduodenectomy: risk factors Med Res 13(2):79–86, 2008.
for postoperative complications, Hepatogastroenterology 55(84):1093– Hickman IJ, Macdonald GA: Impact of diabetes on the severity of liver
1098, 2008. disease, Am J Med 120(10):829–834, 2007.
Fauconnet P, et al: Hepatopulmonary syndrome: The anaesthetic con- Ho V: Current concepts in the management of hepatopulmonary syn-
siderations, Eur J Anaesthesiol 30(12):721–730, 2013. drome, Vasc Health Risk Manag 4(5):1035–1041, 2008.
Fausto N, et al: Liver regeneration, Hepatology 43(2 Suppl 1):S45–S53, Ho CK, et al: Complications of pancreatic surgery, HPB (Oxford)
2006. 7(2):99–108, 2005.
Fayad L, et al: The performance of prognostic models as predictors of Hoetzel A, et al: Anesthetic considerations for the patient with liver
mortality in patients with acute decompensation of cirrhosis, Ann disease, Curr Opin Anaesthesiol 25(3):340–347, 2012.
Hepatol 14(1):83–92, 2015. Holodinsky JK, et al: Risk factors for intra-abdominal hypertension and
Fenger-Eriksen C, et al: Fibrinogen concentrate substitution therapy abdominal compartment syndrome among adult intensive care unit
in patients with massive haemorrhage and low plasma fibrinogen patients: a systematic review and meta-analysis, Crit Care 17(5):R249,
concentrations, Br J Anaesth 101(6):769–773, 2008. 2013.
Fernandez J, et al: Primary prophylaxis of spontaneous bacterial peri- Hoteit MA, et al: Model for end-stage liver disease score versus Child
tonitis delays hepatorenal syndrome and improves survival in cir- score in predicting the outcome of surgical procedures in patients
rhosis, Gastroenterology 133(3):818–824, 2007. with cirrhosis, World J Gastroenterol 14(11):1774–1780, 2008.
Ferral H, et al: Survival after elective transjugular intrahepatic porto- Hou MC, et al: Antibiotic prophylaxis after endoscopic therapy pre-
systemic shunt creation: prediction with model for end-stage liver vents rebleeding in acute variceal hemorrhage: a randomized trial,
disease score, Radiology 231(1):231–236, 2004. Hepatology 39(3):746–753, 2004.
Filicori F, et al: Prognostic criteria for postoperative mortality in 170 Johansson PI: Hemostatic strategies for minimizing mortality in surgery
patients undergoing major right hepatectomy, HBPD INT 11(5):507– with major blood loss, Curr Opin Hematol 16(6):509–514, 2009.
512, 2012. Juakiem W, et al: Nutrition in cirrhosis and chronic liver disease, Clin
Filsoufi F, et al: Early and late outcome of cardiac surgery in patients Liver Dis 18(1):179–190, 2014.
with liver cirrhosis, Liver Transpl 13(7):990–995, 2007. Kalaitzakis E, et al: Malnutrition and diabetes mellitus are related to
Franzetta M, et al: Prognostic factors of cirrhotic patients in extrahepatic encephalopathy in patients with liver cirrhosis, Liver Int
hepatic surgery, Minerva Chir 58(4):541–544, 2003. 27(9):1194–1201, 2007.
García-Pagán JC, et al: Acute variceal bleeding, Semin Resp Crit Care Kalambokis G, et al: The effects of treatment with octreotide, diuretics,
Med 33(1):46–54, 2012. or both on portal hemodynamics in nonazotemic cirrhotic patients
Garcia-Pagán JC, et al: Nadolol plus isosorbide mononitrate alone or with ascites, J Clin Gastroenterol 40(4):342–346, 2006.
associated with band ligation in the prevention of recurrent bleeding: Kato K, et al: Autologous blood storage before hepatectomy for hepa-
a multicentre randomised controlled trial, Gut 58(8):1144–1150, tocellular carcinoma, Hepatogastroenterology 56(91–92):802–807,
2009. 2009.
George R, Shiu MH: Hypophosphatemia after major hepatic resection, Katz SC, et al: Operative blood loss independently predicts recurrence
Surgery 111(3):281–286, 1992. and survival after resection of hepatocellular carcinoma, Ann Surg
Georgiou C, et al: Cirrhosis and trauma are a lethal combination, World 249(4):617–623, 2009.
J Surg 33(5):1087–1092, 2009. Kawut SM, et al: Hemodynamics and survival of patients with porto-
Giannini EG, et al: Incidence of bleeding following invasive procedures pulmonary hypertension, Liver Transpl 11(9):1107–1111, 2005.
in patients with thrombocytopenia and advanced liver disease, Clin Keeffe EB, et al: A treatment algorithm for the management of chronic
Gastroenterol Hepatol 8(10):899–902, 2010. hepatitis B virus infection in the United States: 2008 update, Clin
Gines P, Schrier RW: Renal failure in cirrhosis, N Engl J Med Gastroenterol Hepatol 6(12):1315–1341, quiz 1286, 2008.
361(13):1279–1290, 2009. Khamaysi I, et al: Sub-clinical hepatic encephalopathy in cirrhotic
Gluud LL, et al: Terlipressin for hepatorenal syndrome, Cochrane patients is not aggravated by sedation with propofol compared to
Database Syst Rev 9:CD005162, 2012. midazolam: a randomized controlled study, J Hepatol 54(1):72–77,
Gould MK, et al: Prevention of VTE in nonorthopedic surgical 2011.
patients: Antithrombotic Therapy and Prevention of Thrombosis, Khan J, et al: Ascites: aetiology, mortality and the prevalence of spon-
9th ed: American College of Chest Physicians Evidence-Based taneous bacterial peritonitis, Scand J Gastroenterol 44(8):970–974,
Clinical Practice Guidelines, Chest 141(Suppl 2):e227S–e277S, 2009.
2012. Koretz RL, et al: Nutritional support for liver disease, Cochrane Data-
Gupta D, et al: Pulmonary function changes after large volume para- base Syst Rev 5:CD008344, 2012.
centesis, Trop Gastroenterol 21(2):68–70, 2000. Krowka MJ, et al: Portopulmonary hypertension: results from a 10-year
Hadziyannis SJ, et al: Long-term therapy with adefovir dipivoxil for screening algorithm, Hepatology 44(6):1502–1510, 2006.
HBeAg-negative chronic hepatitis B for up to 5 years, Gastroenterol- Kuiper JJ, et al: Review article: management of ascites and associated
ogy 131(6):1743–1751, 2006. complications in patients with cirrhosis, Aliment Pharmacol Ther
Harimoto N, et al: Significance of stroke volume variation during 26(Suppl 2):183–193, 2007.
hepatic resection under infrahepatic inferior vena cava and portal Kwo PY: Management of hyponatremia in clinical hepatology practice,
triad clamping, Fukuoka Igaku Zasshi 104(10):362–369, 2013. Curr Gastroenterol Rep 16(5):382, 2014.
Harrison SA: Liver disease in patients with diabetes mellitus, J Clin Kwok CS, et al: Albumin reduces paracentesis-induced circulatory dys-
Gastroenterol 40(1):68–76, 2006. function and reduces death and renal impairment among patients
Harrod-Kim P, Waldman DL: Abnormal portal venous flow at sonog- with cirrhosis and infection: a systematic review and meta-analysis,
raphy predicts reduced survival after transjugular intrahepatic por- Biomed Res Int 2013:295153, 2013.
tosystemic shunt creation, J Vasc Interv Radiol 16(11):1459–1464, Lau GK, et al: Peginterferon Alfa-2a, lamivudine, and the combination
2005. for HBeAg-positive chronic hepatitis B, N Engl J Med 352(26):2682–
Hashimoto D, et al: Is an estimation of physiologic ability and surgical 2695, 2005.
stress able to predict operative morbidity after pancreaticoduodenec- Lavayssiere L, et al: Impact of molecular adsorbent recirculating
tomy?, J Hepatobiliary Pancreat Sci 17(2):132–138, 2010. system on renal recovery in type-1 hepatorenal syndrome patients
Hayashida N, et al: Clinical outcome after cardiac operations in with chronic liver failure, J Gastroenterol Hepatol 28(6):1019–1024,
patients with cirrhosis, Ann Thorac Surg 77(2):500–505, 2004. 2013.
Chapter 25 Perioperative critical care in hepatopancreatobiliary patients 444.e3
Levesque E, et al: Respiratory impact of paracentesis in cirrhotic Northup PG, et al: Coagulopathy does not fully protect hospitalized
patients with acute lung injury, J Crit Care 26(3):257–261, 2011. cirrhosis patients from peripheral venous thromboembolism, Am J
Levesque E, et al: Outcome of patients with cirrhosis requiring Gastroenterol 101(7):1524–1528, quiz 1680, 2006.
mechanical ventilation in ICU, J Hepatol 60(3):570–578, 2014. Obstein KL, et al: Association between model for end-stage liver
Lin CS, et al: Postoperative adverse outcomes after non-hepatic surgery disease and spontaneous bacterial peritonitis, Am J Gastroenterol
in patients with liver cirrhosis, Br J Surg 100(13):1784–1790, 2013. 102(12):2732–2736, 2007.
Lo GH, et al: Banding ligation versus nadolol and isosorbide mononi- O’Neill S, et al: Meta-analysis of ischaemic preconditioning for liver
trate for the prevention of esophageal variceal rebleeding, Gastroen- resections, Br J Surg 100(13):1689–1700, 2013.
terology 123(3):728–734, 2002. Pabinger I, et al: Prothrombin complex concentrate (Beriplex P/N) for
Loomba R, et al: Role of fluoroquinolones in the primary prophylaxis emergency anticoagulation reversal: a prospective multinational
of spontaneous bacterial peritonitis: meta-analysis, Clin Gastroenterol clinical trial, J Thromb Haemost 6(4):622–631, 2008.
Hepatol 7(4):487–493, 2009. Paugam-Burtz C, et al: Prospective validation of the “fifty-fifty” criteria
Lorf T, et al: Prognostic value of the monoethylglycinexylidide as an early and accurate predictor of death after liver resection in
(MEGX)-test prior to liver resection, Hepatogastroenterology 55(82– intensive care unit patients, Ann Surg 249(1):124–128, 2009.
83):539–543, 2008. Pavlidis TE, et al: Laparoscopic cholecystectomy in patients with cir-
Luddington RJ: Thrombelastography/thromboelastometry, Clin Lab rhosis of the liver and symptomatic cholelithiasis, JSLS 13(3):342–
Haematol 27(2):81–90, 2005. 345, 2009.
Malbrain ML, et al: Consensus conference definitions and recommen- Pohl J, et al: Antibiotic prophylaxis after variceal hemorrhage reduces
dations on intra-abdominal hypertension (IAH) and the abdominal incidence of early rebleeding, Hepatogastroenterology 51(56):541–
compartment syndrome (ACS)—the long road to the final publica- 546, 2004.
tions, how did we get there?, Acta Clin Belg Suppl 1:44–59, 2007. Polverino F, et al: Recumbent deoxygenation in mild/moderate liver
Malinchoc M, et al: A model to predict poor survival in patients under- cirrhosis: the “clinodeoxia”. The ortho-clino paradigm, Respir Med
going transjugular intrahepatic portosystemic shunts, Hepatology 108(7):1040–1048, 2014.
31(4):864–871, 2000. Poropat G, et al: Enteral nutrition formulations for acute pancreatitis,
Marcellin P, et al: Tenofovir disoproxil fumarate versus adefovir dipiv- Cochrane Database Syst Rev 3:CD010605, 2015.
oxil for chronic hepatitis B, N Engl J Med 359(23):2442–2455, 2008. Porres-Aguilar M, et al: Portopulmonary hypertension and hepatopul-
Marchesini G, et al: Nutritional supplementation with branched-chain monary syndrome: a clinician-oriented overview, Eur Respir Rev
amino acids in advanced cirrhosis: a double-blind, randomized trial, 21(125):223–233, 2012.
Gastroenterology 124(7):1792–1801, 2003. Qamar AA, et al: Incidence, prevalence, and clinical significance of
Martin-Llahi M, et al: Terlipressin and albumin vs albumin in patients abnormal hematologic indices in compensated cirrhosis, Clin Gas-
with cirrhosis and hepatorenal syndrome: a randomized study, Gas- troenterol Hepatol 7(6):689–695, 2009.
troenterology 134(5):1352–1359, 2008. Qi X, et al: Transjugular intrahepatic portosystemic shunt for acute
Mazzeo AT, et al: Hepatopulmonary syndrome: a concern for the anes- variceal bleeding: a meta-analysis, J Clin Gastroenterol 49(6):495–
thetist? Pre-operative evaluation of hypoxemic patients with liver 505, 2015.
disease, Acta Anaesthesiol Scand 48(2):178–186, 2004. Rahnemai-Azar AA, et al: Percutaneous endoscopic gastrostomy: indi-
McCormack L, et al: Hepatic steatosis is a risk factor for postoperative cations, technique, complications and management, World J Gastro-
complications after major hepatectomy: a matched case-control enterol 20(24):7739–7751, 2014.
study, Ann Surg 245(6):923–930, 2007. Rebulla P: Platelet transfusion trigger in difficult patients, Transfus Clin
Melendez JA, et al: Perioperative outcomes of major hepatic resections Biol 8(3):249–254, 2001.
under low central venous pressure anesthesia: blood loss, blood Richter A, et al: Long-term results of partial pancreaticoduodenectomy
transfusion, and the risk of postoperative renal dysfunction, J Am for ductal adenocarcinoma of the pancreatic head: 25-year experi-
Coll Surg 187(6):620–625, 1998. ence, World J Surg 27(3):324–329, 2003.
Moller S, Henriksen JH: Cardiovascular complications of cirrhosis, Gut Richter B, et al: Nutritional support after open liver resection: a sys-
57(2):268–278, 2008. tematic review, Dig Surg 23(3):139–145, 2006.
Monroe DM, Hoffman M: The coagulation cascade in cirrhosis, Clin Ripoll C, et al: Hepatic venous pressure gradient predicts clinical
Liver Dis 13(1):1–9, 2009. decompensation in patients with compensated cirrhosis, Gastroenter-
Montalto P, et al: Bacterial infection in cirrhosis impairs coagulation ology 133(2):481–488, 2007.
by a heparin effect: a prospective study, J Hepatol 37(4):463–470, Robinson SM, et al: N-acetylcysteine administration does not improve
2002. patient outcome after liver resection, HPB (Oxford) 15(6):457–462,
Morris TA, et al: No difference in risk for thrombocytopenia during 2013.
treatment of pulmonary embolism and deep venous thrombosis with Rockey DC, et al: Liver biopsy, Hepatology 49(3):1017–1044, 2009.
either low-molecular-weight heparin or unfractionated heparin: a Rodriquez-Roisin R, et al: Highlights of the ERS Task Force on
metaanalysis, Chest 132(4):1131–1139, 2007. pulmonary-hepatic vascular disorders (PHD), J Hepatol 42(6):924–
Nassar Junior AP, et al: Terlipressin versus norepinephrine in the treat- 927, 2005.
ment of hepatorenal syndrome: a systematic review and meta- Romero-Castro R, et al: Recombinant-activated factor VII as hemo-
analysis, PLoS ONE 9(9):e107466, 2014. static therapy in eight cases of severe hemorrhage from esophageal
Neeff HP, et al: Early mortality and long-term survival after abdominal varices, Clin Gastroenterol Hepatol 2(1):78–84, 2004.
surgery in patients with liver cirrhosis, Surgery 155(4):623–632, Rowe AJ, et al: Perioperative analysis of laparoscopic versus open liver
2014. resection, Surg Endosc 23(6):1198–1203, 2009.
Nguyen GC, et al: The impact of cirrhosis and portal hypertension on Roy JD, et al: A comparison of intrathecal morphine/fentanyl and
mortality following colorectal surgery: a nationwide, population- patient-controlled analgesia with patient-controlled analgesia alone
based study, Dis Colon Rectum 52(8):1367–1374, 2009. for analgesia after liver resection, Anesth Analg 103(4):990–994,
Nicoll A: Surgical risk in patients with cirrhosis, J Gastroenterol Hepatol 2006.
27(10):1569–1575, 2012. Salerno F, et al: Diagnosis, prevention and treatment of hepatorenal
Niedergethmann M, et al: Early and enduring nutritional and func- syndrome in cirrhosis, Gut 56(9):1310–1318, 2007.
tional results of pylorus preservation vs classic Whipple procedure Sam J, Nguyen GC: Protein-calorie malnutrition as a prognostic indica-
for pancreatic cancer, Langenbecks Arch Surg 391(3):195–202, 2006. tor of mortality among patients hospitalized with cirrhosis and portal
Nishiyama T, et al: A comparison of liver function after hepatectomy hypertension, Liver Int 29(9):1396–1402, 2009.
in cirrhotic patients between sevoflurane and isoflurane in anesthesia Sanyal AJ, et al: A randomized, prospective, double-blind, placebo-
with nitrous oxide and epidural block, Anesth Analg 98(4):990–993, controlled trial of terlipressin for type 1 hepatorenal syndrome, Gas-
2004, table. troenterology 134(5):1360–1368, 2008.
Nobre SR, et al: In-hospital mortality in spontaneous bacterial perito- Satoi S, et al: Less morbidity after pancreaticoduodenectomy of
nitis: a new predictive model, Eur J Gastroenterol Hepatol 20(12):1176– patients with pancreatic cancer, Pancreas 33(1):45–52, 2006.
1181, 2008. Scheingraber S, et al: Indocyanine green disappearance rate is the most
Northup PG, Caldwell SH: Coagulation in Liver Disease: A Guide for useful marker for liver resection, Hepatogastroenterology 55(85):1394–
the Clinician, Clin Gastroenterol Hepatol 11(9):1064–1074, 2013. 1399, 2008.
444.e4 PART 3 ANESTHETIC MANAGEMENT, PRE- AND POSTOPERATIVE CARE
Schepke M, et al: Comparison of MELD, Child-Pugh, and Emory Triantos C, et al: Endogenous heparinoids detected by anti-Xa activity
model for the prediction of survival in patients undergoing transjugu- are present in blood during acute variceal bleeding in cirrhosis. A
lar intrahepatic portosystemic shunting, Am J Gastroenterol 98(5): prospective study, J Gastrointestin Liver Dis 23(2):187–194, 2014.
1167–1174, 2003. Tripodi A, et al: Thrombin generation in plasma from patients with
Schiffman SC, et al: Laparoscopic versus open liver resection for meta- cirrhosis supplemented with normal plasma: considerations on the
static colorectal cancer: a metaanalysis of 610 patients, Surgery efficacy of treatment with fresh-frozen plasma, Intern Emerg Med
157(2):211–222, 2015. 7(2):139–144, 2012.
Schilsky ML: Acute liver failure and liver assist devices, Transplant Proc Tripodi A, Mannucci PM: The coagulopathy of chronic liver disease,
43(3):879–883, 2011. N Engl J Med 365(2):147–156, 2011.
Schmidt CM, et al: Pancreaticoduodenectomy: a 20-year experience in Uchida K, et al: Impact of remifentanil introduction on practice pat-
516 patients, Arch Surg 139(7):718–725, discussion 725-717, 2004. terns in general anesthesia, J Anesth 25(6):864–871, 2011.
Senzolo M, et al: Heparin-like effect in liver disease and liver transplan- Uzunoglu FG, et al: Preoperative Pancreatic Resection (PREPARE)
tation, Clin Liver Dis 13(1):43–53, 2009. score: a prospective multicenter-based morbidity risk score, Ann
Serste T, et al: Deleterious effects of beta-blockers on survival in Surg 260(5):857–863, discussion 863-854, 2014.
patients with cirrhosis and refractory ascites, Hepatology 52(3):1017– Vavra P, et al: Radiofrequency-assisted liver resections: comparison of
1022, 2010. open and laparoscopic techniques, Hepatogastroenterology 61(136):
Shah NL, et al: Procoagulant therapeutics in liver disease: a critique 2359–2366, 2014.
and clinical rationale, Nat Rev Gastroenterol Hepatol 11(11):675–682, Verbeeck RK: Pharmacokinetics and dosage adjustment in patients
2014. with hepatic dysfunction, Eur J Clin Pharmacol 64(12):1147–1161,
Shaheen AA, et al: Morbidity and mortality following coronary artery 2008.
bypass graft surgery in patients with cirrhosis: a population-based Vidal MG, et al: Incidence and clinical effects of intra-abdominal
study, Liver Int 29(8):1141–1151, 2009. hypertension in critically ill patients, Crit Care Med 36(6):1823–
Sharma P, et al: An open label, pilot, randomized controlled trial of 1831, 2008.
noradrenaline versus terlipressin in the treatment of type 1 hepatore- Wang ZX, et al: Dexmedetomidine reduces intestinal and hepatic
nal syndrome and predictors of response, Am J Gastroenterol 103(7): injury after hepatectomy with inflow occlusion under general anaes-
1689–1697, 2008. thesia: a randomized controlled trial, Br J Anaesth 112(6):1055–
Silkauskaite V, et al: Hepatic venous pressure gradient measurement 1064, 2014.
in patients with liver cirrhosis: a correlation with disease severity and Welsch T, et al: The need for extended intensive care after pancreati-
variceal bleeding, Medicina (Kaunas) 45(1):8–13, 2009. coduodenectomy for pancreatic ductal adenocarcinoma, Langenbecks
Silver M, et al: Efficacy of recombinant human erythropoietin in criti- Arch Surg 396(3):353–362, 2011.
cally ill patients admitted to a long-term acute care facility: a Woods K, et al: Low-dose oral vitamin K to normalize the international
randomized, double-blind, placebo-controlled trial, Crit Care Med normalized ratio prior to surgery in patients who require temporary
34(9):2310–2316, 2006. interruption of warfarin, J Thromb Thrombolysis 24(2):93–97, 2007.
Singh C, Sager JS: Pulmonary complications of cirrhosis, Med Clin Wright G, Jalan R: Management of hepatic encephalopathy in patients
North Am 93(4):871–883, viii, 2009. with cirrhosis, Best Pract Res Clin Gastroenterol 21(1):95–110, 2007.
Slankamenac K, et al: Development and validation of a prediction Ye B, et al: Ischemic preconditioning provides no additive clinical value
score for postoperative acute renal failure following liver resection, in liver resection of cirrhotic and non-cirrhotic patients under portal
Ann Surg 250(5):720–728, 2009. triad clamping: a prospective randomized controlled trial, Clin Res
Sogaard KK, et al: Risk of venous thromboembolism in patients with Hepatol Gastroenterol 38(4):467–474, 2014.
liver disease: a nationwide population-based case-control study, Am Yeo CJ, et al: Does prophylactic octreotide decrease the rates of pan-
J Gastroenterol 104(1):96–101, 2009. creatic fistula and other complications after pancreaticoduodenec-
Squires MH 3rd, et al: Hypophosphataemia after major hepatectomy tomy? Results of a prospective randomized placebo-controlled trial,
and the risk of post-operative hepatic insufficiency and mortality: an Ann Surg 232(3):419–429, 2000.
analysis of 719 patients, HPB (Oxford) 16(10):884–891, 2014. Zakian KL, et al: Liver regeneration in humans is characterized by
Sussman NL, et al: Hepatic assist devices: will they ever be successful?, significant changes in cellular phosphorus metabolism: assessment
Curr Gastroenterol Rep 11(1):64–68, 2009. using proton-decoupled 31P-magnetic resonance spectroscopic
Swanson KL, et al: Survival in portopulmonary hypertension: Mayo imaging, Magn Reson Med 54(2):264–271, 2005.
Clinic experience categorized by treatment subgroups, Am J Trans- Zambruni A, et al: Endogenous heparin-like activity detected by
plant 8(11):2445–2453, 2008. anti-Xa assay in infected cirrhotic and non-cirrhotic patients, Scand
Teh SH, et al: Risk factors for mortality after surgery in patients with J Gastroenterol 39(9):830–836, 2004.
cirrhosis, Gastroenterology 132(4):1261–1269, 2007. Zhu Y, et al: Ischemic preconditioning versus intermittent clamping of
Terg R, et al: Ciprofloxacin in primary prophylaxis of spontaneous portal triad in liver resection: a meta-analysis of randomized con-
bacterial peritonitis: a randomized, placebo-controlled study, trolled trials, Hepatol Res 44(8):878–887, 2014.
J Hepatol 48(5):774–779, 2008. Zipprich A, et al: Incorporating indocyanin green clearance into the
Tinmouth A: Evidence for a rationale use of frozen plasma for the Model for End Stage Liver Disease (MELD-ICG) improves prog-
treatment and prevention of bleeding, Transfus Apher Sci 46(3):293– nostic accuracy in intermediate to advanced cirrhosis, Gut 59(7):963–
298, 2012. 968, 2010.
Tran KT, et al: Pylorus preserving pancreaticoduodenectomy versus Ziser A, et al: Morbidity and mortality in cirrhotic patients undergoing
standard Whipple procedure: a prospective, randomized, multi- anesthesia and surgery, Anesthesiology 90(1):42–53, 1999.
center analysis of 170 patients with pancreatic and periampullary
tumors, Ann Surg 240(5):738–745, 2004.
CHAPTER 26
Preoperative and postoperative nutrition in
hepatobiliary surgery
Farzad Alemi, D. Owen Young, and William S. Helton
GOALS OF NUTRITIONAL SUPPORT IN in catabolism is thought to be related, in part, to a defective

HEPATOPANCREATOBILIARY SURGERY gastrointestinal (GI) mucosal barrier, which leads to the trans-
location of endotoxin and bacteria from the bowel into the
The objectives of nutritional support (NS) in patients undergo- peritoneal cavity and portal venous system (Helton, 1994). The
ing hepatopancreatobiliary (HPB) surgery are to restore and transmigrated endotoxin stimulates peritoneal mononuclear
preserve health while minimizing postoperative morbidity and phagocytic cells and Kupffer cells in the liver to release proin-
mortality. Nutritional intervention should attempt to maintain flammatory cytokines and mediators that adversely influence
or restore immunity, support anabolism, and attenuate the amino acid, protein, lipid, carbohydrate, and trace mineral
catabolic response to injury. For hepatic surgery, NS should metabolism (Andus et al, 1991).
strive to restore specific hepatocellular function, attenuate Tumor necrosis factor (TNF) is the proximal cytokine signal
hepatocellular injury in response to ischemia and reperfusion produced by hepatic Kupffer cells in response to endotoxin or
(I/R), and optimally support hepatocellular function and regen- I/R injury. TNF initiates a cascade of inflammatory events that
eration after resection, ablation, and/or transplantation. For are important in the pathogenesis of many types of surgically
pancreatic and biliary surgery, NS should strive to replenish induced liver injury that occur with liver resection and trans-
patients with malnutrition before surgery, optimize endocrine plantation. Endotoxemia occurring in response to manipulation
and exocrine function, prevent or attenuate cancer-related of the biliary tree in patients with biliary obstruction also stimu-
anorexia-cachexia syndrome, and restore normal digestion lates the release of TNF, which mediates the septic response
and absorption of nutritional substrates, micronutrients, and and organ failure associated with operating in the setting of
vitamins. biliary obstruction and infection (Nolan, 1981). TNF and
In addition to providing the basic components to correct interleukin-6 (IL-6) cause a reprioritization of hepatic protein
protein-calorie malnutrition, NS strategies have tried to identify synthesis, a process involving accelerated production of acute-
subsets of patients who are at increased risk for nutrition- phase proteins at the expense of constitutive proteins. Indirect
related problems and provide disease-specific nutrient require- evidence suggests that parenteral nutrition (PN) also weakens
ments for individual organs, cells, and tissues. This new the intestinal barrier, allowing endotoxin to escape from the
direction of nutritional science has been termed nutritional phar- gut, where it then stimulates Kupffer cells to release cytokines
macology, and it is based on the observation that individual (Fong et al, 1989).
components of the immunoinflammatory response to operation
and infection can be influenced and manipulated by the timely Antioxidant Nutrient Depletion in the Pathogenesis
administration of specific nutrients, in specific concentrations, of Liver Injury
and by specific routes of feeding. Using the principles of nutri- Patients with liver disease, biliary obstruction, bacterial or viral
tional pharmacology, clinicians and scientists have observed infection, or malnutrition have impaired antioxidant defenses
that health and disease can be influenced by nutritional means. coupled with increased oxidant stresses (Bell et al, 1992; Burra
The first part of this chapter will discuss disease and treatment- et al, 1992). Additional factors that deplete hepatic antioxi-
related physiologic and anatomic factors that predispose dants include smoking, alcohol ingestion, general anesthesia,
patients with HPB disease to malnutrition before and after and surgery (Bulger & Helton, 1998). Animal studies suggest
operation. The latter part of the chapter will focus on NS with that a major pathophysiologic event in hepatocellular injury is
an emphasis not only on who should receive it, but also when, depletion of endogenous antioxidants at the time of increased
how, and in what formulation. oxidative stress from infection (Sugino et al, 1989), liver resec-
tion (Ouchi et al, 1991), or transplantation (Serino et al, 1990).
ALTERATIONS IN LIVER METABOLISM THAT Patients with chronic liver disease have altered bile salt pools
AFFECT NUTRITIONAL STATUS and enterohepatic circulation of bile salts (see Chapter 8). This
leads to impaired micelle formation and consequent malab-
The liver plays a crucial role in the metabolism and assimilation sorption of fat and fat-soluble vitamins A, D, E, and K. Patients
of nutrients, and it is central in the orchestration of protein and with cirrhosis undergoing liver transplantation have even further
carbohydrate metabolism. Patients with advanced liver disease depletion of plasma levels of vitamin E, vitamin A, and caro-
usually are catabolic and have carbohydrate intolerance and tene, and these decrease even further after transplantation
increased muscle proteolysis (Eigler et al, 1979). This increase (Goode et al, 1994) (see Chapters 76 and 112).
445
intestine (Sun et al, 2014) and Kupffer cells of the liver (Billiar
Vitamin E et al, 1988). These observations show that hepatocellular func-
Vitamin E has the ability to scavenge intermediate peroxyl tion before and after HPB operations can be potentially modu-
radicals and therefore interrupt the chain reactions of lipid lated by the administration of specific nutrients and vitamins
peroxidation (Birlouez-Aragon et al, 2003), and it is the most (Helton, 1994) and provide a potential opportunity whereby
important inhibitor of the free-radical chain reaction of lipid the hepatobiliary surgeon might influence patient outcomes by
peroxidation. Studies in rodents show that supplemental nutritional means. ω3FAs have also been implicated in the
vitamin E significantly attenuates liver I/R injury and hepatic maintenance of postoperative hepatic function by increasing
lipid peroxidation (Sugino et al, 1989). In vitro, physiologic liver perfusion. A meta-analysis of 21 randomized controlled
concentrations of vitamin E inhibit lipopolysaccharide- trials showed perioperative administration of ω3FAs to signifi-
stimulated TNF secretion by Kupffer cells, suggesting that cantly reduce hospital stay, infections, and liver functions tests
subnormal tissue or plasma levels of vitamin E may potentiate (Li et al, 2014).
macrophage cytokine release (McClain et al, 1994).
Plasma levels of vitamin E decrease significantly in the first Obstructive Jaundice
hours after surgery. A significant reduction in the levels of liver Patients with jaundice often have anorexia and weight loss
α-tocopherol, an active form of vitamin E, was observed during because of decreased oral intake coupled with malabsorption.
the first hour of reperfusion in a rat model of liver ischemia Approximately 45% to 70% of patients with obstructive jaun-
(Maderazo et al, 1990, 1991). Obstructive jaundice often is dice are seen with malnutrition as evidenced by greater than
associated with endotoxemia, which leads to Kupffer cell pro- 10% weight loss, albumin less than 3 g/dL, decreased triceps
duction of oxygen-free radicals and nitric oxide, which inhibit skin fold, and impaired delayed hypersensitivity reactivity
protein synthesis (Curran et al, 1990). Endotoxemia also (Foschi et al, 1986). The primary nutritional deficit resulting
reduces endogenous levels of the antioxidants glutathione, from obstructive jaundice is malabsorption of fat and fat-
vitamin E, and coenzyme Q (Marubayashi et al, 1986). soluble vitamins. In addition, there is loss of trace minerals,
Vitamin E has a variety of protective effects on the hepato- such as phosphate, calcium, magnesium, and zinc, because of
biliary system. It attenuates endotoxemia (Powell et al, 1991), salt formation from unabsorbed dietary fat (Shronts, 1988).
hepatocellular membrane lipid peroxidation, and cellular Biliary sepsis in patients with obstructive jaundice contrib-
damage after I/R injury (Lee & Clemens, 1992). Large doses utes to malnutrition by shifting protein synthesis from anabolic
of enterally administered vitamin E inhibit the release of TNF protein synthesis to acute-phase protein synthesis (O’Neill
in animal models of infection (Bulger et al, 1997). Under these et al, 1997). This reprioritization of protein synthesis occurs as
conditions, vitamin E supplementation improves survival after a result of endotoxin-stimulated Kupffer cell production of
a septic challenge (Yoshikawa et al, 1984). α-Tocopherol has TNF, IL-6, eicosanoids, and other inflammatory mediators that
also shown protective effects against warm and cold I/R injury directly inhibit protein synthesis (see Chapter 43). Because of
(Eum et al, 2002; Gondolesi et al, 2002). These animal studies these derangements, some authors have advocated preoperative
show a protective effect of vitamin E on liver function and biliary decompression (PBD) in patients undergoing HPB
survival during conditions commonly encountered in patients surgery. However, multicenter trials and Cochrane analysis
undergoing HPB surgery. Despite this evidence, no clinical have demonstrated no evidence to support or refute routine
trials have demonstrated a beneficial effect of vitamin E in biliary drainage and stenting in patients with malignant pancre-
patients undergoing HPB surgery. aticobiliary diseases awaiting surgery (Mumtaz et al, 2007;
Wang et al, 2008) (see Chapters 29, 30, 51, and 52). On the
Vitamin C other hand, PBD should probably be performed in patients
Ascorbic acid (vitamin C) is required as a cofactor for many undergoing extended hepatectomy for hilar cholangiocarci-
enzymes involved in the scavenging of free radicals. Vitamin C noma, if the future liver remnant volume is estimated to be less
deficiency is evident in 50% of patients with alcoholic liver than or equal to 30% (see Chapter 108). To reverse the cata-
disease (Muller, 1995) and is probably even higher in alcoholics bolic effects of chronic endotoxemia and restore hepatic protein
who smoke. Vitamin C recycles α-tocopherol and is intimately synthesis, patients with biliary infection should be treated with
linked to vitamin E’s ability to quench free radical–mediated biliary decompression for at least 4 weeks before major HPB
cellular damage (Sardesai, 1995). The simultaneous adminis- surgery to allow hepatocytes to recover their protein synthetic
tration of ascorbate and α-tocopherol is more effective in inhib- capacity.
iting oxidation than either alone (Niki et al, 1995). The
synergistic protective effects of vitamin C and vitamin E in Cirrhosis and Liver Failure
preventing lipid peroxidation and cellular damage suggest that Patients with cirrhosis have multiple hormonal and metabolic
these vitamins should be administered together for maximal alterations (see Chapter 79). These changes cause loss of fat
potential benefit (Bulger & Helton, 1998). A prospective ran- and muscle mass, growth failure, glucose intolerance, hyperin-
domized study in patients undergoing liver resection by using sulinemia, insulin resistance, increased plasma glucagon and
an infusion containing 10 mg α-tocopherol acetate and 1 g catecholamines, elevated serum free fatty acids, elevated glyc-
ascorbate administered before reperfusion demonstrated less erol, hypoproteinemia, hyperammonemia, hypophosphatemia,
plasma lipid peroxidation and acute liver damage and fewer and alterations in plasma and cerebrospinal fluid amino-acid
postoperative complications (Cerwenka et al, 1999). profiles (Achord, 1987; Henriksen et al, 1985; Petrides & De
Fronzo, 1989; Riggio et al, 1984). These metabolic aberrations
Omega-3 Fatty Acids lead to altered metabolism of fat, protein, and carbohydrate. In
The administration of fish oils rich in omega-3 fatty acids addition, there is an increase in skeletal muscle proteolysis for
(ω3FAs) can influence cytokine and prostanoid release by the energy provision, which leads to eventual muscle wasting. In
Chapter 26 Preoperative and postoperative nutrition in hepatobiliary surgery 447
methionine—but not those of system N (glutamine, histidine,

BOX 26.1 Causes of Malnutrition in Cirrhosis
asparagine) or system ASC (e.g., cysteine) (Fowler et al, 1992).
Poor Dietary Intake Increased system A activity depends on glucagon and insulin
Anorexia, nausea, alcohol abuse, dietary restrictions (protein, fat, secretion into the portal venous system and substrate amino-
sodium, fluid)
acid supply (Dolais-Kitabgi et al, 1981); system N and system
ASC are not similarly regulated (Fowler et al, 1992). This fact
Malabsorption/Maldigestion
Cholestasis, intraluminal bile deficiency, coexisting pancreatic exo- supports the argument that enteral administration of glucose is
crine insufficiency, fat malabsorption the preferred route of feeding because of the effect on insulin
release, which is vital to the function of the regenerating liver
Increased Catabolism (Ozawa, 1992).
Muscle proteolysis John and colleagues (1992) reported that postoperative PN
consisting of 45% fat calories significantly impaired hepatic
Decreased Protein Synthesis regeneration and albumin synthesis, caused cholestasis, and
Decreased hepatocyte growth hormone receptor, IGF-1 and increased mortality compared with the same diet administered
IGF-BP, hepatic transport proteins, fibrinogen, coagulation factors,
by the enteral route after 70% hepatectomy in rats. Mortality
lipoproteins
was 68% in rats fed with total parenteral nutrition (TPN), 9%
Drug Therapy Effects in rats fed enterally, and 8% in rats fed chow. This increased
Neomycin: villous atrophy, diarrhea, zinc deficiency mortality rate in TPN-fed rats could be the result of increased
Lactulose: diarrhea, zinc deficiency bacterial translocation and lipopolysaccharide migration across
Diuretics: potassium, magnesium, zinc deficiency the gut, which overwhelms the limited phagocyte capacity of
Cholestyramine: diarrhea, fat-soluble vitamin deficiency the remnant Kupffer cell mass (van Leeuwen et al, 1991). The
increased mortality rate may also be the result of excessive
IGF, Insulin-like growth factor; BP, binding protein.
administration of intravenous (IV) glucose, which adversely
affects liver substrate metabolism (Ozawa, 1992). The lethal
consequences of PN after liver resection in rats can be amelio-
addition, there is increased peripheral lipolysis with a decreased rated by decreasing the caloric and amino-acid load.
ability to use fat and carbohydrate, which leads to hyperglyce-
mia and hyperlipidemia (Katz, 1986). Compounding the meta- Transplantation
bolic alterations of cirrhosis are issues that predispose patients In the first 6 hours after liver transplantation, glucose use by
with cirrhosis to malnutrition, including decreased dietary the transplanted liver is impaired until the redox state of the
intake due to anorexia, nausea and vomiting, and the common mitochondria improves (Ozaki et al, 1991) (see Chapter 113).
practice of imposing protein restriction in an effort to prevent During this time, the liver preferentially uses fatty acid oxida-
encephalopathy. The practice of routine protein restriction tion for adenosine triphosphate (ATP) generation (Ozaki et al,
should be abandoned in an already malnourished patient 1991; Takada et al, 1993). After 6 hours, normally functioning
because it exacerbates the problems inherently associated liver allografts shift substrate use from fat to glucose, whereas
with malnutrition and prohibits the goal of liver regeneration failing livers continue to use fat. This shift in metabolism can
(Box 26.1). be followed by measuring the plasma concentration of total
ketone bodies and arterial ketone body ratio (Takada et al,
Liver Resection 1993). Based on these observations, Takada and colleagues
Numerous metabolic alterations occur in the regenerating liver suggested that glucose should be administered in small quanti-
after liver resection (see Chapter 6). Krebs cycle activity is ties in the immediate postoperative period without insulin to
depressed, as is the reduction-oxidation state of the hepatic avoid suppressing peripheral fat mobilization. Glucose infusion
mitochondria, with a switch from the use of glucose to fat as should be increased slowly while mitochondria respiration
the preferred source of energy by way of β-oxidation (Nakatoni recovers and the Krebs cycle becomes active. When the redox
et al, 1981). Because hyperglycemia and hyperinsulinemia sup- state of mitochondria is sufficiently improved, amino acids and
press the release of fatty acids from adipose tissue and decrease increased amounts of glucose can be administered. It has been
ketone body production by the liver (Riou et al, 1986), hyper- demonstrated that intraportal insulin administration augments
tonic glucose infusions and insulin administration should be liver regeneration during the first postoperative transplant week
avoided in the immediate (<6 hours) postoperative period by improving hepatic function in live-donor liver transplant
(Ozawa, 1992). These metabolic alterations emphasize the recipients (Xu et al, 2009).
importance of glycemic control before, during, and immedi- The nutritional status of the donor liver may be an impor-
ately after operation, such as with preoperative carbohydrate tant factor affecting posttransplant allograft function. Initially,
loading. it was believed that liver allografts obtained from fasted donors
Provision of adequate protein is important to ensure ade- were more susceptible to anoxic liver injury than allografts
quate liver regeneration. Specific types of protein-supplemented obtained from donors being administered glucose until the time
diets, such as a nucleoside-nucleotide–supplemented PN of liver donation because of an absence of glycogen as a source
solution, may improve protein synthesis after hepatectomy of glucose from glycolysis in the fasted group (Bradford et al,
(Ogoshi et al, 1989). The regenerating liver has an increased 1986). Subsequent work in animal liver transplant models
demand for specific amino acids, and provision of these in the (Astarcioglu et al, 1991) and human liver donors (Cywes et al,
diet accelerates regeneration. Immediately after liver resection, 1992) showed that glucose administration to the donor increases
an abrupt increase is seen in the synthesis of the system A liver glycogen and ATP content before cold preservation and
amino-acid transporter—which transports alanine, serine, and attenuates liver injury after transplantation.
The effects of fasting and feeding on liver allograft function upon chronic pancreatitis, are at risk for profound malnutrition
after cold preservation and reperfusion are complex and incom- and metabolic derangement because of the catabolic effects of
pletely understood. No guidelines currently exist to support critical illness and sepsis (see Chapters 55 and 56). Despite
specific feeding regimens or NS guidelines for liver transplant these collective risks, patients with severe acute pancreatitis are
donors. However, using the principles of pharmaconutrition, best managed with oral and/or enteral nutrition (EN) (Bakker
there is the potential to modulate the response of liver grafts to et al, 2014). A detailed discussion of nutritional support for
I/R. patients with acute pancreatitis is covered in Chapter 54.
Hepatocellular Carcinoma Pancreatic Cancer (See Chapter 62)

Patients with hepatocellular carcinoma (HCC) commonly are Pancreatic cancer (PC) is associated with a severe metabolic
seen with progressive weight loss, anorexia, malnutrition, and derangement referred to as the cancer anorexia-cachexia syn-
sarcopenia because the majority have cirrhosis (see Chapter drome (Inui, 2002; Ryan et al, 1998). This syndrome is associ-
91). Cirrhosis can lead to inefficient energy expenditure ated with anorexia, tissue wasting, malnutrition, weight loss,
resulting from increased Cori cycle activity, increased protein and a loss of compensatory increase in feeding. The pathogen-
turnover rates with altered amino-acid profiles, increased esis is dependent on disorders of carbohydrate, protein, lipid,
gluconeogenesis, and increased lipid oxidation (Holroyde & and energy metabolism mediated by proinflammatory cytokine
Reichard, 1981). If jaundice is present, fat malabsorption often elaboration and an overall increase in leptin (Bruera et al, 2000;
occurs in these patients. Malnourished patients with cirrhosis Tisdale, 2003; Walker, 2001;). Patients with PC have the
or chronic active hepatitis and HCC usually are seen initially highest incidence of cachexia among patients with cancer; as
in an advanced clinical stage. These patients generally should many as 80% of such patients have cachexia at the time of
not undergo operation because of a prohibitive operative mor- initial diagnosis. Increasing severity of anorexia-cachexia in
tality (Bruix & Sherman, 2011). patients with PC interferes with therapy and correlates with
short survival (Bachmann et al, 2008; Martin et al, 2014). For
this reason, patients with PC are perhaps the most likely of
NUTRITIONAL PROBLEMS IN PATIENTS WITH all patients undergoing HPB surgery to benefit from NS
PANCREATIC DISEASES strategies.
Etiology of Malnutrition With Pancreatic Disease Pancreatic Exocrine Insufficiency

There are many factors that contribute to perioperative nutri- Patients with benign and malignant conditions of the pancreas
tional deficits and/or malnutrition in patients with pancreatic can experience biliary and/or pancreatic duct obstruction,
disease (see Chapter 62) (Box 26.2). These factors may be which can lead to malnutrition caused by maldigestion and
related to the disease itself or to its treatment. Nutritional defi- malabsorption. The majority of patients with pancreatic carci-
cits may also arise only after a long time or appear suddenly in noma are seen with malnutrition and weight loss at the time of
response to acute disease or treatment-related complications. diagnosis. Sikkens observed that 75% of patients with PC had
For example, patients with chronic pancreatitis and abdominal pancreatic exocrine insufficiency (PEI) at presentation, and this
pain may have profound malnutrition because of food avoid- increased to 92% of patients within 2 months (Sikkens et al,
ance, dietary restriction, exocrine or endocrine insufficiency, 2014). The etiology of PEI in patients with PC is incompletely
pancreatic and/or biliary duct obstruction, and chronic malab- understood but likely related to several things. First, many
sorption (see Chapters 57 and 58). Patients with severe acute patients have main pancreatic duct obstruction within the pan-
pancreatitis, and particularly those with acute superimposed creatic head, and pancreatic juice cannot flow into the duode-
num. Second, long-standing pancreatic duct obstruction leads
to pancreatic atrophy. Third, pancreatic exocrine senescence
develops with age, and PC occurs predominantly in the elderly.
BOX 26.2 Causes of Malnutrition in Pancreatic Disease
Fourth, pancreatitis develops in many patients with pancreatic
Poor Dietary Intake duct obstruction, which can reduce exocrine function. Matsu-
Anorexia-cachexia syndrome, nausea, food fear of postprandial pain moto and Traverso (2006) observed that the majority of patients
Posprandial abdominal bloating, cramping, diarrhea undergoing a pancreaticoduodenectomy (PD) for pancreatic
head disease had significantly decreased exocrine secretion
Malabsorption/Maldigestion
postoperatively, as measured by decreased fecal elastase 1.
Exocrine insufficiency, cholestasis, intraluminal bile deficiency, fat
malabsorption
Significant PEI leads to steatorrhea, which presents with
Vitamin deficiency greasy, foul-smelling stools, abdominal cramping, bloating,
Endocrine insufficiency, diabetes, insulin insensitivity excess flatus, and diarrhea. Mild PEI may not be associated
with any symptoms and therefore is unrecognized as a cause of
Increased Catabolism maldigestion, malabsorption, and weight loss. Because PEI is
Acute inflammation, infection, tumor effects, sedentary lifestyle so common in patients with pancreatic cancer, an argument can
be made to provide pancreatic enzymes to all symptomatic
Surgical Effects patients at the time of diagnosis. Asymptomatic patients, on the
Ileus, obstruction, pancreatic fistula, delayed gastric emptying, other hand, should be evaluated for PEI by stool analysis to
chylous ascites
look for fat droplets (Congo red stain) and/or a fecal elastase
Drug Therapy Effects
1 assay. The latter test is the easiest and most convenient
Nausea, vomiting, anorexia, diarrhea, obstipation, mucositis way to diagnose PEI-induced steatorrhea (Friess et al, 2009).
The nutritional consequences of untreated or unrecognized
steatorrhea can lead to significant weight loss, malnutrition, of vitamin A, D, E, and K either orally or through a feeding
vitamin deficiencies, poor quality of life, and delays in therapy tube.
(Klapdor et al, 2012). In one study, one third of patients with
chronic pancreatitis admitted for medical rehabilitation had a Pain
body mass index (BMI) less than 20 kg/m2, and one fourth had Patients with pancreatic disease often experience severe
profound steatorrhea (Armbrecht, 2001). Deficits in fat-soluble abdominal pain because of involvement of the celiac plexus by
vitamins (A, D, E, and K) are common in patients with both tumor, pancreatic duct obstruction, and/or pancreatitis. When
jaundice and PEI. Multiple studies have reported that patients abdominal pain is exacerbated by oral intake, patients will
with chronic pancreatitis and pancreatic cancer are also vitamin avoid food intake or modify their diet, which contributes to a
D deficient (Cho et al, 2013; Klapdor et al, 2012). Further- malnourished state (Rasmussen et al, 2013). In many cases,
more, patients with serum 25-hydroxyvitamin D-25(OH)D patients unconsciously avoid certain foods, such as those con-
levels less than 20 ng/mL have a worse prognosis when com- taining fat, which during time leads to weight loss. Patients
pared with those with levels greater than 20 ng/mL (Cho et al, taking large doses of narcotics to control their pain may also
2013). have alterations in GI motility and obstipation that also lead
to reduced food intake.
Biliary Obstruction
Biliary obstruction in patients with pancreatic head disease can Cancer-Related Effects on Metabolism
be associated with nutritional deficits, which are corrected once Direct adverse effects of cancers on nutrition and metabolism
obstruction is relieved either surgically or by nonoperative have been known to exist for years but are poorly understood.
means, such as endoscopic stenting or percutaneous drainage The mechanism(s) by which pancreatic cancer induces cachexia
(see Chapter 8). Patients with recent-onset obstructive jaundice are largely unknown but believed to be modulated, in part,
who are not profoundly malnourished and candidates for major through activation of proinflammatory mediators that stimulate
abdominal operation do not necessarily require preoperative systemic inflammation, the magnitude of which is associated
biliary decompression (PBD). In fact, PBD is to be discouraged with poor prognosis (Fearon et al, 2006; McMillan et al, 2009,
in such patients because it substantially increases the incidence 2013) (see Chapter 11). Protein malabsorption coupled with
of postoperative infectious complications (Mezhir et al, 2009). increased catabolism leads to weight loss and impaired immune
On the other hand, patients with long-standing obstructive surveillance by decreasing lymphocyte counts, ILs, helper T
jaundice associated with profound malnutrition will most likely cells, and T-cell blastogenesis (Kanda et al, 2011). Peripancre-
benefit from a period of nutritional repletion before undergoing atic tissues and tumors also stimulate the release of TNF-α,
major abdominal surgery. This is true for patients undergoing which plays a pivotal role in cachexia and catabolism (Todorov
any type of HPB operation. Such patients are best managed et al, 1996).
with endoscopic PBD as part of a comprehensive nutritional A combination of factors—exocrine insufficiency, tumor
repletion program. Unfortunately, endoscopic drainage is effects, pain, alcohol use—leading to malnutrition in patients
sometimes unsuccessful and requires percutaneous transhe- with pancreatic disease has a detrimental effect on patient out-
patic biliary decompression (PTBD). When this is the case, a comes and treatment. The degree of malnutrition, as measured
rendezvous procedure should be performed to convert the by weight alone, is often more than 25% of preillness weight
PTBD to an internal endoscopic stent either immediately or as (Wigmore et al, 1997). Others have observed that more than
soon as the patient is medically stable so that bile flows back 80% of patients are cachectic at diagnosis; this proportion
into the duodenum (see Chapters 29 and 30). Bile refeeding increases while the disease progresses (Tisdale, 2003) and is
into the proximal small bowel is an important issue, because associated with poor prognosis (Bachmann et al, 2008). These
prolonged PBD that occurs with either complete biliary obstruc- metabolic and nutritional responses negatively impact treat-
tion or with biliary disconnection from a bile duct injury will ment plans and can often delay resection, chemotherapy, and
lead to metabolic derangement and malnutrition because exces- lead to progressive deterioration in performance status.
sive losses of bile can lead to dehydration, metabolic acidosis,
progressive loss of biliary protein, and malabsorption. Refeed- Malnutrition and Pancreatic Resection
ing bile into the small bowel improves the digestion and absorp- Pancreatic resection is associated with a relatively high inci-
tion of fat and reduces the volume of hepatic bile output dence of perioperative morbidity (see Chapters 27 and 66).
through the enterohepatic recirculation of bile salts. Most While strategies have evolved to minimize and effectively
patients can tolerate bolus infusion of bile into their small bowel manage the typical complications following pancreatic resec-
of 150 mL or less every 4 hours, and they can be taught to do tion, such as pancreatic fistula and infection, optimization of
this on their own at home. But if a patient has a feeding jeju- perioperative nutritional status has remained a challenging
nostomy or gastrojejunostomy tube, it is preferable to provide endeavor. Postoperative complications following pancreatic
bile refeeding in a continuous manner. An alternative to enteric operations, particularly PD, frequently exacerbate a preopera-
bile refeeding is to provide oral bile salts in the form of urso- tive chronic malabsorption state and weight loss (Sanford et al,
deoxycholic acid (300 mg, four times daily) to form micelles 2014). Furthermore, neoadjuvant chemotherapy can worsen an
for fat absorption (Tripathy et al, 1996). already marginal preoperative nutritional state and subject an
Malnourished patients with complete biliary obstruction or already malnourished patient to additional risk at the time of
disconnection may require specialized semielemental diets to operation (Attar et al, 2012; Bozzetti et al, 2009). Some have
avoid fat malabsorption. Diets rich in medium-chain triglycer- reported that preventing weight loss during neoadjuvant che-
ides (MCTs) are preferred over diets containing long-chain motherapy is associated with better overall survival (Naumann
triglycerides. These patients are also at risk for fat-soluble et al, 2013). This fact underscores the importance of screening
vitamin deficiency and should be offered water-soluble forms and identifying patients with PC who are at nutritional risk and
taking steps to restore and prevent weight loss before operation 2013). Patients with chylous leaks are at increased risk for
(Sanford et al, 2014). postoperative malnutrition not only because low-fat diets
The indications for pancreatic resection are evolving to provide fewer calories and are nonpalatable, but also because
include more elderly patients with benign and premalignant of protein losses stemming from externally drained chyle. After
diseases such as intraductal papillary mucinous neoplasms. In dietary modification, chyle leaks are usually self limited and
one review, close to 50% of patients undergoing PD in a geri- resolve without the need for further surgical intervention.
atric population did so for benign disease (Sanford et al, 2014). Chylous ascites is significantly associated with early enteral
Because patients are living longer following PD, the necessity feeding following PD (Kuboki et al, 2013). Even though there
to properly manage postoperative nutritional side effects is of is no further morbidity or mortality once conservative manage-
utmost importance. This is illustrated by the observation that ment is initiated, chylous ascites is associated with prolonged
1 in 3 patients identified with nutritional risk after a PD for abdominal drainage and decreased likelihood of hospital dis-
benign disease were dead within 5 years, compared with only charge within 10 days postoperatively (Hilal et al, 2013).
1 in 14 patients without nutritional risk (Sanford et al, 2014).
Recent trials have shown that as many as 88% of preoperative Delayed Gastric Emptying
patients are at medium to high risk of malnutrition (La Torre The most frequent complication following PD is DGE, which
et al, 2013). These data underscore the importance of identify- significantly delays the resumption and tolerance of adequate
ing patients at nutritional risk so that nutritional repletion strat- oral nutrition and exacerbates any preoperative malnourished
egies can be initiated before operative intervention. state. DGE significantly prolongs hospitalization and is associ-
ated with increased postoperative morbidity and mortality
Nutritional Implications of Pancreatic Resection (Hu et al, 2014). The exact etiology of DGE following PD is
Malnourished patients undergoing HPB surgery have increased incompletely understood and likely multifactorial. One cause is
risk for intraabdominal and systemic complications and thought to be related to disruption of the migrating motor
decreased survival (Windsor, 1993). Nutritional deficits lead to complex that originates from the gastric pacemaker. A meta-
impaired wound healing, increased susceptibility to infections, analysis demonstrated that the most likely cause of DGE was
and increased bleeding as a result of capillary wall fragility and a postoperative complication (Hu et al, 2014), in particular
impaired clotting (Kanda et al, 2011). Preoperative malnutri- postoperative pancreatic fistulas and/or abdominal fluid collec-
tion has been shown to be an independent predictor of morbid- tions (Hashimoto et al, 2010; Liu et al, 2014). Patients with
ity in patients undergoing PD and results in a significantly significant DGE often require prolonged nasogastric decom-
higher rate of complications and fourfold longer hospitalization pression and may not have return of normal gastric emptying
(La Torre et al, 2013). Severely malnourished patients under- for weeks or months. When this happens (on average in 20%
going distal pancreatectomy have a 6.0 to 8.8 times increased of patients), patients are best served with a nasoenteric tube
risk for postoperative complications (Sierzega et al, 2007). It is placed distal to the gastrojejunal anastomosis alone or with
intuitive to think that correction of nutritional deficits will lead concomitant percutaneous gastric decompression when persis-
to improved outcomes or mitigate risk. Unfortunately, there is tent emesis occurs. Studies have reported that routine postop-
a paucity of good prospective clinical trial data to support this erative EN both exacerbates (Martignoni et al, 2000) and
contention. One prospective clinical trial showed that preopera- alleviates DGE (Zhu et al, 2013). Many surgeons have used a
tive nutritional support for 7 days decreased complications in variety of mechanisms and techniques to counteract this com-
malnourished patients identified to have high preoperative plication, including preoperative NS, intraoperative placement
nutritional risk who subsequently underwent abdominal surgery of jejunostomy feeding tubes, or postoperative placement of
(Jie et al, 2012). Hence, early nutritional assessment has impli- nasojejunal (NJ) feeding tubes. Zhu demonstrated significant
cations for reducing complications and overall hospital stay decrease in DGE after initiation of early EN via an NJ tube
(Kruizenga et al, 2005). feed and hypothesized that the mechanical effects of the NJ
Like GI surgery, pancreatic resection is fraught with the tube and its presence across the gastrojejunostomy helped
typical complications of infection and ileus. However, a PD or gastric and intestinal motility (Zhu et al, 2013). Others have
distal pancreatectomy also harbors specific subsets of complica- argued that NJ feeding tubes should be placed only if patients
tions that are unique to pancreatic surgery, including pancreatic do not resume adequate oral intake by postoperative day 7.
fistula, delayed gastric emptying (DGE), and chylous ascites. More than one meta-analysis however, failed to identify any
Although these high rates of complications are likely the result particular operative technique or process of care that reduces
of a variety of factors, the interplay between them and nutri- the incidence of DGE (Hu et al, 2014; Wu et al, 2014).
tional status (both preoperatively and postoperatively) must be
appreciated and addressed (Zhu et al, 2013). Pancreatic Fistula
The most dreaded complication of a pancreatic resection is
Chylous Ascites pancreatic fistula (PF). Although there are a multitude of
The incidence of chyle leak after a pancreatic resection occurs factors contributing to PF, including surgical technique, gland
in as many as 11% of operations (Hilal et al, 2013). Chyle leaks texture, pancreatic duct size, and stent placement (Kanda et al,
are treated with simple dietary modification, which includes 2011), its incidence and severity is still associated with malnu-
transition to a low-fat or no-fat diet or a diet enriched with trition. Nutritional risk was the only independent risk factor
MCTs, which shifts the burden of fat digestion away from the identified by Sierzega et al (2007) to be associated with a PF
lymphatic plexus. Some surgeons prefer the institution of PN for patients following distal pancreatectomy. Others have
and administration of somatostatin analogues to further assist reported that patients who are overweight (BMI >25 kg/m2)
with treatment and hasten resolution of the fistula, although actually tended to have increased rates of PF after distal
the evidence to support this practice is weak (Kuboki et al, pancreatectomy (Sledzianowski, 2005). This observation is
supported by Kanda and colleagues (2011), who found that jaundice, and altered stool pattern, which may indicate fat
patients with low BMI tended to have infrequent rates of PF. malabsorption and steatorrhea. There are a variety of ways to
The authors hypothesized that weight loss is related to the assess preoperative nutritional status.
amount of peripancreatic fat and that the resultant changes to
pancreatic gland texture as a result of malnutrition may con- Anthropometric Measurements
tribute to the eventual fistula rates in distal resections. On the Anthropometric tests have the benefit of objectivity, rapidity,
other hand, Sanford and colleagues (2014) reported that and ease of interpretation. A variety of anthropometric mea-
patients older than 65 years with a high Nutritional Risk Score surements have been tested with respect to degree and risk of
(NRS) were at risk for postoperative PF following a PD, malnutrition, such as mid arm circumference and triceps skin-
whereas those younger than 65 years were not. This latter study fold thickness (Durnin & Womersley, 1974). Although these
suggests that it is not just nutritional risk but rather an interac- modes of assessment are cheap and easy to perform, they are
tion between age and nutritional risk that predisposes to PF. rarely, if ever, used for NA in patients undergoing HPB opera-
More studies are necessary to better define the relationship tions. Others have used surrogates of muscle wasting, such as
between nutritional status and PF. hand-grip strength, to assess protein-energy malnutrition
(Alvares-da-Silva et al, 2005). These assessments have largely
IDENTIFICATION OF PATIENTS AT RISK FOR been demonstrated to be incompletely characteristic of nutri-
POSTOPERATIVE COMPLICATIONS tional state, although they can be used in conjunction with
other scoring systems to provide a clinical picture of
Great efforts have been made to develop scoring systems or nutrition.
methods of nutritional assessment (NA) that identify patients
at risk for perioperative morbidity and mortality HPB opera- Bioelectrical Impedance
tions. The ideal test for preoperative NA should incorporate Changes in body cell mass can serve as a marker for nutritional
measures for quantifying the severity of malnutrition and deple- status and measured using bioelectrical impedance (BEI)
tion of lean body mass as well as estimating the magnitude of (Pirlich et al, 2000). BEI directly measures the impedance to
metabolic stress that would be induced by surgical interven- an alternating current within the body. Different body compart-
tions (Schiesser et al, 2009). ments, including body cell mass, body fat, and lean body mass,
NA should begin preoperatively and continue throughout a can be separately measured and used to define a patient’s
patient’s course of treatment. A thorough NA includes: (1) nutritional state. BEI has been shown to reliably predict sur-
clinical and dietary history; (2) physical examination with vival and surgical morbidity in patients undergoing GI opera-
assessment of muscle mass and strength; (3) evaluation of tions (Schiesser et al, 2009).
serum albumin, prealbumin, and C-reactive protein; (4) mea-
surement of vitamin and mineral deficits; and (5) determination Sarcopenia
of nutrient requirements (Box 26.3). Historic questions should BMI is widely used in patient risk models. However, BMI fails
focus on the patient’s nutritional intake and habits, the degree to account for the diversity of body composition; patients of
and rate of weight loss during the previous month, and 6 the same BMI may differ significantly in composition of lean
months (Windsor, 1993), use of alcohol, length of time with muscle and fat mass. A relative dearth of lean muscle mass,
called sarcopenia, is associated with functional impairments,
physical disability, perioperative complications, prolonged hos-
BOX 26.3 Nutritional Assessment pital length of stay following abdominal operations, and poorer
Clinical Evaluation long-term outcomes in cancer patients (Harimoto et al, 2013;
Dietary and medical history Janssen et al, 2002; Lieffers et al, 2012; Tan et al, 2009; van
Physical activity Vledder et al, 2012). The term sarcopenia was originally used
Weight loss during past 6 months to describe age-related declines in muscle mass (Rosenberg,
1989) but is increasingly recognized as an independent risk
Physical Examination factor for poor survival in patients with cancer regardless of
Body weight their age. Sarcopenia should not be confused with cachexia,
Skin changes which is marked by weight loss and muscle wasting over a rela-
Ascites tively short time period. Sarcopenia develops during a long
Muscle mass and strength
period of time and, in the elderly, is not associated with weight
Objective Evaluation loss (Walrand et al, 2011). Thus many patients who fall into a
Bioelectical impendance analysis normal range for weight and BMI may have unrecognized
Dual-energy x-ray absorptiometry sarcopenia and be at risk for perioperative complications and
Computed tomography muscle mass analysis poor outcomes. Many obese patients may not appear to be
malnourished but have significant occult depletion of their skel-
Protein Synthetic Function etal muscle mass, putting them at particularly high risk for
Albumin, transferrin postoperative complications because excess weight coupled
Prealbumin, prothrombin time with reduced muscle mass leads to impaired mobility, poor
C-reactive protein respiratory function, and prolonged rehabilitation (Lieffers
et al, 2012). Patients who are both obese and sarcopenic are
Vitamin, Mineral, Trace Element Deficits
Vitamins A, D, E, K, B6, niacin, folate, B12 now categorized with “sarcopenic obesity.” Sarcopenic obesity
Thiamine, magnesium, potassium, zinc, iron, phosphorus has been reported as one of the most powerful independent
predictors of poor survival for patients with cancer and is
associated with impaired functional status, decreased ability to obtainable through simple blood tests. The rapidity and ease
tolerate chemotherapy, and surgery and other invasive therapies of measurement have led to the use of serum markers in predic-
(Martin et al, 2014; Prado et al, 2008). Because the prevalence tive models to facilitate rapid surgical judgment and therapeutic
of obesity continues to rise, recognition of occult sarcopenia by decisions.
diagnostic imaging is an increasingly important method that
should be used to identify patients at nutritional risk. Albumin
The pathophysiology of sarcopenia is not well understood. Serum albumin has been extensively studied as a marker for
It is unclear if significant loss of muscle mass is a natural and nutritional status, and low levels have been shown to be a sensi-
immutable consequence of aging in some individuals or if ther- tive predictor of adverse surgical outcomes (Billingsley et al,
apeutic or lifestyle interventions can slow or reverse it (Walrand 2003; Gibbs et al, 1999; Lai et al, 2011). Gibbs analyzed more
et al, 2011). Age-related sarcopenia is thought to result from than 50,000 surgical patients and demonstrated a significant
factors related to muscle cell senescence and extrinsic factors increase in surgical morbidity and mortality (especially sepsis
such as decreases in growth hormone and insulin-like growth and infections) with worsening hypoalbuminemia (Gibbs et al,
factor-1 (Kalyani et al, 2014). Sarcopenia results in preferential 1999). In HPB surgery, serum albumin has been studied as a
atrophy of type II skeletal muscle fibers, resulting in a relative marker of perioperative risk. Used in conjunction with other
loss of muscle strength over endurance. Cancer, autoimmune markers of malnutrition, albumin was found to be predictive of
diseases, human immunodeficiency virus infection, cirrhosis, mortality after PD (Venkat, 2011). In liver surgery, serum
and endocrine dysfunction are among the many chronic disease albumin is harder to interpret as a nutritional marker in patients
states that can lead to rapid muscle loss. Poor nutrition, altera- with cirrhosis, given its correlation with intrinsic liver function
tions in hormonal and other signaling pathways, as well as (Piquet et al, 2006). Likewise, with liver transplant, some have
inflammatory factors and cytokines are thought to be the prin- suggested a correlation between preoperative serum albumin
cipal mechanisms behind the development of sarcopenia in the and posttransplant complications (Fusai et al, 2006). However,
chronically ill (Kalyani et al, 2014; McMillan et al, 2009). serum albumin should not be used as a sole prognosticator
Several techniques have been used to identify the presence of nutritional status, given its relationship as an acute-phase
and magnitude of sarcopenia. The most clinically applicable reactant.
method uses computed tomography (CT) imaging to estimate
lean muscle mass (Janssen et al, 2002; Mourtzakis et al, 2008). Prealbumin
The method estimates lean muscle mass by measuring the area Also known as transthyretin, prealbumin is synthesized by the
of the psoas muscle of a single CT image at the L3 level nor- liver and serves as a transport protein. Unlike albumin, which
malized for height (total psoas area [mm2]/height [m2]). Patients has a half-life of approximately 20 days, prealbumin is rapidly
are sarcopenic if these values are less than 385 mm2/m2 in processed and is thought to be a better indicator of nutritional
women or less than 545 mm2/m2 in men. In retrospective status, given its half-life of 48 hours. As such, short-term fluc-
studies based on CT imaging, sarcopenic patients who under- tuations of nutritional standing can be more accurately assessed
went liver transplantation, liver resection of primary liver using serum prealbumin rather than albumin (Beck et al, 2002).
cancers, or colorectal metastases and pancreas resection had Unfortunately, like albumin, prealbumin is also an acute-phase
increased perioperative complications, increased postoperative reactant, and its measurement can vary with chronic disease
mortality, and worse overall and recurrence-free survival states. Huang et al (2012) studied the outcomes of liver resec-
(Englesbe et al, 2010; Harimoto et al, 2013; Itoh et al, 2014; tion in Child-Pugh Class A patients and found that serum pre-
Peng et al, 2012; Valero et al, 2015; van Vledder et al, 2012; albumin was significantly correlated with postoperative liver
Voron et al, 2015). A study in patients with liver transplants failure. Whether this relationship was the result of nutritional
found that serum albumin and Model for End-Stage Liver depletion or intrinsic liver disease remains unknown.
Disease (MELD) score correlated poorly with sarcopenia,
underscoring the fact that traditional measures of poor nutrition Nutritional Scoring Systems
may underestimate risk in these patients (Englesbe et al, 2010). Given the limitations of anthropometric and biochemical
Some studies have identified sarcopenia as an independent assays, more accurate methods of NA have incorporated key
risk factor for mortality in patients with cirrhosis (Meza-Junco features of both to create predictive nutritional scoring systems.
et al, 2013; Montano-Loza et al, 2011; Montano-Loza, 2014). Although a number of different schemas exist, the more fre-
In one study of cirrhotics, 70% of those who underwent tran- quently used include the Nutritional Risk Index (NRI), NRS,
sjugular intrahepatic portosystemic stenting had reversal of their Subjective Global Assessment (SGA), Malnutrition Universal
sarcopenia, which correlated with a lower mortality compared Screening Tool (MUST), and Glasgow Prognostic Score
with those whose sarcopenia did not improve or worsened (GPS).
(9.8% vs. 43.5%) (Tsien et al, 2013) (see Chapter 79). Others
have reported that exercise combined with supplemental protein Nutritional Risk Index
or amino acids can reverse sarcopenia in the elderly (Evans et al, Developed by the Veteran’s Affair Total Parenteral Nutrition
2014; Yang et al, 2012). These observations provide support for Group, the NRI (Box 26.4) calculates a risk score based on
the practice of prescribing exercise along with supplemental serum albumin and the ratio of present to usual weight. Its use
protein in patients identified to be sarcopenic prior to subjecting as a scoring system to reflect nutritional risk and predict adverse
them to HPB operations (Evans et al, 2014). outcomes has been validated (Cohendy et al 1999). The NRI
is simple to use and can define a high-risk subgroup of patients
Serum Biochemical Markers with obstructive jaundice considered for operation. In one
Although no single laboratory value is by itself indicative of study, patients with obstructive jaundice and a NRI less than
nutritional insufficiency, many have the benefit of being easily 83.5 were found to have a significant increased risk for
TABLE 26.1 Nutritional Risk Score

Impaired Nutritional Status Severity of Disease (≈Increase in Requirements)
Absent: Score 0 Normal nutritional status Absent: Score 0 Normal nutritional requirements
Mild: Score 1 Weight loss >5% in 3 mo or food intake Mild: Score 1 Hip fracture,* chronic patients, in
below 50%-85% of normal requirement particular with acute complications:
in preceding week cirrhosis, COPD,* chronic
hemodialysis, diabetes, oncology
Moderate: Score 2 Weight loss >5% in 2 mo or BMI 18.5-20.5 Moderate: Score Major abdominal surgery,* stroke,*
+ impaired general condition or food 2 severe pneumonia, hematologic
intake 25%-60% of normal requirement malignancy
in preceding week
Severe: Score 3 Weight loss >5% in 1 mo (>15% in 3 mo) Severe: Score 3 Head injury,* bone marrow
or BMI <18.5% + impaired general transplantation* Intensive care
condition or food intake 0%-25% of patients (APACHE >10)
normal requirement in preceding week
Score: + Score: = Total Score
Age: If ≥70 years, add 1 to total score above: = age-adjusted total score
Score ≥3: The patient is nutritionally at risk, and a nutritional care plan is initiated.
Score <3: Weekly rescreening of the patient. If the patient, e.g., is scheduled for a major operation, a preventive nutritional care plan is
considered to avoid the associated risk status.
*indicates that a trial directly supports the categorization of patients with that diagnosis. From Kondrup J, et  al: Nutritional risk screening (NRS 2002): a new method based on an analysis of controlled
clinical trials, Clin Nutr 22(3):321. e36, 2003.
APACHE, Acute Physiology and Chronic Health Evaluation; BMI, body mass index; COPD, chronic obstructive pulmonary disease.
were present: BMI greater than 18.5 kg/m2, albumin less than

BOX 26.4 Nutritional Risk Index
3.0 g/dL, or mean weight loss greater than 10% of total body
Nutritional risk index = (1.519 × serum albumin [g/dL]) + (41.7 × weight in the preceding 6 months (Sanford et al, 2014). In
present weight [kg]/usual body weight* [kg]) geriatric patients, severe nutritional risk was independently
>100: Not malnourished
associated with significantly poorer survival on both univariate
97.5-100: Mild malnourishment
and multivariate analysis. Using these guidelines, geriatric
83.5-97.5: Moderate malnourishment
<83.5: Severe malnourishment patients with severe nutritional risk who underwent PD for
benign or premalignant indications had significantly decreased
*Usual weight is defined as the stable weight 6 mo or more before illness. overall 5-year survival compared with their nongeriatric coun-
terparts (65% vs. 93%, P < .001) (Sanford et al, 2014). The
authors concluded that patients with benign disease and low
postoperative death and longer duration of hospital stay but not NRS should be offered nonoperative therapy or no therapy
an increased complication rate (Clugston et al, 2006). Shinkawa until their NRS improves to a level that predicts a more favor-
and colleagues (2013) reviewed a series of more than 60 able outcome following operation.
patients undergoing PD and demonstrated the NRI to be an
independent risk factor for surgical-site infection on multivari- Subjective Global Assessment
ate analysis. The SGA analyzes patient-reported facets of nutritional health
to provide a NA (Detsky et al, 1987). The elements used in
Nutritional Risk Score the SGA include changes in dietary intake, weight loss, func-
First developed as a screening methodology by the European tional capacity, signs of muscle wasting, lower extremity edema,
Society of Parenteral and Enteral Nutrition in 2002, the NRS and GI symptoms. Severity of malnutrition is scored by a
(Table 26.1) encompasses patient age, food intake in the week trained examiner and used to classify perioperative risk. The
before surgery, weight loss, BMI, and severity of underlying SGA is a widely used assessment tool and has the advantage of
disease (Kondrup et al, 2003). The NRS has proved to be reli- not being influenced by fluid retention or ascites (Gunsar et al,
able and reproducible in retrospective analysis of more than 2006). However, the SGA has the disadvantage of subjectivity
100 controlled nutrition trials (Schiesser et al, 2008). When and interobserver variability, which is related to the memory of
analyzed prospectively in a cohort of more than 600 patients the patient and training of the examiner. When used in combi-
undergoing GI surgery, the NRS maintained an ability to iden- nation with other NA tools or biochemical markers of nutrition,
tify those deemed as malnourished with significantly more com- the SGA was able to predict patients who experienced postop-
plications as well as longer hospital stays (Schiesser et al, 2008). erative pancreatic surgery complications ranging from 82% to
The NRS has been adapted to preoperatively stratify geriatric 90% of instances (La Torre et al, 2013).
patients by the American College of Surgeons National Surgical
Quality Improvement Program. Preoperative screening is rec- Malnutrition Universal Screening Tool
ommended for geriatric and cancer patients using serum The MUST score is highly consistent with SGA scores to
albumin, BMI, and weight loss (Chow et al, 2012). Severe identify patients at nutritional risk (Loh et al, 2012). The
nutritional risk was shown to increase the risk of perioperative MUST is calculated by using BMI, weight loss, and an acute
morbidity and mortality when any of the following features illness component lasting longer than 5 days, during which the
BMI score Weight loss score Acute disease effect score
Unplanned weight
BMI kg/m2 Score loss in past 3-6 months If patient is acutely ill and
>20 (>30 obese) =0 there has been or is likely to
% Score be no nutritional intake
18.5-20 =1
<18.5 =2 <5 =0 for >5 days
5-10 =1 Score 2
>10 =2
If unable to obtain height and weight, see Acute disease effect is unlikely to
reverse for alternative measurements and apply outside hospital. See “MUST”
use of subjective criteria Step 4 Explanatory Booklet for further
Overall risk of malnutrition information
Add scores together to calculate overall risk of malnutrition

Score 0: low risk Score 1: medium risk Score 2 or more: medium risk
FIGURE 26.1. Malnutrition Universal Screening Tool (MUST). BMI, Body mass index.
BOX 26.5 Glasgow Prognostic Score BOX 26.6 General Guidelines for Postoperative Nutritional
CRP level (>1.0 mg/dL) and hypoalbuminemia (<3.5 g/dL) = 2 points Support
CRP level (>1.0 mg/dL) or hypoalbuminemia (<3.5 g/dL) = 1 point Oral route > enteral > PICC > TPN
CRP level (<1.0 mg/dL) and normoalbuminemia (>3.5 g/dL) = 0 point kcal: 25-30 kcal/kg/day
Nitrogen: 1-1.5 g/kg/day
CRP, C-reactive protein.
Glucose: ≤5 mg/kg/min
Fat calories: ≤30% of total calories
patient is likely to have no nutritional intake (Fig. 26.1). Unlike PICC, Percutaneous indwelling central catheter; TPN, total parenteral nutrition.
NRI or SGA, the MUST is simple to use, can be performed
rapidly (usually <5 minutes), and is less expensive (Tu et al,
2012). When compared with SGA, MUST has good predictive BOX 26.7 Management of Perioperative Nutrition Complications
ability for perioperative mortality, hospital cost, and length of Management of Perioperative Nutritional Complications
stay (Kyle et al, 2006). Loh and colleagues (2012) demon- Anorexia: appetite stimulants
strated that three quarters of patients had a high calculated risk Chylous ascites: low-fat diet/medium-chain triglycerides
of malnutrition despite normal BMI because of recent weight Delayed gastric emptying: frequent, small meals; jejunostomy tube
loss. This study emphasizes the importance of unintentional feeds; TPN
weight loss, which was found to be the only significant predictor Pancreatic insufficiency: pancreatic enzymes
of adverse outcomes. La Torre et al (2013) reported a cohort Intraluminal bile deficiency: bile refeeding, oral ursodeoxycholic acid
of 143 patients who underwent pancreatic surgery for malig- TPN, total parenteral nutrition.
nancy and used MUST for nutritional screening. They identi-
fied that patients deemed to have high risk of malnutrition by
MUST had a fivefold higher surgical-site infection rate, four- procedure. An analysis of more than 120 patients with advanced
fold longer postoperative hospital course, and a morbidity rate PC demonstrated a significant association between median
of 53% compared with patients at low nutritional risk. Malnu- overall survival and GPS; patients scoring 2 points had an
trition defined by MUST was an independent predictor of overall survival of 1.8 months compared with 8.3 months for
overall morbidity. patients with 0 points (Martin et al, 2014). Although the exact
interplay between nutrition and systemic inflammation as mea-
Glasgow Prognostic Score sured by GPS is incompletely understood, it appears to be a
The GPS (Box 26.5) was created to quantify the magnitude of good measure of outcome in patients undergoing HPB surgery.
the systemic inflammatory response seen with malignancy and
was introduced initially as a marker of survival. The GPS clas- NUTRITIONAL SUPPORT OF
sifies patients with 0, 1, or 2 points depending on the presence PATIENTS UNDERGOING
of hypoalbuminemia (<35 g/L) or elevated C-reactive protein
(>10 mg/L) (Forrest et al, 2003). The GPS has been shown to
HEPATOPANCREATOBILIARY SURGERY
be predictive of poor survival in patients with cancer and is The primary goal for NS in patients before and after HPB
strongly associated with malnutrition, increased weight loss, surgery is to restore health and function as quickly as possible
increased perioperative morbidity, and poor performance status (Boxes 26.6 and 26.7). There are three principal ways to
(McMillan et al, 2013). Glen and colleagues (2006) demon- achieve this goal. The first is to enable patients to resume oral
strated the GPS to be a significant predictor of survival intake as soon as possible; this is facilitated by the use of
(independent of tumor-nodes-metastasis stage) for those enhanced recovery after surgery (ERAS) protocols. The second
patients with inoperable PC who underwent a palliative bypass is to support and/or restore normal digestion and intestinal
absorption. The third is to identify nutritional deficits preop- Heidegger et al, 2007). In a few situations, specially formulated
eratively and to correct them. No data support the routine use diets may be beneficial for patients with hepatobiliary disease
of NS in well-nourished patients undergoing HPB surgery. In or PEI. For example, patients with diarrhea caused by malab-
fact, prospective, randomized trials have shown that the admin- sorption of fat may have improved tolerance to feeds with a diet
istration of PN to well-nourished patients undergoing surgery enriched with MCTs, which do not require micelle formation
is associated with a higher incidence of postoperative complica- for absorption or processing. In theory, dampening the proin-
tions (Brennan et al, 1994; Buzby, 1991). Conversely, patients flammatory response to illness and operation by administration
who are profoundly malnourished or deficient in specific vita- of water-soluble fatty acids alone or in combination with high
mins probably benefit from preoperative and postoperative NS doses of antioxidants has the potential to improve patient out-
(Buzby, 1991; Evans et al 2014; Halliday et al, 1988). comes (Braga et al, 2012; Ok et al, 2003; Persson et al, 2005).
The majority of patients with HPB disease who are malnour- Clinical evidence to support this theory is beginning to emerge,
ished do not require specialized NS to correct their nutritional but additional study is necessary to identify which patients are
deficits because, in most circumstances, their malnutrition is most likely to benefit (Wu et al, 2012).
the consequence of inadequate caloric intake during prolonged
periods of time, rather than maldigestion and malabsorption or Parenteral Nutrition
specific vitamin deficiency. Correction of malnutrition in these In the rare event when a patient fails to meet their total nutri-
patients only requires supplying sufficient quantity of appropri- tional needs by combined oral and enteral routes, the addition
ate substrate (fat, protein, and carbohydrate) (Klein, et al, of PN is appropriate. When this is the case, the initiation of PN
1997). There are no consensus guidelines supporting a certain should wait until at least 7 to 10 days after surgery to avoid
length of time of nutritional repletion to achieve a certain level adverse consequences (Casaer et al, 2011; Heidegger et al,
of risk reduction of operative complications. Experts recom- 2007; Marik, 2011; Martindale et al, 2009). The increased
mend at least 7 days and preferably 2 to 3 weeks of oral nutri- infectious complications and costs associated with PN mandate
tional repletion in patients who have profound malnutrition or that it is used only in patients with anatomic abnormalities of
until such time that the serum prealbumin rises to a normal the GI tract in whom EN is not feasible or fails (Braunschweig
range (Evans et al, 2014; Jie et al, 2012; Klein et al, 1999). et al, 2001; Heyland et al, 2003; Kreymann et al, 2006; Mar-
tindale et al, 2009; Meier et al, 2006; Plauth et al, 2006;
Routes of Feeding Weimann et al, 2006). High-quality studies have not found
There are several ways to provide patients NS preoperatively routine PN to be beneficial to patients undergoing HPB opera-
and postoperatively. Most common and physiologic is an oral tions in the first 2 weeks postoperatively (Bengmark, 2012).
diet tailored to the patient’s needs. If the patient is unable to Prolonged PN can cause hepatocellular steatosis and choles-
consume an oral diet or meet their estimated caloric needs tatic liver dysfunction (Bauer and Kiehntopf, 2014). In pancre-
orally, a nasogastric or nasoduodenal feeding tube should be atic surgery, routine postoperative PN after PD is associated
placed, even in the presence of known esophageal varices. If with increased morbidity (Brennan et al, 1994; Gianotti et al,
the patient is unable or unwilling to accept a nasoenteric feeding 2000). It is important that when PN is used, patients must not
tube, placing a feeding jejunostomy or gastrostomy tube using be given excess calories (e.g., they should receive <30 kcal/kg/
laparoscopic, interventional radiologic, or percutaneous endo- day), and glycemic control (serum glucose 100 to 150) should
scopic approaches should be considered. Profoundly malnour- be maintained with insulin.
ished patients with malabsorption who experience diarrhea in
response to enteral feeding can receive partial EN with the Enteral and Parenteral Nutrition
balance of calories, protein, fat, trace elements, and vitamins Some clinicians have tried a combination of EN and PN to
administered by the peripheral IV route until their intestinal more quickly meet patients’ estimated total caloric needs. Zhu
digestive and absorptive capacity adapts and allows them to a nd colleagues (2013) reported their outcomes comparing
tolerate higher amounts of EN (Kondrup et al, 2003). post-PD patients who were administered as combination EN
The physiologic beneficial effects of EN over PN are main- and PN with those who only received PN in the postoperative
tenance of intestinal immunity and gut integrity, prevention of period. Compared with the PN-only group, the combination
intestinal microbial translocation, reduced postoperative infec- group had a significant decrease in complications, hospital stay,
tions and sepsis, and less expense (Hilal et al, 2013; Kuboki and infections. Moreover, PD-specific complications such as
et al, 2013; Zhu et al, 2013). However, jejunal feeding can DGE and hyperglycemia were also reduced.
contribute to and/or exacerbate post-PD DGE. Martignoni and
colleagues (2000) showed a significant increase in nasogastric
Nutritional Support as Part of Enhanced Recovery after
intubation and prolonged hospital stay with postoperative Surgery Programs
jejunal EN. Perioperative feeding regimens have changed substantially
An enteral formula should be chosen based on the patient’s during the last decade with the advent of ERAS (Kehlet, 1997).
ability to digest and absorb nutrients. For most patients, a ERAS is a multimodal and multidisciplinary management
standard enteral diet is sufficient. The general guidelines for pathway whose goal is to more quickly restore health and func-
feeding listed in Box 26.6 are sufficient for most patients. There tion to patients by mitigating the catabolic stress response to
is no clinical evidence to support the administration of super- operation. A central tenet of ERAS is to limit preoperative
physiologic amounts of substrate (e.g., more than 5 mg/kg/min fasting and restore oral intake as soon as possible postopera-
of glucose or more than 2 g protein/kg/day. Conversely, there tively. Current ERAS guidelines recommend (grade A evi-
is evidence to suggest that “overfeeding” malnourished patients, dence) the use of preoperative oral carbohydrate supplementation
particular with carbohydrates and/or with diets rich in omega-6 2 to 3 hours before anesthesia (Lassen et al, 2012; Viganò et al,
fatty acids may be associated with harm (Buzby et al, 1988; 2012; Weimann, et al, 2006). This practice has been shown to
reduce psychological and physiologic stress, improve glycemic surgery or transplantation (Evans et al, 2014; Langer et al,
control during and after surgery, reduce skeletal muscle prote- 2012). Nevertheless, many believe that there is value to the
olysis and weakness, restore bowel motility, and reduce DGE administration of oral IMN and supplements to all patients
after PD (Viganò et al, 2012). The feasibility, safety, and utility undergoing high-risk surgery (Evans et al, 2014; Martindale
of ERAS in patients undergoing HPB operations has recently et al, 2009), and many centers using ERAS protocols incorpo-
been established and leads to reduced morbidity and shorter rate IMN as part of a multimodal perioperative care plan
hospital length of stay (Balzano et al, 2008; Braga et al; 2014; (Burden et al, 2012).
Coolsen et al 2013; Gerritsen A et al, 2014; Hendry et al 2010;
Hughes et al, 2014; Jones et al, 2013; Kagedan et al, 2015; Vitamins
Lassen et al, 2012; Revie et al 2012; Wong-Lun-Hing et al, Vitamins are considered to be essential components of periop-
2014). Although some ERAS programs have included the erative nutrition. Vitamin D, for instance, has been noted to be
administration of an immune-enhancing dietary supplement for deficient in three quarters of patients with pancreatic cancer at
5 to 7 days preoperatively, the value of such dietary supple- diagnosis, and this deficiency is associated with poor survival
ments to patients undergoing HPB operations within the (Cho et al, 2013). A 2005 placebo-controlled double-blind
context of an ERAS program have not been established. study of 47 patients who underwent partial liver resection dem-
The majority of patients managed within an ERAS protocol onstrated that the group receiving infusions containing vitamin
who do not experience a major complication, have rapid return E had a significantly shorter length of stay in the intensive care
of bowel function and tolerate ingestion of a normal diet within unit (ICU) and showed improvement in aspartate aminotrans-
24 to 48 hours of surgery independent of the type of operation ferase, alanine aminotransferase, and glutamic dehydrogenase
(Hughes et al, 2014; Kagedan et al, 2014, Revie et al, 2012; van (Bartels et al, 2004). Vitamin C is clinically important in peri-
Dam et al, 2008). Conversely, patients who fall off of an ERAS operative healing (Cevikel et al, 2008). Patients who are active
pathway, (e.g., do not tolerate an oral diet within a few days of smokers have significant reductions in total body and circulat-
surgery) usually have an underlying major complication, such as ing vitamin C and should receive preoperative supplementa-
liver failure, intraabdominal collection, fistula, or systemic infection. In the perioperative period, vitamin C concentrations are
tion. In a systematic review of the literature, Kagedan et al significantly reduced and can fall even further with prolonged
(2015) reported on 10 ERAS studies in patients undergoing PD. hospitalization and complications (Fukushima et al, 2010).
Nasogastric tubes were discontinued either at the end of opera- This decrease in vitamin C can adversely affect other compo-
tion or on the first postoperative day. In the majority of studies, nents of perioperative recovery, such as muscle weakness,
patients were allowed access to and tolerated solid food on the fatigue, and physical performance (Lukaski, 2004). Although
first or second postoperative day. A systematic review of five randomized clinical trials have yet to show that preoperative
feeding routes after PD concluded that an oral diet is the pre- administration of vitamins A, D, E, and C reduce perioperative
ferred route of feeding and that there is no evidence to support morbidity, it makes physiologic sense to do so and is safe (Evans
routine EN or PN after PD (Gerritsen et al, 2013). Lassen et al et al, 2014). When considering the fact that vitamins are cheap,
(2012) observed that it was safe to allow patients undergoing safe, and readily available, and that patients undergoing HPB
major upper GI surgery, including all types of HPB surgery, to operations are at risk for oxygen radical–mediated injury, it is
eat normal food at will beginning on the first day after surgery. reasonable to consider administration of vitamins E and C for
ERAS regimens in patients undergoing liver surgery routinely several days preoperatively in patients at risk for postoperative
allow patients a regular diet at will beginning the night of surgery complications.
(Coolsen et al, 2012; Dunne et al, 2014; Hughes et al, 2014;
Jones et al, 2013; van Dam et al, 2008). These studies demon- Probiotics and Synbiotics
strate that patients undergoing PD or liver resection do not need A recent area of interest within NS science is the use of both
prolonged gastric decompression, nor do they have to wait for prebiotics and probiotics (synbiotics) in combination with EN
return of bowel function to tolerate an oral diet. However, there to enhance GI immunity. A prospective, randomized, double-
are patients who will not tolerate early oral feeding. Unfortu- blind trial treated postoperative liver transplant recipients with
nately, there is no reliable way to identify these patients. Never- EN and a combination of lactic acid bacteria and fiber, and it
theless, until new clinical studies demonstrate that an early reduced bacterial infection rates from 48% to 3% following
postoperative oral diet is harmful, patients should be allowed to liver transplantation (Rayes et al, 2005). This resulted in
eat normal food at will shortly after major HPB operations. shorter ICU and overall hospital stays and duration of antibiotic
treatment. Zhang et al (2013) retrospectively reviewed 67 post
Immune-Modulating Nutrition liver transplant recipients who took probiotics and fiber and
The potential benefit of administering immune-modulating showed a significant reduction in infection rate from 30% to
nutrition (IMN) and antioxidants to patients undergoing HPB 9%, compared with controls. Sugawara and colleagues (2006)
surgery is controversial, although of increasing interest (Evans randomly assigned more than 100 patients who underwent
et al, 2014). Several meta-analyses concluded that specialized hepatic resection for biliary cancers to preoperative and post-
oral supplements enriched with ω3FAs, arginine, and gluta- operative synbiotics or standard postoperative supplements and
mine given orally for 7 to 14 days preoperatively reduced the found that the symbiotic group had a lower incidence of post-
risk of postoperative complications, particularly infection, fol- operative infectious complications (12% vs. 30%).
lowing upper GI surgery (Braga et al, 2013; Burden et al, 2012;
Marik et al, 2010, Marimuthu et al, 2012). ω3FAs adminis- Appetite Stimulants
tered parenterally preoperatively have also been shown to One strategy to improve perioperative malnutrition has used
decrease levels of IL-6 and TNF in patients undergoing hepa- dietary stimulants. The most common approach is through the
tectomy for HCC (Wu et al, 2012). Other reviews have reported administration of oral megestrol acetate (MA), a derivative of
no conclusive evidence to support the use of IMN in GI tract progesterone. The mechanism by which MA increases appetite
is unknown but is hypothesized to inhibit the action of TNF

and other proinflammatory cytokines. A Cochrane Review ana- Cirrhosis and Hepatic Failure
lyzing more than 1000 studies concluded that MA increases Hospitalized patients with cirrhosis (see Chapters 79 and 80)
appetite in approximately 25% of patients, results in weight have substantial morbidity and mortality (Rice et al, 1997).
gain in approximately 10%, but does not improve quality of Because malnutrition is a major problem in many of these
life. Coupled with the increased risk of blood clots, fluid reten- patients, NS should be undertaken when surgery is being con-
tion, and death, some authors are cautious about the routine templated. Prospective, randomized clinical trials have shown
use of MA (Ruiz Garcia et al, 2013). that these patients do not meet their nutritional needs with
Tetrahydrocannabinol (THC), the main psychoactive sub- aggressive daily dietary counseling and have reduced hospital
stance found in the Cannabis sativa plant, has also been inves- mortality and morbidity when fed via nasogastric or duodenal
tigated as an appetite stimulant. Synthetic THC is known as tubes (Cabré et al, 1990). Hence it is recommended that all
dronabinol and is available as a prescription medication as hospitalized cirrhotic patients undergoing emergency operation
Marinol. The exact mechanism of action is unknown, but it is be fed via the enteral route unless otherwise contraindicated.
thought that cannaboid receptors in the hypothalamus are Use of PN in these patients is not indicated if the GI tract is
involved in the motivational and reward aspects of increased functional (Braga et al, 2009; Braunschweig et al, 2001;
nutritional intake (Fride et al, 2005). In cancer patients, the Heyland DK et al; 2003; Kreymann et al, 2006; Martindale
effect of THC on appetite and weight maintenance have been RB et al, 2009).
mixed. A multicenter randomized double-blind trial in 243 Because the metabolic derangement of advanced liver
cancer patients showed no statistically significant appetite disease affects all three macronutrients, any dietary intervention
improvement using Cannabis (Strasser et al, 2006). Additional needs to supplement all three. The caloric needs of patients
studies are necessary to establish a beneficial role of cannaboids with liver disease are often increased (35 to 40 kcal/kg body
in patients with hepatobiliary cancers. weight/day), particularly if they have ascites (McCullough &
Tavill, 1991). Protein needs depend on the patient’s current
nutritional status, ability to tolerate protein, and history or
DISEASE- AND CONDITION-SPECIFIC presence of encephalopathy. In a nonencephalopathic patient,
MANAGEMENT 1 to 1.5 g protein/kg dry weight is recommended. If encepha-
lopathy is present, a thorough diet history should be taken to
Nonalcoholic Fatty Liver Disease and determine the tolerable amount of protein with recommenda-
Nonalcoholic Steatohepatitis tions to supply at least a minimum 0.5 to 0.7 g/kg dry weight
The obesity epidemic is rapidly contributing to the develop- per day. Fluid and sodium restriction is often necessary,
ment of nonalcoholic fatty liver disease (NAFLD) and nonal- depending on the severity of ascites and edema. A level of at
coholic steatohepatitis (NASH) (see Chapter 71). Their least 2 g of sodium per day is recommended because lower
incidence is growing and affect more than 6% of patients in the levels are unpalatable and usually decrease oral dietary intake
West (Torres et al, 2012), although approximately 30% of further, exacerbating the patient’s malnutrition. Treatment of
Western patients have excess fat accumulation within the liver, intractable ascites should rely more on aggressive diuretics and
which increases to 80% in the morbidly obese (Zelber-Sagi paracentesis rather than drastic limits on fluid intake and salt
et al, 2011). NAFLD and NASH are intimately associated with restriction. Patients with cirrhosis and those with hepatic failure
risk factors such as diabetes mellitus and obesity, and are commonly have micronutrient deficiencies of phosphate, mag-
known to progress to liver cirrhosis and HCC (Nishikawa et al, nesium, and zinc; these should be monitored and aggressively
2013). Increased fat content within the liver has been shown to replaced as necessary by the IV route (Shronts, 1988).
impair postoperative recovery. Animal models of fatty liver The oral administration of branched-chain amino acids
disease have shown decreased liver regenerative capacity and (BCAAs) to patients with decompensated liver disease improves
DNA synthesis after hepatectomy, resulting in increased mor- event-free survival, serum albumin concentration, and quality
bidity and mortality (Dai et al, 2006). Hoppe et al (2015) of life (Bianchi et al, 2005; Marchesini et al, 2003; Muto et al,
demonstrated impaired liver function and recovery in a cohort 2005). Changes in MELD and Child-Turcotte-Pugh scores
of posthepatectomy patients with NAFLD and NASH com- were smaller in a BCAA-administered group than in a control
pared with normal controls. Compared with their counterparts group, and serum total bilirubin and serum albumin were better
with viral hepatitis B, patients with fatty liver disease who preserved. The incidence of cirrhotic-related complications was
underwent hepatectomy for HCC had significantly higher mor- also reduced in those receiving BCAA (Kawamura et al, 2009).
bidity (59% vs. 30%) and mortality (12% vs. 1.4%) (Wakai BCAAs may also reduce inflammation in hepatitis C–positive
et al, 2011). patients with cirrhosis, reduce the production of oxidative
The hepatotoxic effects of intrahepatic fat in NAFLD and stress, and possibly decrease the occurrence of HCC (Ohno
NASH are reversible with weight loss. Taitano (2015) comet al, 2008). A Cochrane Review concluded there was no
pared preoperative and postoperative liver biopsy specimens in benefit from BCAA administration on postoperative morbidity,
morbidly obese patients who underwent bariatric surgery and mortality, or infection in patients undergoing HPB surgery
found that steatosis and lobular inflammation resolved in 75% (Koretz et al, 2012).
of patients, and steatohepatitis resolved in greater than 90%.
These observations highlight the importance of diet, nutrition, Liver Resection
and weight loss in obese patients, not only in the prevention of Most patients in Western countries undergoing liver resection
liver disease but also on surgical outcomes. Clearly, patients (see Chapters 103 and 108) do not have cirrhosis and no need
being considered for liver resection who are suspected or proven for specialized NS preoperatively or postoperatively. Most can
to have NAFLD or NASH should undergo dietary counseling take solids on the evening of surgery, be carbohydrate loaded
and medically supervised weight loss. per ERAS regimens, and begin an oral diet safely on the first
postoperative day. Almost all patients are able to tolerate a full much difficulty or patient discomfort. The routine use of PN
regular diet within 5 days. Malnourished patients being consid- in malnourished patients undergoing liver transplantation
ered for elective liver resection should receive nutritional reple- (Reilly et  al, 1990) is no longer considered appropriate
tion preoperatively and postoperatively by the oral and/or because the majority of patients can be fed EN successfully
enteral route. and safely by a nasoduodenal or NJ tube within hours after
It is strongly recommended that patients not receive hyper- transplantation (Mehta et  al, 1995; Sekido et  al, 2003). In
tonic glucose infusions in the immediate (<6 hours) postopera- several studies, early initiation of EN after operation (<24
tive period because this can interfere with the liver’s preferential hours) was associated with a postoperative infection rate that
use of fatty acids for energy and hepatic regeneration. If NS is was lower than expected (Hasse et  al, 1995; Martin et  al,
desired within hours after liver resection, and it is anticipated 1993; Mehta et  al, 1995; Wicks et  al, 1994). In a prospective,
that the patient will not be able to take an oral diet (e.g., in the randomized study reported by Hasse and colleagues (1995),
event they are intubated), then intraoperative placement of a patients receiving EN had no viral infections in the first 2
nasoduodenal or jejunal tube should be considered. postoperative weeks. Bacterial and overall infection rates also
Several prospective randomized clinical trials have reported were decreased in the EN group (14% vs. 29%, and 21% vs.
that the administration of PN enriched with BCAAs is of 41%, respectively). Postoperative liver transplant patients fed
benefit to patients undergoing liver resection. Fan and col- immediately via the enteral route also have a decreased inci-
leagues (1994) showed a benefit of preoperative and postopera- dence of postoperative ileus (5% vs. 28%; P < .01) and started
tive BCAA-enriched PN in cirrhotic patients undergoing an oral diet much sooner (8 days vs. 12 days; P = .07) than
elective liver resection for HCC. The PN-supplemented group patients given PN (Mehta et  al, 1995). BCAA supplementa-
had reduced septic complications compared with controls tion has been shown to improve metabolic and nutritional
(17% vs. 37%), less deterioration of liver function, and less abnormalities in the early posttransplant period by increasing
ascites. Postoperative mortality in the PN group was also 45% albumin synthesis and improving liver function (Yoshida et  al,
less than in the control group. In two other studies, Okabayashi 2012).
and colleagues (2008, 2011) reported that oral supplementa-
tion with carbohydrate- and BCAA-enriched nutrients reduced Pancreatic Cancer
postoperative complications and improved postoperative The most common presenting conditions of patients with pan-
quality of life in patients with chronic liver disease undergoing creatic head cancers (see Chapter 62) are PEI and steatorrhea.
hepatic resection for HCC. There are no data to support the Sikkens et al (2014) reported the prevalence of PEI was 66%
use of BCAAs in patients without liver disease who undergo in a cohort of pancreas cancer patients, which increased to 92%
liver resection. after 2 months. Fortunately, this nutritional deficiency can be
easily corrected with pancreatic enzyme replacement therapy,
Liver Transplantation which should be taken with every meal and snack (Domínguez-
Clinical and experimental observations show that nutrition Muñoz, 2011; Sikkens et al, 2010). Even in cases of surgical
plays an extremely important role in just about all aspects of or medical futility, relief of exocrine insufficiency can signifi-
liver transplantation (see Chapter 112). The status of the cantly palliate symptoms and improve quality of life. Pancreatic
donor, the function of the transplanted liver, the overall health head cancers can also lead to biliary obstruction. As neoadju-
of the recipient, the rejection process, the absorption and vant chemotherapy is more frequently used as part of a multi-
metabolism of immunosuppressive drugs, and a variety of disciplinary treatment approach for pancreatic cancer, enteric
metabolic pathways are influenced by nutrients and how they flow of bile is preferable and can be achieved with use of endo-
are administered. Clinical observations show that malnour- scopic stenting or percutaneous drainage and refeeding. In fact,
ished patients undergoing liver transplantation have a signifi- many chemotherapeutic agents cannot be administered safely
cantly poorer outcome than well-nourished patients (Porayko in the setting of biliary obstruction.
et al, 1991). They have an incidence of infectious morbidity
and 3-month mortality rates twice that of well-nourished SUMMARY
patients undergoing liver transplant (Harrison et al, 1997).
Pretransplant nutritional status and disease severity are inde- Careful attention must be paid to the HPB patient’s overall
pendently associated with infectious episodes during hospital- clinical condition, inadequacy of diet, degree of malnutrition,
ization (Meril et al, 2009). Degree of malnutrition also metabolic derangements, and impaired digestion and absorp-
correlates significantly with days spent in the ICU, days spent tion to determine the appropriate level and type of NS. Inap-
mechanically ventilated, and length of hospital stay, and also propriate nutritional support, particularly the provision of
correlates significantly with mortality after transplantation excess calories via PN and poor glycemic control, are poten-
(Meril et al, 2009; Pikul et al, 1994). tially harmful. Increasing evidence shows that specific nutrients
Patients awaiting transplantation who are malnourished administered under specific conditions to select patients can
should receive aggressive nutritional counseling and support modulate metabolic events in patients with hepatobiliary
(Cabré & Gassull, 2001; Hasse et  al, 1995; Mehta et  al, 1995). disease undergoing operation. Before prescribing nutritional
Nutritional supplements, including vitamins, should be taken therapy to a patient, a clinician or dietician with experience and
in an effort to attenuate further nutritional deficits and to expertise in liver metabolism should be consulted to evaluate
replenish losses. Patients who ingest inadequate amounts of the patient, provide nutritional recommendations, and monitor
calories should be fed EN via a soft nasal feeding tube. In the patient’s course.
contrast to what most people and physicians believe, these
tubes are well tolerated and can be used for months without References are available at expertconsult.com.
Chapter 26 Preoperative and postoperative nutrition in hepatobiliary surgery 458.e1
REFERENCES Bulger E, et al: Enteral vitamin E supplementation inhibits the cytokine

response to endotoxin, Arch Surg 132:1337–1341, 1997.
Achord J: Malnutrition and the role of nutritional support in alcoholic Bulger E, Helton W: Nutrient antioxidants in gastrointestinal disease,
liver support, Am J Gastroenterol 82:1–7, 1987. Gastroenterol Clin North Am 27:403–419, 1998.
Alvares-da-Silva MR, Reverbel da Silveira T: Comparison between Burden S, et al: Pre-operative nutrition support in patients undergoing
handgrip strength, subjective global assessment, and prognostic gastrointestinal surgery, Cochrane Database Syst Rev (11):CD008879,
nutritional index in assessing malnutrition and predicting clinical 2012.
outcome in cirrhotic outpatients, Nutrition 21(2):113–117, 2005. Burra P, et al: Hepatic malondialdehyde and glutathione in end stage
Andus T, et al: Effects of cytokines on the liver, Hepatology 13:364–375, chronic liver disease [abstract], Hepatology 16(4 Pt 2):266, 1992.
1991. Buzby G, et al: Prognostic nutritional index in gastrointestinal surgery,
Armbrecht U: Chronic pancreatitis: weight loss and poor physical Am J Surg 139:160–167, 1988.
performance–experience from a specialized rehabilitation centre, Buzby GP: Veterans Affairs Total Parenteral Nutrition Cooperative
Rehabilitation (Stuttg) 40:332–336, 2001. Study Group: perioperative total parenteral nutrition in surgical
Astarcioglu I, et al: High levels of glycogen in the donor liver improve patients, N Engl J Med 325:525–532, 1991.
survival after liver transplantation in rats, Transplant Proc 23:2465– Cabré E, Gassull MA: Nutritional aspects of liver disease and trans-
2466, 1991. plantation, Curr Opin Clin Nutr Metab Care 4:581–589, 2001.
Attar A, et al: Malnutrition is high and underestimated during chemo- Cabré E, et al: Effect of total enteral nutrition on the short-term
therapy in gastrointestinal cancer: an AGEO prospective cross- outcome of severely malnourished cirrhotics, Gastroenterology
sectional multicenter study, Nutr Cancer 64(4):535–542, 2012. 98:715–720, 1990.
Bachmann J, et al: Cachexia worsens prognosis in patients with resect- Casaer MP, et al: Early versus late parenteral nutrition in critically ill
able pancreatic cancer, J Gastrointest Surg 12(7):1193–1201, 2008. adults, N Engl J Med 365(6):506–517, 2011.
Bakker OJ, et al: Dutch Pancreatitis Study Group: Early versus Cerwenka H, et al: Normothermic liver ischemia and antioxidant treat-
on-demand nasoenteric tube feeding in acute pancreatitis, N Engl J ment during hepatic resections, Free Radic Res 30:463–469, 1999.
Med 371(21):1983–1993, 2014. Cevikel MH, et al: Supplementation with high-dose ascorbic acid
Balzano G, et al: Fast-track recovery programme after pancreatico- improves intestinal anastomotic healing, Eur Surg Res 40:29–33,
duodenectomy reduces delayed gastric emptying, Br J Surg 95(11): 2008.
1387–1393, 2008. Cho M, et al: Vitamin D deficiency and prognostics among patients
Bartels M, et al: Pilot study on the effect of parenteral vitamin E on with pancreatic adenocarcinoma, J Transl Med 11:206, 2013.
ischemia and reperfusion-induced liver injury: a double-blind, ran- Chow WB, et al: ACS NSQIP/AGS Best Practice Guidelines: Optimal
domized, placebo-controlled trial, Clin Nutr 23(6):1360–1370, Preoperative Assessment of the Geriatric Surgical Patient 2012. Available
2004. at http://site.acsnsqip.org/wp-content/uploads/2011/12/ACS-NSQIP
Bauer M, Kiehntopf M: Shades of yellow: monitoring nutritional needs -AGS-Geriatric-2012-Guidelines.pdf.
and hepatobiliary function in the critically ill, Hepatology 60(1):26– Clugston A, et al: Nutritional risk index predicts a high-risk population
29, 2014. in patients with obstructive jaundice, Clin Nutr 25(6):949–954,
Beck FK, Rosenthal TC: Prealbumin: a marker for nutritional evalua- 2006.
tion, Am Fam Physician 65:1575–1578, 2002. Cohendy R, et al: Preoperative nutritional evaluation of elderly patients:
Bell H, et al: Reduced concentration of hepatic alpha tocopherol in the Mini Nutritional Assessment as a practical tool, Clin Nutr
patients with alcoholic liver cirrhosis, Alcohol Alcohol 27:39–46, 18(6):345–348, 1999.
1992. Coolsen MM, et al: Systematic review and meta-analysis of enhanced
Bengmark S: Nutrition of the critically ill—emphasis on liver and pan- recovery after pancreatic surgery with particular emphasis on pan-
creas, Hepatobiliary Surg Nutr 1(1):25–52, 2012. creaticoduodenectomies, World J Surg 37(8):1909–1918, 2013.
Bianchi G, et al: Update on branched-chain amino acid supplementa- Curran R, et al: Multiple cytokines are required to induce hepatocyte
tion in liver diseases, Curr Opin Gastroenterol 21:19–20, 2005. nitric oxide production and inhibit protein synthesis, Ann Surg
Billiar T, et al: Fatty acid intake and Kupffer cell function: fish oil alters 212:460–471, 1990.
eicosanoid and monokine production to endotoxin stimulation, Cywes R, et al: Effect of intraportal glucose infusion on hepatic glyco-
Surgery 104:343–349, 1988. gen content and degradation and outcome of liver transplantation,
Billingsley KG, et al: Outcome after pancreaticoduodenectomy for Ann Surg 216:235–247, 1992.
periampullary cancer: an analysis from the Veterans Affairs National Dai XF, Chen DF: Liver regenerative capacity after partial hepatec-
Surgical Quality Improvement Program, J Gastrointest Surg 7:484– tomy in rats with nonalcoholic fatty liver disease, Zhonghua Gan Zang
491, 2003. Bing Za Zhi 14(8):597–601, 2006.
Birlouez-Aragon I, et al: Antioxidant vitamins and degenerative pathol- Detsky AS, et al: What is subjective global assessment of nutritional
ogies: a review of vitamin C, J Nutr Health Aging 7:103–109, 2003. status?, JPEN J Parenter Enteral Nutr 11(1):8–13, 1987.
Bozzetti F, SCRINIO Working Group: Screening the nutritional status Dolais-Kitabgi J, et al: Effect of insulin and glucagon on amino acid
in oncology: a preliminary report on 1,000 outpatients, Support Care transport in isolated hepatocytes after partial hepatectomy in the rat,
Cancer 17(3):279–284, 2009. Endocrinology 109:868–875, 1981.
Bradford BU, et al: New simple models to evaluate zone-specific Domínguez-Muñoz JE: Pancreatic exocrine insufficiency: diagnosis
damage due to hypoxia in the perfused rat liver: time course and and treatment, J Gastroenterol Hepatol 26(Suppl 2):12–16, 2011.
effect of nutritional state, J Pharmacol Exp Ther 236:262–268, 1986. Dunne DF, et al: Enhanced recovery in the resection of colorectal liver
Braga M, et al: Oral preoperative antioxidants in pancreatic surgery: a metastases, J Surg Oncol 110(2):197–202, 2014.
double-blind, randomized, clinical trial, Nutrition 28:160–164, 2012. Durnin JV, Womersley J: Body fat assessed from total body density
Braga M, et al: Enhanced recovery after surgery pathway in patients and its estimation from skinfold thickness: measurements on 481
undergoing pancreaticoduodenectomy, World J Surg 38:2960–2966, men and women aged from 16 to 72 years, Br J Nutr 32(1):77–97,
2014. 1974.
Braga M, et al: Clinical evidence for pharmaconutrition in major elec- Eigler N, et al: Synergistic interactions of physiologic increments of
tive surgery, JPEN J Parenter Enteral Nutr 37(Suppl 5):66S–72S, glucagon, epinephrine and cortisol in the dog: a model for stress-
2013. induced hyperglycemia, J Clin Invest 63:114–123, 1979.
Braunschweig CL, et al: Enteral compared with parenteral nutrition: a Englesbe MJ, et al: Sarcopenia and mortality after liver transplantation,
meta-analysis, Am J Clin Nutr 74:534–542, 2001. J Am Coll Surg 211(2):271–278, 2010.
Brennan M, et al: A prospective randomized trial of total parenteral Eum HA, et al: Trolox C ameliorates hepatic drug metabolizing dys-
nutrition after major pancreatic resection for malignancy, Ann Surg function after ischemia/reperfusion, Arch Pharm Res 25:940–945,
220:436–444, 1994. 2002.
Bruera E, et al: Cachexia and asthenia in cancer patients, Lancet Oncol Evans DC, et al: Nutrition optimization prior to surgery, Nutr Clin
1:138–147, 2000. Pract 29(1):10–21, 2014.
Bruix J, Sherman M; American Association for the Study of Liver Fan S, et al: Perioperative nutritional support in patients undergoing
Diseases: Management of hepatocellular carcinoma: an update, hepatectomy for hepatocellular carcinoma, N Engl J Med 331:1547–
Hepatology 53(3):1020–1022, 2011. 1552, 1994.
Fearon KC, et al; Cancer Cachexia Study Group: Definition of cancer Hu HL, et al: Factors influencing delayed gastric emptying after
cachexia: effect of weight loss, reduced food intake, and systemic pancreaticoduodenectomy—a meta-analysis, Hepatogastroenterology
inflammation on functional status and prognosis, Am J Clin Nutr 61(134):1539–1545, 2014.
83(6):1345–1350, 2006. Huang L, et al: Prealbumin is predictive for postoperative liver insuf-
Fong Y, et al: Total parenteral nutrition and bowel rest modify the ficiency in patients undergoing liver resection, World J Gastroenterol
metabolic response to endotoxin in humans, Ann Surg 210:449–457, 18(47):7021–7025, 2012.
1989. Hughes MJ, et al: Enhanced recovery following liver surgery: a system-
Forrest LM, et al: Evaluation of cumulative prognostic scores based on atic review and meta-analysis, HPB (Oxford) 16:699–706, 2014.
the systemic inflammatory response in patients with inoperable non- Inui A: Cancer anorexia-cachexia syndrome: current issues in research
small-cell lung cancer, Br J Cancer 89:1028–1030, 2003. and management, CA Cancer J Clin 52:72–91, 2002.
Foschi D, et al: Hyperalimentation of jaundiced patients on percutane- Itoh S, et al: Effect of body composition on outcomes after hepatic
ous transhepatic biliary drainage, Br J Surg 73(9):716–719, 1986. resection for hepatocellular carcinoma, Ann Surg Oncol 21(9):3063–
Fowler F, et al: Characterization of sodium-dependent amino acid 3068, 2014.
transport activity during liver regeneration, Hepatology 16:1187– Janssen I, et al: Low relative skeletal muscle mass (sarcopenia) in older
1194, 1992. persons is associated with functional impairment and physical dis-
Fride E, et al: Endocannabinoids and food intake: newborn suckling ability, J Am Geriatr Soc 50:889–896, 2002.
and appetite regulation in adulthood, Exp Biol Med (Maywood) Jie B, et al: Impact of preoperative nutritional support on clinical
230(4):225–234, 2005. outcome in abdominal surgical patients at nutritional risk, Nutrition
Friess H, Michalski CW: HPB (Oxford) 11(Suppl 3):3–10, 2009. 28(10):1022–1027, 2012.
Fukushima R, Yamazaki E: Vitamin C requirement in surgical patients, John J, et al: Seventy percent hepatectomy: comparative effects of par-
Curr Opin Clin Nutr Metab Care 13(6):669–676, 2010. enteral and enteral nutrition on liver regeneration, morbidity, and
Fusai G, et al: Incidence and risk factors for the development of pro- mortality, Surg Forum 43:124–125, 1992.
longed and severe intrahepatic cholestasis after liver transplantation, Jones C, et al: Randomized clinical trial on enhanced recovery versus
Liver Transpl 12(11):1626–1633, 2006. standard care following open liver resection, Br J Surg 100:1015–
Gerritsen A, et al: Early oral feeding after pancreatoduodenectomy 1024, 2013.
enhances recovery without increasing morbidity, HPB (Oxford) Kagedan DJ, et al: Enhanced recovery after pancreatic surgery: a sys-
16:656–664, 2014. tematic review of the evidence, HPB (Oxford) 17:11–16, 2015.
Gianotti L, et al: Artificial nutrition after pancreaticoduodenectomy, Kalyani RR, et al: Age-related and disease-related muscle loss: the
Pancreas 21:344–351, 2000. effect of diabetes, obesity, and other diseases, Lancet Diabetes Endo-
Gibbs J, et al: Preoperative serum albumin level as a predictor of opera- crinol 2(10):819–829, 2014.
tive mortality and morbidity: results from the National VA Surgical Kanda M, et al: Nutritional predictors of postoperative outcome in
Risk Study, Arch Surg 134(1):36–42, 1999. pancreatic cancer, Br J Surg 98(2):268–274, 2011.
Glen P, et al: Evaluation of an inflammation-based prognostic score in Katz N: Metabolism of lipids. In Thurman R, et al, editors: Regulation
patients with inoperable pancreatic cancer, Pancreatology 6(5):450– of hepatic metabolism: intra- and intercellular compartmentalization,
453, 2006. New York, 1986, Plenum Press, pp 237–252.
Gondolesi GE, et al: Reduction of ischemia-reperfusion injury in Kawamura E, et al: A randomized pilot trial of oral branched-
parenchymal and nonparenchymal liver cells by donor treatment chain amino acids in early cirrhosis: validation using prognostic
with DL-alpha-tocopherol prior to organ harvest, Transplant Proc markers for pre-liver transplant status, Liver Transpl 15(7):790–797,
34:1086–1091, 2002. 2009.
Goode J, et al: Reperfusion injury, antioxidants and hemodynamics Kehlet H: Multimodal approach to control post operative pathophysi-
during orthotopic liver transplantation, Hepatology 19:354–359, ology and rehabilitation, Br J Anesthesth 78:606–617, 1997.
1994. Klapdor S, et al: Vitamin D status and per-oral vitamin D supplemen-
Gunsar F, et al: Nutritional status and prognosis in cirrhotic patients, tation in patients suffering from chronic pancreatitis and pancreatic
Aliment Pharmacol Ther 24(4):563–572, 2006. cancer disease, Anticancer Res 32(5):1991–1998, 2012.
Halliday A, et al: Nutritional risk factors in major hepatobiliary surgery, Klein S, et al: Nutrition support in clinical practice: review of published
JPEN J Parenter Enteral Nutr 12:43–48, 1988. data and recommendations for future research directions. National
Harimoto N, et al: Sarcopenia as a predictor of prognosis in patients Institutes of Health, American Society for Parenteral and Enteral
following hepatectomy for hepatocellular carcinoma, Br J Surg Nutrition, and American Society for Clinical Nutrition, JPEN J
100(11):1523–1530, 2013. Parenter Enteral Nutr 21:133–156, 1997.
Harrison J, et al: A prospective study on the effect of recipient nutri- Kondrup J, et al: Nutritional risk screening (NRS 2002): a new method
tional status on outcome in liver transplantation, Transpl Int 10:369– based on an analysis of controlled clinical trials, Clin Nutr 22(3):321.
374, 1997. e36, 2003.
Hashimoto Y, Traverso LW: Incidence of pancreatic anastomotic Koretz RL, et al: Nutritional support for liver disease, Cochrane Data-
failure and delayed gastric emptying after pancreatoduodenectomy base Syst Rev (5):CD008344, 2012.
in 507 consecutive patients: use of a web-based calculator to improve Kreymann KG, et al; DGEM (German Society for Nutritional Medi-
homogeneity of definition, Surgery 147(4):503–515, 2010. cine); ESPEN (European Society for Parenteral and Enteral Nutri-
Hasse J, et al: Early enteral nutrition support in patients undergoing tion): ESPEN Guidelines on Enteral Nutrition: Intensive care, Clin
liver transplantation, JPEN J Parenter Enteral Nutr 19:437–443, Nutr 25(2):210–223, 2006.
1995. Kruizenga HM, et al: Effectiveness and cost-effectiveness of early
Heidegger CP, et al: Is it now time to promote mixed enteral and screening and treatment of malnourished patients, Am J Clin Nutr
parenteral nutrition for the critically ill patient?, Intensive Care Med 82:1082–1089, 2005.
33:963–969, 2007. Kuboki S, et al: Chylous ascites after hepatopancreatobiliary surgery,
Helton W: Nutritional issues in hepatobiliary surgery, Semin Liver Dis Br J Surg 100(4):522–527, 2013.
14:386–391, 1994. Kyle UG, et al: Comparison of tools for nutritional assessment and
Henriksen J, et al: Sympathetic nervous activity in cirrhosis: a survey screening at hospital admission: a population study, Clin Nutr
of plasma catecholamine studies, J Hepatol 1:55–65, 1985. 25:409–417, 2006.
Heyland KD, et al: Canadian clinical practice guidelines for nutrition Lai CC, et al: Low preoperative serum albumin in colon cancer: a risk
support in mechanically ventilated, critically ill adult patients, JPEN factor for poor outcome, Int J Colorectal Dis 26:473–481, 2011.
J Parenter Enteral Nutr 27:355–373, 2003. Langer G, et al: Nutritional interventions for liver-transplanted
Hilal M, et al: Postoperative chyle leak after major pancreatic resections patients, Cochrane Database Syst Rev (8):CD007605, 2012.
in patients who receive enteral feed: risk factors and management Lassen K, et al; International Association for Surgical Metabolism and
options, World J Surg 37(12):2918–2926, 2013. Nutrition: Guidelines for perioperative care for pancreaticoduode-
Holroyde C, Reichard A: Carbohydrate metabolism in cancer cachexia, nectomy: Enhanced Recovery After Surgery (ERAS) Society recom-
Cancer Treat Rep 65(Suppl 5):55–59, 1981. mendations, Clin Nutr 31(6):817–830, 2012.
Hoppe S, et al: Nonalcoholic steatohepatits and liver steatosis modify La Torre M, et al: Malnutrition and pancreatic surgery: prevalence and
partial hepatectomy recovery, J Invest Surg 28(1):24–31, 2015. outcomes, J Surg Oncol 107(7):702–708, 2013.
Lee S, Clemens M: Effect of alpha-tocopherol on hepatic mixed func- Montano-Loza AJ, et al: Muscle wasting is associated with mortality
tion oxidases in hepatic ischemia/reperfusion, Hepatology 15:276– in patients with cirrhosis, Clin Gastroenterol Hepatol 10:166–173,
281, 1992. 2011.
Li NN, et al: Does intravenous fish oil benefit patients post-surgery? A Montano-Loza AJ: Muscle wasting: a nutritional criterion to prioritize
meta-analysis of randomized controlled trials, Clin Nutr 33:226–239, patients for liver transplantation, Curr Opin Clin Nutr Metab Care
2014. 17:219–225, 2014.
Lieffers JR, et al: Sarcopenia is associated with postoperative infection Mourtzakis M, et al: A practical and precise approach to quantification
and delayed recovery from colorectal cancer resection surgery, Br J of body composition in cancer patients using computed tomography
Cancer 107:931–936, 2012. images acquired during routine care, Appl Physiol Nutr Metab
Liu QY, et al: Risk factors of delayed gastric emptying following pan- 33:997–1006, 2008.
creaticoduodenectomy, ANZ J Surg 86(1–2):69–73, 2014. Muller M: Malnutrition in cirrhosis, J Hepatol 23(Suppl 1):31–35,
Loh KW, et al: Unintentional weight loss is the most important indica- 1995.
tor of malnutrition among surgical cancer patients, Neth J Med Mumtaz K, et al: Endoscopic retrograde cholangiopancreaticography
70(8):365–369, 2012. with or without stenting in patients with pancreaticobiliary malig-
Lukaski HC: Vitamin and mineral status: effects on physical perfor- nancy, prior to surgery, Cochrane Database Syst Rev (3):CD006001,
mance, Nutrition 20:632–644, 2004. 2007.
Maderazo E, et al: Additional evidence of auto-oxidation as a possible Muto Y, et al: Effects of oral branched-chain amino acid granules on
mechanism of neutrophil locomotory dysfunction in blunt trauma, eventfree survival in patients with liver cirrhosis, Clin Gastroenterol
Crit Care Med 18:141–147, 1990. Hepatol 3(7):705–713, 2005.
Maderazo E, et al: A randomized trial of replacement antioxidant Nakatoni T, et al: Changes in predominant energy substrate after hepa-
vitamin therapy for neutrophil locomotory dysfunction in blunt tectomy, Life Sci 28:257–264, 1981.
trauma, J Trauma 31:1142–1150, 1991. Naumann P, et al: Outcome after neoadjuvant chemoradiation and
Marchesini G, et al: Nutritional supplementation with branched-chain correlation with nutritional status in patients with locally advanced
amino acids in advanced cirrhosis: a double-blinded, randomized pancreatic cancer, Strahlenther Onkol 189(9):745–752, 2013.
trial, Gastroenterology 124:1792–1801, 2003. Niki E, et al: Interaction among vitamin C, vitamin E, and beta caro-
Marik PE: Death by total parenteral nutrition: the saga continues, Crit tene, Am J Clin Nutr 62(Suppl):1322S–1326S, 1995.
Care Med 39(6):1536–1537, 2011. Nishikawa H, Osaki Y: Non-B, non-C hepatocellular carcinoma
Marimuthu K, et al: A meta-analysis of the effect of combinations of [review], Int J Oncol 43(5):1333–1342, 2013.
immune modulating nutrients on outcome in patients undergoing Nolan J: Endotoxin, reticuloendothelial function, and liver injury,
major open gastrointestinal surgery, Ann Surg 255:1060–1068, 2012. Hepatology 1:458–465, 1981.
Martignoni ME, et al: Enteral nutrition prolongs delayed gastric emp- Ockenga J: Importance of nutritional management in diseases with
tying in patients after Whipple resection, Am J Surg 180:18–23, exocrine pancreatic insufficiency, HPB (Oxford) 11(Suppl 3):11–15,
2000. 2009.
Martin M, et al: The benefit of early enteral feeding in patients under- Ogoshi S, et al: Effect of a nucleoside-nucleotide mixture on protein
going liver transplantation [abstract], Hepatology 18(Pt 2):337A,1993. metabolism in rats after seventy percent hepatectomy, Nutrition
Martin HL, et al: Prognostic value of systemic inflammation-based 5:173–178, 1989.
markers in advanced pancreatic cancer, Intern Med J 44(7):676–682, Ohno T, et al: Suppressive effect of oral administration of branched-
2014. chain amino acid granules on oxidative stress and inflammation in
Martindale RG, et al; American College of Critical Care Medicine; HCV-positive patients with liver cirrhosis, Hepatol Res 38(7):683–
A.S.P.E.N. Board of Directors: Guidelines for the provision and 688, 2008.
assessment of nutrition support therapy in the adult critically ill Ok E, et al: Use of olive oil-based emulsions as an alternative to
patient: Society of Critical Care Medicine and American Society for soybean oil-based emulsions in total parenteral nutrition and their
Parenteral and Enteral Nutrition: executive summary, Crit Care Med effects on liver regeneration following hepatic resection on rats, Ann
37(5):1757–1761, 2009. Nutr Metab 47:221–227, 2003.
Marubayashi S, et al: Role of free radicals in ischemic rat liver cell Okabayashi T, et al: Effects of branched-chain amino acids-enriched
injury: prevention of damage by alpha-tocopherol administration, nutrient support for patients undergoing liver resection for hepato-
Surgery 99:184–192, 1986. cellular carcinoma, J Gastroenterol Hepatol 12:1869–1873, 2008.
Matsumoto J, Traverso LW: Exocrine function following the Whipple Okabayashi T, et al: Oral supplementation with carbohydrate- and
operation as assessed by stool elastase, J Gastrointest Surg 10:1225– branched-chain amino acid-enriched nutrients improves postopera-
1229, 2006. tive quality of life in patients undergoing hepatic resection, Amino
McClain C, et al: Antioxidants depress monocyte/Kupffer cell tumor Acids 40(4):1213–1220, 2011.
necrosis factor production, JPEN J Parenter Enteral Nutr 18:23S, O’Neill S, et al: Obstructive jaundice in rats resulting in exaggerated
1994. hepatic production of tumor necrosis factor-alpha and systemic and
McCullough A, Tavill A: Disordered energy and protein metabolism tissue necrosis factor-alpha levels after endotoxin, Surgery 122:281–
in liver disease, Semin Liver Dis 11:265–277, 1991. 287, 1997.
McMillan DC: Systemic inflammation, nutritional status and survival Ouchi K, et al: Hepatic resection under Pringle maneuver induces
in patients with cancer, Curr Opin Clin Nutr Metab Care 12:223–226, endotoxemia and lipoperoxidative attack in the jaundiced rat liver,
2009. Nippon Shokakibyo Gakkai Zasshi 88:1216–1220, 1991.
McMillan DC: The systemic inflammation-based Glasgow Prognostic Ozaki N, et al: Ketone body ratio as an indicator of early graft survival
Score: a decade of experience in patients with cancer, Cancer Treat in clinical liver transplantation, Clin Transplant 5:48–54, 1991.
Rev 39(5):534–540, 2013. Ozawa K: Perioperative care based on the redox theory. In Ozawa K,
Mehta P, et al: Nutritional support following liver transplantation: a editor: Liver surgery approached through the mitochondria: the redox
comparison of jejunal versus parenteral routes, Clin Transplant theory in evolution, Tokyo, 1992, Medical Tribune, pp 96–100.
5:364–369, 1995. Peng P, et al: Impact of sarcopenia on outcomes following resection of
Meier R, et al: ESPEN (European Society for Parenteral and Enteral pancreatic adenocarcinoma, J Gastrointest Surg 16:1478–1486,
Nutrition): ESPEN guidelines on enteral nutrition: pancreas, Clin 2012.
Nutr 25(2):275–284, 2006. Persson C, et al: Impact of fish oil in patients with advanced gastroin-
Meril M, et al: Nutritional status: its influence on the outcome of testinal cancer: a randomized pilot study, Nutrition 21:170–178,
patients undergoing liver transplantation, Liver Int 30:208–214, 2005.
2009. Petrides A, De Fronzo R: Glucose and insulin metabolism in cirrhosis,
Meza-Junco J, et al: Sarcopenia as a prognostic index of nutritional J Hepatol 8:107–114, 1989.
status in concurrent cirrhosis and hepatocellular carcinoma, J Clin Pikul J, et al: Degree of preoperative malnutrition is predictive of post-
Gastroenterol 47:861–870, 2013. operative morbidity and mortality in liver transplant recipients,
Mezhir JJ, et al: A matched case-control study of preoperative biliary Transplantation 57:469–472, 1994.
drainage in patients with pancreatic adenocarcinoma: routine drain- Piquet MA, et al: Nutritional indices in cirrhotic patients, Nutrition
age is not justified, J Gastrointest Surg 13:2163–2169, 2009. 22(2):216–217, 2006.
Pirlich M, et al: Bioelectrical impedance analysis is a useful bedside Cannabis-In-Cachexia-Study-Group, J Clin Oncol 24(21):3394–
technique to assess malnutrition in cirrhotic patients with and 3400, 2006.
without ascites, Hepatology 32(6):1208–1215, 2000. Sugawara G, et al: Perioperative synbiotic treatment to prevent post-
Plauth M, et al: ESPEN (European Society for Parenteral and Enteral operative infectious complications in biliary cancer surgery: a ran-
Nutrition): ESPEN guidelines on enteral nutrition: liver disease, Clin domized controlled trial, Ann Surg 244(5):706–714, 2006.
Nutr 25(2):285–294, 2006. Sugino K, et al: Changes in the levels of endogenous antioxidants in
Porayko M, et al: Impact of malnutrition and its therapy on liver trans- the liver of mice with experimental endotoxemia and the protective
plantation, Semin Liver Dis 11:305–314, 1991. effects of antioxidants, Surgery 105:200–206, 1989.
Powell R, et al: Effect of oxygen-free radical scavengers on survival in Sun Y, et al: Perioperative nutritional support and fluid therapy in
sepsis, Am J Surg 57:86–88, 1991. patients with liver diseases, Hepatobiliary Surg Nutr 3(3):140–148,
Prado CMM, et al: Prevalence and clinical implications of sarcoenic 2014.
obesity in patients with solid tumours of the respoiratory and gas- Taitano AA, et al: Bariatric surgery improves histological features of
trointestinal tracts: a population-based study, Lancet Oncol 9:629– nonalcoholic fatty liver disease and liver fibrosis, J Gastrointest Surg
635, 2008. 19(3):429–436, 2015.
Rasmussen HH, et al: Nutrition in chronic pancreatitis, World J Gas- Takada Y, et al: Arterial ketone body ratio and glucose administration
troenterol 19(42):7267–7275, 2013. as an energy substrate in relation to changes in ketone body concen-
Rayes N, et al: Supply of pre- and probiotics reduces bacterial infection tration after liver-related liver transplantation in children, Transplan-
rates after liver transplantation: a randomized, double-blind trial, Am tation 55:1314–1319, 1993.
J Transplant 5(1):125–130, 2005. Tan BH, et al: Sarcopenia in an overweight or obese patient is an
Reilly J, et al: Nutritional support after liver transplantation: a random- adverse prognostic factor in pancreatic cancer, Clin Cancer Res
ized prospective study, JPEN J Parenter Enteral Nutr 14:386–391, 15:6973–6979, 2009.
1990. Tisdale MJ: Pathogenesis of cancer cachexia, J Support Oncol 1:159–
Revie EJ, et al: Randomized clinical trial of local infiltration plus 168, 2003.
patient-controlled opiate analgesia vs. epidural analgesia following Todorov P, et al: Characterization of a cancer cachectic factor, Nature
liver resection surgery, HPB (Oxford) 14(9):611–618, 2012. 379:739–742, 1996.
Rice H, et al: Morbid prognostic features in chronic liver failure Torres DM, Harrison SA: Nonalcoholic steatohepatitis and noncir-
patients undergoing non-hepatic surgery, Am J Surg 132:880–884, rhotic hepatocellular carcinoma: fertile soil, Semin Liver Dis 32:30–
1997. 38, 2012.
Riggio O, et al: Total and individual free fatty acid concentration in Tripathy U, et al: Preoperative bile salt administration versus bile
liver cirrhosis, Metabolism 333:646–651, 1984. refeeding in obstructive jaundice, Natl Med J India 9(2):66–69,
Riou J, et al: Antiketogenic effect of glucose per se in vivo in man and 1996.
in vitro in isolated rat liver cells, Metabolism 35:608–613, 1986. Tsien C, et al: Reversal of sarcopenia predicts survival after a transjugu-
Rosenberg I: The epidemiologic and methodologic problems in deter- lar intrahepatic portosystemic stent, Eur J Gastroenterol Hepatol
mining nutritional status of older persons. (Summary comments), 25(1):85–93, 2013.
Am J Clin Nutr 50:1231–1233, 1989. Tu MY, et al: Using a nutritional screening tool to evaluate the nutri-
Ruiz Garcia V, et al: Megestrol acetate for treatment of anorexia- tional status of patients with colorectal cancer, Nutr Cancer 64:323–
cachexia syndrome, Cochrane Database Syst Rev (3):CD004310, 330, 2012.
2013. Valero V, et al: Sarcopenia adversely impacts postoperative complica-
Ryan DP, et al: Pancreatic cancer: local success and distant failure, tions following resection or transplantation in patients with primary
Oncologist 3:178–188, 1998. liver tumors, J Gastrointest Surg 19(2):272–281, 2015.
Sanford DE, et al: Severe nutritional risk predicts decreased long-term van Dam RM, et al; Enhanced Recovery After Surgery (ERAS) Group:
survival in geriatric patients undergoing pancreaticoduodenectomy Initial experience with a multimodal enhanced recovery programme
for benign disease, J Am Coll Surg 219(6):1149–1156, 2014. in patients undergoing liver resection, Br J Surg 95(8):969–975,
Sardesai V: Role of antioxidants in health maintenance, Nutr Clin Pract 2008.
10:19–25, 1995. van Leeuwen P, et al: Hepatic failure and coma after liver resection is
Schiesser M, et al: Assessment of a novel screening score for nutritional reversed by manipulation of gut contents: the role of endotoxin,
risk in predicting complications in gastro-intestinal surgery, Clin Nutr Surgery 110:169–175, 1991.
27(4):565–570, 2008. van Vledder MG, et al: Body composition and outcome in patients
Sekido H, et al: Impact of early enteral nutrition after liver transplanta- undergoing resection of colorectal liver metastases, Br J Surg
tion for acute hepatic failure: report of four cases, Transplant Proc 99:550–557, 2012.
35:369–371, 2003. Venkat R: Predicting the risk of perioperative mortality in patients
Serino F, et al: Coenzyme Q alpha tocopherol graft uptake and lipid undergoing pancreaticoduodenectomy: a novel scoring system, Arch
peroxidation in human liver transplantation: evidence of energy Surg 146(11):1277–1284, 2011.
metabolism protection in University of Wisconsin (UW) solution- Viganò J, et al: Effects of preoperative oral carbohydrate supplementa-
preserved livers, Transplant Proc 22:2194–2197, 1990. tion on postoperative metabolic stress response of patients undergo-
Shinkawa H, et al: Nutritional risk index as an independent predictive ing elective abdominal surgery, World J Surg 36(8):1738–1743,
factor for the development of surgical site infection after pancreati- 2012.
coduodenectomy, Surg Today 43(3):276–283, 2013. Voron T, et al: Sarcopenia impacts on short- and long-term results of
Shronts EP: Nutritional assessment of adults with end-stage hepatic hepatectomy for hepatocellular carcinoma, Ann Surg 261(6):1173–
failure, Nutr Clin Pract 3:113–119, 1988. 1183, 2015.
Sierzega M, et al: Nutritional status affects the rate of pancreatic fistula Walker PK: The anorexia/cachexia syndrome, Primary Care Cancer
after distal pancreatectomy: a multivariate analysis of 132 patients, 21:13–17, 2001.
J Am Coll Surg 205(1):52–59, 2007. Walrand S, et al: Physiopathological mechanism of sarcopenia, Clin
Sikkens EC, et al: A prospective assessment of the natural course of Geriatr Med 27(3):365–385, 2011.
the exocrine pancreatic function in patients with a pancreatic head Wakai T, et al: Surgical outcomes for hepatocellular carcinoma in non-
tumor, J Clin Gastroenterol 48(5):e43–e46, 2014. alcoholic fatty liver disease, J Gastrointest Surg 15(8):1450–1458,
Sikkens EC, et al: Pancreatic enzyme replacement therapy in 2011.
chronic pancreatitis, Best Pract Res Clin Gastroenterol 24(3):337–347, Wang P, et al: Risk factors for ERCP-related complications: a prospec-
2010. tive multicenter study, Am J Gastroenterol 104(1):31–40, 2008.
Sledzianowski JF, et al: Risk factors for mortality and intraabdominal Weimann A, et al: ESPEN (European Society for Parenteral and
morbidity after distal pancreatectomy, Surgery 137:180–185, Enteral Nutrition): ESPEN guidelines on enteral nutrition: surgery
2005. including organ transplantation, Clin Nutr 25(2):224–244, 2006.
Strasser F, et al: Comparison of orally administered Cannabis extract Wicks C, et al: Comparison of enteral feeding and total parenteral
and delta-9-tetrahydrocannabinol in treating patients with cancer- nutrition after liver transplantation, Lancet 344:837–840, 1994.
related anorexia-cachexia syndrome: a multicenter, phase III, ran- Wigmore SJ, et al: Changes in nutritional status associated with unre-
domized, double-blind,placebo-controlled clinical trial from the sectable pancreatic cancer, Br J Cancer 75:106–109, 1997.
Windsor J: Underweight patients and the risk of major surgery: sym- Yoshida R, et al: Branched-chain amino acid-enriched nutrients
posium on perioperative care, World J Surg 17:165–172, 1993. improve nutritional and metabolic abnormalities in the early post-
Wong-Lun-Hing EM, et al: Is current perioperative practice in hepatic transplant period after living donor liver transplantation, J Hepato-
surgery based on enhanced recovery after surgery (ERAS) princi- biliary Pancreat Sci 19(4):438–448, 2012.
ples?, World J Surg 38(5):1127–1140, 2014. Yoshikawa T, et al: Endotoxin-induced disseminated intravascular
Wu W, et al: The effect of pylorus removal on delayed gastric emptying coagulation in vitamin E-deficient rats, Toxicol Appl Pharmacol
after pancreaticoduodenectomy: a meta-analysis of 2,599 patients, 74:173–178, 1984.
PLoS ONE 9(10):e108380, 2014. Zelber-Sagi S, et al: Nutrition and physical activity in NAFLD: an
Wu Z, et al: Omega-3 fatty acid improves the clinical outcome of hepa- overview of the epidemiological evidence, World J Gastroenterol
tectomized patients with hepatitis B virus (HBV)-associated hepato- 17(29):3377–3389, 2011.
cellular carcinoma, J Biomed Res 26:395–399, 2012. Zhang Y, et al: Probiotic use in preventing postoperative infection in
Xu MQ, et al: Initial clinical effect of intraportal insulin administration liver transplant patients, Hepatobiliary Surg Nutr 2(3):142–147,
on liver graft regeneration in adult patients who underwent living 2013.
donor right lobe liver transplantation, Liver Transpl 15(10):1368– Zhu XH, et al: Effect of early enteral combined with parenteral nutri-
1369, 2009. tion in patients undergoing pancreaticoduodenectomy, World J Gas-
Yang Y, et al: Resistance exercise enhances myofibrillar protein synthe- troenterol 19(35):5889–5896, 2013.
sis with graded intakes of whey protein in older men, Br J Nutr
108:1780–1788, 2012.
CHAPTER 27
Postoperative complications requiring intervention:
diagnosis and management
Stephen B. Solomon, James F. Griffin, Matthew J. Weiss, and Franz Edward Boas
OVERVIEW the surgical mortality from 25% to less than 2% at some high-
volume centers, whereas morbidity has remained essentially
During the last 2 decades, the number of complex pancreatic unchanged (Cameron & He, 2015) (see Chapter 66). The
and hepatic resections performed in the United States has reported morbidity and mortality following pancreatic resection
steadily increased along with a significant decrease in associated in several large series reported since 2000 is presented in Table
perioperative mortality (Allen, 2007; Cameron & He, 2015; 27.1. For pancreaticoduodenectomy (PD), overall periopera-
Dimick et al, 2004; Jarnagin et al, 2002; Nathan et al, 2012; tive mortality ranges from 1.6% to 9%, whereas reported mor-
Winter et al, 2006). This is due in part to a trend toward surgi- bidity ranges from 35% to 45% (Büchler et al, 2000; Cameron
cal specialization and centralization of care at high-volume & He, 2015; Mezhir et al, 2009; Muscari et al, 2006; Winter
institutions, where the inverse volume-outcomes relationship et al, 2006).
has been well described for many complex procedures (Finks Pancreatic resection is associated with a host of complica-
et al, 2011; Nathan et al, 2009; Stitzenberg et al, 2009; Valero tions, many of which are procedure specific and occur at the
et al, 2015). Despite the observed decrease in postoperative operative site (hemorrhage, leak, obstruction), whereas others
mortality, morbidity has remained essentially unchanged, with occur remotely as a consequence of major surgery rather than
most large series reporting a 35% to 45% major complication the specific type of surgery (deep venous thrombosis, cardiac
rate following either pancreatic or hepatic resection (Cameron arrhythmia, pneumonia). This chapter is primarily concerned
& He, 2015; Kneuertz et al, 2012; Nathan et al, 2009; Vin with the most common severe complications directly related to
et al, 2008). the technical aspects of the operation itself. In the case of pan-
This widening gap between morbidity and mortality implies creatic resection, this includes delayed gastric emptying (DGE),
an improved ability to manage postoperative complications intraabdominal fluid collection or abscess, pancreatic fistula,
when they occur, thereby rescuing patients who previously and postoperative hemorrhage. The reported frequencies of
would have died (Bassi et al, 2001; Sohn, 2003). This is par- these complications are listed in The reported frequencies for
tially due to earlier and more effective detection based on famil- these complications in several large series since 2000 are pre-
iarity with procedure-specific complications and advances in sented in Table 27.2 and a comprehensive list of complications
high-quality imaging. Another reason is the shift in management from the Johns Hopkins series (Winter et al, 2006) is presented
from operative reexploration to interventional and endoscopic in Table 27.3.
techniques that are effective with much less added morbidity
(Sohn, 2003; Vin et al, 2008). Interventional radiologists now
Defining and Grading Complications
make up an indispensible part of the multidisciplinary team After Pancreatectomy
caring for patients undergoing complex hepatopancreatobiliary Postoperative complications are notoriously difficult to study
procedures and their expertise in the minimally invasive treat- due to the inconsistent definitions and imprecise terminology
ment of anastomotic leaks, intraabdominal abscesses, and post- used to designate them (see Chapter 66). Before the formal
operative bleeding is responsible for significantly improved definition of postoperative pancreatic fistula (POPF) (Bassi
patient outcomes (Mezhir, 2013; Sohn, 2003). et al, 2005), pancreatic leaks, fistulae, and intraabdominal
This chapter will provide an overview of the most commonly abscesses were indicative of this problem (Mezhir, 2013). They
reported postoperative complications associated with major comprise three overlapping manifestations of the same underly-
pancreatic and hepatic resections, including a review of the ing issue, which is breakdown and leakage at the pancreatic
current definitions and grading systems used for reporting. It anastomosis (or pancreatic closure site for distal and central
will then provide a detailed discussion of the appropriate man- pancreatectomies). For example, an intraabdominal fluid col-
agement of each complication according to the most current lection requiring drain placement could reasonably be reported
literature and with an emphasis on interventional techniques. as an abscess. If the drainage was studied and found to be
amylase rich, it could be considered a leak. If drainage from
PANCREATECTOMY the leak persisted during a period of time, it could also be
considered a fistula. However, the absence of any standardized,
Before the 1980s, the high morbidity and mortality rates associ- quantifiable measures for reliably differentiating these from
ated with pancreatic resection led many surgeons to all but one another meant that the same complication could be
abandon the procedure. Since that time, significant improve- reported in three different ways. This inconsistency in reporting
ments in surgical technique and postoperative care have reduced was demonstrated by Bassi and colleagues (2004), who
459
TABLE 27.1 Morbidity and Mortality Rates According to Pancreatectomy Procedure

Study Years Inclusive N Morbidity N (%) Mortality N (%)
Pancreaticoduodenectomy
Büchler et al, 2000 1993-1999 331 127 (38) 7 (2)
Muscari et al, 2006 1995-1999 300 117 (39) 28 (9)
Winter et al, 2006 1970-2006 1175 415 (35) 26 (2)
Mezhir et al, 2009* 2000-2005 340 147 (43) 9 (3)
Braga et al, 2011 2002-2010 700 432 (67) 27 (4)
Cameron and He, 2015 1969-2012 2000 894 (45) 31 (1.6)
Distal Pancreatectomy
Kooby et al, 2008*† 2002-2006 342 334 (98) 1 (0.3)
Kleef et al, 2007 1993-2006 302 105 (35) 6 (2)
Goh et al, 2008 1986-2006 232 107 (46) 7 (3)
Nathan et al, 2009 1984-2006 704 232 (33) 7 (1)
Central Pancreatectomy
Sauvanet et al, 2002† 1990-1998 53 22 (42) 1 (2)
Roggin et al, 2006 1993-2005 10 6 (60) 0 (0)
Crippa et al, 2007† 1990-2005 100 58 (58) 9 (9)
Adham et al, 2008 1987-2005 50 18 (36) 0 (0)
Ocuin et al, 2008 2000-2007 13 12 (92) 1 (8)
Hirono et al, 2009 1999-2008 24 12 (50) 1 (4)
Total Pancreatectomy
Karpoff et al, 2001 1983-1998 35 15 (54) 1 (3)
Billings et al, 2005‡ 1985-2002 99 32 (32) 5 (5)
Müller et al, 2007 2001-2006 100 38 (38) 6 (6)
Reddy et al, 2009 1970-2007 100 69 (69) 8 (8)
Watanabe et al, 2015‡ 1990-2013 44 14 (32) 2 (5)
Epelboym et al, 2013 1994-2011 77 37 (49) 2 (3)
*Matched case-control study.
†
Multiinstitutional study.
‡
Includes completion pancreatectomy patient(s).
systematically reviewed the definitions used for reporting pan- complications after pancreatectomy. Note that the grade of the
creatic fistulae. They generated four consensus definitions rep- complication increases only when a new, more invasiveness or
resentative of those most commonly used in the literature and aggressive treatment is required for management.
applied each to the same control group of pancreatectomies. An example of this in the surgical literature is the develop-
Using these four definitions, the incidence of pancreatic fistulae ment of the Clavien-Dindo classification of surgical com
varied within the same group of patients from 10% to 29%. plications. Clavien and colleagues (1992) recognized the
This lack of consistency is reflected in the literature, where the inconsistency in reporting surgical complications in 1992 and
actual reported incidence of pancreatic fistulae varies similarly proposed a definition and classification system for grading them
from 2% to 29% (Braga et al, 2011; Büchler et al, 2000; in a standardized fashion. A complication, as defined by
Muscari et al, 2006; Reid-Lombardo et al, 2007). Clavien, is any deviation from the normal postoperative course
Regardless of the difficulty inherent in defining complica- and is differentiated from a sequela, which is an “after-effect”
tions, accurate measurement and documentation in a standard- of surgery inherent to the procedure (e.g., inability to walk after
ized manner is vital to improving quality of care by allowing a leg amputation), and a failure to achieve the primary objective
meaningful comparisons between procedures, institutions, and of the surgery, which is a “failure to cure” (e.g., residual tumor
even individual surgeons. With this in mind, several institutions after surgery). The system was updated in 2004 as the Clavien-
have developed classification and grading systems for surgical Dindo classification, which is a graded system of increasing
complications based on the Common Terminology Criteria for severity determined by the presumed long-term consequences
Adverse Events (CTCAE), a system used by the National Insti- of a given complication as well as the invasiveness, level of care,
tutes of Health Cancer Therapy Evaluation Program for clinical and resources required for the intervention needed to correct
trials since 1982 (Clavien et al, 1992, 2009; DeOliveira et al, it (Dindo et al, 2004). The system is made up of five grades,
2006; Grobmyer et al, 2007). This approach supplies a simpli- two of which have subgroups, and a suffix “d” category that
fied, unifying definition for each complication, with grades of can be appended to any grade. Grade I and II complications
increasing severity based upon the clinical impact of the com- are minor and need no specific therapy (I) or pharmacologic
plication and the invasive of the required treatment. Table 27.4 treatment only (II). Complications requiring invasive interven-
demonstrates an example of this approach applied to common tion (endoscopic, radiologic, or surgical) are classified as grade
Chapter 27 Postoperative complications requiring intervention: diagnosis and management 461
TABLE 27.2 Common Complications Associated With Pancreatectomy

Intraabdominal
Patients With Collection or Hemorrhage Wound Infection Reoperation
Study N Complications N (%) Fistula N (%) Abscess N (%) DGE N (%) N (%) N (%) Rate N (%)
Pancreaticoduodenectomy
Büchler et al, 331 127 (38) 7 (2) 4 (1) 54 (16) 12 (4) 13 (4) 13 (4)
2000
Muscari et al, 300 117 (39) 50 (17) 15 (5) — 18 (6) — —
2006
Winter et al, 1175 415 (35) 52 (4) 38 (3) 161 (14) — 91 (8) 35 (3)
2006
Reid- 1507 — 196 (13) 97 (6) 187 (12) 54 (4) — 53 (4)
Lombardo
et al, 2007*
Mezhir et al, 340 147 (43) 20 (6) 23 (7) 29 (9) 12 (4) 46 (14) —
2009†
Braga et al, 700 432 (67) 200 (29) 42 (6) 105 (15) 36 (5) 58 (8) 56 (8)
2011
Cameron and 2000 894 (45) 295 (15) — 410 (21) 32 (2) 222 (11) 69 (3.5)
He, 2015
Distal Pancreatectomy
Kooby et al, 342 170 (50) 99 (29) — — — 36 (11) 6 (2)
2008*†
Kleef et al, 302 105 (35) 35 (12) 14 (5) 14 (5) 10 (3) 8 (3) 26 (9)
2007
Goh et al, 232 107 (46) 72 (31) — — — 17 (7) 11 (5)
2008
Nathan et al, 704 232 (33) 203 (29) 36 (5) — — — 40 (6)
2009
Central Pancreatectomy
Sauvanet et al, 53 22 (42) 16 (30) 3 (6) 1 (2) 2 (4) 2 (4) 3 (6)
2002*
Roggin et al, 10 6 (60) 3 (30) — — 1 (10) — 1 (10)
2006
Crippa et al, 100 58 (58) 44 (44) 15 (15) 2 (2) 1 (1) — 0 (0)
2007*
Adham et al, 50 18 (36) 4 (8) 7 (14) — 3 (6) — 6 (12)
2008
Hirono et al, 24 12 (50) 15 (63) 1 (4) 1 (4) 0 (0) 1 (4) 0 (0)
2009
Total Pancreatectomy
Karpoff et al, 35 15 (54) NA — — — — —
2001
Billings et al, 99 32 (32) NA 6 (6) 8 (8) — 4 (4) 2 (2)
2005‡
Müller et al, 100 38 (38) NA 2 (2) 8 (8) 2 (2) 2 (2) 15 (15)
2007
Reddy et al, 100 69 (69) NA — 11 (11) 14 (14) 18 (18) —
2009
Epelboym 77 37 (49) NA — — — — —
et al, 2013
Watanabe 44 14 (32) NA — 4 (9) — 2 (5) —
et al, 2015‡
*Multiinstitutional study.
†
Matched case-control study.
‡
Includes completion pancreatectomy patient(s).
—, Not available; DGE, delayed gastric emptying; NA, not applicable.
the need for a universally accepted, objective definition of pan-

TABLE 27.3 Postoperative Complications Following
creatic anastomotic leak as well as a standardized scale for
Pancreaticoduodenectomy for Ductal Adenocarcinoma (n = 1175) grading it (Bassi et al, 2005). They used the CTCAE approach
Category n (%) to outline a definition simple and broad enough to unify the full
spectrum of associated complications—leaks, collections, and
Perioperative Mortality 26 (2)
fistulae—while allowing for differentiation between the three
1970s (n = 23) 7 (30)*
complications based on clinical impact. Postoperative pancre-
1980s (n = 65) 3 (5)* atic fistula (POPF) was defined as “drain output of any measur-
1990s (n = 514) 10 (2) able volume of fluid on or after postoperative day 3, with
2000s (n = 573) 6 (1) amylase content greater than three times that of normal serum
Perioperative Morbidity 415 (38) amylase” (Bassi et al, 2005). Scenarios meeting this initial quali-
1970s (n = 23) No data fying criterion may then be graded as A, B, or C based on nine
1980s (n = 65) 7 (30) criteria, including clinical conditions, specific treatment, ultra-
sound or computed tomography findings, persistent drainage,
1990s (n = 514) 158 (31)*
death related to POPF, signs of infection, sepsis, and readmis-
2000s (n = 573) 250 (45)
sion (Table 27.5). Grade A fistulae have no significant impact
Reoperation Rate (During Index 35 (3) and require no interventions, whereas grade C represents a clini-
Admission)
cally unstable patient who may require parenteral nutritional
Specific Complications support, ICU monitoring, and invasive drainage procedures.
Delayed gastric emptying 161 (15) Since its inception, many institutions have implemented the
Wound infection 91 (8) ISGPF definition and grading system for reporting the rates
Pancreatic fistula 52 (5) and outcomes of pancreatic fistulae, and independent prospec-
Cardiac morbidity 27 (4) tive validation has been reported. Pratt and colleagues (2006)
Abdominal abscess 38 (4) prospectively analyzed postoperative complications in 176 con-
Cholangitis 26 (2) secutive patients after PD. There were 53 patients with con-
Sepsis 19 (2) firmed fistulae (30% incidence), of which 26 (49%) were grade
Bile leak 16 (2)
A, 21 (40%) were grade B, and 6 (11%) were grade C. No
deaths were attributed to fistula in this study. Patients with
Lymph leak 11 (1)
grade A fistulae had shorter hospital stays and fewer secondary
Urinary tract infection 11 (1)
complications than patients with grade B and C fistulae,
Peptic ulcer 10 (1) although the overall rate of secondary complications was not
Pneumonia 10 (1) significantly different between grades B and C (76% vs. 100%,
Acute pancreatitis 5 (1) respectively; P = .2). Patients with grade C fistulae had a higher
Small bowel obstruction 3 (0.3) rate of ICU transfer, more blood transfusions, and longer hos-
Length of Stay (Days), Median (Range) 9 (4-375) pital stays compared with those with grade B fistulae. Also, 5
1980s 16 (10-51)* of the 6 patients with a grade C fistula were discharged to a
1990s 11 (7-373)* rehabilitation facility, which was significantly more frequently
2000s 8 (4-375) than patients with grade A and B fistulae. In terms of cost
burden, a positive correlation between increased health care
*P < .05 compared to the 2000s.
costs and increase in fistula grade was identified. Overall, this
Modified from Winter J, et al, 2006: 1423 pancreaticoduodenectomies for pancreatic cancer: a study validated the ISGPF classification scheme by demon-
single-institution experience. J Gastrointest Surg 10(9):1199-1210; discussion 1210-1211.
strating no clinical consequence of grade A fistulae, whereas
grade B and C fistulae led to more complications, longer hos-
III and differentiated as IIIa or IIIb depending on the need for pital stays, and higher health care costs.
general anesthesia. Grade IV complications are life threatening
and require escalation of care to an intermediate or intensive Management of Postpancreatectomy Complications
care unit (ICU). Grade IV is further divided into IVa (single- Postoperative Pancreatic Fistula
organ dysfunction) and grade IVb (multiorgan dysfunction). POPF, as defined by the ISGPF (see Table 27.5), is the most
Patient death is considered a grade V complication. Finally, the common and costly complication associated with pancreatec-
suffix “d” can be applied to any of the first four grades to tomy. Suspicion for it begins with any deviation from the
indicate the presence of disability at the time of discharge. In normal postoperative course, including nonspecific signs and
recent years, this system has become widely used in the surgical symptoms such as abdominal discomfort, nausea, tachycardia,
literature to help improve the consistency with which surgical and fever. It has a range of presentations that include transient
complications are reported. Moreover, many surgical subspe- anastomotic leak, intraabdominal abscess, and frank fistula for-
cialties have adapted this strategy to meet the specific needs of mation. Given the significant morbidity associated with patient
identifying and reporting complications within their respective outcomes once a fistula has developed, a great deal of research
fields. has been devoted to developing strategies for prevention. One
of the most extensively studied is pharmacologic therapy using
Postoperative Pancreatic Fistula: A Unifying System somatostatin analogues, which has yielded mixed results.
for Reporting Somatostatin is an inhibitory peptide hormone that decreases
The International Study Group of Pancreatic Fistula (ISGPF), exocrine secretions from the pancreas, as well as the gastroin-
an international panel of pancreatic surgery experts, recognized testinal (GI) and biliary tracts. Klempa and colleagues (1979)
TABLE 27.4 Definitions and Grading of Complications According to CTCAE: Example

Adverse Event and Definition Grade 1 Grade 2 Grade 3 Grade 4 Grade 5
Gastroparesis
An incomplete paralysis of Mild nausea, early Moderate symptoms; Weight loss refractory to — —
the muscles of the satiety, and able to maintain medical intervention;
stomach wall, resulting bloating; able to nutrition with dietary unable to maintain
in delayed gastric maintain caloric and lifestyle nutrition orally
emptying into the small intake on regular modifications; may
intestine diet need pharmacologic
intervention
Abdominal Infection
An infectious process — — IV antibiotic indicated; Life-threatening Death
involving the abdominal radiologic or operative consequences;
cavity intervention indicated urgent intervention
indicated
Intraabdominal Hemorrhage
Bleeding in the abdominal – Medical intervention or Transfusion, radiologic, Life-threatening Death
cavity minor cauterization endoscopic, or consequences;
indicated elective operative urgent intervention
intervention indicated indicated
Wound Infection
An infectious process — Localized; local IV antibiotic indicated; Life-threatening Death
involving the wound intervention indicated radiologic or operative consequences;
(e.g., topical antibiotic) intervention indicated urgent intervention
indicated
Pancreatic Fistula
An abnormal Asymptomatic; Symptomatic; altered GI Severely altered GI Life-threatening Death
communication between clinical or function function; tube feeding consequences;
the pancreas and diagnostic or TPN or urgent operative
another organ or observations only; hospitalization intervention
anatomic site intervention not indicated; elective indicated
indicated operative intervention
indicated
CTCAE, Common Terminology Criteria for Adverse Events; IV, intravenous; GI, gastrointestinal; TPN, total parenteral nutrition.
Antibiotic intervention includes antifungal and antiviral therapy.
Modified from http://www.ctep.info.nih.gov.
TABLE 27.5 International Study Group of Pancreatic Fistula Grading System

Clinical Variable Grade A Grade B Grade C
Clinical condition Well Often well Ill-appearing/bad

Supportive treatment: PPN/TPN/enteral No Yes/no Yes
nutrition; antibiotics; somatostatin analogue
Imaging results (if obtained) Negative Negative/positive Positive
Persistent drainage ≥ 3 weeks (±drain in place) No Usually yes Yes
Reoperation No No Yes
Death related to fistula No No Possibly yes
Clinical signs of infection No Yes Yes
Sepsis No No Yes
Readmission No Yes/no Yes/no
PPN, Partial parenteral nutrition; TPN, total parenteral nutrition.Pancreatic fistula is defined as drain output of any measurable volume of fluid on or after postoperative day 3 with amylase content greater
than three times normal serum activity. Once this criterion has been met, the grading system is applied.
Modified from Bassi C, et al: Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery 138(1):8-13, 2005.
first introduced the idea of exocrine inhibition for the preven- 1995; Montorsi et al, 1995; Pederzoli et al, 1994). Unfortu-
tion of POPF in 1979 when they reported lower complication nately, American trials did not recapitulate these early findings
rates following perioperative continuous somatostatin infusion. and instead showed no reduction in rates of POPF (Lowy et al,
A long-acting synthetic version of somatostatin called octreotide 1997; Yeo et al, 2000). A new somatostatin analogue called
has been studied for use in pancreatic surgery, and several pasireotide, which has a longer half-life and broader binding
European studies initially reported decreased overall complica- profile compared with octreotide, has shown promise recently
tion rates, including POPF (Büchler et al, 1992; Friess et al, in reducing rates of POPF. Allen and colleagues (2014)
conducted a single-center, randomized, double-blind trial of tion following upper GI tract surgery, particularly after pancre-
perioperative subcutaneous pasireotide in 300 patients under- atic surgery. In many cases, DGE is associated with concurrent
going either PD or distal pancreatectomy and observed a sta- intraabdominal abscesses or POPF and results in significantly
tistically significant decrease in POPF compared with placebo longer hospital stays and increased health care costs (Beane
(9% vs. 21%; P = .006). These results will have to be validated et al, 2014; Yeo et al, 2002). According to the studies included
with additional prospective trials, but there is currently opti- in Table 27.2, incidence of DGE following PD ranges from 9%
mism for a viable new therapy on the horizon. to 21%, although many reports are even higher, depending on
Once POPF develops, by definition, only grades B and the definition used. In 2007, the International Study Group for
C are of any clinical significance warranting specific inter Pancreatic Surgery (ISGPS) proposed a consensus definition
vention. Uncomplicated low-output fistulae can be managed that classified DGE into grades A, B, and C according to clini-
conservatively, which includes adequate drainage of intraab- cal impact determined by nasogastric tube (NGT) require-
dominal collections (see Image-Guided Abdominal Drainage) ment, ability to tolerate oral intake, and need for prokinetic
(see Chapter 30), restriction of oral intake, and nutritional therapy (Tables 27.6A and B) (Wente et al, 2007a). Further-
support (see Chapter 26). Empiric antibiotic therapy may be more, it stipulated that the patient must first undergo an upper
initiated if there is any concern for infection, and all drains GI contrast series or endoscopic evaluation to rule out mechan-
should be kept in place with outputs measured regularly. ical obstruction, such as an early postoperative small bowel
With conservative management, spontaneous fistula closure obstruction or stricture at the anastomosis. Subsequent valida-
will occur in as many as 90% of cases, usually within 4 weeks tion by several studies reported postoperative DGE rates of
(Machado, 2012). 14% to 45%, with grade-specific rates of 6% to 27% (grade A),
A key element of the conservative approach to management 7% to 11% (grade B), and 1% to 12% (grade C) (Akizuki et al,
of POPF is avoidance of oral intake, which prevents food- 2009; Chong et al, 2015; Malleo et al, 2010; Park et al, 2009;
induced stimulation of pancreatic secretions. However, pro- Welsch et al, 2010). The new definition actually increased
longed fasting results in rapid nutritional depletion and impaired reported rates of DGE for some series that previously used
wound healing, so two forms of nutritional support are advo- much more stringent definitions (Akizuki et al, 2009; Malleo
cated for maintaining patients until resolution of their fistulae: et al, 2010).
total parenteral nutrition (TPN) and enteral nutrition (EN) The management of DGE generally consists of supportive
delivered via a distally positioned nasointestinal tube or feeding care, beginning with discontinuation of oral intake and reinser-
jejunostomy (Malleo et al, 2014) (see Chapter 26). tion of an NGT for gastric decompression. Once an upper GI
TPN has traditionally been the mainstay in nutritional series or endoscopy has ruled out mechanical obstruction,
support for POPF because it avoids stimulation of GI secretions patients may be started on a prokinetic agent to improve symp-
and is usually able to meet all necessary nutritional require- toms and shorten time to recovery. If symptoms persist for an
ments. However, long-term complications associated with TPN extended period, the patient may also require some form of
are significant and include sepsis, metabolic derangements, cir- nutritional support and warrants a cross-sectional imaging
rhosis, and atrophy of the GI tract with gut barrier dysfunction evaluation to look for secondary causes of DGE, such as an
(Klek et al, 2011; Malleo et al, 2014). Enteral feeding avoids intraabdominal abscess.
these complications because it does not require a long-term The underlying cause of DGE is unclear and likely multi-
central venous catheter, and it has a trophic effect on the GI factorial, but it may result in part from decreased circulating
tract. Initially, there were concerns that enteral feeding would levels of motilin, a prokinetic peptide hormone located in the
increase exocrine pancreatic secretion and prolong or worsen enterochromaffin cells of the duodenum and proximal small
POPF. In a randomized comparison of TPN and EN in patients intestine (Yeo et al, 1993). Erythromycin, a macrolide antibi-
with grade B POPF, EN was associated with a greater than otic, acts as a motilin agonist and binds its receptors to stimu-
twofold increased probability of fistula closure, with shortened late gastric emptying. Randomized, controlled trials evaluating
times to closure and lower overall cost (Klek et al, 2011) (see erythromycin for use in pancreatic surgery have shown
Chapter 26). decreased incidence of DGE, decreased rates of NGT reinser-
tion, shorter duration of NGT insertion, and shorter time to
Delayed Gastric Emptying oral feeding (Ohwada et al, 2001; Yeo et al, 1993). Based on
Delayed gastric emptying (DGE) is a functional gastroparesis these results, erythromycin is now often used in the treatment
without mechanical obstruction that may occur as a complica- of postoperative DGE.
TABLE 27.6A ISGPS Consensus Definition of DGE After Pancreatic Surgery

Unable to Tolerate Solid
DGE Grade NGT Required Oral Intake by Pod Vomiting/Gastric Distension Use of Prokinetics
A 4-7 days or reinsertion > POD 3 7 ± ±

B 8-14 days or reinsertion > POD 7 14 + +
C >14 days or reinsertion > POD 14 21 + +
DGE, Delayed gastric emptying; ISGPS, International Study Group of Pancreatic Surgery; NGT, nasogastric tube; POD, postoperative day.
To exclude mechanical causes of abnormal gastric emptying, the patency of either the gastrojejunostomy or the duodenojejunostomy should be confirmed by endoscopy or upper gastrointestinal
gastrographin series.
Modified from Wente MN, et al: Delayed gastric emptying (DGE) after pancreatic surgery: a suggested definition by the International Study Group of Pancreatic Surgery (ISGPS). Surgery 142(5):761-768,
2007.
TABLE 27.6B ISGPS Parameters for Grading DGE

Parameter Grade A Grade B Grade C
Clinical condition Well Often well/minor discomfort Ill/bad/severe discomfort (increased

overall risk owing to complications
and procedures)
Comorbidities No Possibly yes (pancreatic leak Possibly yes (pancreatic leak or
or fistula, intraabdominal fistula, intraabdominal abscess)
abscess)
Specific treatment Possibly yes (prokinetic drugs) Yes (prokinetic drugs, potential Yes (prokinetic drugs, NGT)
reinsertion of NGT)
Nutritional support (enteral Possibly yes (slower return to Yes (PPN) Yes (TPN or EN via NGT, prolonged,
or parenteral) solid food intake) i.e., >3 weeks postoperatively)
Diagnostic evaluation No Possibly yes (endoscopy, Yes (endoscopy, upper GI contrast
upper GI contrast study, CT) study, CT)
Interventional treatment No No Possibly yes (e.g., abscess drainage,
relaparotomy for complication,
relaparotomy for DGE)
Prolongation of hospital stay Possibly yes Yes Yes
CT, Computed tomography; DGE, delayed gastric emptying; EN, enteral nutrition; GI, gastrointestinal; ISGPS, International Study Group of Pancreatic Surgery; NGT, nasogastric tube; PPN, partial
parenteral nutrition; TPN, total parenteral nutrition.
Modified from Wente MN, et al: Delayed gastric emptying (DGE) after pancreatic surgery: a suggested definition by the International Study Group of Pancreatic Surgery (ISGPS). Surgery 142(5):761-
768, 2007.
If patients are unable to resume oral feeding within 7 to 10

TABLE 27.7A ISGPS Consensus Definition of
days postoperatively, they should receive nutritional support in
the form of EN or parenteral nutrition (PN) (Braga et al, Postpancreatectomy Hemorrhage
2009). Beane and colleagues (2014) reported that when supple-
Time of Onset
mental nutrition was initiated by postoperative day 10, patients
• Early hemorrhage (≤24 hr after the end of the index operation)
resumed a regular diet sooner (day 24 vs. day 36; P = .05) and • Late hemorrhage (>24 hr after the end of the index operation)
were readmitted less often (25% vs. 65%; P < .01). EN is pre-
Location
ferred if a feeding jejunostomy was placed intraoperatively;
• Intraluminal (intraenteric, e.g., anastomotidc suture line at
otherwise, PN in the form of TPN should be used. In the stomach or duodenum, or pancreatic surface at anastomosis,
setting of prolonged DGE, endoscopic placement of a feeding stress ulcer, pseudoaneurysm)
jejunostomy, followed by EN feeding may be necessary to • Extraluminal (extraenteric, bleeding into the abdominal cavity,
reduce the risk of the infectious, metabolic, and GI complica- e.g., from arterial or venous vessels, diffuse bleeding from
resection area, anastomosis suture lines, pseudoaneurysm)
tions associated with long-term TPN.
Severity of Hemorrhage
Postpancreatectomy Hemorrhage Mild
Major postoperative bleeding is one of the most feared and • Small- or medium-volume blood loss (from drains, nasogastric
potentially lethal complications of pancreatic surgery. It is tube, or on ultrasonography, decrease in hemoglobin
concentration <3 g/dL)
reported to occur after 1% to 8% of all pancreatic resections • Mild clinical impairment of the patient, no therapeutic
and accounts for 11% to 38% of overall mortality (Wente et al, consequence, or at most the need for noninvasive treatment
2007b). Its presentation is extremely heterogeneous because with volume resuscitation or blood transfusions (2-3  U packed
bleeding may occur from a number of potential sites with vari- cells within 24 hr of end of operation or 1-3  U if later than
able severity and time of onset. The ISGPS has attempted to 24 hr after operation)
standardize the evaluation and reporting of postoperative bleed- • No need for reoperation or interventional angiographic
embolization; endoscopic treatment of anastomotic bleeding
ing by creating a unifying consensus definition under the name
may occur provided the other conditions apply
postpancreatectomy hemorrhage (PPH). The definition is divided
Severe
into grades A, B, and C based on time of onset (early or late),
location (intraluminal or extraluminal), and severity (mild or • Large-volume blood loss (drop of hemoglobin level by ≥3 g/dL)
• Clinically significant impairment (e.g., tachycardia, hypotension,
severe) (Wente et al, 2007b). The specific definition and oliguria, hypovolemic shock), need for blood transfusion (>3 U
grading criteria are presented in Tables 27.7A and B. packed cells)
• Need for invasive treatment (interventional angiographic
POSTPANCREATECTOMY HEMORRHAGE: DEFINITIONS, TERMINOLOGY, embolization, or relaparotomy)
AND GRADING ISGPS, International Study Group of Pancreatic Surgery.
The overall grade is based primarily on the time of onset and Modified from Wente MN, et al: Delayed gastric emptying (DGE) after pancreatic surgery: a
severity such that grade A consists of all early mild bleeding suggested definition by the International Study Group of Pancreatic Surgery (ISGPS). Surgery
events, and grade C consists of all late severe hemorrhages. 2007;142(5):761-768, 2007.
Grade B is the area of overlap and consists of the two remaining
groups, early severe and late mild bleeding episodes. All grades
may be subdivided further into intraluminal or extraluminal
TABLE 27.7B ISGPS Grading System for Postpancreatectomy Hemorrhage

Grade Onset, Location, Severity, and Clinical Impact Clinical Condition Diagnostic Consequence Therapeutic Consequence
A Early, intraluminal Well Observation, blood count, No

or extraluminal, ultrasonography and, if
mild necessary, computed
tomography
B Early, intra luminal Late, intraluminal or Often well/ Observation, blood count, Transfusion of fluid/blood,
or extraluminal, extraluminal, mild* intermediate, ultrasonography, intermediate care unit (or
severe very rarely life computed tomography, ICU), therapeutic
threatening angiography, endoscopy,† embolization,
endoscopy† relaparotomy for early PPH
C Late, intraluminal or Severely impaired, Angiography, computed Localization of bleeding,
extraluminal, life threatening tomography, angiography and
severe endoscopy† embolization, (endoscopy†)
or relaparotomy, ICU
ICU, Intensive care unit; ISGPS, International Study Group of Pancreatic Surgery; PPH, postpancreatectomy hemorrhage.
*Late, intraluminal or extraluminal, mild bleeding may not be immediately life threatening to patient but may be a warning sign for later severe hemorrhage (“sentinel bleed”) and is therefore grade
B.†Endoscopy should be performed when signs of intraluminal bleeding are present (melena, hematemesis, or blood loss via nasogastric tube).
Modified from Wente MN, et al: Postpancreatectomy hemorrhage (PPH): an International Study Group of Pancreatic Surgery (ISGPS) definition. Surgery 142(1):20-25, 2007.
sources of bleeding, which influences the type of interventions extraluminal source of bleeding may present as intraluminal
available for treatment. This classification system is meant to bleeding should it originate in the setting of an anastamotic
generate groups of increasing clinical impact such that grade A failure, such as a pancreatic dehiscence.
does not result in significant deviation in the normal clinical
course, grade B requires specific therapy and a prolonged hos- POSTPANCREATECTOMY HEMORRHAGE VALIDATION STUDIES
pital stay, and grade C is potentially life threatening and requires Despite the ISGPS authors’ attempt at formulating an explicit
invasive treatment (Wellner et al, 2014). set of guidelines to standardize the definition of PPH, validation
Time of onset may be within 24 hours of surgery (early) or studies to date have reported mixed results. This is partly
anytime thereafter (late). Early hemorrhage is usually indicative because databases used for retrospective analyses often do not
of a technical failure to achieve adequate hemostasis in the capture clinically insignificant mild to moderate bleeding
operating room, whereas late hemorrhage usually occurs in the events. Moreover, the definition relies on nonspecific surrogate
setting of inflammatory intraabdominal processes (pancreatic markers for bleeding, which do not necessarily correlate with
leaks, abscesses, fistulae, and anastomotic ulceration) that actual or surgery-specific bleeding. With this in mind, several
cause vascular erosion or the formation of arterial pseudoaneu- validation studies narrowed the scope of their evaluations by
rysms (Rajarathinam et al, 2008; Wellner et al, 2014; Welsch applying the definition only to severe bleeding, delayed bleeding,
et al, 2011). Late hemorrhages are the most deadly because and/or bleeding from the surgical site (Correa-Gallego et al, 2012;
they can occur unexpectedly days to weeks after surgery— Grützmann et al, 2012; Rajarathinam et al, 2008; Yehuda et al,
between postoperative days 7 and 32 in one study (Choi et al, 2014). These studies report an overall incidence of PPH
2004)—and often present as abrupt, massive bleeding. ranging from 3% to 10.4%, with individual grades ranging from
The severity of hemorrhage is classified as mild or severe 0% (A), 0.6% to 1.7% (B), and 1.2% to 10.4% (C) (Correa-
depending on the degree to which it impacts the postoperative Gallego et al, 2012; Grützmann et al, 2012; Rajarathinam
course. Mild hemorrhage has minimal clinical impact and is et al, 2008; Wellner et al, 2014; Yehuda et al, 2014). Mortality
defined as a drop in hemoglobin concentration of less than among those affected ranges from 3% to 29%, of which the vast
3 g/dL requiring no more than three U of packed red blood majority (71% to 100%) occurs in association with grade C
cells (PRBCs) and no invasive interventions. Alternatively, hemorrhage.
severe hemorrhage is life threatening and characterized by a Two additional studies, one prospective (Ricci et al, 2012)
drop in hemoglobin of greater than or equal to 3 g/dL, with and one retrospective (Welsch et al, 2011), applied the consen-
clinical signs of significant volume loss (tachycardia, hypoten- sus definition strictly as written and reported much higher rates
sion, oliguria) requiring more than three U of PRBCs and/or of PPH, from 27% to 29%. The incidence for each grade was
invasive intervention. 1.8% to 4.8% (A), 15.2% to 20.4% (B), and 5.4% to 9.2%
Location refers to the source of bleeding, which is either intra- (C), with mortality ranging from 7.3% to 19.4%. These studies
luminal (bleeding into the GI tract) or extraluminal (bleeding demonstrate a higher incidence of PPH with a lower associated
into the abdominal cavity). Intraluminal hemorrhage usually mortality, which reflects a broader inclusion of mild to moder-
originates from anastomotic sites, marginal or stress ulcers, and ate bleeding events. In many cases, patients had bleeding,
the cut surface of the pancreas and presents clinically as melena/ transfusions, and mortality unrelated to their surgery that
hematochezia, hematemesis, or blood loss through the NGT. resulted in frequent false positives. For example, patients
Extraluminal bleeding originates from peripancreatic vessels, requiring transfusion for underresuscitation in the operating
pseudoaneurysms, external staple/suture lines, and raw surfaces room, fluid and blood-product management in the setting of
in the operative field. This form of bleeding may be evident sepsis, higher hemoglobin goals due to cardiovascular comor-
from the wound and abdominal drains, or there may be no bidities, or bleeding from unrelated causes or procedures would
external signs of bleeding. It is important to note that an still be considered PPH based on the broad surrogate markers
used for inclusion (Ricci et al, 2012; Welsch et al, 2011). Hounsfield U) collections. Active arterial bleeding can be seen
Despite these shortcomings, both studies reported that the as extravasation of contrast on arterial-phase contrast-enhanced
definition did effectively segregate patients based on overall CT scans. On delayed-phase images, if obtained, the extrava-
disease severity, clinical impact, and mortality. sated contrast continues to spread if there is active bleeding.
The most important factor in distinguishing types of PPH is On the other hand, a pseudoaneurysm, which can bleed inter-
the time of onset, which correlates with two different mecha- mittently, is an enhancing structure next to an artery that main-
nisms of bleeding and, subsequently, two different mortality tains its shape on delayed CT images (Fig. 27.2).
rates. Current literature shows that early bleeding accounts for Magnetic resonance cholangiopancreatography (MRCP) is
9% to 36% of all PPH and has a better prognosis, regardless a fluid-sensitive magnetic resonance imaging (MRI) sequence
of severity or location, compared with late hemorrhage (Correa- that clearly shows the pancreatic duct, bile ducts, fistulae, and
Gallego et al, 2012; Rajarathinam et al, 2008; Ricci et al, 2012; fluid collections. The MRCP is typically reconstructed into
Wellner et al, 2014; Welsch et al, 2011). In a study of 796 PDs axial and coronal slices, as well as three-dimensional (3D)
by Welsch and colleagues (2011), 35% (81/232) of PPH was images. The site of a pancreatic fistula can be identified in 75%
early with a mortality rate of 1.2%, compared with a mortality of patients on CT, compared with 93% on MRCP (O’Toole
rate of 10.6% among the late hemorrhages (16/151). Older et al, 2007) (see Chapter 19).
studies (and those that do not apply the ISGPS definition
stringently) demonstrate much higher mortality rates of 18%
Interventional Radiology Techniques After
to 64% for delayed hemorrhage (Choi et al, 2004; Rajarathi- Pancreatectomy
nam et al, 2008; Yekebas et al, 2007). This is because early and Many postpancreatectomy complications are managed using
late hemorrhages result from different causes, which have image-guided percutaneous interventions, reducing the need
implications for overall survival. Early bleeding is usually an for reoperation. The most common interventional radiology
isolated issue of inadequate hemostasis due to a technical error. (IR) procedures after PD are: intraabdominal abscess drainage
Late hemorrhage, on the other hand, generally occurs in sicker (72%), percutaneous biliary drainage (PBD) (18%), and angi-
patients with additional underlying complications such as ography with or without embolization (10%) (Baker et al,
POPF, infection/sepsis, and DGE. Each of these additional 2008) (see Chapters 20, 21, and 30).
complications carries its own individual morbidity and mortal-
ity burden, which contributes to the overall mortality observed Image-Guided Abdominal Drainage
in late PPH. To this point, Yekebas and colleagues reported Postoperative abscesses can be drained percutaneously, using
that all 14 PPH-associated deaths in their study were late hem- ultrasound or CT guidance (see Fig. 27.1). Small abscesses
orrhages, 13 (93%) had concomitant POPF, and bleeding was (<3 cm) can usually be treated with antibiotics alone, but larger
the direct cause of death in only 2 (14%) cases (Yekebas et al, collections require both antibiotics and drainage. Complica-
2007). The other 12 patients succumbed to various forms of tions of image-guided drainage of fluid collections are
critical illness, including sepsis, liver failure, and pulmonary infrequent but include bleeding, sepsis, and peritonitis.
complications. Abscesses are typically drained using the Seldinger tech-
nique. A needle is advanced into the collection under ultra-
sound or CT guidance. After aspirating fluid, a stiff guidewire
IMAGING AND IMAGE-GUIDED THERAPY OF is placed through the needle into the collection. The needle is
COMPLICATIONS AFTER PANCREATECTOMY then removed, and a drain is placed over the wire into the
collection. Typically, 8- to 10-Fr drains are placed in thin,
Imaging After Pancreatectomy serous collections, and 10- to 12-Fr drains are placed in thick,
Image-guided interventions to manage postoperative complica- bloody, or purulent collections. Larger drains are available for
tions are common after pancreatic surgery. After PD, 12% very thick collections, up to 20 Fr for locking-loop drains, and
to 22% of patients require percutaneous intervention (Sohn, 36 Fr for straight drains. A retrospective study showed that
2003; Zink et al, 2009) (see Chapter 66). Postoperative imaging 7- and 14-Fr drains worked equally well for abdominal
should be performed if there is suspicion for complications abscesses (Röthlin et al, 1998). Biliary-type drains, which
based on the clinical picture or abnormal laboratory values. have additional side holes, may be helpful for long multilocu-
Computed tomography (CT) is the most common lated collections.
imaging modality for evaluation of the pancreas after surgery. After catheter placement, the abscess is emptied, and speci-
Complications that can be detected on CT include anastomotic mens are sent for gram stain and culture. The fluid can also be
leak, abscess, fistula, and bleeding (Scialpi et al, 2005) (see sent for amylase (to evaluate for pancreatic leak) and bilirubin
Chapter 18). (to evaluate for bile leak). A drain fluid–to–serum bilirubin ratio
Fluid collections such as seromas, abscesses, and pancreatic greater than 5 indicates a bile leak (Darwin et al, 2010), and
pseudocysts can be identified on CT (Fig. 27.1). Rim enhance- drain amylase–to–serum ratio greater than 5 indicates a pan-
ment suggests abscess or pseudocyst. Gas within a collection creatic leak (Shinchi et al, 2006).
suggests infection or enteric leak. Enteric leaks can have a thin
tract containing gas and fluid, extending from an enteric anas- Drain Management
tomosis to an abscess. Oral contrast given prior to the CT can The drain should be flushed with normal saline two to three
leak into the abscess, which is diagnostic of an enteric leak (see times per day to prevent clogging. Drainage from loculated
Chapter 18). collections can be improved by injecting tissue plasminogen
Postoperative bleeding can be evaluated using a CT angio- activator (tPA) into the tube. One report showed an 89%
gram, which should include noncontrast and arterial phases. success rate using tPA to drain abscesses refractory to simple
Hematomas are visible on noncontrast CT as high-density (>20 catheter drainage (Beland et al, 2008). In a postsurgical patient,
A B
C
FIGURE 27.1 A 77-year-old female patient after distal pancreatectomy for neuroendocrine tumor. A, Postoperative development of a large
rim-enhancing fluid collection in the operative bed, suspicious for abscess or pseudocyst (asterisk). B, Computed tomography (CT)-guided drainage
of the fluid collection with a 10-Fr drainage catheter. Rust-colored fluid was aspirated and sent for analysis, which showed elevated amylase and
negative culture, consistent with a pancreatic leak. The catheter was removed after the output had diminished. C, CT follow-up 10 months after
drain removal showed resolution of the fluid collection (asterisk).
the benefit of tPA should be weighed against the risk of persistent pancreatic leaks is discussed later (see Interventional
bleeding. Management of Pancreaticocutaneous Fistulae).
If the drain output remains high, this suggests an ongoing Minimal output from the drain indicates that the drainage
pancreatic or enteric leak. The drain should be positioned is complete, or that the drain is clogged or malpositioned. Pus
adjacent to the leak for optimal drainage. When output from leaking around a drain suggests that the drain is clogged.
the drain decreases, the locking-loop drain can be exchanged Clogged drains can be exchanged over a guidewire.
for a straight drain, to collapse the abscess cavity adjacent to An abscess drain can typically be removed when the output
the leak. The straight drain can be slowly pulled back during is less than 20 mL/day, the patient has no fever or pericatheter
days or weeks in an attempt to close the fistula. Occasionally, leakage, and the drain flushes easily. An abscessogram or a CT
a persistent fistula is seen on abscessogram, even after the scan can be performed prior to drain removal. An abscessogram
patient is doing well clinically, with no residual abscess cavity shows if the tube is clogged or malpositioned, as well as the size
and no output from the drain. This is called a one-way fistula, of the residual collection and any fistulae. A CT scan shows
and the drain can usually be safely removed. Management of the position of the tube and any undrained collections. On CT,
A B
C D
FIGURE 27.2 A 63-year-old female patient had clinical evidence of bleeding 3 weeks after pancreaticoduodenectomy for pancreatic
adenocarcinoma. A, Computed tomography angiogram showed a large hematoma (arrowheads) in the subhepatic space, with highly enhancing
components within this hematoma consistent with a pseudoaneurysm (arrows) at the origin of the gastroduodenal artery. B, Catheter angiography
confirmed a pseudoaneurysm (arrow) of the gastroduodenal artery, which was treated by occluding both the proper hepatic (outflow) and common
hepatic arteries (inflow) with stainless steel coils. C, Postembolization angiogram showed coils in the hepatic arteries (arrows) with no enhancement
of the hepatic arteries or the pseudoaneurysm of the gastroduodenal artery. D, A 46-year-old male had gastroduodenal artery stump bleeding after
pancreaticoduodenectomy, which was treated by placement of a covered stent (between arrows) into the hepatic artery. Covered stent placement
preserves hepatic arterial flow.
a residual collection around an abscess drain indicates a clogged infarction due to collateral arterial supply to the spleen (Yama-
or poorly functioning drain. moto et al, 2008).
The median drainage time is 11 days for sterile fluid collec- Pseudoaneurysms and arterial extravasation can also be
tions, 29 days for abscesses, and 30 days when there is a pan- treated using a covered stent to exclude the pseudoaneurysm
creatic leak (Zink et al, 2009). while preserving distal flow (Heiss et al, 2008; Suzuki et al,
2009). Covered stents are particularly helpful for gastroduode-
Interventional Management of Pancreaticocutaneous Fistulae nal artery stump blowouts (in which coil embolization might
Pancreatic leaks with pancreaticocutaneous fistulae can develop not be technically possible), distal splenic artery pseudoaneu-
after pancreatic surgery. Most fistulae resolve after 1 month of rysms (where embolization has a higher risk of splenic infarc-
conservative therapy, including jejunal feeding, somatostatin tion), common and proper hepatic artery pseudoaneurysms (to
analogue injection, pseudocyst drains, and endoscopic stent preserve arterial flow to the liver), and superior mesenteric
placement in the pancreatic duct (Cabay et al, 1998; Klek et al, artery pseudoaneurysms (to preserve flow to bowel).
2011; Voss & Pappas, 2002). The fistula is likely to persist if Examples of embolization and covered stent placement for
there is complete transection of the pancreatic duct, if there is bleeding after pancreatic surgery are shown in Figure 27.2.
a downstream ductal stricture, or if it is a high-output fistula
(Voss & Pappas, 2002). HEPATECTOMY
Several percutaneous approaches have been described for
reconnecting the pancreas to the GI tract, to divert pancreatic Hepatic resection is widely performed for the treatment of
fluid away from the fistula and allow it to heal. Cystgastrostomy many liver diseases, including malignant and benign tumors,
(surgical, endoscopic, or percutaneous) can be performed if calculi in the intrahepatic ducts, hydatid disease, and abscesses.
there is a pseudocyst associated with the fistula. If the pancre- In the landmark paper by Foster and Berman from 1977,
atic duct is dilated (>4 mm), it can be punctured percutane- overall mortality in a multicenter analysis of 621 hepatic resec-
ously, allowing placement of a drain from the pancreatic duct tions was 13%, and even greater (20%) in patients undergoing
to the stomach or bowel (Cope et al, 2001). However, the major hepatic re section. Mortality for hepatic resections
pancreatic duct is frequently nondilated, as it is decompressed has improved significantly since that time with and currently
into the cutaneous tract. If the pancreatic duct is not dilated, ranges from 0% to 3% at high-volume centers (Table 27.8)
then a snare can be placed into the duct via the cutaneous (Adam et al, 2006; Cescon et al, 2009; Huang et al, 2009;
fistula, providing a target for percutaneous puncture and drain- Hyder et al, 2013; Imamura et al, 2003; Jarnagin et al, 2002;
age of the duct into the stomach (Boas et al, 2015). Koffron et al, 2007; Palavecino et al, 2010; Poon et al, 2004;
Rees et al, 2008). Unfortunately, other outcomes have not
Percutaneous Biliary Drainage improved in parallel with mortality such that overall morbidity
See “Interventional Management of Bilomas and Bile Leaks,” is still reported in the range of 14% to 45% (see Chapter 103).
and “Interventional Management of Biliary Strictures,” later. Since the time of Foster and Berman, the field of hepatobili-
ary surgery has benefited from a better understanding of liver
Angiography, Embolization, and Covered Stent Placement anatomy, advances in cross-sectional imaging technology,
(See Chapters 21 and 30) expanded use of parenchymal-sparing approaches, improved
Hemorrhage is seen in less than 10% of patients after pancre- anesthetic techniques, and better patient selection. These
atectomy but is associated with high mortality (Puppala et al, improvements are reflected in a recent multi-institutional study
2011). Major bleeding is seen on average 19 days after surgery of 2056 patients undergoing hepatic resection between 1990
and is usually preceded by a smaller sentinel bleed (Otah et al, and 2011 (Hyder et al, 2013). In this study by Hyder and col-
2002). Therefore even a small amount of increased bleeding leagues, 90-day mortality was 1.6%, overall morbidity was
from surgical drains or GI bleeding more than 3 days after 19.3%, and liver-specific complications included postoperative
pancreatic surgery should be evaluated immediately. Bleeding ascites (2.5%), biliary leak (3.2%), bleeding (0.9%), abscess
can be due to vessel injury during surgery or due to pancreatic (0.7%), and liver insufficiency/failure (0.5%). Of note, whereas
fluid eroding the vessel wall. bleeding accounts for a minor percentage of overall morbidity
In hemodynamically stable patients, CT angiography can in the current series, 30 years ago it was the primary cause of
identify the bleeding vessel. A CT angiogram is most likely overall mortality following hepatectomy (Foster & Berman,
to be positive if performed when the patient is actively 1977). A list of common complications reported after major
bleeding. Hemodynamically unstable patients should proceed hepatic resections in several large series since 2000 is presented
directly to catheter angiography and intervention, or to the in Table 27.9.
operating room.
Pseudoaneurysms and arterial extravasation after pancreatic Complications Associated With Hepatic Resections
surgery occur in the gastroduodenal artery most commonly, Posthepatectomy Liver Failure
followed by the hepatic artery, superior mesenteric artery, and Posthepatecomy liver failure (PHLF) is one of the most severe
splenic artery (Tien et al, 2008) (see Chapter 124). Selective complications following hepatectomy, and its incidence varies
coil embolization across the pseudoaneurysm is successful in widely in the literature, between 1.2% to 32% (Rahbari et al,
approximately 85% of patients (Tsai et al, 2007). In an other- 2011a) (see Chapter 103). The International Study Group of
wise normal liver, the hepatic arteries on the right or left side Liver Surgery (ISGLS) defined it as postoperatively acquired
can be safely coil embolized, because the embolized lobe of the deterioration in the ability of the liver to maintain its synthetic,
liver will be supplied by the portal vein and intrahepatic arterial excretory, and detoxifying functions, which are characterized
collaterals (Nicholson et al, 1999). The proximal splenic artery by an increased international normalized ratio (INR) and con-
can also be safely coil embolized without complete splenic comitant hyperbilirubinemia on or after postoperative day 5
TABLE 27.8 Morbidity and Mortality Rates After Hepatic Resection From Large, Single-Institution Series
Primary Hepatic or Major
Years Biliary Tract Metastatic Benign Hepatectomy Morbidity Mortality
Study Inclusive N Cancer N (%) Disease N (%) Disease N (%) N (%)* N (%) N (%)
Primary and Metastatic Disease

Jarnagin et al, 1991-2001 1803 375 (21) 1249 (69) 161 (9) 1568 (87) 817 (45) 55 (3)
2002
Imamura et al, 1994-2002 1056 522 (49) 188 (18) 205 (19) 305 (29) 208 (20) 0 (0)
2003
Poon et al, 1989-2003 1222 878 (72) 156 (13) 188 (15) 757 (62) 396 (32) 39 (3)
2004
Koffron et al, 2001-2006 300 43 (14) 60 (20) 177 (59) 119 (40) 27 (9) NR
2007†
Huang et al, 1986-2005 2008 1663 (83) 128 (6) 834 (42) 638 (32) 281 (14) 12 (0.6)
2009
Cescon et al, 1985-2007 1500 673 (45) 532 (36) 295 (20) 479 (32) 337 (23) 45 (3)
2009
Palavecino et al, 1998-2007 1557 247 (16) 1249 (80) 61 (4) 958 (62) 407 (26) 15 (1)
2010
Metastatic Disease Only
Rees et al, 2008 1987-2005 929 Colorectal — — 605 (65) 242 (26) 14 (2)
metastases
Adam et al, 1983-2004 1452 NCNN — — 799 (55) 320 (22) 32 (2)
2006 metastases
*Hepatectomy or extended hepatectomy
†
Laparoscopic resections only
NCNN, noncolorectal nonneuroendocrine; NR, not reported.
TABLE 27.9 Common Complications Associated With Hepatectomy

Sterile
Patients With Required Liver Perihepatic Bile Leak/ Perihepatic Pulmonary Pleural Wound
Complications Transfusion Failure Abscess Biloma Collection Complication Effusion Infection
Study N N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%) N (%)
Jarnagin 1803 817 (45) 880 (49) 99 (6) 110 (6) 47 (3) 97 (5) 344 (19) 154 (9) 94 (5)
et al, 2002
Imamura 1056 208 (20) 230 (22) 1 (.1) 87 (8) 97 (9) NR 274 (26) 194 (18) 43 (4)
et al, 2003
Poon et al, 1222 396 (32) 414 (34) 47 (4) 33 (3) NR NR 164 (13) 63 (5) 115 (9)
2004
Cescon et al, 1500 337 (23) 548 (37) 86 (6) 41 (3) 56 (4) NR NR 97 (7) NR
2009
Hyder et al, 2056 396 (19) NR 9 (.5) 15 (.7) 65 (3) NR 19 (.9) 47 (2) 7 (.3)
2013
NR, Not reported.
(Rahbari et al, 2011a). This may range from a transient decline fail conservative management ultimately require liver transplan-
in liver function requiring no specific intervention (grade A) or tation to survive (Kauffmann & Fong, 2014). In the study
progress to fulminant liver and multisystem organ failure (grade by Hyder and colleagues (2013), only two patients (0.1%)
C) (Kauffmann & Fong, 2014). Mortality for PHLF patients required extracorporeal liver support followed by “rescue” liver
rises dramatically based on grade with rates for A, B, and C transplantation.
reported to be 0%, 12%, and 54%, respectively (Reissfelder
et al, 2011). Risk factors include large-volume blood loss, pro- Bile Leak
longed operative time, and major resection of greater than 50% Bile leakage is a common complication following hepatic resec-
of liver volume (Kauffmann & Fong, 2014; Nonami et al, tions, with an incidence ranging from 3.6% to 12% (Erdogan
1999). PHLF presents with persistent hyperbilirubinemia and et al, 2008). It occurs at the cut surface of the liver remnant,
coagulopathy in the setting of progressive multisystem organ from the closed stumps of hepatic ducts, and from injury to the
failure. Management is largely supportive, and severe cases extrahepatic bile duct. When present, bile leaks may prolong
require transfer to an ICU where intubation, dialysis, pressors, hospital stay, delay drain removal, or even require return to the
and transfusions can be coordinated for multiple failing organ operating room in severe cases. The ISGLS defined a bile leak
systems. Unfortunately, patients who progress to this point and as the drainage of intraabdominal fluid with an increased
bilirubin concentration (at least three times the serum bilirubin Chapter 24). Multivariate analysis performed by Jarnagin and
concentration) on or after postoperative day 3 (Koch et al, colleagues (2002) identified the number of hepatic segments
2011). This may be determined from the intraabdominal drains resected and operative blood loss as factors predictive of overall
placed at the time of surgery or through invasive interventions morbidity and mortality. Moreover, blood loss, transfusion, and
(i.e., percutaneous drain placement or relaparotomy). Grade A extent of resection have been correlated specifically with
bile leaks are transient with little to no clinical impact, grade B increased rates of postoperative liver failure and bile leaks
leaks require additional diagnostics and potentially percutane- (Brooke-Smith et al, 2015; Kauffmann & Fong, 2014; Nonami
ous drainage, and grade C leaks require relaparotomy for bile et al, 1999). In recognition of these associations, there has been
peritonitis. A prospective, multiinstitutional validation study of great interest in developing techniques aimed at reducing blood
949 hepatic resection demonstrated a 7.3% overall incidence loss and transfusions and limiting resections. One example of
of bile leak, of which 45%, 46%, and 9% were classified as this is the more widespread use of parenchymal-sparing
grades A, B, and C, respectively (Brooke-Smith et al, 2015). approaches, which several studies have reported to have less
Mortality was 7.2% in the setting of bile leak compared with perioperative morbidity without compromising oncologic
1.1% (P < .001) without it. Overall, 16 patients (23%) under- results (de Haas et al, 2008; Gold et al, 2008) (see Chapter
went endoscopic retrograde cholangiopancreatography (ERCP) 108). Another major advance in liver surgery has been the
or percutaneous transhepatic cholangiography, and 30 patients concept of low CVP anesthesia during hepatic resection. Mul-
(43.5%) required additional drainage, either percutaneously or tiple high-quality studies have demonstrated that maintaining
surgically. Interestingly, use of intraoperative drains was identi- a low CVP (<5 mm Hg) reduces blood loss and transfusion
fied as an independent predictor of bile leak but did not affect rates, decreases operative times, decreases renal dysfunction,
the chances of requiring an intervention. Moreover, 100% and and lowers overall morbidity (Li et al, 2014; Melendez et al,
91% of grade A and B leaks, respectively, were detected in cases 1998; Wang et al, 2006) (see Chapter 24).
with intraoperative drains. This suggests that most bile leaks, Techniques and devices intended to improve intraoperative
as defined by the ISGLS, are self-resolving with little to no efficiency and hemostasis during parenchymal transection
clinical impact, and the presence of intraoperatively placed have been studied extensively to determine their effect, if any,
drains simply allows a biochemical diagnosis. on postoperative complications (Pamecha et al, 2009; Poon,
It should be noted that bile leakage rates after complex 2007). Among the numerous techniques studied in random-
resections that also involve biliary reconstruction, for example, ized trials—crush-clamping, stapling systems, vessel sealing
hilar cholangiocarcinoma, are higher than after standard systems (i.e., LigaSure, Covidien, Boulder, CO), and multiple
resections and often involve more complex management (see options for ultrasonic dissection—there does not appear to be
Chapters 51, 52, and 103C). a single “best” option. This may be due to the fact that these
trials are usually performed at high-volume centers where the
Posthepatectomy Hemorrhage skill among experienced liver surgeons could make it difficult
Posthepatectomy hemorrhage (PHH) is defined by the ISGLS to demonstrate small differences among techniques (Arita
as any of the following in the setting of confirmed bleeding et al, 2005; Ikeda et al, 2009; Palavecino et al, 2010;
noted in drains or on imaging: (1) drop in hemoglobin level Patrlj et al, 2010). In a recent Cochrane Review analyzing
of greater than 3 g/dL, (2) any postoperative transfusion of seven randomized controlled trials that compared methods
packed red blood cells (PRBCs) for a falling hemoglobin, or for parenchymal division, the authors concluded that the
(3) the need for invasive intervention (e.g., embolization or crush-clamp technique was the procedure of choice (Pamecha
relaparotomy) (Rahbari et al, 2011b). As with other consensus et al, 2009). Their decision was based on the fact that paren-
definitions, PHH is graded from A to C such that grade A chymal transection devices were slower and significantly more
hemorrhage may require transfusion of up to 2 U of PRBCs, expensive, while offering no advantages in time reduction or
grade B requires more than 2 U of PRBCs, and grade C neces- reduced blood loss compared with crush-clamping (see
sitates invasive intervention. PHH usually occurs within 48 Chapter 103).
hours of surgery as venous bleeding from the raw surfaces of
the liver and diaphragm. This may be exacerbated by altera-
tions in outflow resulting from surgery, elevated central venous IMAGING AND IMAGE-GUIDED THERAPY OF
pressure (CVP), and valsalva maneuvers such as coughing and COMPLICATIONS AFTER HEPATECTOMY
defecating (Kauffmann & Fong, 2014; Lim et al, 2014). It is
also possible, albeit rare, for delayed injury to arterial struc- Imaging After Hepatectomy
tures to manifest postoperatively and for vascular ties or clips Ultrasound, CT, hepatobiliary iminodiacetic acid (HIDA)
to detach from short hepatic vein stumps along the vena cava. scan, and MRI/MRCP are noninvasive diagnostic procedures
Internal validation of the definition for PHH performed on a widely applied to the liver and biliary system. Postoperative
series of 835 hepatic resections demonstrated an incidence of bilomas, hematomas, and abscesses can be detected with these
3%, of which 14% were grade A, 43% grade B, and 43% grade modalities.
C (Rahbari et al, 2011b). The in-hospital mortality for grades The typical appearance of biloma on imaging is that of an
A, B, and C was 0%, 17%, and 50%, respectively. encapsulated fluid collection adjacent to the liver resection
plane. Rim enhancement of a fluid collection suggests infection,
Techniques for Reducing Intraoperative Blood Loss but this is not a specific finding. In some cases, aspiration may
Intraoperative blood loss, transfusion of PRBCs, and extent of be required to distinguish infected and noninfected collections.
hepatic resection have all been reported in numerous studies as A postoperative biloma, hematoma, or seroma might initially
major contributors to mortality and morbidity of all types (Jar- be sterile but can progress to become an abscess. Gas within a
nagin et al, 2002; Rahbari et al, 2011c; Wei et al, 2003) (see collection suggests infection or enteric leak. Topical hemostatic
agents used during surgery can also contain gas bubbles (see
Interventional Radiology Procedures Posthepatectomy
Chapters 18 and 19).
A HIDA scan, also known as hepatobiliary scintigraphy, is Intrahepatic Abscess Drainage
a nuclear medicine scan that can show biliary leaks (Sandoval Small pyogenic liver abscesses can be successfully treated with
et al, 1997). Although technetium-99m–labeled HIDA has antibiotics alone. Large (>3 cm) unilocular abscesses can be
largely been replaced by other radiotracers (which have treated by percutaneous drainage and antibiotics. Large multi-
improved liver uptake), the term “HIDA scan” remains in loculated abscesses have a lower success rate with percutaneous
common use. A normal HIDA scan initially shows radiotracer drainage, and might require surgery (Hope et al, 2008) (see
uptake in liver, followed by excretion of radiotracer into the bile Chapter 12). Intracavitary tPA can help drain multiloculated
ducts, gallbladder, and small bowel. Pooling of tracer elsewhere collections that are refractory to simple percutaneous drainage
indicates a bile leak. HIDA scans are typically two-dimensional (Beland et al, 2008). When biliary obstruction is present, relief
images acquired with a gamma camera, but fused 3D images of the obstructed biliary tree is mandatory for successful abscess
can also be obtained, using a single-photon emission computed treatment. Abscesses located near the dome of the liver may be
tomography. The 3D images can be helpful for more precise technically more difficult to drain without transgressing the
anatomic localization (see Chapter 17). pleura. Transpleural abscess drains carry a risk of empyema.
MRCP is an MRI protocol optimized for seeing the bile
ducts (see Chapter 19). It can show the anatomy of the biliary Interventional Management of Bilomas and Bile Leaks
tree and any associated bilomas. MRCP can also show a biliary After liver tumor resection, bile can leak from a bile duct injury,
leak, when it is performed using an intravenous contrast agent biliary-enteric anastomosis, or the cut surface of the liver. Bile
that is excreted into the bile ducts, such as gadoxetate (Eovist) leakage can cause bile peritonitis, as well as bilious fluid collec-
(Aduna et al, 2005). (Typically, MRCP is performed without tions that can become infected. These fluid collections can be
intravenous contrast.) MRCP has higher resolution than a drained percutaneously, under CT or ultrasound guidance
HIDA scan and shows the biliary tree more clearly than CT. (Fig. 27.3). When an infected biloma is drained, the fluid can
Acute and subacute hematomas appear as high-density col- initially appear purulent, and then it may turn bilious if there
lections on CT. In addition, active bleeding can be demon- is a continued bile leak after the infection clears. Ideally, the
strated by contrast media extravasation on contrast-enhanced drain should be placed near the bile leak to provide optimal
CT (see Chapter 18). drainage (see Chapter 42).
*
*
A B
FIGURE 27.3 A 46-year-old male patient with hepatitis B infection, status post partial left hepatectomy for hepatocellular carcinoma.
A, Postoperative computed tomography showed a fluid collection at the liver resection margin (asterisk), which was treated with percutaneous
catheter placement. B, Contrast injection into the cavity under fluoroscopy showed a small biloma cavity (asterisk) communicating with the common
hepatic bile duct (arrow), with contrast then flowing into the duodenum (arrowhead). Endoscopic retrograde cholangiopancreatography was per-
formed, and a plastic stent was placed into the common bile duct to decrease the output of the fistula.
The amount of drain output allows monitoring of the has side holes both above and below the obstruction. An
amount of bile leak over time. Small bile leaks can resolve internal/external biliary drain can be capped if there is no leak,
spontaneously (Viganò et al, 2008). Persistent drainage greater infection, or significant blood in the bile, and should be flushed
than 100 mL/day (at 10 days’ posthepatectomy) may be treated daily with 10 mL normal saline to maintain patency.
with endoscopic or percutaneous biliary drain placement, Biliary drains are typically exchanged every 3 months to
which decompresses the biliary system and diverts bile flow prevent clogging but are exchanged more frequently if cholan-
away from the defect in the bile ducts, allowing the leak to heal gioplasty or other interventions are planned. An over-the-wire
(see Fig. 27.3). Complete transection of a bile duct at the hilum cholangiogram can be performed through a sheath that does
typically requires surgical repair (see Chapter 42). not cross the bile duct injury, to evaluate for persistent leak or
Percutaneous biliary drainage is the preferred treatment for stenosis. If the bile duct injury has resolved on the cholangio-
high bile duct injuries (at or above the bifurcation), and biliary- gram, then an external biliary drain can be placed to maintain
enteric anastomotic leaks, both of which are difficult to access to the bile ducts. This external drain should be capped
access endoscopically. ERCP is less invasive than PBD, and is for 2 weeks without flushing. If the patient passes the capping
the preferred treatment for leaks from the cut surface of trial (no fever, no significant leakage around the tube, no rise
the liver and for accessible common duct leaks (see Chapters in bilirubin), then the drain can be safely removed.
29 and 30). Nonsurgical management of benign biliary strictures are
Percutaneous biliary drainage leads to resolution of 88% to typically managed with endoscopic placement of a plastic biliary
100% of postoperative bile leaks, after an average of 1 to 3 stent, or percutaneous internal/external biliary drainage. Chol-
months of drainage (Cozzi et al, 2006; Ernst et al, 1999). angioplasty can be performed using a high-pressure balloon
Intractable bile leaks can be managed with surgery, portal vein during biliary drain placement or exchange. High pressures and
embolization (Hai et al, 2012), or bile duct embolization with prolonged cholangioplasty (as long as 15 minutes) are typically
N-butyl cyanoacrylate glue (Vu et al, 2006). required to overcome the dense fibrous tissue around biliary
strictures. An 8-mm balloon can be used for intrahepatic stric-
Interventional Management of Biliary Strictures tures and a 10- to 12-mm balloon for common duct strictures.
Postoperative biliary strictures can occur after hepatectomy (see Cholangioplasty can be repeated at 2- to 14-day intervals
Chapter 103), cholecystectomy (see Chapters 38 and 42), cho- (Cantwell et al, 2008; Zajko et al, 1995). For benign biliary
ledochojejunostomy, liver transplant (see Chapter 120), and strictures, cholangioplasty and internal/external biliary drainage
other procedures. These strictures can be caused by direct have a long-term (25 years) primary success rate of 59% and a
biliary injury from surgery, ischemia, or recurrent tumor. secondary success rate of 80% (Cantwell et al, 2008). Failure
Benign and malignant strictures can be difficult to distinguish of these approaches in otherwise healthy patients should prompt
on imaging, and biopsy may be required. Biliary strictures can consideration of operative management (see Chapter 42).
cause jaundice as well as cholangitis. Malignant biliary obstruction can be relieved with a biliary
Low bile duct obstruction (common bile duct or common drainage catheter, to treat cholangitis, pruritus, or to lower bili-
hepatic duct not involving the bifurcation) can be relieved via rubin for chemotherapy. A metal biliary stent can be placed
ERCP and placement of a plastic or metal biliary stent across (percutaneously or endoscopically) for palliation of unresect-
the obstruction (see Chapter 29). High bile duct obstruction able symptomatic biliary obstruction in patients with limited
and biliary-enteric anastomotic strictures, both of which are life expectancy. Metal biliary stents placed for malignant
difficult to access endoscopically, can be treated with percuta- obstruction remain patent for an average of 11 months
neous biliary drain or metal stent placement (see Chapter 30). (Dahlstrand et al, 2009).
Metal stents are typically only used for malignant obstruction,
as they have a limited patency rate—30 months on average Bleeding Complications
when used for benign disease (Tesdal et al, 2005). See “Angiography, Embolization, and Covered Stent Place-
If the biliary stricture cannot be crossed percutaneously, an ment,” earlier.
external biliary drain can be placed. If the biliary stricture can
be crossed, an internal/external biliary drain is placed, which References are available at expertconsult.com.
Chapter 27 Postoperative complications requiring intervention: diagnosis and management 474.e1
REFERENCES Cope C, et al: Percutaneous management of chronic pancreatic duct

strictures and external fistulas with long-term results, J Vasc Interv
Adam R, et al: Hepatic resection for noncolorectal nonendocrine liver Radiol 12(1):104–110, 2001.
metastases: analysis of 1,452 patients and development of a prog- Correa-Gallego C, et al: Contemporary experience with postpancre-
nostic model, Ann Surg 244(4):524–535, 2006. atectomy hemorrhage: results of 1,122 patients resected between
Adham M, et al: Central pancreatectomy: single-center experience of 2006 and 2011, J Am Coll Surg 215(5):616–621, 2012.
50 cases, Arch Surg 143:175–180, discussion 180–181, 2008. Cozzi G, et al: Percutaneous transhepatic biliary drainage in the man-
Aduna M, et al: Bile duct leaks after laparoscopic cholecystectomy: value agement of postsurgical biliary leaks in patients with nondilated
of contrast-enhanced MRCP, Abdom Imaging 30(4):480–487, 2005. intrahepatic bile ducts, Cardiovasc Interv Radiol 29(3):380–388,
Akizuki E, et al: Reconsideration of postoperative oral intake tolerance 2006.
after pancreaticoduodenectomy: prospective consecutive analysis of Crippa S, et al: Middle pancreatectomy: indications, short- and long-
delayed gastric emptying according to the ISGPS definition and the term operative outcomes, Ann Surg 246:69–76, 2007.
amount of dietary intake, Ann Surg 249(6):986–994, 2009. Dahlstrand U, et al: Primary patency of percutaneously inserted self-
Allen PJ: Pancreatic adenocarcinoma: putting a hump in survival, J Am expanding metallic stents in patients with malignant biliary obstruc-
Coll Surg 205(Suppl 4):S76–S80, 2007. tion, HPB (Oxford) 11(4):358–363, 2009.
Allen PJ, et al: Pasireotide for postoperative pancreatic fistula, N Engl Darwin P, et al: Jackson Pratt drain fluid-to-serum bilirubin concentra-
J Med 370(21):2014–2022, 2014. tion ratio for the diagnosis of bile leaks, Gastrointest Endosc 71(1):99–
Arita J, et al: Randomized clinical trial of the effect of a saline-linked 104, 2010.
radiofrequency coagulator on blood loss during hepatic resection, Br De Haas RJ, et al: R1 resection by necessity for colorectal liver metas-
J Surg 92(8):954–959, 2005. tases: is it still a contraindication to surgery? Ann Surg 248(4):626–
Baker TA, et al: Role of interventional radiology in the management 637, 2008.
of complications after pancreaticoduodenectomy, Am J Surg DeOliveira ML, et al: Assessment of complications after pancreatic
195(3):386–390, discussion 390, 2008. surgery: a novel grading system applied to 633 patients undergoing
Bassi C, et al: Pancreatic fistula rate after pancreatic resection. The pancreaticoduodenectomy, Ann Surg 244(6):931–937, discussion
importance of definitions, Dig Surg 21(1):54–59, 2004. 937–939, 2006.
Bassi C, et al: Postoperative pancreatic fistula: an international study Dimick JB, et al: National trends in the use and outcomes of hepatic
group (ISGPF) definition, Surgery 138(1):8–13, 2005. resection, J Am Coll Surg 199(1):31–38, 2004.
Bassi C, et al: Management of complications after pancreaticoduode- Dindo D, et al: Classification of surgical complications: a new proposal
nectomy in a high volume centre: results on 150 consecutive patients, with evaluation in a cohort of 6336 patients and results of a survey,
Dig Surg 18(6):453–457, discussion 458, 2001. Ann Surg 240(2):205–213, 2004.
Beane JD, et al: Optimal management of delayed gastric emptying after Epelboym I, et al: Quality of life in patients after total pancreatectomy
pancreatectomy: an analysis of 1,089 patients, Surgery 156(4):939– is comparable with quality of life in patients who undergo a partial
946, 2014. pancreatic resection, J Surg Res 187:189–196, 2014.
Beland MD, et al: Complex abdominal and pelvic abscesses: efficacy Erdogan D, et al: Incidence and management of bile leakage after
of adjunctive tissue-type plasminogen activator for drainage, Radiol- partial liver resection, Dig Surg 25(1):60–66, 2008.
ogy 247(2):567–573, 2008. Ernst O, et al: Biliary leaks: treatment by means of percutaneous tran-
Billings BJ, et al: Quality-of-life after total pancreatectomy: is it really shepatic biliary drainage, Radiol 211(2):345–348, 1999.
that bad on long-term follow-up? J Gastrointest Surg 9:1059–1066, Finks JF, et al: Trends in hospital volume and operative mortality for
discussion 1066–1067, 2005. high-risk surgery, N Engl J Med 364(22):2128–2137, 2011.
Boas FE, et al: Targeted transgastric drainage of isolated pancreatic Foster JH, Berman MM: Solid liver tumors, Major Probl Clin Surg
duct segments to cure persistent pancreaticocutaneous fistulas from 22:1–342, 1977.
pancreatitis, J Vasc Interv Radiol 26(2):247–251, 2015. Friess H, et al: Randomized controlled multicentre study of the preven-
Braga M, et al: A prognostic score to predict major complications after tion of complications by octreotide in patients undergoing surgery
pancreaticoduodenectomy, Ann Surg 254(5):702–707, discussion for chronic pancreatitis, Br J Surg 82(9):1270–1273, 1995.
707–708, 2011. Goh BKP, et al: Critical appraisal of 232 consecutive distal pancreatec-
Braga M, et al: ESPEN Guidelines on Parenteral Nutrition: surgery, tomies with emphasis on risk factors, outcome, and management of
Clin Nutr (Edinburgh) 28(4):378–386, 2009. the postoperative pancreatic fistula: a 21-year experience at a single
Brooke-Smith M, et al: Prospective evaluation of the International institution, Arch Surg 143:956–965, 2008.
Study Group for Liver Surgery definition of bile leak after a liver Gold JS, et al: Increased use of parenchymal-sparing surgery for bilat-
resection and the role of routine operative drainage: an international eral liver metastases from colorectal cancer is associated with
multicentre study, HPB (Oxford) 17(1):46–51, 2015. improved mortality without change in oncologic outcome: trends in
Büchler M, et al: Role of octreotide in the prevention of postoperative treatment over time in 440 patients, Ann Surg 247(1):109–117,
complications following pancreatic resection, Am J Surg 163(1):125– 2008.
130, discussion 130–131, 1992. Grobmyer SR, et al: Defining morbidity after pancreaticoduodenec-
Büchler MW, et al: Pancreatic fistula after pancreatic head resection, tomy: use of a prospective complication grading system, J Am Coll
Br J Surg 87(7):883–889, 2000. Surg 204(3):356–364, 2007.
Cabay JE, et al: Percutaneous catheter drainage of external fistulas of Grützmann R, et al: Evaluation of the International Study Group of
the pancreatic ducts, Eur Radiol 8(3):445–448, 1998. Pancreatic Surgery definition of post-pancreatectomy hemorrhage in
Cameron JL, He J: Two thousand consecutive pancreaticoduodenec- a high-volume center, Surgery 151(4):612–620, 2012.
tomies, J Am Coll Surg 220(4):530–536, 2015. Hai S, et al: Portal vein embolization for an intractable bile leakage
Cantwell CP, et al: Thirty years’ experience with balloon dilation of after hepatectomy, Clin J Gastroenterol 5(4):287–291, 2012.
benign postoperative biliary strictures: long-term outcomes, Radiol- Heiss P, et al: Delayed visceral arterial hemorrhage following Whipple’s
ogy 249(3):1050–1057, 2008. procedure: minimally invasive treatment with covered stents, Ann
Cescon M, et al: Trends in perioperative outcome after hepatic resec- Surg Oncol 15(3):824–832, 2008.
tion: analysis of 1500 consecutive unselected cases over 20 years, Hirono S, et al: A central pancreatectomy for benign or low-grade
Ann Surg 249(6):995–1002, 2009. malignant neoplasms, J Gastrointest Surg 13:1659–1665, 2009.
Choi SH, et al: Delayed hemorrhage after pancreaticoduodenectomy, Hope WW, et al: Optimal treatment of hepatic abscess, Am Surg
J Am Coll Surg 199(2):186–191, 2004. 74(2):178–182, 2008.
Chong CC-N, et al: Risk factors for delayed gastric emptying after Huang Z, et al: Hepatic resection: an analysis of the impact of operative
pancreatoduodenectomy, Surg Pract 19(1):22–28, 2015. and perioperative factors on morbidity and mortality rates in 2008
Clavien PA, et al: The Clavien-Dindo classification of surgical compli- consecutive hepatectomy cases, Chin Med J 122(19):2268–2277,
cations: five-year experience, Ann Surg 250(2):187–196, 2009. 2009.
Clavien PA, et al: Proposed classification of complications of surgery Hyder O, et al: A risk model to predict 90-day mortality among patients
with examples of utility in cholecystectomy, Surgery 111(5):518–526, undergoing hepatic resection, J Am Coll Surg 216(6):1049–1056,
1992. 2013.
Ikeda M, et al: The vessel sealing system (LigaSure) in hepatic Nonami T, et al: Blood loss and ICG clearance as best prognostic
resection: a randomized controlled trial, Ann Surg 250(2):199–203, markers of post-hepatectomy liver failure, Hepatogastroenterology
2009. 46(27):1669–1672, 1999.
Imamura H, et al: One thousand fifty-six hepatectomies without mor- Ocuin LM, et al: Comparison of central and extended left pancreatec-
tality in 8 years, Arch Surg 138(11):1198–1206, discussion 1206, tomy for lesions of the pancreatic neck, Ann Surg Oncol 15:2096–
2003. 2103, 2008.
Jarnagin WR, et al: Improvement in perioperative outcome after Ohwada S, et al: Low-dose erythromycin reduces delayed gastric emp-
hepatic resection: analysis of 1,803 consecutive cases over the past tying and improves gastric motility after Billroth I pylorus-preserving
decade, Ann Surg 236(4):397–406, discussion 406–407, 2002. pancreaticoduodenectomy, Ann Surg 234(5):668–674, 2001.
Karpoff HM, et al: Results of Total Pancreatectomy for Adenocarci- Otah E, et al: Visceral artery pseudoaneurysms following pancreato-
noma of the Pancreas, Arch Surg 136:44, 2001. duodenectomy, Arch Surg 137(1):55–59, 2002.
Kauffmann R, Fong Y: Post-hepatectomy liver failure, Hepatobiliary O’Toole D, et al: Diagnosis and management of pancreatic fistulae
Surg Nutr 3(5):238–246, 2014. resulting in pancreatic ascites or pleural effusions in the era of helical
Kleef J, et al: Distal pancreatectomy: risk factors for surgical failure in CT and magnetic resonance imaging, Gastroenterol Clin Biol 31(8–9
302 consecutive cases, Ann Surg 245:573–582, 2007. Pt 1):686–693, 2007.
Klek S, et al: Enteral and parenteral nutrition in the conservative treat- Palavecino M, et al: Two-surgeon technique of parenchymal transec-
ment of pancreatic fistula: a randomized clinical trial, Gastroenterol- tion contributes to reduced transfusion rate in patients undergoing
ogy 141(1):157–163, 163.e1, 2011. major hepatectomy: analysis of 1,557 consecutive liver resections,
Klempa I, et al: Prevention of postoperative pancreatic complications Surgery 147(1):40–48, 2010.
following duodenopancreatectomy using somatostatin, Chirurg Pamecha V, et al: Techniques for liver parenchymal transection: a
50(7):427–431, 1979. meta-analysis of randomized controlled trials, HPB (Oxford) 11(4):
Kneuertz PJ, et al: Risk of morbidity and mortality following hepato- 275–281, 2009.
pancreato-biliary surgery, J Gastrointest Surg 16(9):1727–1735, Park JS, et al: Clinical validation and risk factors for delayed gastric
2012. emptying based on the International Study Group of Pancreatic
Koch M, et al: Bile leakage after hepatobiliary and pancreatic surgery: Surgery (ISGPS) Classification, Surgery 146(5):882–887, 2009.
a definition and grading of severity by the International Study Group Patrlj L, et al: Combined blunt-clamp dissection and LigaSure ligation
of Liver Surgery, Surgery 149(5):680–688, 2011. for hepatic parenchyma dissection: postcoagulation technique, J Am
Koffron AJ, et al: Evaluation of 300 minimally invasive liver resections Coll Surg 210(1):39–44, 2010.
at a single institution: less is more, Ann Surg 246(3):385–392, dis- Pederzoli P, et al: Efficacy of octreotide in the prevention of complica-
cussion 392–394, 2007. tions of elective pancreatic surgery. Italian Study Group, Br J Surg
Kooby DA, et al: Left-sided pancreatectomy: a multicenter comparison 81(2):265–269, 1994.
of laparoscopic and open approaches, Ann Surg 248:438–446, 2008. Poon RTP: Current techniques of liver transection, HPB (Oxford)
Li Z, et al: Controlled low central venous pressure reduces blood loss 9(3):166–173, 2007.
and transfusion requirements in hepatectomy, World J Gastroenterol Poon RT, et al: Improving perioperative outcome expands the role of
20(1):303–309, 2014. hepatectomy in management of benign and malignant hepatobiliary
Lim C, et al: Reoperation for post-hepatectomy hemorrhage: increased diseases: analysis of 1222 consecutive patients from a prospective
risk of mortality, Langenbecks Arch Surg 399(6):735–740, 2014. database, Ann Surg 240(4):698–708, discussion 708–710, 2004.
Lowy AM, et al: Prospective, randomized trial of octreotide to prevent Pratt W, et al: Postoperative pancreatic fistulas are not equivalent after
pancreatic fistula after pancreaticoduodenectomy for malignant proximal, distal, and central pancreatectomy, J Gastrointest Surg
disease, Ann Surg 226(5):632–641, 1997. 10(9):1264–1278, discussion 1278–1279, 2006.
Machado NO: Pancreatic fistula after pancreatectomy: definitions, risk Puppala S, et al: Hemorrhagic complications after Whipple surgery:
factors, preventive measures, and management—review, Int J Surg imaging and radiologic intervention, AJR Am J Roentgenol 196(1):
2012:602478, 2012. 192–197, 2011.
Malleo G, et al: Delayed gastric emptying after pylorus-preserving pan- Rahbari NN, et al: Posthepatectomy liver failure: a definition and
creaticoduodenectomy: validation of International Study Group of grading by the International Study Group of Liver Surgery (ISGLS),
Pancreatic Surgery classification and analysis of risk factors, HPB Surgery 149(5):713–724, 2011a.
(Oxford) 12(9):610–618, 2010. Rahbari NN, et al: Post-hepatectomy haemorrhage: a definition and
Malleo G, et al: Diagnosis and management of postoperative pancre- grading by the International Study Group of Liver Surgery (ISGLS),
atic fistula, Langenbecks Arch Surg 399(7):801–810, 2014. HPB (Oxford) 13(8):528–535, 2011b.
Melendez JA, et al: Perioperative outcomes of major hepatic resections Rahbari NN, et al: The predictive value of postoperative clinical risk
under low central venous pressure anesthesia: blood loss, blood scores for outcome after hepatic resection: a validation analysis in
transfusion, and the risk of postoperative renal dysfunction, J Am 807 patients, Ann Surg Oncol 18(13):3640–3649, 2011c.
Coll Surg 187(6):620–625, 1998. Rajarathinam G, et al: Post pancreaticoduodenectomy haemorrhage:
Mezhir JJ: Management of complications following pancreatic outcome prediction based on new ISGPS Clinical severity grading,
resection: an evidence-based approach, J Surg Oncol 107(1):58–66, HPB (Oxford) 10(5):363–370, 2008.
2013. Reddy S, et al: Total pancreatectomy for pancreatic adenocarcinoma:
Mezhir JJ, et al: A matched case-control study of preoperative biliary evaluation of morbidity and long-term survival, Ann Surg 250:282–
drainage in patients with pancreatic adenocarcinoma: routine drain- 287, 2009.
age is not justified, J Gastrointest Surg 13(12):2163–2169, 2009. Rees M, et al: Evaluation of long-term survival after hepatic resection
Montorsi M, et al: Efficacy of octreotide in the prevention of pancreatic for metastatic colorectal cancer: a multifactorial model of 929
fistula after elective pancreatic resections: a prospective, controlled, patients, Ann Surg 247(1):125–135, 2008.
randomized clinical trial, Surgery 117(1):26–31, 1995. Reid-Lombardo KM, et al: Pancreatic anastomotic leakage after pan-
Müller MW, et al: Is there still a role for total pancreatectomy? Ann creaticoduodenectomy in 1,507 patients: a report from the Pancre-
Surg 246:966–974, discussion 974–975, 2007. atic Anastomotic Leak Study Group, J Gastrointest Surg 11(11):
Muscari F, et al: Risk factors for mortality and intra-abdominal com- 1451–1458, discussion 1459, 2007.
plications after pancreatoduodenectomy: multivariate analysis in 300 Reissfelder C, et al: Postoperative course and clinical significance of
patients, Surgery 139(5):591–598, 2006. biochemical blood tests following hepatic resection, Br J Surg 98(6):
Nathan H, et al: The volume-outcomes effect in hepato-pancreato- 836–844, 2011.
biliary surgery: hospital versus surgeon contributions and specificity Ricci C, et al: Late postpancreatectomy hemorrhage after pancreatico-
of the relationship, J Am Coll Surg 208(4):528–538, 2009. duodenectomy: is it possible to recognize risk factors? JOP 13(2):
Nathan H, et al: National trends in surgical procedures for hepatocel- 193–198, 2012.
lular carcinoma: 1998-2008, Cancer 118(7):1838–1844, 2012. Roggin KK, et al: Central pancreatectomy revisited, J Gastrointest Surg
Nicholson T, et al: Hepatic artery angiography and embolization for 10:804–812, 2006.
hemobilia following laparoscopic cholecystectomy, Cardiovasc Interv Röthlin MA, et al: Percutaneous drainage of abdominal abscesses: are
Radiol 22(1):20–24, 1999. large-bore catheters necessary? Eur J Surg 164(6):419–424, 1998.
Chapter 27 Postoperative complications requiring intervention: diagnosis and management 474.e3
Sandoval BA, et al: Cholescintigraphy in the diagnosis of bile leak after Welsch T, et al: Evaluation of the International Study Group of Pan-
laparoscopic cholecystectomy, Am Surg 63(7):611–616, 1997. creatic Surgery definition of delayed gastric emptying after pancre-
Sauvanet A, et al: Medial pancreatectomy: a multi-institutional retro- atoduodenectomy in a high-volume centre, Br J Surg 97(7):1043–1050,
spective study of 53 patients by the French Pancreas Club, Surgery 2010.
132:836–843, 2002. Welsch T, et al: Critical appraisal of the International Study Group of
Scialpi M, et al: Imaging evaluation of post pancreatic surgery, Eur J Pancreatic Surgery (ISGPS) consensus definition of postoperative
Radiol 53(3):417–424, 2005. hemorrhage after pancreatoduodenectomy, Langenbecks Arch Surg
Shinchi H, et al: The usefulness of drain data to identify a clinically 396(6):783–791, 2011.
relevant pancreatic anastomotic leak after pancreaticoduodenec- Wente MN, et al: Delayed gastric emptying (DGE) after pancreatic
tomy? J Gastrointest Surg 10(4):490–498, 2006. surgery: a suggested definition by the International Study Group of
Sohn T: Pancreaticoduodenectomy role of interventional radiologists Pancreatic Surgery (ISGPS), Surgery 142(5):761–768, 2007a.
in managing patients and complications, J Gastrointest Surg 7(2):209– Wente MN, et al: Postpancreatectomy hemorrhage (PPH): an Interna-
219, 2003. tional Study Group of Pancreatic Surgery (ISGPS) definition,
Stitzenberg KB, et al: Centralization of cancer surgery: implications for Surgery 142(1):20–25, 2007b.
patient access to optimal care, J Clin Oncol 27(28):4671–4678, 2009. Winter JM, et al: 1423 pancreaticoduodenectomies for pancreatic
Suzuki K, et al: Stent-graft treatment for bleeding superior mesenteric cancer: a single-institution experience, J Gastrointest Surg 10(9):1199–
artery pseudoaneurysm after pancreaticoduodenectomy, Cardiovasc 1210, discussion 1210–1211, 2006.
Interv Radiol 32(4):762–766, 2009. Yamamoto S, et al: Transcatheter coil embolization of splenic artery
Tesdal IK, et al: Metallic stents for treatment of benign biliary obstruc- aneurysm, Cardiovasc Interv Radiol 31(3):527–534, 2008.
tion: a long-term study comparing different stents, J Vasc Interv Yehuda A, et al: Delayed post pancreatectomy hemorrhage: incidence,
Radiol 16(11):1479–1487, 2005. clinical characteristics, risk factors and management, J Vasc Med Surg
Tien Y-W, et al: Angiography is indicated for every sentinel bleed after 2:3, 2014.
pancreaticoduodenectomy, Ann Surg Oncol 15(7):1855–1861, 2008. Yekebas EF, et al: Postpancreatectomy hemorrhage: diagnosis and
Tsai C-C, et al: Transcatheter arterial coil embolization of iatrogenic treatment: an analysis in 1669 consecutive pancreatic resections,
pseudoaneurysms after hepatobiliary and pancreatic interventions, Ann Surg 246(2):269–280, 2007.
Hepatogastroenterology 54(73):41–46, 2007. Yeo CJ, et al: Erythromycin accelerates gastric emptying after pancre-
Valero V, et al: A novel risk scoring system reliably predicts readmission aticoduodenectomy. A prospective, randomized, placebo-controlled
after pancreatectomy, J Am Coll Surg 220(4):701–713, 2015. trial, Ann Surg 218(3):229–237, discussion 237–238, 1993.
Viganò L, et al: Bile leak after hepatectomy: predictive factors of spon- Yeo CJ, et al: Does prophylactic octreotide decrease the rates of pan-
taneous healing, Am J Surg 196(2):195–200, 2008. creatic fistula and other complications after pancreaticoduodenec-
Vin Y, et al: Management and outcomes of postpancreatectomy fistula, tomy? Results of a prospective randomized placebo-controlled trial,
leak, and abscess: results of 908 patients resected at a single institu- Ann Surg 232(3):419–429, 2000.
tion between 2000 and 2005, J Am Coll Surg 207(4):490–498, 2008. Yeo CJ, et al: Pancreaticoduodenectomy with or without distal gastrec-
Voss M, Pappas T: Pancreatic fistula, Curr Treat Options Gastroenterol tomy and extended retroperitoneal lymphadenectomy for periampul-
5(5):345–353, 2002. lary adenocarcinoma, part 2: randomized controlled trial evaluating
Vu DN, et al: Biliary duct ablation with N-butyl cyanoacrylate, J Vasc survival, morbidity, and mortality, Ann Surg 236(3):355–366, dis-
Interv Radiol 17(1):63–69, 2006. cussion 366–368, 2002.
Wang W-D, et al: Low central venous pressure reduces blood loss in Zajko AB, et al: Transhepatic balloon dilation of biliary strictures in
hepatectomy, World J Gastroenterol 12(6):935–939, 2006. liver transplant patients: a 10-year experience, J Vasc Interv Radiol
Watanabe Y, et al: Long-Term Outcomes After Total Pancreatectomy: 6(1):79–83, 1995.
Special Reference to Survivors’ Living Conditions and Quality of Zink SI, et al: Pancreaticoduodenectomy: frequency and outcome
Life, World J Surg 39(5):1231–1239, 2015. of post-operative imaging-guided percutaneous drainage, Abdom
Wei AC, et al: Risk factors for perioperative morbidity and mortality Imaging 34(6):767–771, 2009.
after extended hepatectomy for hepatocellular carcinoma, Br J Surg
90(1):33–41, 2003.
Wellner UF, et al: Postpancreatectomy hemorrhage–incidence, treat-
ment, and risk factors in over 1,000 pancreatic resections, J Gastro-
intest Surg 18(3):464–475, 2014.
CHAPTER 28
Quality of life and hepatobiliary tumors
Julie N. Leal, Piera Marie Cote Robson, and Michael I. D’Angelica
QUALITY OF LIFE AND HEALTH-RELATED associated with improved survival, were severely toxic, and it
QUALITY OF LIFE: WHAT LIES BENEATH THE became apparent that assessing the “quality of survival” was
SURVIVAL CURVE? essential (Izsak & Medalie, 1971). However, most early studies
reporting on HRQoL in oncology simply focused on a single
Quality of life (QoL) is a complex construct that varies during parameter of QoL as assessed by functional indices, such as the
time and across cultures, and cannot be determined by a single Karnofsky Performance Score (KPS) (Karnofsky & Burchnel,
parameter or measurement. In the very broadest of terms, QoL 1948) and the Eastern Cooperative Oncology Group (ECOG)
has been defined as individuals’ perceptions of their position in performance status (Oken et al, 1982).
life in the context of the culture and value systems in which These initial studies lacked multidimensional assessment,
they live, and in relation to their goals, expectations, standards, with no patient-specific interpretation, and by current stan-
and concerns (World Health Organization [WHO], 1993). QoL dards are not true representations of HRQoL.
is multidimensional and influenced by many factors outside the Since these early ventures, the field of qualitative analysis in
realm of health/disease, including financial security, job satis- medicine has advanced significantly. Review of published
faction, personal freedom, and so forth. In the field of medi- HRQoL data from 1990-2004 by Efficace and colleagues
cine, and more specifically surgery, the true interest lies in the (2007) reveal a significant learning curve in the conduct and
aspects of QoL that are influenced by both underlying health/ reporting of HRQoL. Based on a checklist of minimal criteria,
disease status and treatment. Rudimentary to our understand- 39.3% of studies conducted before 2000 were considered
ing of the concept of health-related quality of life (HRQoL) is robust compared with 64.3% of studies published after 2000.
not only comprehension of what QoL is but also an essential It is encouraging that the observed increase in the volume of
appreciation for what “health” is. In the 1946 WHO constitu- published HRQoL studies is paralleled by improved outcome
tion, “health” is defined as a state of complete physical, mental, measurement and reporting.
and social well-being, not merely the absence of disease or These improvements in the study of HRQoL have led to
infirmity (WHO, 1946). Although no globally accepted defini- increasing acceptance of its importance as a valid measureable
tion exists, one can infer from QoL and health definitions that outcome in oncology and surgery. The true magnitude of the
HRQoL refers to the effect of illness and its treatments on value of HRQoL assessment is best highlighted by a policy
functional status, disease-related symptomatology, and psycho- statement from the American Society of Clinical Oncology in
logical and social well-being as perceived by the individual 1996, in which it was said that “QoL as a clinical endpoint is
(Roila & Cortesi, 2001). second in importance only to survival” (Sloan & Dueck, 2004).
Physician concern regarding the subjective impact of disease Nearly two decades since this initial policy statement, the
and its treatment on patients has always been an integral part majority of physicians and surgeons would agree that QoL
of clinical medicine. Since the time of Hippocrates, the simple evaluation is important, and most have a basic understanding
mandate “first, do no harm” has been indoctrinated into the of what is meant by HRQoL outcomes. However, skeptics
medical profession. In the past, the paternalistic nature of med- remain. Critics argue that HRQoL measurement is too subjec-
icine did not ignore patients’ subjective experience of disease, tive, and reported outcomes are esoteric with a lack of clarity
but rather, its interpretation and use in decision making was in terms of the clinical implication/meaning of numeric changes,
thought to be implicit and at the discretion of the doctor. It was limiting any reasonable clinical application (Whalen & Ferrans,
not until the patient-centered model of medical care emerged 2001).
that explicit and standardized measurement of HRQoL moved To address such concerns, Wilson and Cleary (1995) devel-
to the forefront of many clinical research efforts. oped a conceptual framework in which the relationships among
Formal evaluation of HRQoL is rooted in the study of different health outcomes are outlined. This model delineates
chronic diseases (i.e., diabetes, arthritis, claudication, etc.), in these relationships as a continuum from basic biologic pro-
which cure is not the therapeutic goal and patients are often cesses to symptoms to psychological and social implications.
required to adjust and accommodate to long-term functional The model flow is likened to the idea of moving from the cel-
limitations. As is common in any new field of study, initial lular level, to the individual level, to the individual in the
endeavors were fraught with complications, from ambiguity of context of society. In addition, it highlights the fact that indi-
the definition of QoL to the use of an ad hoc questionnaire that vidual patient characteristics and environmental factors will
lacked standardization, validation, and overall generalizability. most certainly contribute to the interpretation of symptoms,
In the field of oncology, early chemotherapeutics, although functional status, and patients’ perception of health. Although
475
the linkages in this model are simplistic, it provides a basic physical functioning domain have been shown to be an inde-
template to begin evaluating the complex interrelatedness of pendent predictor of survival for multiple different cancers
clinical parameters and HRQoL. It is hoped that characterizing (Gotay et al, 2008; Quinten et al, 2011, 2014).
and elucidating these relationships will improve clinical applica- More specific to hepatobiliary surgery, global QoL, social
tion of HRQoL outcomes. well-being, and physical functioning have all been shown to
Recently, much press has been given to the importance of independently predict survival outcomes in the patient with
“patient-reported outcomes” (PRO). This term is not synony- colorectal liver metastases (CRLMs) and to add to the prog-
mous with HRQoL, but rather encompasses a much wider nostic value of survival models, including standard biomedical
spectrum of outcomes. Basically, PROs include any aspect of data (Earlam et al, 1996; Efficace et al, 2006; Maisey et al,
the patient’s health status reported by the patient without inter- 2002). Similar findings have been observed among patients
pretation by the clinician or anyone else (U.S. Food and Drug with hepatocellular carcinoma (HCC) (Yeo et al, 2006).
Administration, 2006). Furthermore, instruments such as the Whether HRQoL simply reflects a highly sensitive measure of
KPS or the ECOG performance score, although useful and well patients’ overall health status not evaluated elsewhere, or
validated, are not truly HRQoL measurements. They simply whether HRQoL impacts other important areas, such as self-
measure one functional domain without evaluation of impor- care/treatment adherence and thus survival, is unknown. What
tant factors such as disease-related symptoms and psychological is clear is that measurement of HRQoL to aid informed consent
and/or social functioning. Furthermore, the KPS and ECOG and clinical decision making, to facilitate physician-patient rela-
scores are typically reported based on the clinician’s perception tionship and manage expectations, and to help improve prog-
of the patient’s status, which has been shown to differ signifi- nostication is requisite to optimize care of surgical patients.
cantly from patients’ perceptions (Schag et al, 1984). Many As with any outcome measure, it is essential to identify the
symptom scales also exist and have been reported as measure- clinical scenarios in which assessment is appropriate. Conse-
ments of HRQoL; however, they fail to place symptoms in the quently, it is incumbent upon surgeons to define these relevant
context of individual patients. Bona fide HRQoL assessment settings where HRQoL end points will result in clinically
requires multidimensional evaluation with patient input as to meaningful/actionable outcomes. It is suggested that the most
the impact of the disease/illness experience in the context of his pertinent scenarios for surgical patients include evaluation of
or her own life. Only then does it truly represent “what lies palliative interventions or procedures in which survival out-
beneath the survival curve.” comes are thought to be equivocal and/or survival outcomes
similar, but morbidity/side-effect profiles differ significantly
HEALTH-RELATED QUALITY OF LIFE (Blazeby et al, 2006; Bruner et al, 2004; Byrne et al, 2007;
AS AN OUTCOME MEASURE IN SURGERY: Langenhoff et al, 2001; Sajid et al, 2008; Velanovich, 2001).
In fact, the intention of palliative surgery is to mitigate physical
WHY AND WHEN? symptoms in patients with noncurable disease, with the primary
Surgery is unique; it is immediate and, for the most part, irre- goal of improving or maintaining HRQoL (Hofmann et al,
versible. In this high-stakes setting, evaluation of the impact of 2005). Inclusion of HRQoL measurements in studies in which
surgery on HRQoL is essential, and integration of the informa- these specific scenarios are encountered will maximize the clini-
tion gleaned into the informed consent process and clinical cal applicability of the HRQoL outcomes observed.
decision making is the ultimate goal. A recent review of 33
randomized surgical oncology trials, in which HRQoL data SURGEON, PATIENT, OR CAREGIVER: WHO
were collected, indicated that in two-thirds of trials, the HRQoL SHOULD GATHER HRQOL INFORMATION?
information influenced clinical decision making and/or facili-
tated the surgical consent process. Integration of the HRQoL For practical reasons, initial investigations used the external
data was more common among later trials, reflecting the pro- judgment of observers, typically nurses or clinicians, to evaluate
gressively increased value attached to these outcomes by sur- treatment impact on patient QoL. Typically, this included
geons (Blazeby et al, 2006). On a similar note, it has been evaluation of symptoms and overall QoL. This method of
shown that communicating both technical procedural-related obtaining HRQoL data is inherently biased by the observer’s
information and QoL information to the patient and their own internal standards. Multiple studies evaluating the degree
family facilitates an improved physician-patient relationship, of agreement between the proxy ratings of QoL by observers to
reconciles patient expectations, and improves patient satisfac- patient assessments have consistently found very little correla-
tion (Chen et al, 2013; D’Angelica et al, 1998; Detmar et al, tion between the two measures (Choiniere et al, 1990; Kahn
2002; Passik & Kirsh, 2000; Velikova et al, 2004). Further- et al, 1992; Presant, 1984; Slevin et al, 1988; Sprangers &
more, while measurement tools have evolved from generic to Aaronson, 1992; Stephens et al, 1997). Given these findings,
disease/site specific, and subsequently become more sensitive, it is generally accepted that validity of QoL instruments and
focus on HRQoL as a predictive/prognostic variable of out- rigor of results is contingent on patients’ reporting. Unfortu-
comes such as morbidity, mortality, and survival has increased. nately, this requisite self-reporting can become an issue while
Meta-analysis of 30 randomized clinical trials with HRQoL patients are followed during time; disease status and symptoms
data from the European Organization for the Research and worsen, and patients are less likely to fill out instruments,
Treatment of Cancer (EORTC) Quality-of-Life Questionnaire leading to nonrandom missing data, which is a significant chal-
(QLQ)-C30 (core 30 items) revealed that addition of HRQoL lenge in QoL research (Bernhard et al, 1998; Hahn et al,
parameters (physical functioning, pain, anorexia) to sociode- 1998). Feinstein (1987) coined the term “sensibility” of
mographic and clinical variables provides prognostic value and HRQoL instruments to denote the practical issues related to
significantly improves predictive accuracy of models of survival implementation of HRQoL measurements. Many surveys or
(Quinten et al, 2009). Moreover, patient scores within the instruments are impractical and can be overwhelming in length
Chapter 28 Quality of life and hepatobiliary tumors 477
and time required to complete, leading to “questionnaire burn- tools of varying quality exist and can be accessed through the
out” and subsequent patient drop-out. Other issues, such as Patient-Reported Outcome and Quality-of-Life Instruments
literacy and native language, may prohibit patients from com- Database (PROQOLID) (http://www.proqolid.org). The sheer
pleting instruments independently. Recent HRQoL measure- number of measurement tools highlights the expansion of the
ment tools were developed and validated for completion by the HRQoL field; however, it can also make research and interpre-
patient. However, to optimize accrual and retention in HRQoL tation of the literature challenging. The following section briefly
studies, certain situations may dictate that physicians, nurses, outlines general concepts of measurement and instruments,
caregivers, research administrators, or even translators com- whereas specific instruments commonly used in surgical trials
plete questionnaires on the patient’s behalf. are represented in Table 28.1A and Table 28.1B.
For HRQoL results to be respected on the same level as
HRQOL MEASUREMENT TOOLS, other measured clinical variables, the scientific rigor of mea-
INSTRUMENTS, AND INTERPRETATION surement tools is essential and depends upon three primary
concepts: reliability, validity, and responsiveness/sensitivity.
Early in the history of HRQoL assessment in cancer clinical Reliability is the degree to which an instrument is free from
trials, no standard instrument was used, and the default was random error and includes both reproducibility and internal
simply ad hoc trial-specific questionnaires that prohibited com- consistency. Validity is the degree to which an instrument mea-
parison across trials and often even among patients with the sures what it is suppose to measure and consists of content,
same disease (de Haes & van Knippenberg, 1985). These find- criteria,and construct validity. Responsiveness is the ability of
ings prompted a new era in QoL research heralded by the an instrument to detect and measure changes during time and
development of multiple standardized tools/instruments to treatments. A good instrument is stable when nothing has
measure QoL. To date, more than 800 QoL clinical outcome changed and capable of detecting even small changes when they
TABLE 28.1A Generic and Disease-Specific Health-Related Quality-of Life-Measurement Instruments

HPB-
Author, Year, Type of Description of Quality-of- Specific
Instrument Country Instrument Life Measures Interpretation Modules Comments
Medical Ware, 1992, Generic 36 items 8 health-status scale No Comparison across

Outcomes U.S. profile 8 health-status scales: scores broad range of
Study 36-Item based General health Items in each scale are disease states
Short Form perceptions summed and and treatments
(SF-36) Physical functioning averaged to give a Less sensitive to
Role limitations due to single score (0-100) changes in
physical problems for each health disease/health
Role limitations due to status status
emotional problems Higher scores = better
Social functioning HRQoL
Bodily pain
Vitality (energy/fatigue)
General mental health
Nottingham Hunt, 1985, Generic 38 items 6 domain scores No Evaluates perceived
Health Profile U.K. profile 6 domains: Items in each domain distress across
(NHP) based Physical mobility are weighted, populations
Pain summed and
Social isolation averaged to give a
Emotional reactions single score (0-100)
Energy for each domain
Sleep Higher scores = worse
HRQoL
Sickness Impact Bergner, 1981, Generic 136 items, 2 domains, Each item is scored No Time consuming to
Profile (SIP) U.S. profile 12 categories: and weighted to give complete (20-30
based Physical domain: overall score (%), 2 min)
ambulation, mobility, domain scores, 12
body care, category scores
movement Higher scores = worse
Psychosocial domain: HRQoL
social interaction,
communication,
alertness behavior,
emotional behavior,
sleep, eating, home
management,
recreation,
employment
Continued
TABLE 28.1A Generic and Disease Specific Health Related Quality of Life Measurement Instruments—cont’d
HPB-
EuroQoL EQ EuroQoL Generic EQ-5D-5L (descriptive Each dimension given No Self-administered
5D-5L Group, value/ system): a 1-digit number with minimal
1990, UK preference 5 dimensions, 5 levels expressing the level responder bias
based of severity: Mobility, selected Revision of
self-care, usual Numbers from each EQ-5D-3L,
activities, pain/ dimension are increased levels
discomfort, anxiety/ combined to give of severity from
depression 5-digit number 3-5 to improve
corresponding to a sensitivity and
specific health state; reduce ceiling
preference value is effect
then assigned based
on empirically
derived valuations
(0-1)
Spitzer Spitzer, 1980, Generic EQ-VAS (visual VAS: quantitive No 3 types of data
Quality-of-Life Australia index analogue scale) measure of overall produced:
Index (QLI) 5 items: health perceived by 1. Profile indicating
Activity respondent extent of the
Daily living (self-rated) problems in
Health Higher score = better each domain
Support HRQoL 2. Population-
Outlook weighted health
index
3. Self-rated
assessment of
health status
Quality of Fanshel & Generic 3 scales assessing Each item scored 0-2 No Completed by
Well-being Bush, 1970, value/ daily functioning, 3 and summed to give surgeon, not
Scale (QWB) U.S. preference levels (recall period total score (0-10) patient; newer
*
Multiple based is 6 days): Higher score = better versions modified
adaptations Mobility HRQoL for self-
Physical activity Using empirically administration
Social activity + derived preference Reliable in patients
symptom/problem values: function with definite
complexes levels for each scale physical disease
combined with the Interviewer
most undesirable administered (26
reported symptoms/ symptom/
problems to give a problem
single well-being complexes): long,
score: 0 (death) to 1 complex, difficult
(fully functional) to administer,
Higher score = better and requires
QoL specific training
Self- administered
version (QWB-
SA, 58
symptom/
problem
complexes)
FACT-G Cella, 1993, Disease 27 items, 4 domains: Each item scored 0-4 Yes Comparable across
(Functional U.S. specific: Physical (7) on Likert scale; different cancers
Assessment cancer Social/family (7) items in given Moderate sensitivity
of Cancer Emotional (6) domain are summed to changes over
Treatment– Functional (7) to give overall score time
General) for each domain Published data
Scores for each exists regarding
domain summed to clinically
give overall score important
Higher score = worse changes over
HRQoL time, allowing
improved
interpretation
TABLE 28.1A Generic and Disease Specific Health Related Quality of Life Measurement Instruments—cont’d
HPB-
EORTC Aaronson, Disease 30 items; 9 multiitem Each item in a scale is Yes Comparable across
QLQ-C30 1993, the specific: scales + 6 single scored, and an different cancers
(European Netherlands cancer items: overall scale score Sensitive to
Organization 5 functional scales: from 0-100 is cancer- related
for Research Physical reported changes
and Role Higher-score overtime
Treatment of Social functional/global Published data
Cancer-Core Cognitive scales = better exists regarding
Quality-of-Life Emotional HRQoL clinically
Questionnaire) 1 Global Health/QoL important
Scale changes during
3 symptom scales: time, allowing
Pain improved
Fatigue interpretation
Nausea/vomiting
Functional Living Schipper, Disease 22 items, 5 domains: Higher-score symptom No Single score lacks
Index-Cancer 1984, specific: Physical well-being and scales/single items = sensitivity to
(FLIC) Canada cancer ability worse HRQoL change
Emotional state
Sociability
Family situation
Nausea
Gastrointestinal Eypasch, Disease 36 items, 5 response Scores from all No Assesses QoL
Quality-of-Life 1995, specific: categories (recall questions are pertaining to a
Index (GQLI) Germany any GI over 2 wk): summed and a variety of
and Canada disease Core symptoms items single score is diseases of the
Physical items reported liver, pancreas,
Psychological items Higher score = better and biliary
Social items quality of life system
Disease-specific items Scores from all items Single score lacks
are summed, and a sensitivity to
single score is change
reported
Higher score = better
quality of life
occur. Responsiveness is dependent on the number of items in combining domain scores to give a single value as an indicator
a questionnaire as well as the number of potential responses. of overall HRQoL is not standard, nor is it valid.
From a practical standpoint, the number of items and associ- To date, no “gold-standard” instrument exists for the evalu-
ated responses must be optimized to ensure sufficient respon- ation of HRQoL. Measurement tools are either health-profile
siveness without creating overly cumbersome questionnaires in based—provide a variety of subscores representing each domain
which respondents experience “question fatigue,” leading to of HRQoL—or preference based—provide a single number
increased rates of nonresponse and missing data. A fourth (health index score) from 0 (death) to 1 (perfect health) that is
parameter, “sensibility,” is also essential and refers to the need representative of a patient’s subjective health status at a single
for strategic balance between practicality and the basic neces- time point. Preference values, typically developed from
sity of instruments to be reliable, valid, and sensitive (Feinstein, population-based studies, are applied to the patient-reported
1987). In addition to choosing a reliable, valid, sensitive, and health states to give a single HRQoL score/index. Measurement
practical instrument, a thorough understanding of how a given tools may be further categorized into three clinically relevant
disease progresses and the expected effects of intervention is categories: generic QoL, disease-specific HRQoL, and modular
requisite (Avery & Blazeby 2006). HRQoL tools.
In general, from a content perspective, HRQoL instruments Generic instruments used in many surgical trials include the
typically consist of a variable number of domains containing Medical Outcomes Survey 36-item Short Form (SF-36), the
information focused on a specific area of health/disease (i.e., Nottingham Health Profile (NHP), and the Sickness Impact
physical, social, emotional, role, global functioning). In turn, Profile (SIP). Each of these measurement tools produces a
each domain contains a number of items (questions or state- health profile, with multiple domain-specific subscores, and
ments) for which respondents must provide an answer, either provides a broad overview of the patient’s HRQoL. The generic
categorically (i.e., Likert scale) or in visual analogue form. Each nature of these measures allows comparison across disease
item is individually scored, and scores from all items in a given states and treatment types. However, these surveys can be time
domain are summed to give each domain a specific score consuming to complete and may not provide adequate sensitiv-
(Fraser, 1993). For most multidimensional instruments, ity to identify changes in HRQoL specifically related to different
TABLE 28.1B Modular Disease Specific Health Related Quality of Life Measurement Instruments
Description of Quality-of-Life
Core Questionnaire Module Author, Year, Country Measurement Comments
Functional Hepatobiliary Heffernan et al, 2002, FACT-G + 18-item subscale, HS can be combined with
Assessment of subscale (HS) U.S. including HPB-specific physical and functional
Cancer Therapy FACT + HS = concerns related to: domain scores of FACT-G
(FACT) FACT-Hep Jaundice to give Trial Outcome
GI obstruction Index (TOI) 7 additional
Fatigue/energy questions (nonscored) at
Each item scored (0-4) and end of HS, addressing
summed to give overall HAIP and biliary drainage
subscale score
FACT Hepatobiliary Yount et al, 2002, 8 key symptoms from Brief index with good
Symptom Index-8 USA 18-item HS scale correlation with scores on
(FHSI-8) including: FACT-G and FACT-Hep
Pain Capacity to discriminate
Nausea patients based on
Fatigue (× 2 items) performance status/
Jaundice treatment; status
Weight loss adequate but not as
Back Pain good as FACT-G
Stomach Pain/discomfort subscales or HS
*Developed in response to Developed primarily based
clinician concern regarding on pancreatic cancer
the time and resources
required to complete and
interpret multidimensional
QoL assessments
European Pancreatic Fitzsimmons et al, 26 -item subscale, including Examples:
Organization for carcinoma 1999, UK items concerning Pain (abdominal, back,
Research and (QLQ-PAN26) pancreas-specific: positional, night)
Treatment of Symptoms Dietary restrictions
Cancer–Quality-of- Treatments GI symptoms
Life Emotional issues Cachexia
Questionnaire– Weight loss
Core 30 (EORTC Jaundice
QLQ-C30) Pruritus
Ascites
Colorectal liver Kavadas et al, 21-item subscale. including Examples:
metastases 2003, UK items concerning Pain (abdominal,back)
(QLQ-LMC21) CRLM-specific: Eating problems (early
Symptoms satiety)
Treatment Fatigue
Emotional issues Lethargy
Jaundice
Taste
Tingling hands and feet
Stress
Loss of enjoyment
Hepatocellular Blazeby et al, 18-item subscale, including Examples:
carcinoma 2004, UK items concerning Fatigue
(QLQ-HCC18) HCC-specific: Body image Jaundice
Symptoms Nutrition
Treatments Pain
Emotional Issues Fevers
Cholangiocarcinoma, Friend et al, 21-item subscale, including Examples:
Gallbladder 2011, UK items concerning CCA/ Eating
Carcinoma GBCA-specific: Satiety
(QLQ-BIL21) Symptoms Jaundice
Treatments Fatigue
Emotional issues Anxiety
Pain
Stress
Worry
CCA, Cholangiocarcinoma; CRLM, colorectal liver metastases; GBCA, gallbladder carcinoma; GI, gastrointestinal; HAIP, hepatic arterial infusion pump; HPB, hepatopancreatobiliary; HRQoL, health-related
quality of life; U.K., United Kingdom; U.S., United States.
diseases and/or treatments. Preference-based instruments such HRQoL; however, their practicality makes them attractive
as the Quality of Well-being Scale (QWB) and EuroQoL-5D- alternatives.
3L/5L (EQ-5D-3L) produce a single score/index/utility. These HRQoL is more than the sum of its parts, and it is likely
instruments, which tend to be insensitive to changes, are less that a global assessment in the form of an index is complemen-
frequently used in therapeutic surgical trials focusing on tary to, rather than an alternative to, individual domain evalu-
HRQoL. However, these preference-based methods are essen- ation. It is generally accepted that a single assessment of
tial and often used in the setting of health economics and patients’ perceptions of overall HRQoL is improved over aggre-
policy. When a preference value for given health state is com- gation of individual domain scores, where each domain is
bined with time values (how long a patient spends in a given given equal importance and a “mean” QoL is estimated (Aar-
disease state) the quality-adjusted life-year (QALY) can be gen- onson, 1988; Osoba, 1994). Ideally, optimal HRQoL assess-
erated. This can then be used in conjunction with the cost of ment strategies should include both global- and domain-specific
treatment/procedure to determine the cost per QALY, and deci- measurements.
sions regarding the economically utility of interventions can be The vast array and potential combinations of available mea-
assessed. At present, the cost of health care is exorbitant, and surement tools can be overwhelming and make interpretation/
while we forge ahead toward more and more precision/designer evaluation of HRQoL literature difficult. To critically appraise
medicine, this will only increase, and cost-utility assessment of HRQoL literature, it is essential to understand the tools used
intervention will be essential. to determine HRQoL outcomes so that judgment as to the
Disease-specific tools focus on a single disease state. In the robustness of the results can be made. Table 28.2 outlines
realm of hepatobiliary surgery and cancer, the two most com- HRQoL levels of evidence based on the psychometric proper-
monly used are the Functional Assessment of Cancer Therapy– ties of the measurement tool(s) used. Low level refers to de
General (FACT-G, 27 items) and the EORTC QLQ-C30 facto questionnaires developed for a single study, midlevel to
(Fig. 28.1). These types of instruments focus more on expected questionnaires not developed or validated in study group of
changes specifically related to cancer and its treatment. Con- interest (psychological assessment, depression scales, etc.), and
sequently, they are more sensitive to HRQoL changes and are high level is represented by the aforementioned instruments
good for comparisons across different cancer diagnoses but are that assess QoL in a multidimensional disease-specific manner.
not generalizable to other disease states. The modular approach Two excellent review articles have been published regarding
is specific to patients with cancer. It refers to the use of a core how to interpret/evaluate studies involving HRQoL (Guyatt
cancer questionnaire in conjunction with validated modules et al, 1997; Wu et al, 2014). Both present a standardized
for specific disease sites (i.e., pancreas, liver, biliary modules) approach to critical appraisal and provide a means by which
(Langenhoff et al, 2001). This approach is based on the HRQoL can be integrated into evidence-based medicine.
premise that, although there are similar effects of disease/
treatment across cancers, each primary tumor site is also asso- HEALTH-RELATED QUALITY-OF-LIFE STUDIES
ciated with a unique set of HRQoL concerns. Use of a modular IN HEPATOBILIARY CANCER
approach increases the sensitivity to detect small, yet clinically
relevant, changes in HRQoL. This method has been popular- Measurement and reporting of HRQoL has increased exponen-
ized by the EORTC, in which the core EORTC QLQ-C30 tially during the past decade. The development of reliable, valid,
questionnaire is paired/supplemented with a site- and/or and sensitive measurement tools for use in cancer patients and,
symptom-specific instrument. Site-specific modules validated more specifically, in patients with hepatobiliary malignancy, has
for use among hepatopancreatobiliary (HPB) tumors include led to a vast increase in the number and quality of studies
the following: pancreatic cancer module (PAN26), colorectal assessing HRQoL in patients with HPB-related malignancy.
liver metastases module (LM21), hepatocellular carcinoma The following sections highlight the current status of
module (HCC18), and cholangiocarcinoma/gallbladder carci- HRQoL as it pertains to definitive surgical management as well
noma module (BIL21). A second modular approach, devel- as palliative interventions in patients with advanced HPB
oped and validated in North America, uses the FACT-G as the malignancies. Tables 28.3 to 28.7 summarize the most recent
core questionnaire, which is supplemented with site-/symptom- reports in the literature as well as landmark studies on HRQoL
specific modules, such as the hepatobiliary subscale (FACT- in the setting of HPB malignancy.
Hep; Fig. 28.2). Despite the seeming similarity and potential
overlap of the FACT-G plus supplemented modules with the Pancreatic Resection
EORTC QLQ-C30 plus supplemental models, the two instru- Despite advances in surgical technique and perioperative care
ments have been compared and found to measure different following pancreatic resection, morbidity remains common
aspects of HRQoL (Kemmler et al, 1999). The comprehensive (20% to 30%) (see Chapters 27, 62, and 66). Furthermore, in
nature of these tools provides increased sensitivity; however, the setting of pancreatic cancer, even with complete resection,
practicality, in terms of time and resources required to com- long-term survival is rare. It is therefore not surprising, that
plete, may be prohibitive in some settings. Based on this, with high morbidity and less than optimal survival outcomes,
attempts to develop brief scales that adequately correlate with the impetus for robust HRQoL evaluation among patients
the more extensive evaluations have been attempted. Yount undergoing pancreatic resection has gained significant momen-
and colleagues (2002) developed and tested an eight-item tum. As a consequence of this increased interest, study design
symptom scale from the FACT hepatobiliary subscale that cor- and methodology have improved greatly and appreciably
related well with overall FACT-G scores; however, the ability enhanced the quality of this expanding body of knowledge
to discriminate patients based on performance status/treatment (Table 28.3).
status was limited. Overall, these simplified indices are not as Patients with pancreatic cancer often experience consider-
sensitive or reliable as longer multidimensional evaluations of able symptoms and are at increased risk for impaired HRQoL
ENGLISH
EORTC QLQ-C30 (version 3)

We are interested in some things about you and your health. Please answer all of the questions yourself
by circling the number that best applies to you. There are no "right" or "wrong" answers. The information
that you provide will remain strictly confidential.
Please fill in your initials:

Your birthdate (Day, Month, Year):
Today's date (Day, Month, Year): 31
Not at A Quite Very

All Little a Bit Much
1. Do you have any trouble doing strenuous activities,
like carrying a heavy shopping bag or a suitcase? 1 2 3 4
2. Do you have any trouble taking a long walk? 1 2 3 4
3. Do you have any trouble taking a short walk outside of the house? 1 2 3 4
4. Do you need to stay in bed or a chair during the day? 1 2 3 4
5. Do you need help with eating, dressing, washing

yourself or using the toilet? 1 2 3 4
During the past week: Not at A Quite Very

6. Were you limited in doing either your work or other daily activities? 1 2 3 4
7. Were you limited in pursuing your hobbies or other

leisure time activities? 1 2 3 4
8. Were you short of breath? 1 2 3 4
9. Have you had pain? 1 2 3 4
10. Did you need to rest? 1 2 3 4
11. Have you had trouble sleeping? 1 2 3 4
12. Have you felt weak? 1 2 3 4
13. Have you lacked appetite? 1 2 3 4
14. Have you felt nauseated? 1 2 3 4
15. Have you vomited? 1 2 3 4
16. Have you been constipated? 1 2 3 4
Please go on to the next page
FIGURE 28.1. European Organization for the Research and Treatment of Cancer Quality-of-life Questionnaire (EORTC QLQ-C30). A user’s agree-
ment is required to use this scale and can be accessed through the website www.eortc.be/home/qol/downloads/index.html.
ENGLISH
During the past week: Not at A Quite Very

17. Have you had diarrhea? 1 2 3 4
18. Were you tired? 1 2 3 4
19. Did pain interfere with your daily activities? 1 2 3 4
20. Have you had difficulty in concentrating on things,

like reading a newspaper or watching television? 1 2 3 4
21. Did you feel tense? 1 2 3 4
22. Did you worry? 1 2 3 4
23. Did you feel irritable? 1 2 3 4
24. Did you feel depressed? 1 2 3 4
25. Have you had difficulty remembering things? 1 2 3 4
26. Has your physical condition or medical treatment

interfered with your family life? 1 2 3 4

interfered with your social activities? 1 2 3 4

caused you financial difficulties? 1 2 3 4
For the following questions please circle the number between 1 and 7 that best applies to you
29. How would you rate your overall health during the past week?
1 2 3 4 5 6 7
Very poor Excellent
30. How would you rate your overall quality of life during the past week?
1 2 3 4 5 6 7
Very poor Excellent
© Copyright 1995 EORTC Quality of Life Group. All rights reserved. Version 3.0
FIGURE 28.1, cont’d
TABLE 28.2 Levels of Evidence in Health-Related Quality-of-Life Evaluation

Level of Evidence Methodology/Approach Explanation
Low Single items (symptoms/performance/VAS) Often developed for a single study

Typically not psychometrically validated
Middle (A) Conversion of preexisting tools Adaptation of tools from traditional psychology and psychiatry
Assumes measures are reliable and valid in cancer patients (example:
Beck Depression Inventory)
Middle (B) Assessment of a single HRQoL domain Measurement of multiple single items within a single domain of HRQoL
Typically focused on physical well-being and/or measurement of
treatment toxicities/side effects (example: performance scores)
High Multidimensional assessments Highest level of HRQoL evidence
Multiple subscales evaluating multiple domains (physical, social,
emotional, role, spiritual, etc.)
Disease specific (example: FACT, EORTC QLQ-C30, FLIC)
EORTC-QLQ-C30, European Organization for the Research and Treatment of Cancer Quality-of-Life Questionnaire–Core 30; FACT, Functional Assessment of Cancer Therapy; FLIC, Functional Living
Index-Cancer; HRQoL, health-related quality of life; VAS, visual analogue scale.
From Table 1 in: Passik, et al: The importance of quality of life endpoints in clinical trials to the practicing oncologist. Hematol Oncol Clin N Am 14(4):877-886, 2000.
FACT-Hep (Version 4)
Below is a list of statements that other people with your illness have said are important. Please circle
or mark one number per line to indicate your response as it applies to the past 7 days.
Not A little Some- Quite Very

PHYSICAL WELL-BEING at all bit what a bit much
GP1 I have a lack of energy................................................... 0 1 2 3 4

GP2 I have nausea................................................................ 0 1 2 3 4
GP3 Because of my physical condition, I have trouble
meeting the needs of my family..................................... 0 1 2 3 4
GP4 I have pain..................................................................... 0 1 2 3 4
GP5 I am bothered by side effects of treatment.................... 0 1 2 3 4
GP6 I feel ill............................................................................ 0 1 2 3 4
GP7 I am forced to spend time in bed.................................... 0 1 2 3 4
SOCIAL/FAMILY WELL-BEING Not A little Some- Quite Very

at all bit what a bit much
GS1 I feel close to my friends................................................ 0 1 2 3 4

GS2 I get emotional support from my family.......................... 0 1 2 3 4
GS3 I get support from my friends......................................... 0 1 2 3 4
GS4 My family has accepted my illness................................. 0 1 2 3 4
GS5 I am satisfied with family communication about my
illness............................................................................. 0 1 2 3 4
GS6 I feel close to my partner (or the person who is my
main support) ................................................................ 0 1 2 3 4
Q1 Regardless of your current level of sexual activity,
please answer the following question. If you prefer not
to answer it, please mark this box and go to the
next section.
GS7 I am satisfied with my sex life........................................ 0 1 2 3 4
English (Universal) 16 November 2007

Copyright 1987, 1997 Page 1 of 3
FIGURE 28.2. Functional Assessment of Cancer Treatment–Hepatobiliary Questionnaire (FACT-Hep). (Courtesy Dr. David Cella at www.FACIT.org;
copyright 1987, 1997. Permission for use must be obtained by contacting Dr. David Cella at www.FACIT.org or d-cella@northwestern.edu.)
Please circle or mark one number per line to indicate your response as it applies to the past 7
days.
EMOTIONAL WELL-BEING Not A little Some- Quite Very

GE1 I feel sad........................................................................... 0 1 2 3 4
GE2 I am satisfied with how I am coping with my illness......... 0 1 2 3 4
GE3 I am losing hope in the fight against my illness................ 0 1 2 3 4
GE4 I feel nervous.................................................................... 0 1 2 3 4
GE5 I worry about dying........................................................... 0 1 2 3 4
GE6 I worry that my condition will get worse............................ 0 1 2 3 4
FUNCTIONAL WELL-BEING Not A little Some- Quite Very

GF1 I am able to work (include work at home)......................... 0 1 2 3 4
GF2 My work (include work at home) is fulfilling...................... 0 1 2 3 4
GF3 I am able to enjoy life........................................................ 0 1 2 3 4
GF4 I have accepted my illness................................................ 0 1 2 3 4
GF5 I am sleeping well............................................................. 0 1 2 3 4
GF6 I am enjoying the things I usually do for fun..................... 0 1 2 3 4
GF7 I am content with the quality of my life right now.............. 0 1 2 3 4

FIGURE 28.2, cont’d Continued
compared with population norms (Belyaev et al, 2013; Chan Early postoperative evaluation of HRQoL typically repre-
et al, 2012; Ljungman et al, 2011; Muller-Nordhorn et al, sents a transient state, primarily reflective of a patient’s recovery
2006). Results of HRQoL evaluation following pancreatic from the acute effects of surgery. These findings certainly allow
surgery are divergent and heavily dependent on the cohort surgeons to better inform patients, improve the consent process,
studied and the timing of evaluation relative to surgery. Among and help tailor expectations. However, longer-term follow up
well-selected patients with localized, resectable disease, HRQoL postpancreatectomy is requisite to unveil the true clinical
at 3 months following surgery was reportedly improved (Crippa impact of surgical resection on HRQoL outcomes. In a recent,
et al, 2008; Yeo et al, 2012). Unfortunately, the beneficial find- prospective longitudinal study, HRQoL 12 months after pan-
ings reported in these studies appear to be the exception, with creaticoduodenectomy (PD) for periampullary malignancies
the great majority of investigations reporting stability and/or was evaluated using the SF-36 survey. Despite the heterogene-
decline in HRQoL scores following pancreatic resection ity in cancer subtypes included, and initial decrements, signifi-
(Belyaev et al, 2013; Crippa et al, 2008; Ljungman et al, 2011; cant improvements were observed in the majority of subscales
Mbah et al, 2012; Muller-Nordhorn et al, 2006; Nieveen when measured at 12 months (Chan et al, 2012). Similarly,
van Dijkumet al, 2005; Schniewind et al, 2006). The hetero- significant improvements in HRQoL domains assessed with the
geneity observed in these studies is likely multifactorial and EORTC QLQ-C30 were observed among a group of patients
related, but not limited, to differences in disease stage, HRQoL undergoing partial pancreatectomies for pancreatic adenocar-
instrument used, timing/frequency of measurement, and length cinoma at 6 and 12 months postoperatively. Although baseline
of follow-up. measurements were obtained postoperatively, and absolute
Please circle or mark one number per line to indicate your response as it applies to the past 7
days.
ADDITIONAL CONCERNS Not at A little Some- Quite Very

all bit what a bit much
C1 I have swelling or cramps in my stomach area .................. 0 1 2 3 4
C2 I am losing weight............................................................... 0 1 2 3 4
C3 I have control of my bowels................................................ 0 1 2 3 4
C4 I can digest my food well .................................................... 0 1 2 3 4
C5 I have diarrhea (diarrhoea) ................................................. 0 1 2 3 4
C6 I have a good appetite ........................................................ 0 1 2 3 4
Hep I am unhappy about a change in my appearance............... 0 1 2 3 4

1
CNS I have pain in my back ....................................................... 0 1 2 3 4

7
Cx6 I am bothered by constipation ............................................ 0 1 2 3 4
H17 I feel fatigued ..................................................................... 0 1 2 3 4
An7 I am able to do my usual activities..................................... 0 1 2 3 4
Hep I am bothered by jaundice or yellow color to my skin......... 0 1 2 3 4

2
Hep I have had fevers (episodes of high body temperature) ..... 0 1 2 3 4

3
Hep I have had itching ............................................................... 0 1 2 3 4

4
Hep I have had a change in the way food tastes ....................... 0 1 2 3 4

5
Hep I have had chills ................................................................. 0 1 2 3 4

6
HN My mouth is dry .................................................................. 0 1 2 3 4

2
Hep I have discomfort or pain in my stomach area .................... 0 1 2 3 4

8

FIGURE 28.2, cont’d
HRQoL improvement is most certainly overestimated in this jejunostomy (PJ), for a mix of benign and malignant indications
later study, the overall conclusions remain the same (Halloran (Billings et al, 2005; Epelboym et al, 2014; Han et al, 2007;
et al, 2011). Following pancreatectomy, patients tend to expe- Nguyen et al, 2003; Schmidt et al, 2005; Shaw et al, 2005; You
rience surgery-related decline in HRQoL, although the over- et al, 2007). In general, patients undergoing partial pancreatec-
whelming majority of data support a gradual improvement/ tomy for benign disease, regardless of the type (PD, RPD,
return to baseline within the first postoperative year. PPPD, and PSP) had similar long-term global ratings of
To date, several studies have compared HRQoL of patients HRQoL as well as comparable social, functional, and role
undergoing different types of partial pancreatomy—classic scores on the EORTC QLQ-C30. Furthermore, studies of
pancreaticoduodenectomy (PD) versus radical pancreati long-term HRQoL among patients with completely resected
coduodenectomy (RPD) versus pylorus-preserving pancreati- malignant tumors also suggest that HRQoL trends follow the
coduodenectomy (PPPD) versus parenchyma-preservation/ same postoperative trajectory independent of the type of resec-
-sparing pancreatectomy (PSP), with different reconstruction tion and/or reconstruction.
techniques—pancreaticogastrostomy (PG) versus pancreatico- Text continued on p. 494

TABLE 28.3 Studies of Pancreatic Surgery and Health-Related Quality of Life

Study Purpose/ Study Design, N, and
Quality-of-Life End HRQoL Assessment Quality-of-Life
First Author, Journal, Year Population (Patient With) Point Time Points Instrument Quality-of-Life Results Comments
Pancreatic Resection
Nguyen, Journal of Periampullary cancer Primary outcome: Cross-sectional, N = FACT-Hep Assessments were
Gastrointestinal Surgery, randomized to PD or a postoperative 105 (55 PD and completed at mean 2.1 yr
2003 RPD (PD + HRQoL 50 RPD); HRQoL after surgery. No
retroperitoneal lymph assessed at a differences at long-term
node dissection) single, follow-up in HRQoL and
nonstandard GI functional status
postoperative between PD and RPD.
time point RPD patients had a
significantly higher
postoperative complication
rates (43% to 29%, P =
.01)
Van Heek, Annals of Surgery, Unresectable periampullary Primary outcome: Prospective EORTC 2/36 HJ + GJ and 12/29 HJ Completion rates were
2003 cancer randomized to development of randomized trial, QLQ-C30 + patients developed GOO >90% in both groups for
undergo a double GOO N = 65 (36 HJ + PAN26 (P < .01). After an initial the first 4 mo
bypass (HJ + GJ) or a Secondary end GJ, 29 HJ alone) module decline with surgery, postoperatively and 75%
single bypass (HJ) point: HRQoL HRQoL assessed HRQoL returned to in the last 2 mo
preoperatively, at baseline by 4 mo
discharge, and postoperatively and was
q1mo for 6 mo not different between
groups
Billings, J Gastrointestinal Periampullary disease Primary outcome: Cross-sectional SF-36 Health Mean follow-up time: 7.5 yr Only evaluated TP patients
Surgery, 2005 (benign and malignant) longitudinal matched design, Survey, Audit postoperatively. TP cohort alive with no disease at
requiring TP HRQoL N = 27 TP; TP of Diabetes- had lower SF-36 role, and time of survey
cohort age/ Dependant physical and general administration (mean,
gender matched HRQoL (ADD health scores (P < .05). 7.5 yr), “healthy-
1 : 1 with IDDM HRQoL), ADD HRQoL scores were survivor” bias
cohort; HRQoL EORTC decreased but not different
assessed at a Pan26, from IDDM controls
single, nonvalidated
nonstandard, institutional
postoperative survey
time point
Nieveen van Dijkum, British Resectable pancreatic Primary outcome: Prospective Medical After an initial decline,
Journal of Surgery, 2005 tumors (benign and longitudinal longitudinal, N = outcomes HRQoL scores returned to
malignant) undergoing HRQoL 114 (PD = 72, HJ study baseline at 3 mo
PD compared with + GJ = 42); (MSO-24), postoperatively in both
unresectable pancreatic assessed gastrointestinal groups. Rapid decline in
tumors undergoing preoperatively, HRQoL index HRQoL was observed in
double bypass (HJ + postoperatively, (GIQLI) both groups in the 8th wk
GJ) and at 1.5-, 3-, prior to death
6-, 9-, and
12-mo follow-up
Chapter 28 Quality of life and hepatobiliary tumors
Continued
487
TABLE 28.3 Studies of Pancreatic Surgery and Health-Related Quality of Life—cont’d
488

Schmidt, Annals of Surgical Pancreatic tumors (benign Primary outcome: Cross-sectional, N = EORTC 104/133 surviving patients Only evaluated subset of
Oncology, 2005 and malignant) HRQoL and 104 (PG = 63, PJ QLQ-C30 + responded to patients alive at the time
undergoing PD with PG long- term = 41); HRQoL PAN26 questionnaires (mean of survey administration
or PJ reconstruction morbidity assessed at a module, assessment time, 6.4 yr (mean, 6.5 yr), “healthy-
single, nonvalidated postoperatively). Global survivor” bias
nonstandardized, institutional HRQoL was the same
postoperative evaluation of between PJ and PG
time point GI symptoms groups. PG group had
significant decrease in
multiple GI symptoms but
increase in steatorrhea,
early satiety, and food
aversion. PJ patients had
no change in GI
symptoms but did report
reduced jaundice-related
symptoms
Shaw, Pancreatic tumors (benign Primary outcome: Cross-sectional EORTC Mean assessment at 42 mo
Hepaticogastroenterology, and malignant) longitudinal matched design, QLQ-C30 + postprocedure. Global
2005 undergoing PD HRQoL N = 49 PD; PD PAN26 health status was similar
cohort age/ module between PD and matched
gender matched controls. Patients with
1 : 1 with patients malignant indication for PD
undergoing open had decreased physical
cholecystectomy and role functioning as
PART 3 ANESTHETIC MANAGEMENT: PREOPERATIVE AND POSTOPERATIVE CARE
for benign well as more symptoms of

disease; HRQoL fatigue, muscle weakness,
assessed at a and failure to gain weight
single, compared with patients
nonstandardized undergoing PD for benign
postoperative disease
time point
Wehrmann, European Journal Unresectable pancreatic Primary outcome: Prospective Spitzer QLI Successful procedure in Completion rates:
of Gastroenterology and head cancer with HRQoL and longitudinal, N = 19/20 patients. Patients preoperative 100%,
Hepatology, 2005 pancreatic duct pain control 20; HRQoL reported a significant 4 wk 100%, 12 wk
obstruction and assessed improvement in pain and 16/19 (84%)
postprandial epigastric preprocedure and HRQoL at 4 wk (P < .01)
pain undergoing 4 and 12 wk but decreased again at
endoscopic pancreatic postprocedure 12 wk to preprocedure
duct stenting values
Muller-Nordhorn, Digestion, Consecutive patients Primary outcome: Cross sectional EORTC Global HRQoL as well as all
2006 admitted to hospital HRQoL in matched design, QLQ-C30 and subscale measures were
with various stages of patients with N = 45 (44% EuroQol decreased among patients
pancreatic cancer pancreatic stage 4) HRQoL (EQ-5D-3L) with pancreatic cancer.
cancer assessed at a Fatigue and pain were
Secondary single time point significantly associated
outcome: following with worse HRQoL
relation of admission to
symptoms to hospital
HRQoL

Schniewind, British Journal of Pancreatic head cancer Primary outcome: Prospective EORTC HRQoL, global as well as all Completion rates:
Surgery, 2006 undergoing PD or PPPD longitudinal longitudinal, N = QLQ-C30 + subscales, declined preoperative 56%,
HRQoL 91 (PPD = 34, PAN26 postoperative but returned 3 mo, 72% of survivors;
PPPD = 57); module to baseline by 3-6 mo in 24 mo, 56% of survivors
HRQoL assessed both groups. Among
preoperatively survivors, no differences
and at 3, 6, 12, were observed in HRQoL
and 24 mo
postoperatively
Han, Hepatogastroenterology, Pancreatic mass (benign Primary outcome: Cross-sectional, N = EORTC PD patients reported less
2007 and malignant) treated long-term 67 (23 PD and QLQ-C30 steatorrhea and diabetic
with PD or PPPD and gastrointestinal 44 PPPD); symptoms but reported
alive without at ≥3 yr functional HRQoL assessed more flatus, diarrhea, and
outcomes and at a single, fatigue compared with
HRQoL nonstandardized PPPD patients. Mean
postoperative scores on global HRQoL
time point subscales were higher in
the PPPD group than the
PD group (P > .05).
You, Surgery, 2007 Patients with MEN-1 Primary outcome: Cross-sectional, N = EORTC Assessed HRQoL at mean
undergoing PD for perioperative 28; HRQoL QLQ-C30 and 5.4 yr postoperatively.
pancreaticoduodenal outcomes and assessed at a a unique Global measures of
neoplasm survival single, symptom HRQoL among the treated
Secondary nonstandard index group were no different
outcome: postoperative than those of the general
HRQoL time point population; however,
symptoms of nausea/
vomiting, diarrhea, and
appetite loss were greater
(P ≤ .02). Compared with
asymptomatic patients,
symptomatic patients
reported a significantly
lower HRQoL (P = .05)
Crippa, Journal of Various stages of newly Primary outcome: Prospective FACT-Hep Median OS for the entire Survival poor overall, but
Gastrointestinal Surgery, diagnosed pancreatic HRQoL and longitudinal, N = cohort, 9.8 mo. In group 1 HRQoL improved with
2008 cancer (treatment survival by 92 divided into 3 (resected), HRQoL was surgery
course at enrollment stage/treatment groups: significantly improved (P =
unknown) Group 1: localized .03). Group 2 (locally
disease treated advanced, various
surgically (28, treatments) experienced
30.5%) no change, and group 3
Group 2: locally (metastatic disease,
advanced various treatments)
disease (34, experienced a persistent
37%) decline over time
Group 3: metastatic
disease (30,
32.5%)
HRQoL assessed at
enrollment, then
3- and 6-mo
follow-up
Continued
489
490

Kostro, Acta Chirugica Potentially resectable Primary outcome: Prospective EORTC No differences in global Completion rates:
Belgica, 2008 pancreatic cancer, longitudinal longitudinal, N = QLQ-C30 + HRQoL between the preoperatively 100%,
subsequently HRQoL 54 (26 PD, 17 PAN26 patient groups. Palliative 1 mo 46/54 (85%),
undergoing curative PD DBP, 11 HJ + GJ was associated 2 mo 42/54 (78%),
or palliative double laparotomy); with increased symptoms 3 mo 31/54 (57%),
bypass (HJ + GJ) or HRQoL assessed immediately 6 mo 22/54 (41%)
laparotomy alone preoperatively postoperatively but
and at 1, 2, 3, acceptable HRQoL. Eight
and 6 mo wk before death, all
postoperatively subscales and global
HRQoL declined rapidly
Halloran, Pancreatology, Patients undergoing partial Primary outcome: Prospective EORTC Overall, HRQoL increased at HRQoL was not measured
2011 pancreatectomy for pancreatic longitudinal, N = QLQ-C30 6 (P = .03) and 12 mo (P preoperatively; baseline
pancreatic cancer exocrine 40; HRQoL < .01) = postoperative (6 wk);
insufficiency assessed only Physical and role functioning improvements in HRQoL
Secondary postoperatively at were increased at 3 mo (P may be inflated
outcome: 1.5, 3, 6, and = .03); social functioning
HRQoL, 12 mo was improved at 6 and
nutrition, 12 mo (P = .03 and P <
symptoms .01, respectively)
Trend toward worse HRQoL
in patients who
experienced exocrine
insufficiency
Ljungman, World Journal of Resectable periampullary Primary outcome: Retrospective SF-36 Health HRQoL index lower in study HRQoL data available for
Surgery, 2011 cancer cost utility of cohort, N = 119; Survey and patients compared with only 58 patients (37
curative HRQoL assessed Utility Index controls at all time points early and 27 late); late
treatment for preoperatively, (SF-36-6D (P < .05) data subject to
pancreatic early preference- HRQoL index (0.69) at “healthy-survivor bias”
cancer postoperatively based utility long-term follow-up was
Secondary (<1 yr), or late index, scored not different from index
outcome: postoperatively 0 = death to 1 scores preoperatively
HRQoL (1-5 yr), = perfect (0.65) or early
compared with health) postoperatively (0.63) in
age-matched the treated group
healthy controls
Walter, European Journal of Advanced pancreatic Primary outcome: Prospective EORTC Outside of worse scores on Completion rates: baseline
Surgical Oncology, 2011 cancer undergoing longitudinal longitudinal, N = QLQ-C30 physical functioning 100%, discharge 78/86
palliative resection (PR) HRQoL 86 (61 HJ + GJ, subscales at 6 mo, the HJ (91%), 3 mo 58/86
or double bypass (HJ + between PR 25 PR); HRQoL + GJ group had (67%), 6 mo 41/86
GJ) and HJ + GJ assessed significantly better HRQoL (48%)
preoperatively, at and improved symptoms
discharge, and 3 compared with PR group
and 6 mo
postoperatively

Chan, Journal of Periampullary cancer Primary outcome: Prospective SF-36 Health One mo postprocedure, Completion rates:
Gastrointestinal Surgery, undergoing PD longitudinal longitudinal, N = Survey significant decline in preoperative 100%,
2012 HRQoL 37; HRQoL physical function (P < 1 mo 29/37 (78%),
assessed 0.01) and emotional role 3 mo 26/37 (70%),
preoperatively, 1, (P < .03). At 3 mo, mental 6 mo 28/34 survivors
3, 6, and 12 mo health increased (82%), 12 mo 28/28
postoperatively significantly (P = .02); survivors (100%)
6 mo, physical role (P <
.01), physical pain (P =
.01), social function (P =
.01) improved significantly.
At 12 mo, these changes
were sustained as well as
significant improvement in
vitality (P = .02) and
emotional role (P < .01)
Mbah, Journal of the Pancreatic cancer Primary outcome: Prospective EORTC Following an initial HRQoL data collected as
Pancreas, 2012 undergoing HRQoL longitudinal, N = QLQ-C30 and postoperative decline, part of a prospective
pancreatectomy following 34 HRQoL FACT-Anemia HRQoL returned to trial examining the safety
pancreatectomy assessed baseline at 6 wk. Overall, of intraoperative
in patients with preoperatively complication rate was autotransfusion during
and without and at 2-3 wk, 21%, with no difference in oncologic resections;
complication 6 wk, 3, and HRQoL scores among HRQoL stratified by
6 mo those who did and did not anemia at discharge
postoperatively have complications (P = was also evaluated
.11). At 6 mo, scores on
cognitive functioning
significantly declined (P =
.02).
Belyaev, Langenbecks Pancreatic disease (benign Primary outcome: Retrospective SF-36 Health HRQoL was worse Completion rates:
Archives of Surgery, 2013 and malignant) longitudinal cohort, N = 174 Survey compared with population preoperatively 100%,
undergoing pancreatic HRQoL (105 malignant, norms at all time points, P 3 mo 133/174 (76%),
resection comparison by 69 benign); = .03. Early postoperative 24 mo 83/174 (48%)
procedure and HRQoL assessed HRQoL was best in
diagnosis preoperatively patients undergoing distal
and at 3 and pancreatectomy and worst
24 mo with TP. Patients with
postoperatively, cancer, as opposed to
included benign diagnosis, had
comparison to lower postoperative
age-matched HRQoL and persistent
population norms decline to 24 mo. Patients
with benign tumors/
pancreatitis had initial
drop, followed by slow
trend toward recovery at
24 mo
Continued
491
492

Roberts, HPB (Oxford), 2014 Pancreatic tumors (benign Primary outcome: Cross sectional EORTC Physical (P < .01), cognitive 88/123 TP were dead at
or malignant) treated overall HRQoL matched design, QLQ-C30 (P < .01), and social time of survey
with TP and diabetes- N = 28 TP; TP +PAN26 functioning (P = .02), as administration; 28 of the
specific cohort age/ module and well as working ability (P = remaining 33 returned
problems gender matched Problem Areas .01), were worse in TP surveys
between TP to IDDM group; in Diabetes compared with IDDM
and matched HRQoL assessed Scale groups. Symptoms of
IDDM at a single, fatigue, nausea/vomiting,
nonstandardized and insomnia were also
postoperative worse (P ≤ .01); however,
time point no differences in diabetes-

specific problems were
observed
Epelboym, Journal of Surgical Pancreatic tumors treated Primary outcome: Cross sectional EORTC HRQoL assessed at median TP patients had more
Research, 2014 with TP morbidity/ matched design, QLQ-C30 + 45 mo postoperatively. hypoglycemic events (2
mortality N = 17 TP; TP PAN26 Overall and pancreas- requiring hospitalization),
post-TP cohort matched module and specific measures of but given small
Secondary on age/gender/ Audit of HRQoL were not different numbers, this was not
outcome: operative Diabetes- between TP and PD + DM statistically significant
overall, indication to Dependent groups. Diabetes
pancreas- patients who Quality of Life negatively impacted
specific and underwent PD (ADD-QOL) HRQoL and was not
diabetes- and had or different between TP and
specific HRQoL developed PD + DM groups.
following TP diabetes
postoperatively
(PD + DM);
HRQoL assessed
at a single,
nonstandardized
postoperative
time point

Serrano, International Journal Resectable or borderline Primary outcome: Prospective EORTC After 2 cycles of HRQoL assessed as part
of Radiation Oncology resectable pancreatic HRQoL during multiinstitutional QLQ-C30 + chemoradiation HRQoL of a Phase II clinical trial
Biology Physics, 2014 cancer undergoing and after Phase II clinical PAN26 was not different from of neoadjuvant
neoadjuvant chemoradiation trial, N = 55; module and baseline. Global HRQoL chemoradiation
chemoradiation and surgery HRQoL assessed FACT-Hep was not statistically (gemcitabine +
at baseline, after (EORTC QLQ-C30) or oxaliplatin + radiation) in
2 cycles of clinically (FACT-G) different patients with resectable
neoadjuvant from baseline at the or borderline resectable
therapy, after conclusion of pancreatic cancer;
surgery, at 6 mo chemoradiation; however, compliance in
from initiation of pancreatic pain was nonoperated group poor
therapy, and improved, scores of and not included in
6-mo intervals for physical function declined, longer-term analysis
2 yr and diarrhea symptoms
increased significantly.
Among patients who
underwent resection
global, QoL and most
subscales returned to
baseline measures by
6 mo
Hartwig, Annals of Surgery, Locally advanced or Primary outcome: Prospective EORTC No difference in global 596 TP or completion
2015 centrally located perioperative longitudinal, N = QLQ-C30 + HRQoL scores between pancreatectomy were
pancreatic tumors morbidity/ 81 (25 benign PAN26 TP patients and matched included in primary
undergoing TP mortality and disease, 56 controls; however, all analysis; HRQoL
survival malignant functional scale scores assessment was
Secondary disease); TP (physical, emotional, available in only 81
outcome: cohort age/ social, role, cognitive) were patients who had
longitudinal gender matched significantly lower at all follow-up at study
HRQoL overall to healthy time points in the TP center
and compared controls; HRQoL group. Global HRQoL and
with healthy assessed functional scales were not
controls preoperatively different based on
and 1, 2, 3, 4, indication for operation
and 5 yr (benign vs. malignant), but
symptoms were greater
among patients
undergoing TP for
malignant disease at yr 1
and 2 (P < .01)
DBP, double bypass; DM, diabetes mellitus; EORTC QLQ-C30, European Organization for the Research and Treatment of Cancer Quality-of-Life Questionnaire–Core 30; FACT, Functional Assessment of Cancer Therapy; GJ, gastrojejunostomy; GOO, gastric outlet
obstruction; HJ, hepaticojejunostomy; HRQoL, health-related quality of life; IDDM, insulin-dependent diabetes mellitus; MEN-1, multiple endocrine neoplasia-1; MSO, Medical Outcomes Study; OS, overall survival; PAN26, pancreatic cancer module; PD,
pancreaticoduodenectomy; PG, pancreaticogastrostomy; PJ, pancreaticojejunostomy; PPPD, pylorus-preserving pancreaticoduodenectomy; QLI, Quality-of-Life Index; RPD, radical pancreaticoduodenectomy; SF-36, 36-item Short Form; TP, total pancreatectomy.
493
Total pancreatectomy (TP) is a procedure associated with been suggested as an alternative to prophylactic surgical bypass
substantial and irreversible long-term consequences related to (HJ plus GJ), with the potential added benefit of reducing pain
pancreatic insufficiency. Patients undergoing TP are dependent associated with advanced pancreatic malignancy. When directly
on orally administered exogenous pancreatic enzymes to compared, no differences have been shown in terms of opera-
prevent malabsorptive syndromes/chronic diarrhea as well as tive morbidity/mortality and/or overall survival between pallia-
strict insulin administration regimens to balance blood sugars tive resection and bypass. Furthermore, despite the idea that
and mitigate long-term diabetic complications. Not surpris- resection improves pain control, and potentially HRQoL, it has
ingly, TP has the potential to substantially influence many been observed that patients treated with palliative resection
HRQoL domains. In a longitudinal study of 32 patients under- actually experience significantly greater declines in emotional,
going TP for locally advanced central tumors (benign and cognitive, and social-functioning scales as assessed by the
malignant), HRQoL was assessed postoperatively and annually EORTC QLQ-C30 compared with surgical bypass alone
for 4 years thereafter. Interestingly, no difference in the mean (Walter et al, 2011). In combination, these findings suggest
global HRQoL scores between healthy norms and the entire that resection provides no advantage in terms of survival or
TP population was observed; however, physical functioning HRQoL. Consequently, use of resection as a means of pallia-
among the TP group was significantly lower. Compared with tion for advanced pancreatic head tumors should be used
benign counterparts, patients undergoing TP for malignant sparingly.
disease had higher symptom scale scores, suggesting a greater To reiterate, the primary goal of palliative care is symptom
negative impact on HRQoL among this group. The number of management and maintenance of QoL. To this end, pancreatic
patients evaluable at each time point in this study was low, and stenting for the relief of intractable malignant pain has been
longer-term outcomes are subject to the “health survivor” bias evaluated. Wehrmann and colleagues assessed overall QoL
(Hartwig et al, 2015). using the Spitzer Quality-of-life Index (QLI), as well as pain
Recently, Billings and colleagues (2005), compared HRQoL control and opioid analgesic use, at baseline and every 4 weeks
outcomes of 27 long-term survivors (mean, 7.5 years) treated following intervention in patients with unresectable pancreatic
with TP for malignancy with age- and gender-matched con- carcinoma. Stenting was associated with significantly improved
trols. TP patients experienced a negative impact on their QoL pain control at 4 weeks, which persisted out to 12 weeks. QLI
and health status compared with the general population; scores paralleled pain scores and were improved versus baseline
however, compared with patients with insulin-dependent dia- at the 4-week time point, with return to baseline levels at 12
betes mellitus (IDDM) from other causes, the differences were weeks (Wehrmann et al, 2005). Stenting of the pancreatic duct
no longer observed. Similarly, Epelboym and colleagues (2014) in select patients with advanced pancreatic malignancy and
compared HRQoL between patients treated with TP versus PD significant pain transiently improves pain control and HRQoL
with IDDM. Patients in the TP group had a slightly higher and is most certainly a reasonable option for palliation of these
number of hypoglycemic events, but no significant differences patients (see Chapter 69).
were observed in global HRQoL scores. A subsequent study
matched TP patients to IDDM patients and compared global Hepatic Resection
HRQoL scores using the EORTC-QLQ-C30 plus the PAN26 In the early eras of hepatic surgery, mortality ranged from 30%
module and diabetes-specific QoL scores. Overall, TP patients to 60%, and elective intervention was rare. However, similar to
reported worse global HRQoL compared with matched IDDM pancreatic resection, improvement in surgical technique, peri-
patients; however, when using the diabetes-specific measure- operative care, and multidisciplinary team management has led
ment tool, there were no significant differences (Roberts et al, to significant decreases in morbidity and mortality (from >50%
2014). These findings suggest that although TP impacts down to 2-3%) associated with hepatic resection (Jarnagin et al,
HRQoL, the changes observed are comparable to those seen in 2002) (see Chapter 103). As a consequence, hepatic resection
patients with IDDM alone and/or other types of pancreatic has become the standard of care for definitive management of
resection. This suggests that if TP is indicated, concerns regard- many hepatic tumors. The increased use of hepatectomy for
ing the impact of pancreatic insufficiency on HRQoL should both primary and metastatic hepatic cancers has undoubtedly
not obviate intervention in well-selected patients. improved survival outcomes; however, questions regarding the
Many patients with pancreatic cancer are seen late with impact of hepatic resection on HRQoL remain. At present, our
disease that is not amenable to curative resection. Although less increasing ability to measure HRQoL in a psychometrically
common in the era of percutaneous and endoscopic interven- valid and clinically applicable fashion has resulted in a vast
tions, some patients may require or may be best served with increase in the number and quality of studies addressing
surgical palliation of gastric and/or biliary obstructions (see HRQoL as it pertains to hepatic resection and treatment of
Chapter 69). In a randomized study, Van Heek and colleagues hepatic tumors (Table 28.4).
compared HRQoL following prophylactic double bypass with HRQoL in the early postoperative period primarily reflects
hepaticojejunostomy (HJ) plus gastrojejunostomy (GJ) to single procedure-related factors (i.e., incisional pain, mobility limita-
biliary bypass (HJ) alone. In terms of HRQoL measures, no tions, complications, etc.). To date, multiple studies have eval-
differences were observed between groups. However, an 18% uated the short-term impact of hepatic resection on HRQoL
reduction in the risk of gastric outlet obstruction and need for (Chen et al, 2004; Eid et al, 2006; Langenhoff et al, 2006;
repeat intervention was observed in the HJ-plus-GJ group. Martin et al, 2007; Miller et al, 2013; Wiering et al, 2011).
HRQoL was not reassessed or compared among those who did Among patients undergoing hepatectomy for primary hepatic
and did not develop obstructive symptoms requiring secondary cancer (HCC/intrahepatic cholangiocarcinoma [ICC]), a sig-
intervention. This is unfortunate because this may have influ- nificant decline in mean scores on the Gastrointestinal Quality-
enced longer-term HRQoL (Van Heek et al, 2003). Palliative of-life Index (GQLI) was observed 2 to 10 weeks postoperatively
resection for advanced carcinoma of the pancreatic head has (Chen et al, 2004). The initial decline was followed by a

TABLE 28.4 Studies of Hepatic Resection and Health-Related Quality of Life

Study Purpose/
First Author, Journal, Population (Patient Quality-of-Life End Study Design, N, and Quality-of-Life
Year With:) Point Assessment Time Points Instrument Quality-of-Life Results Comments
Hepatic Resection
Poon, Archives of HCC Primary outcome: Prospective longitudinal FACT-G, translated At 3 mo, global HRQoL, PWB, EWB, and Disease recurrence
Surgery, 2001 undergoing longitudinal N = 76 (66 HR, 10 into Chinese SWB significantly improved over was treated with
HR HRQoL controls—unresectable baseline and were maintained out to TACE, systemic
HCC treated with 2 yr among the HR group; no changes chemotherapy,
TACE); HRQoL were observed in HRQoL measures and/or BSC
assessed q3mo for among unresected controls at 3 mo,
2 yr and decline was observed starting at
9 mo. Disease recurrence was
associated with a significant decline in
mean HRQoL (P < .001)
Chen, Hepatobiliary Primary hepatic Primary outcome: Prospective longitudinal, Gastrointestinal QoL Mean GQLI was decreased at 2-10 wk 47% (17/36) patients
and Pancreatic cancer short-term and N = 36; HRQoL Index (GQLI) postoperatively, followed by a gradual died by 9-mo
Disease undergoing longitudinal assessed return to baseline by 4 mo, and at follow-up
International, 2004 HR HRQoL preoperatively; at 2, 5, 9 mo. Mean GQLI scores were
and 10 wk; 4, 6, and increased above baseline measures.
9 mo; and 1, 1.5, and Disease recurrence was associated with
2 yr postoperatively a steady decline in HRQoL
Langenhoff, British Hepatic Primary outcome: Prospective cohort, N = EORTC QLQ-C30 and Group 1: decrease in global and functional
Journal of Surgery, malignancy short-term and 97 (group 1 = HR [n = EQ-5D-3L + HRQoL domains; decrease in EQ-VAS
2006 undergoing longitudinal 60], group 2 = EQ-VAS and symptoms at 2 wk, with return to
HR (CRLM, HRQoL unresectable at baseline at 3 mo; stable or improved
CCA, HCC, laparotomy [n = 19], out to 12 mo
and other group 3 = Group 2: decrease in global and functional
hepatic unresectable at HRQoL domains; decrease in EQ-VAS
metastases) presentation [n = 20]); and increased symptoms at 2 wk, with
HRQoL assessed at continued decline and ongoing
baseline and 0.5-, 3-, symptoms
6-, and 12-mo Group 3: no change in global, functional,
follow-up or symptoms, HRQoL domains, or
EQ-VAS at 2 wk or 3 mo, decrease in
global and functional domain-specific
QoL at 6 mo
Eid, Cancer, 2006 Hepatic Primary outcome: Prospective longitudinal, FACT-Hep, FHSI-8, Major hepatectomy associated with
malignancy comparison of N = 40 (24 major POMS, EORTC decrease in physical and functional
undergoing HRQoL based hepatectomy, 8 minor QLQ-C30 + PAN26 domain scores on FACT-Hep at 6 wk
HR or on type of hepatectomy, 8 module; global compared with minor resection or RFA.
surgical RFA surgical surgical RFA); HRQoL rating of change No differences in HRQoL measures at 3
(CRLM, CCA, intervention assessed at baseline, scales (6 domains, and 6 mo were observed between
HCC, other discharge, first scales −7 to +7) interventions. A similar trend was
metastases) postoperative visit, observed for all HRQoL measures
6 wk, 3 mo, and 6 mo (POMS, EORTC QLQ-C30/PAN36,
FHSI-8, global rating scales)
Continued
495
TABLE 28.4 Studies of Hepatic Resection and Health-Related Quality of Life—cont’d
496
Study Purpose/
First Author, Journal, Population (Patient Quality-of-Life End Study Design, N, and Quality-of-Life
Year With:) Point Assessment Time Points Instrument Quality-of-Life Results Comments
Lee, Journal of HCC Primary outcome: Cross-sectional cohort, N WHOQoL-BREF and Compared with population norms, HCC 70% Hepatitis B
Surgical Oncology, HRQoL in HCC = 161 (121 HR, 31 EORTC QLQ-C30 was associated with decrease in social
2007 compared with TACE, 8 PEI, 1 BSC) (Taiwanese and psychological domains and
population HRQoL assessed at translation) improved environment domains; HR
norms and single time point was associated with improved HRQoL
between compared with TACE/PEI/BSC
treatment type
Martin, Surgery, 2007 CRLM, CCA, or Primary outcome: Prospective longitudinal FACT-HEP, FHSI-8, Major hepatectomy associated with Major hepatectomy
HCC time to return to N = 32 (24 major EORTC QLQ-C30 + decline in all measures of HRQoL = ≥3 Couinaud
undergoing baseline HRQoL hepatectomy, 8 minor PAN26 module, postoperatively; HRQoL nadir scores segments
HR hepatectomy); HRQoL POMS; global observed at 6 wk, with return to
assessed at consent, rating scale (6 baseline values by 3 mo. Minor
discharge, first domains, scales −7 resection was associated with
postoperative visit, to +7) decrements in all measures at
6 wk, then 3, 6, and discharge; nadir HRQoL scores
12 mo observed at initial postoperative visit
and returned to baseline by 6 wk
postoperatively
Dasgupta, British Hepatic Primary outcome: Prospective longitudinal, EORTC QLQ-C30 Decrease from baseline to 6 mo in 44/103 (43%) alive
Journal of Surgery, malignancy longitudinal N = 103 (74 CRLM, 9 physical functioning domain and at last follow-up
2008 undergoing HRQoL CCA, 8 HCC, 12 increase in dyspnea/fatigue. At 12 mo,
HR (CRLM, other); HRQoL physical function and fatigue return to
CCA, HCC, assessed at baseline, baseline but dyspnea persistent.
other 6, 12, and 36-48 mo Survivors with no recurrence at
metastases) 36-48 mo had improved global HRQoL
over patients with disease recurrence

Banz, World Journal HR for benign Primary outcome: Cross-sectional cohort, N EORTC QLQ-C30 + Patients who underwent hepatic resection HRQoL assessment
of Surgery, 2009 or malignant impact of = 135 (89 malignant LM21 module for malignant disease had similar at a mean of
disease at postoperative disease, 46 benign general, global, and self assessed 27 mo
least 6 mo diagnosis disease); HRQoL HRQoL relative to those with benign postoperatively
prior to (benign/ assessed at a single, diagnoses. However, physical function
analysis malignant) on nonstandardized scores and pain, fatigue, and social
HRQoL postoperative time function scores were worse in the
point malignant group
Bruns, World HR for benign Primary outcome: Cross-sectional cohort, N SF-12 (mental MCS significantly lower among patients HRQoL assessed
Journal of or malignant HRQoL and = 96 (76 malignant [21 component scale with benign vs. malignant diagnosis as once between
Gastroenterology, disease at identification of primary, 55 [MCS] + physical well as primary vs. metastatic cancers 3-36 mo
2010 least 3 mo variables metastases], 20 component scale (P < .05). No difference in QoL based postoperatively
prior to associated with/ benign); HRQoL [PCS]), ad hoc on sex, age, or postoperative
analysis predictive of assessed at a single, symptom and pain complications. Increase in symptoms/
decrease nonstandardized scale pain and decreases in daily activities
HRQoL postoperative time were associated with worse PCS/MCS
point

Wiering, British ≤4 CRLM with Primary outcome: Prospective longitudinal, EQ-5D-3L + EQ-VAS Overall, decline in global HRQoL at 3 and
Journal of Surgery, no EHD short-term and N = 138 (117 curative 6 wk postoperatively. Curative HR was
2011 undergoing longitudinal HR, 19 unresectable associated with return of global HRQoL
HR HRQoL at laparotomy); HRQoL to baseline at 3 mo and stabilization to
assessed at baseline, 3 yr. Patients found to be unresectable
then 3 wk, 6 wk, and at laparotomy had persistent decline in
q3mo-3 yr QoL during the study period. Disease
recurrence associated with decline in
HRQoL during time. Patients with
recurrence amenable to re-resection
had global HRQoL scores similar to
patients never experiencing recurrence
Rees, Journal of CRLM Primary outcome: Prospective longitudinal, EORTC QLQ-C30 + Twenty percent of patients had severe
Clinical Oncology, undergoing longitudinal N = 232; HRQoL LM21 module symptoms at baseline. At 3 mo
2012 HR HRQoL assessed postoperatively, all functional scales
preoperatively and at decreased, proportion with severe
3, 6, and 12 mo symptoms increased, and functional
postoperatively scales returned to baseline at 6 mo and
stabilized thereafter. Ten percent of
patients continued to have severe
problems with pain and sexual function
at 1 yr postoperatively
Miller, American Hepatic Primary outcome: Prospective longitudinal, EORTC QLQ-C30, Social and functional domains scores Anemia defined at
Journal of Surgery, malignancy impact of N = 41 (16 CRLM, 9 FACT-Anemia, lower in anemic group at all time points. discharge as
2013 undergoing anemia and HCC, 4 CCA, 12 global change Major complications associated with <10 g/dL; 34%
HR of > 2 postoperative other); HRQoL rating scale (−7 to increased pain at 6 wk and 6 mo but overall
segments complications assessed +7) no difference in any domain-specific complication rate,
(CRLM, CCA, on short term preoperatively, at first QoL scales 25% major
HCC, other HRQoL postoperative visit, complication rate
metastases) then 1.5, 3, and 6 mo
postoperatively
Mise, World Journal HCC Primary outcome: Prospective longitudinal, SF-36 Health Survey PCS did not change following surgery. At
of Surgery, 2014 undergoing longitudinal N = 69; HRQoL 9 mo, MCS were significantly improved
HR HRQoL assessed versus baseline and population norms;
Secondary preoperatively and female sex, age > 70 yr,
outcome: q3mo-1 yr thoracoabdominal incisions, tumors >
identification of 5 cm, and ICGR-15 < 10% were
perioperative associated with worse PCS at 3 mo; no
predictors of clinical variables were predictive of MCS
HRQoL at 3 mo
Rees, British Journal CRLM treated Primary outcome: Long-term follow-up of EORTC QLQ-C30 + Overall, scores in all domains excellent at HRQoL assessment
of Surgery, 2014 with HR ≥ long-term prospective cohort, N LM21 module long-term follow-up and were at median of 8
5 yr prior to HRQoL = 68; HRQoL significantly improved from baseline; (6.9-9.2) yr
assessment assessed at single, <5% of patients reported severe postoperatively
nonstandardized time symptoms; persistent symptoms
point ≥ 5 yr included peripheral neuropathy, sexual
postoperatively dysfunction, constipation, and diarrhea
BSC, Best supportive care; CCA, cholangiocarcinoma; CRLM, colorectal liver metastases; EHD, extrahepatic disease; EORTC QLQ-C30, European Organization for the Research and Treatment of Cancer Quality-of-Life Questionnaire–Core 30; EWB, emotional well-being;
FACT-G, Functional Assessment of Cancer Therapy–General; FHSI-8, FACT-Hepatobiliary Symptom Index-8; HCC, hepatocellular carcinoma; HR, hepatic resection; HRQoL, health-related quality of life; ICGR-15, indocyanine green retention rate at 15 minutes; LM21,
colorectal liver metastases module; PAN26, pancreatic cancer module; PEI, percutaneous ethanol injection; POMS, profile of mood states; PWB, physical well-being; RFA, radiofrequency ablation; SF-12, 12-item Short Form; SWB, social well-being; TACE, transarterial
chemoembolization; VAS, visual analogue scale; WHOQoL-BREF, World Health Organization Quality-of-Life–BREF (abbreviated) assessment.
497
gradual return to baseline by 4 months, and in patients without CRLMs. However, in a recent study by Rees and colleagues
recurrence, mean GQLI was greater than baseline assessment (2012, using the EORTC QLQ-C30 in conjunction with the
by 9 months. This postoperative trend in HRQoL following liver-specific module, found that as many as 10% of patients
hepatic resection has been observed in multiple studies (Eid had persistent severe symptoms of pain and decrements in
et al, 2006; Langenhoff et al, 2006; Martin et al, 2007; Miller sexual function at 1 year following resection of CRLMs.
et al, 2013; Wiering et al, 2011) and appears similar regardless Longer-term follow-up of this study cohort suggests that, even
of underlying tumor type (primary/metastatic) or HRQoL mea- at a median of 8 years follow-up, 5% of patients still have sig-
surement tool used (EORTC QLQ-C30, FACT-G, FACT- nificant symptoms (Rees et al, 2014). This study did not
Hep, GQLI, etc.). Although the extent of resection may alter account for the use of preoperative or postoperative chemo-
the degree/duration of HRQoL decline postoperatively, in the therapy or further treatments. Consequently, it is difficult to
absence of disease recurrence, the overall trend remains the discern if the reported persistent symptoms are related to
same (Chen et al, 2004; Eid et al, 2006; Martin et al, 2007). CRLMs, hepatic resection, previous surgery for primary tumor,
In direct contrast to patients undergoing curative resection, or ongoing oncologic therapy.
patients found to have unresectable disease at the time of lapa- Overall survival is increased among patients with resectable
rotomy experience a steady decline in global HRQoL and all CRLMs. Unfortunately, disease recurrence requiring further
functional subscales, with persistence of symptoms throughout oncologic therapy is the norm and may influence HRQoL. In
follow-up (Langenhoff et al, 2006; Wiering et al, 2011). Lapa- a well-designed prospective study by Wiering and colleagues
rotomy for hepatic malignancy, without resection of disease (2011), patients undergoing hepatic resection and having less
negatively impacts both early and longer-term HRQoL out- than or equal to four CRLMs were followed for 3 years and
comes. As such, meticulous preoperative evaluation is essential HRQoL assessed at predefined time points. Overall, disease
to avoid subjecting patients to futile intervention. recurrence was associated with a decline in HRQoL as mea-
Compared with short-term postoperative HRQoL, longitu- sured by the EQ-5D-3L plus EQ-VAS (visual analogue scale).
dinal evaluation (≥3 to 6 months) following hepatic resection However, this decline in HRQoL was mitigated by the ability
is less reflective of the technical procedure and more likely to undergo repeat curative resection of recurrent disease.
related to patient- and disease-specific factors (Banz et al, HRQoL scores of patients who underwent complete re-resection
2009; Bruns et al, 2010). Overall, patients with HCC have were similar to those of patients who never experienced a recur-
worse psychological and social functioning compared with rence. Dasgupta and colleagues (2008) also report a decrement
population norms (Lee et al, 2007), and not surprisingly, in global HRQoL on the EORTC QLQ-C30 in the setting of
patients amenable to complete resection have improved recurrence. Similar decrements in HRQoL have been observed
HRQoL compared with HCC treated with transarterial che- following recurrence of HCC (Chen et al, 2004; Poon et al,
moembolization (TACE), percutaneous ethanol injection, and/ 2001), and following repeat resection of recurrent HCC,
or best supportive care (Toro et al, 2012) (see Chapters 91, 96, HRQoL trends parallel to what is seen with CRLMs (Tanabe
and 98). In a long-term evaluation of hepatic resection for et al, 2001). At present, it is reasonable to suggest that recur-
HCC in patients with persevered hepatic function, significant rence of disease following hepatic resection for malignancy
improvements in mean HRQoL and subscale scores were portends worse HRQoL; however, this decrement may be tran-
observed at 3 months and persisted out to 2 years follow-up sient if complete surgical reexcision can be achieved.
(Poon et al, 2001). Similar improvement was not observed in
a control group with HCC undergoing TACE. Chen and col- Locoregional Treatment of Hepatic Tumors
leagues (2004) also observed similar long-term trends in Locoregional therapy (LRT), including hepatic arterial embo-
patients undergoing hepatic resection for HCC. However, find- lization and ablation, may be used with curative and/or pallia-
ings from these studies contrast with those of Mise and col- tive intent for both primary and metastatic tumors of the liver
leagues (2014), in which no change in the physical component (see Chapters 96 to 98). These therapies treat disease/symptoms
scale on SF-36 and minimal change on the mental component using percutaneous methods, thereby avoiding the morbidity of
scale were observed following hepatectomy for HCC. The an open or laparoscopic surgical procedure. In addition to
measurement tools used in these studies differed (disease- bland embolization (hepatic artery embolization), chemother-
specific FACT-G vs. generic SF-36) in terms of sensitivity to apy and radiotherapy may be delivered to the liver locally via
change and may account for the observed differences. It is TACE and yttrium-90 (90Y) microsphere radioembolization
likely that longer-term HRQoL assessment in patients with (see Chapter 96). Ablative therapies vary by the agent/energy
HCC is heavily dependent on the type and severity of underly- source used to induce cell damage/death and include thermal
ing hepatic disease as well as other associated comorbidities. (radiofrequency, microwave, or cryotherapy) (see Chapters 98B
Future studies stratifying for these potential confounders are and 98C), electrical (irreversible electroporation) (see Chapter
necessary. 98C), and chemical (alcohol injection) (see Chapter 98D) tech-
CRLMs are the most common indication for hepatic resec- niques. The use of LRT to treat hepatic tumors has increased
tion, and multiple studies have assessed longer-term HRQoL substantially, and the impact on morbidity/mortality and sur-
outcomes among these patients (Banz et al, 2009; Bruns et al, vival has been well documented. In general, these types of
2010; Diouf et al, 2014; Earlam et al, 1996; Langenhoff et al, “hard” clinical outcomes are similar across locoregional thera-
2006; Rees et al, 2014) (see Chapter 92). Based on the litera- pies; as such, studies evaluating HRQoL and LRT for primary
ture, there is general consensus that following an initial post- and metastatic hepatic tumors are rapidly emerging (Table
operative decline, global HRQoL tends to return to baseline 28.5).
and at times exceeds baseline scores during longer-term HCC is the most common primary hepatic tumor (see
follow-up. Likewise, disease-specific symptoms and symptom Chapter 91). Because HCC typically arises in the setting of
scales tend to follow a similar trajectory after hepatectomy for underlying cirrhosis, treatment decisions are dependent on

TABLE 28.5 Studies of Locoregional Therapy for Hepatic Tumors and Health-Related Ouality of Life
Study Design, N, and
First Author, Journal, Population (Patient Study Purpose/Quality-of-Life HRQoL Assessment Time Quality-of-Life
Year With) End Point Points Instrument Quality-of-Life Results Comments
Locoregional Treatments
Steel, HCC treated with Primary outcome: HRQoL Prospective cohort, N = FACT-Hep Overall, HRQoL and scores on all Overall, survival at 6 mo
Psychooncology, TACE (cisplatin) and survival 28; HRQoL assessed subscales declined from baseline not different between
2004 or 90Y preprocedure and 3, to 6 mo for both groups. At 3 mo, groups; only 14
radioembolization 6, and 12 mo FWB and overall HRQoL scores patients evaluable at
postprocedure were higher in 90Y group (P < .01). 6 mo
At 6 mo, FWB remained
significantly improved over TACE,
but overall HRQoL was not
different
Ruers, Annals of Nonresectable Primary outcome: Prospective EuroQoL-5D-3L Baseline HRQoL similar between All patients underwent
Surgical Oncology, CRLM treated comparison of longitudinal, N = 201 and EORTC groups. All groups experienced an initial laparotomy; no
2006 with surgical longitudinal HRQoL (117 HR, 45 surgical QLQ-C30 initial decline in HRQoL and difference in median
ablation ± and survival ablation, 39 CTX) physical function at 3 wk OS between ablation
resection vs. survival data; HRQoL postprocedure. Ablation and HR and CTX groups (31
systemic assessed groups return to baseline at 3 mo, and 26 mo,
chemotherapy preprocedure and at whereas CTX did not. HR was respectively)
(CTX) 3, 6, 9, and 12 mo associated with higher HRQoL
for randomly than both ablation and CTX;
selected members of however, HRQoL in ablation group
the cohort; N = 109 was significantly higher than CTX
(53 HR, 29 surgical group from 3-12 mo
ablation, 27 CTX)
Wang, Quality of Life HCC treated with Primary outcome: Prospective FACT-G (Chinese TACE + RFA group higher overall
Research, 2007 TACE or TACE+ short-term longitudinal, N = 83 translation) FACT-G, SWB, and FWB scores
percutaneous postprocedure HRQoL (43 TACE, 40 TACE at 3 mo postprocedure (P < .01).
RFA + RFA); HRQoL Predictors of HRQoL were
assessed Child-Pugh class and tumor
preprocedure and recurrence
3 mo postprocedure
Bonnetain, Quality of Advanced HCC Primary outcome: Retrospective Spitzer QLI Higher Spitzer QLI at baseline Combined HRQoL data
Life Research, treated with determine the value of evaluation of 2 associated with increased median from 2 RCTs of
2008 tamoxifen, TACE, baseline HRQoL scores randomized clinical OS (P < .01). In addition to palliative treatment
or BSC in the in predicting overall trials, N = 489; common variables used in for HCC: (1)
palliative setting survival HRQoL assessed at classification systems (Okuda/ tamoxifen to best
(primarily baseline only CLIP/BCLC), HRQoL was the supportive care (N =
alcohol-related single best prognostic factor for 416), (2) TACE to
cirrhosis) survival tamoxifen only (N =
122)
Kalinowski, Unresectable liver; Primary outcome: safety Prospective EORTC Overall, an initial decrease in mean Late decline in HRQoL
Digestion, 2009 only and efficacy of 90Y longitudinal, N = 9; QLQ-C30 + HRQoL scores occurred following was associated with
90
neuroendocrine Secondary outcome: HRQoL assessed LMC21 Y. At 6 mo, the majority of tumor recurrence/
tumor longitudinal HRQoL preprocedure, and 3, module patients had improvement in disease progression
metastases 6, 9, 12 mo HRQoL scores, followed by a
treated with 90Y postprocedure trend toward decline HRQoL at
12 mo
Continued
499
TABLE 28.5 Studies of Locoregional Therapy for Hepatic Tumors and Health-Related Ouality of Life—cont’d
500

First Author, Journal, Population (Patient Study Purpose/Quality-of-Life HRQoL Assessment Time Quality-of-Life
Year With) End Point Points Instrument Quality-of-Life Results Comments
Kuroda, Journal of HCC treated with Primary outcome: Prospective cohort, N = SF-36 Health BCAA group experienced Hepatitis C patients
Gastroenterology percutaneous comparison of hepatic 49; HRQoL assessed Survey improvement in overall health as only
and Hepatology, RFA with and function and nutritional preprocedure and well as physical and social
2010 without status following RFA 1 yr functioning (P < .05 for all
postprocedure with and without BCAA subsets); HRQoL did not change
BCAA supplementation in no-BCAA group following RFA
supplementation Secondary outcome:
HRQoL
Wible, Journal of Previously untreated Primary outcome: HRQoL Prospective SF-36 Health Baseline HRQoL significantly lower Completion rates: 4 mo
Vascular and HCC undergoing Secondary outcome: longitudinal, N = 73 Survey compared with population norms. 48/73 (66%), 8 mo
Interventional TACE HRQoL in patients (≥3 TACE N = 23); Mental health score at 4 mo 38/73 (52%), 12 mo
Radiology, 2010 treated with ≥3 TACE HRQoL assessed postprocedure, but not at 8 or 28/73 (38%)
procedures preprocedure and 4, 12 mo, was improved (P = .05).
8, and 12 mo No other changes in HRQoL were
postprocedure observed in study cohort over
time. The subset of patients
undergoing ≥3 TACE procedures
with improved mental health after
first and second procedures,
whereas bodily pain improved (P
< .01, P = .05, respectively), and
vitality scores declined (P = .04)
after first procedure only
Eltawil, HPB Unresectable, Primary outcome: HRQoL Prospective WHOQoL-BREF Overall HRQoL did not decline in the
(Oxford), 2012 nonablatable and treatment efficacy longitudinal, N = 48 12 mo following TACE. After third
primary hepatic patients; HRQoL TACE procedure (~1 yr), a trend

tumors (HCC and assessed before first toward a decline in the physical
ICC) undergoing TACE and prior to health domain was observed (P =
TACE each subsequent .08) and coincided with increasing
procedure (3-4 mo) AFP and tumor size
Shun, Oncologist, HCC treated with Primary outcome: Prospective cohort, N = SF-12 Health Mean HRQoL was improved from
2012 TACE short-term HRQoL 89; HRQoL assessed Survey, discharge to 8 wk postprocedure.
following single TACE within 3 days prior to Symptom At all time points, physical
procedure discharge, and 4 and distress Scale component scores were lower
Secondary outcome: 8 wk (SDS), than mental component scores.
variables associated Hospital Factors associated with lower
with changes in Anxiety and physical component scores: age,
postprocedure HRQoL Depression new diagnosis, higher levels of
Scale (HADS) symptom distress, and
depression. Factors associated
with lower mental component
scores: male sex, recurrent
disease, higher levels of anxiety,
and depression

Toro, Surgical HCC undergoing Primary outcome: HRQoL Prospective FACT-Hep (Italian Postoperatively, all domains of Percentage change in
Oncology, 2012 HR, TACE, in all patients with HCC longitudinal, N = 51 translation) HRQoL significantly improved in HRQoL from baseline
percutaneous regardless of treatment (14 HR, 15 TACE, 9 the HR group and were was calculated at
RFA, or NT RFA, 13 NT); HRQoL significantly higher compared with each time point. The
assessed all other treatments. Mean PWB average percentage
preprocedure and 3, and EWB following RFA was change in each
6, 12, and 24 mo improved compared with TACE group was used for
postprocedure and NT comparison
90
Salem, Clinical HCC treated with Primary outcome: Prospective cohort, N = FACT-Hep No difference in overall HRQoL Y patients had greater
Gastroenterology TACE or 90Y short-term 56 (27 TACE, 29 between treatment groups. At tumor burden and
90
and Hepatology, postprocedure HRQoL Y); assessed 4 wk, 90Y had a greater more advanced
2013 preprocedure and 2 improvement in SWB, FWB, disease; ESS
and 4 wk embolotherapy-specific score consists of the
postprocedure (ESS), and a trend toward better following items: pain,
overall HRQoL and TOI compared bothered by
with TACE group treatment side
effects, ability to
work, diarrhea, and
appetite
Hamdy, Journal of HCC treated with Primary outcome: Prospective cohort SF-36 Health At baseline, RFA and TACE groups Hepatitis C patients
the Egyptian percutaneous short-term design, N = 120 (40 Survey had significantly worse physical only
Society of RFA or TACE postprocedural HRQoL TACE, 40 RFA, 40 function, energy/fatigue, pain, and
Parasitology, 2013 HCV without HCC general health compared with
[controls]); HRQoL controls. Significant improvements
assessed in all HRQoL domains were
preprocedure and observed at 1 mo in the RFA
1 mo postprocedure group (P < .05) but not the TACE
group
Huang, British Small HCC (<3 cm) Primary outcome: Prospective FACT-Hep Baseline HRQoL was similar Hepatitis B patients
Journal of treated with HR longitudinal HRQoL longitudinal, N = 389 between groups. For both groups, only; no difference in
Surgery, 2014 or percutaneous and survival (HR, RFA); HRQoL following an initial postprocedural DFS and OS
RFA assessed decline, HRQoL consistently between groups;
preprocedure, and 3, improved and exceeded baseline survey completion
6, 12, 24, and values at 6-12 mo. Compared rates >95% at all
36 mo with HR, RFA group had time points
postprocedure significantly higher overall HRQoL
scores at all time points. On
multivariate analysis, HR was
independently associated with
worse HRQoL
Kolligs et al, Liver Unresectable Primary outcome: safety Pilot randomized FACT-Hep Efficacy of treatment for local tumor SIRTACE was an
International, 2015 liver- only HCC, and short-term controlled trial, N = control was similar between open-label
with preserved postprocedure HRQoL 18 (8 SIRT, 10 TACE) groups (73% disease control multicenter
hepatic function HRQoL assesses TACE vs. 77% disease control randomized control
treated with SIRT preprocedure and 1, SIRT). At baseline, SIRT patients pilot study; TACE at
or TACE 6, and 12 wk had significantly worse PWB 6-wk intervals until
postprocedure scores ; however, no differences fully treated or
were observed between groups in disease progression
overall HRQoL or subscales at 6 (gold standard) vs.
or 12 wk postprocedure single session SIRT
AFP, α-Fetoprotein; BCAA, branched-chain amino acids; BCLC, Barcelona Clinic Liver Cancer Classification; BSC, best supportive care; CLIP, Cancer of the Liver Italian Program HCC (heptocellular carcinoma) Classification system; CRLM, colorectal liver metastases;
CTX, systemic chemotherapy; DFS, disease-free survival; EORTC QLQ-C30, European Organization for the Research and Treatment of Cancer Quality-of-Life Questionnaire–Core 30; EWB, emotional well-being; FACT, Functional Assessment of Cancer Therapy; FWB,
functional well-being; HAE, hepatic artery embolization; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; Hep, hepatobiliary; HR, hepatic resection; HRQoL, health-related quality of life; ICC, intrahepatic cholangiocarcinoma; NT, no treatment; Okuda, Okuda
Prognostic Classification; OS, overall survival; PWB, physical well-being (FACT); QLI, Quality-of-Life Index; RCTs, randomized controlled trials; RFA, radiofrequency ablation; SF-36, 36-item Short Form; SIRT, selective internal radiotherapy; SWB, social well-being (FACT);
501
TACE, transarterial chemoembolization; TOI, Trial Outcome Index; WHOQoL-BREF, World Health Organization Quality-of-Life project (short form of WHOQoL-100); 90Y, yttrium-90 radioembolization.
tumor characteristics/disease burden, hepatic function, and actually one of the strongest predictors of overall survival
performance status. Transplant and hepatic resection are the (Bonnetain et al, 2008).
primary curative therapies; however, these are limited to a To date, no single LRT modality has consistently been
minority of patients (see Chapters 103 and 115). As a conse- shown to be superior to others in terms of HRQoL outcomes.
quence, LRT has emerged as the mainstay of treatment for All of the studies, excluding a very small (n = 18) pilot study,
many patients with HCC. Although some patients may be are nonrandomized and thus subject to inherent disease- and
rendered disease free, LRT in the setting of HCC is typically patient-related selection bias. Review of the literature suggests
noncurative and offered to provide local disease control and that in general, 90Y radioembolization and RFA are associated
ameliorate symptoms. As a consequence, interest in HRQoL with better HRQoL compared with TACE (Hamdy et al, 2013;
associated with LRT has focused primarily on this population Salem et al, 2013; Steel et al, 2004; Toro et al, 2012), and
of patients. Although, LRT is used in the treatment of hepatic TACE in combination with RFA is yields better outcomes than
metastases from neuroendocrine tumors (NET) and CRLMs, TACE alone (Wang et al, 2007). TACE and 90Y radioemboliza-
limited data are available regarding the impact of LRT on tion are typically used to treat multifocal unilobar/segmental
HRQoL in these populations (Kalinowski et al, 2009; Ruers disease, whereas RFA is indicated for small solitary lesions or
et al, 2006). low-volume multifocal disease; as such, comparison between
Compared with published population norms, patients with treatments without randomization or matching for extent of
HCC have lower overall HRQoL (Hamdy et al, 2013). Mul- disease/hepatic function is limited. A single, small, randomized
tiple studies have evaluated short-term (0 to 3 months) post- pilot study of TACE versus selective internal radiotherapy was
procedural HRQoL among patients with HCC treated with recently completed. HRQoL was a secondary end point of the
LRT (Hamdy et al, 2013; Kolligs et al, 2015; Salem et al, trial, and when using the FACT-Hep scale, no significant dif-
2013; Shun et al, 2012; Wang et al, 2007). In general, regard- ferences were observed between the treatments arms at 6 or 12
less of modality (TACE/90Y radioembolization/radiofrequency weeks of follow-up.
ablation [RFA]) and/or HRQoL instrument used, most patients Despite the limitations of the available literature, small
experience a transient decrement in overall HRQoL very early sample sizes, heterogenous patient populations, and lack of
(1 to 3 weeks) after intervention. This is typically followed by randomization, it is reasonable to conclude that HRQoL follow-
an improvement in most subscales/domains (physical, emo- ing LRT for HCC is improved in the short term, whereas
tional, social, mental/psychological), global HRQoL, and longer-term outcomes vary and are likely strongly influenced by
symptoms from 4 to 12 weeks. patient- and disease-specific variables. Furthermore, the clinical
The impact of LRT on longer-term HRQoL has also been indications and patient populations amenable to TACE versus
extensively evaluated (Eltawil et al, 2012; Huang et al, 2014; RFA versus 90Y radioembolization are inherently different. At
Kuroda et al, 2010; Steel et al, 2004; Toro et al, 2012; Wible present, no one form of LRT has consistently been associated
et al, 2010). Similar to short-term evaluations most longitudi- with superior HRQoL. Overall, in patients with unresectable,
nal studies report an initial early decline in HRQoL regardless nontransplantable liver, only HCC treatment with LRT pro-
of LRT modality. However, because the natural history of vides local disease control and likely improves patients overall
HCC differs based on the etiology of underlying hepatic HRQoL during the first 6 to 12 months following intervention.
disease (i.e., hepatitis C virus cirrhosis vs. hepatitis B virus cir-
rhosis), it is not surprising that reported longitudinal HRQoL Palliative Treatments
outcomes with LRT treatment of HCC vary (see Chapters 70 Surgery for tumors of the pancreas, liver, and bile ducts has
and 76). Steel and colleagues (2004) found a persistent improved over time; however, as many as 85% of patients with
decrease in global HRQoL and all subscales from baseline to 6 HPB malignancies are seen late with advanced-stage disease
months following TACE (cisplatin) or 90Y radioembolization. that is not amenable to surgical intervention. In this setting,
Interestingly, in this study, despite having greater disease significant complications related to the precarious location of
burden and lower reported HRQoL at treatment initiation, these tumors are common and often require intervention to
patients undergoing 90Y radioembolization had better func- ameliorate associated symptoms. The goal of treatment in this
tional well-being at 3 and 6 month follow-up when compared setting is palliative. It is therefore critical that interventions
with those treated with TACE. Other studies of TACE (Eltawil minimize morbidity and improve symptoms to maintain or
et al, 2012; Wible et al, 2010) as well as RFA (Kuroda et al, improve HRQoL.
2010) for unresectable HCC have reported no significant
changes in HRQoL parameters during longer-term follow-up. Gastric Outlet Obstruction
Toro and colleagues (2012) evaluated longitudinal HRQoL Malignant gastric outlet obstruction (GOO) occurs when the
among patients undergoing hepatic resection, RFA, TACE, or pyloric channel or duodenum is externally compressed by
no treatment (NT). With the exception of the NT group, all tumor and/or internally obstructed because of local tumor inva-
patients (resection, RFA, TACE) reported persistent improve- sion (see Chapter 69). It is a debilitating complication of
ments in HRQoL from 3 to 24 months. In this study, com- advanced malignancies, occurring in 20% to 30% of periampul-
pared with TACE and NT groups, patients treated with RFA lary cancers. The predominant symptoms associated with GOO
had significantly better physical and emotional well-being. include early satiety, bloating, nausea, vomiting, reflux, and
Improved longer-term HRQoL following LRT was also abdominal pain, often leading to anorexia, weight loss, and
observed in a recent study comparing hepatic resection with cachexia. Treatment of GOO is not only essential to improve
RFA in patients with small (<3 cm) HCC (Huang et al, 2014). symptoms but also to allow suitable candidates to return to or
HRQoL is not only important as a clinical outcome, but cor- initiate other oncologic therapy. Before the development
relational studies have shown that among patients undergoing of endoscopic and minimally invasive techniques, treatment
transarterial embolic therapy for HCC, baseline HRQoL is of GOO required laparotomy and open surgical bypass
(gastrojejunostomy [GJ]). More recently, advancements in postprocedure time point); as such, the conclusions regarding
both laparoscopic and endoscopic techniques has led to the HRQoL outcomes are difficult to interpret.
development of less invasive means of treating GOO, including Overall, studies reporting on HRQoL following palliative
laproscopic gastrojejunostomy (LGJ), percutaneous endoscopic procedures for GOO are limited in size, follow-up, and most
gastrostomy/jejunostomy (PEG/PEJ), or endoscopic/radiologic importantly, response rates. Reported median overall survival
duodenal stenting (DS). At present, determining which inter- of patients with advanced/end-stage malignancy and GOO
vention to use requires a balanced assessment of life expec- ranges from 7 to 20 weeks; therefore even short-term follow-up
tancy, risks and benefits of each intervention, the potential need can be difficult, and results are inherently biased in that patients
for repeat intervention, and the overall impact on HRQoL. In returning questionnaires/surveys represent a “healthy survivor”
the context of palliative interventions for GOO, HRQoL has subgroup. Furthermore, HRQoL is most commonly reported
only been assessed as the primary outcome in one study; as a secondary outcome of safety and efficacy trials; as such,
however, it has been assessed extensively as a secondary minimal HRQoL data is reported (i.e., single global measures
outcome (Table 28.6). It should be noted that many studies of reported as means, little evaluation of subscales, etc.), and the
HRQoL related to treatment of GOO have used a tool designed robustness of analysis and methodology is suboptimal. Despite
specifically for this patient population, called the Gastric Outlet this, available literature would suggest that relief of obstruction
Obstruction Scoring Scale (GOO-SS) (Dolz et al, 2011; van obviates the main symptoms associated with GOO, leading to
den Berg et al, 2013; van Hooft et al, 2009). Although the scale a stabilization of HRQoL. The palliative nature of intervention
only evaluates a single domain (physical), it is often reported for GOO demands assessment of HRQoL. At present, addi-
as a surrogate for HRQoL. tional prospective studies with HRQoL as a primary outcome
To date, a single randomized trial of 27 patients with malig- measure are necessary to define the true relationship between
nant GOO (56% pancreatic cancer) treated with DS (n = 13) treatment of malignant GOO and HRQoL.
versus LGJ (n = 14) has been completed (Mehta et al, 2006).
The underlying fragility of this group of patients is highlighted Malignant Biliary Obstruction
by the high rate of hospital mortality (DS at 17% and LGJ Malignant biliary obstruction (MBO) is common in the setting
at 23%). Not surprisingly, LGJ was associated with longer of pancreaticobiliary cancers (see Chapters 49, 51, and 65). In
hospital stay (LOS) and greater postprocedural complications fact, jaundice is often the initial clinical presentation of these
compared with DS. However, the procedural success rate in tumors. Hyperbilirubinemia is associated with increased periop-
the DS group (77%) was lower than in the LGJ group (100%). erative risk in certain situations and is a contraindication to many
Physical and mental HRQoL scores on the SF-36 at baseline chemotherapy regimens. Furthermore, biliary obstruction may
were similar between groups. Among evaluable patients at be associated with symptoms of anorexia, weight loss, pruritus,
1 month follow-up (n = 13, 6 LGJ and 7 DS), patients treated and malabsorption, leading to significant functional impair-
with DS had improved physical HRQoL scores, whereas no ment. As such, intervention to alleviate MBO is undertaken for
changes in HRQoL were observed in the LGJ group. The two primary reasons: to allow initiation of definitive oncologic
authors concluded that DS provided improved palliation versus therapy (surgery and/or chemotherapy) and/or to obviate associ-
LGJ based on decreased LOS, morbidity, and increased physi- ated symptoms. In the past, surgical bypass with hepaticojeju-
cal HRQoL at 1 month; however, this must be viewed in light nostomy (HJ) was the mainstay of treatment for MBO. However,
of the decreased procedural success rates of DS and need for in the current era, ES and radiologically guided percutaneous
potential repeat intervention. Similarly, Schmidt and colleagues biliary drainage (PBD) are the most common procedures used
(2009) observed improvements in global HRQoL, physical to relieve jaundice (Artifon et al, 2006; Mihalache et al, 2010)
and role functioning, and a decrease in symptoms following (see Chapters 29 and 30). Most recently, intraluminal brachy-
palliative intervention (endoscopic stenting [ES] and surgical therapy (ILBT) has also been proposed as a mechanism to treat
bypass) among 50 patients with malignant GOO at 3 months MBO. Initial reports regarding ILBT for MBO suggest improve-
postprocedure. ment in HRQoL and are encouraging. (Aggarwal et al, 2013).
DS is the most recent addition to the armamentarium of Decisions regarding the type of intervention for MBO are
procedures available to palliate GOO, and prospective trials complex and must take into account not only anatomic location
evaluating safety and efficacy have also included an evaluation but also life expectancy, intended future oncologic therapy,
of HRQoL outcomes and/or symptom relief (Dolz et al, 2011; safety/efficacy, and HRQoL. To date, multiple studies have
van den Berg et al, 2013; van Hooft et al, 2009 ). In terms of evaluated management strategies for MBO, many of which have
symptom relief, evaluated with GOO-SS, placement of ES was included analysis of HRQoL outcomes (Table 28.7).
consistently associated with symptom improvement and Several studies have assessed different methods of biliary
increased oral intake across studies. In two of three studies drainage and compared subsequent impact on HRQoL. Surgi-
reporting HRQoL outcomes, no changes in global measures of cal bypass for MBO is uncommon in developed countries and
HRQoL were observed at 1 month follow-up (Dolz et al, 2011; data pertaining to its impact on HRQoL are limited. In a study
van Hooft et al, 2009). Although no untreated GOO control from Brazil, where surgical bypass remains common, a com-
group was included, and the HRQoL in this population was not parison was made between ES and surgical bypass (choledo-
documented, it is likely that HRQoL would decline with chojejunostomy or cholecystojejunostomy) in patients with
progressive/untreated GOO and that stability of global measures metastatic pancreatic cancer. Regardless of treatment type,
represents a positive HRQoL outcome. Conversely, the most global HRQoL improved from baseline at 30- and 60-
recent study by van den Berg and colleagues (2013) reported days follow-up, respectively, but returned to preprocedure
significant improvements in global HRQoL following ES. This values by 120 days. However, global HRQoL scores in patient
study compared preprocedural HRQoL measure with the mean treated with ES were significantly higher at all time points,
HRQoL for the entire follow-up period (not standardized compared with the surgery group. No differences in procedural
504
TABLE 28.6 Studies of Palliative Intervention for Malignant Gastric Outlet Obstruction (GOO) and Health-Related Quality of Life
First Author, Journal, Population (Patient Study Purpose/ HRQoL Assessment Time Quality-of-Life
Year With) Quality-of-Life End Point Points Instrument Quality-of-Life Results Comments
Gastric Outlet Obstruction

Mehta, Surgical Malignant GOO Primary outcome: Randomized clinical SF-36 Health 13/14 successful LGJ vs. 10/13 In-hospital mortality, rates
Endoscopy, 2006 randomly morbidity, pain and trial, N = 27 (14 LGJ, Survey successful DS. LGJ had more high for both groups: 23%
assigned to DS short-term HRQoL 13 DS); HRQoL postprocedure complications (3/13) LGJ and 17% (2/12)
or LGJ (56% assessed and longer LOS. Mean DS
pancreatic preprocedure and physical health score in the DS
cancer) 1 mo postprocedure group was improved at 1 mo
(P < .01), whereas no changes
in HRQoL were observed in

LGJ group
Schmidt, American Malignant GOO Primary outcome: Prospective cohort, N = EORTC QLQ-C30 At 3 mo, both ES and surgery Median OS = 64 days; only
Journal of undergoing short-term HRQoL 47 (16 surgical + STO22 module groups had improved global 10 patients completed
Surgery, 2009 palliative bypass, 7 PEJ/PEG, HRQoL, physical and role HRQoL questionnaires at
intervention: 24 ES); HRQoL functioning, with decreased all time points (5 ES and 5
surgical bypass, assessed GOO symptoms compared surgery); initial procedural
PEG, PEJ, or ES preprocedure and 1 with baseline (statistically success rates: 70% ES vs.
and 3 mo significant only in ES group). In 100% surgical bypass
postprocedure the surgical group, there was a
significant decline in physical
functioning at 1 mo; however,
this rebounded by 3 mo. No
changes were observed in the
PEG/PEJ group
Van Hooft, Malignant GOO Primary outcome: Multicenter single-arm EORTC QLQ-C30, GOO symptoms (GOO-SS) Median OS = 62 days;
Gastrointestinal undergoing symptom palliation prospective trial, N = EQ-5D-3L+VAS consistently improved technical success ES 98%.
Endoscopy, 2009 palliative ES (GOO-SS) 51; HRQoL assessed postprocedure, whereas global Clinically, stent dysfunction
Secondary outcome: preprocedure, at HRQoL (QL2 subscale and occurred in 7 patients
safety, efficacy, 4 wk, and bimonthly EQ-VAS) was unchanged (14%), migration in 1
and global HRQoL following ES patient (2%)

Van Hooft, Malignant GOO Primary outcome: Combined data from 2 EORTC QLQ-C30, WHO performance status (HR, Median OS = 82 days,
Scandinavian undergoing identify predictors multicenter single- EQ-5D-3L + 2.63; 95% CI, 1.68 to 4.12, P although statistically
Journal of palliative ES of survival from arm prospective EQ-VAS < .01), prescription narcotic significant the HR for
Gastroenterology, baseline evaluation trials, N = 101; use (HR, 2.42; 95% CI, 1.38 EORTC pain score =1.01,
2010 baseline HRQoL to 4.25, P < .01), and pain clinical relevance
assessment only score of the EORTC (HR, questionable
1.01; 95% CI, 1.00 to 1.01, P
= .04) independently
associated with worse OS
Dolz, Malignant GOO Primary outcome: Multicenter, single-arm EuroQoL - 5D-3L Symptoms of GOO were Median OS = 91 days, 29/71
Gastroenterology undergoing procedural prospective trial N = significantly improved (40%) completed both
Hepatology, 2011 palliative ES technical success, 71 (38 duodenal postprocedure, whereas no preprocedure and
clinical efficacy, tumors, 15 recurrent change in HRQoL was postprocedure
and short-term tumors localized to observed. Stent efficacy was assessments
HRQoL GJ anastomosis, 18 dependent on tumor location
antral tumors); (GJ anastomosis [87%] >
HRQoL assessed duodenal [70%] > antral [29%])
preprocedure and
1 mo postprocedure
Van den Berg, Malignant GOO Primary outcome: Multicenter, single-arm, EORTC QLQ-C30 Global QoL reflected by health Median OS = 87 days,
Endoscopy, 2013 undergoing procedural prospective trial, N = and EQ-5D-3L + status (QL2) and EQ-VAS technical success = 89%,
palliative ES technical success 46; HRQoL assessed EQ-VAS demonstrated significant clinical success = 72%,
(54% pancreatic and clinical efficacy preprocedure and improvement between baseline periprocedural
cancer) Secondary outcome: bimonthly until death and the mean scores during complication = 57% (one
HRQoL total follow-up. Data from or more complications)
EORTC scales not presented
DS, Duodenal stent; EORTC QLQ-C30, European Organization for the Research and Treatment of Cancer Quality-of-Life Questionnaire–Core 30; EQ-VAS, EuroQoL Visual Analogue Scale; ES, endoscopic stent; GOO-SS, gastric outlet obstruction symptom scale; HR,
hazard ratio; HRQoL, health-related quality of life; LGJ, laparoscopic gastrojejunostomy; LOS, length of stay; OS, overall survival; PEG, percutaneous endoscopic gastrostomy; PEJ, percutaneous endoscopic jejunostomy; QL, quality of life; SF-36, 36-item Short Form;
STO22, EORTC gastric specific QoL Module.
505
TABLE 28.7 Studies of Malignant Biliary Obstruction and Health-Related Quality of Life
506

First Author, Journal, Quality-of-Life End HRQoL Assessment Quality-of-Life
Year Population (Patient With) Point Time Points Instrument Quality-of-Life Results Comments
Malignant Biliary Obstruction

Abraham, MBO undergoing palliative Primary outcome: Prospective cohort, SF-36 Health Overall, decrease in T-bili from baseline to 84% of patients
Gastrointestinal intent ES (66% distal short-term N = 50; HRQoL Survey 1 mo postprocedure was associated experienced at least a
Endoscopy, 2002 obstruction, 12.5% HRQoL assessed with significant improvements in social 33% reduction in T-bili
mid-CBD obstruction, preprocedure function and mental health; however, if postprocedure; 25%
21.5% hilar or and at 1 mo baseline T-bili was very high (≥14 mg/ mortality at 1 mo
intrahepatic obstruction) postprocedure dL), no improvement in social function
was observed
Chan, Journal of MBO undergoing ES as Primary outcome: Randomized SF-36 Health Baseline HRQoL was similar between Median OS and stent
Gastrointestinal initial drainage stent patency double-blinded Survey groups. Compared with baseline, at patency were not
Surgery, 2005 procedure Secondary controlled trial of 1-mo follow-up, social functioning different between
outcome: ciprofloxacin vs. scores improved in the ciprofloxacin groups; however,
morbidity, placebo before group and were significantly different cholangitis was less
mortality, and ES for MBO, N from placebo (P = .05). All other common in the
short-term = 94 (50 subscales and summative HRQoL ciprofloxacin group
HRQoL placebo, 44 measures were not different between
ciprofloxacin); groups
HRQoL
assessed
preprocedure
and 1 mo
postprocedure
Artifon, American Metastatic pancreatic Primary outcome: Prospective SF-36 Health For the entire cohort, HRQoL scores Median OS and
Journal of cancer with MBO cost of care randomized trial, Survey improved at 30 and 60 days postprocedural
Gastroenterology, undergoing surgical Secondary N = 30 (15 postprocedure, followed by decline to morbidity were not
2006 bypass outcome: surgery , 15 ES); preprocedure levels at 120 days. ES different between
(choledochojejunostomy short-term HRQoL was associated with significantly higher groups; overall cost of
or HRQoL assessed HRQoL than surgery at 30 days (P = care from treatment to
cholecystojejunostomy) preprocedure .04) and 60 days (P = .05) death was significantly
or ES and 30, 60, and reduced with ES
120 days ($4270 USD) vs. HJ
postprocedure ($8320 USD)
Saluja, Clinical Unresectable gallbladder Primary outcome: Prospective WHOQoL– At 1 mo postprocedure, no change was Median OS was not
Gastroenterology carcinoma with hilar short-term randomized trial, BREF-26 observed in global HRQoL for either different between
and Hepatology, obstruction undergoing HRQoL and N = 54 (27 ES, and group; however, symptom scale scores groups. Clinical
2008 ES or PBD with plastic clinical 27 PBD); EORTC were improved in both PBD and ES. At procedural success
stent effectiveness of HRQoL QLQ-C30 3 mo, both groups showed a trend was higher in the PBD
biliary drainage assessed toward improvement in all domains of group (89%) vs. ES
preprocedure HRQoL assessed. PBD was associated (41%), and cholangitis
and 1 and 3 mo with improved physical and was less common with
postprocedure psychological scores compared with ES PBD (11%) compared
(P = .02) as well as greater improvement with ES (48%)
in symptom scores at 3 mo
Robson, Annals of MBO undergoing PBD Primary outcome: Prospective cohort, FACT-Hep VASP scores (pruritus symptoms) Median OS = 4.7 mo,
Surgical short-term N = 109; HRQoL and VASP significantly improved during time; mortality 10% at 1 wk,
Oncology, 2010 HRQoL, assessed however, overall HRQoL as measured 28% by 8 wk; 100%
symptom control, preprocedure by mean FACT-Hep scores were technical procedural
morbidity/ and 1 and 4 wk significantly decreased at 1 and 4 wk success but 50% major
mortality and postprocedure postprocedure complications
procedural
efficacy

Larssen, Surgical Malignant GI obstruction Primary outcome: Prospective cohort, EORTC Global HRQoL as well as nausea/vomiting, In addition to biliary
Endoscopy, 2011 undergoing ES SEMS short-term N = 162 (40 QLQ-C30 and appetite loss, and pruritus obstruction, study
HRQoL [25%] biliary); ± STO22 ± symptoms were significantly improved in cohort included
Secondary HRQoL PAN26 patients with biliary obstruction treated patients undergoing
outcome: assessed with SEMS. Both patients and esophageal, duodenal,
comparison of preprocedure physicians reported symptom and colonic SEMS
physician- and and 2 wk improvement; however, degree of
patient-rated postprocedure improvement reported by physicians
treatment effect was higher than that of patients
Artifon, Journal of MBO with history of a Primary outcome: Prospective SF-36 Health No difference in HRQoL measured at 7 or One hundred percent
Clinical failed ERCP undergoing clinical and randomized trial, Survey 30 days postprocedure for either group clinical and technical
Gastroenterology, PBD or EUS-CD technical N = 25 (13 success in both
2012 procedural EUS-CD, 12 groups; complication
success PBD); HRQoL rates and procedural
Secondary assessed costs not different
outcome: cost, preprocedure between groups
morbidity, and and 7 and 30
HRQoL days
postprocedure
Aggarwal, MBO undergoing palliative Primary outcome: Prospective EORTC Significant improvement in overall HRQoL Median OS = 8.7 mo
Brachytherapy, ILBT with access via symptom control, longitudinal, N = QLQ-C30 post-PBD, post-ILBT, and at last
2013 PBD HRQoL, and 18; HRQoL follow-up. Physical and social
survival assessed functioning scores were improved at all
pre-PBD, time points, as was insomnia symptom
post-PBD, scale. The percentage of patients who
pre-ILBT, reported pruritus and icterus declined at
post-ILBT, and each time point
follow-up
Barkay, Journal of MBO undergoing ES Primary outcome: Prospective FACT-G Significant improvement in global HRQoL
Clinical longitudinal longitudinal, N = and PWB, EWB, and FWB subscales at
Gastroenterology, HRQoL 164; HRQoL 30 and 180 days. Weight loss, appetite,
2013 assessed pruritus, and pain improved significantly
preprocedure, at at both time points
30 and 180
days
postprocedure
Moses, World Infrahilar MBO undergoing Primary outcome: Multicentered SF-36 Health At baseline, SEMS group had significantly Time to stent failure
Journal of ES randomly assigned time to stent prospective Survey worse physical functioning scores; no significantly longer
Gastroenterology, to pcSEMS or PS failure randomized trial, differences were observed between among the pcSEMS
2013 Secondary N = 85 enrolled: groups thereafter. At 1 mo, vitality was group, and cholangitis
outcome: 74 with HRQoL improved, and at 6 mo physical was less common; no
morbidity, data (36 PS, 38 functioning was improved in the difference in OS
mortality, KPS, pcSEMS); pcSEMS group. PS was associated
HRQoL HRQoL with significant improvement in
assessed subscales of bodily pain, social
preprocedure at functioning, and mental health at 1 mo
baseline and 1,
3, 6, 9, 12, and
15 mo
CBD, Common bile duct; EWB, emotional well-being; EORTC QLQ-C30, European Organization for the Research and Treatment of Cancer Quality-of-Life Questionnaire–Core 30; ERCP, endoscopic retrograde cholangiopancreatography; ES, endoscopic stenting; EUS-CD,
endoscopic ultrasound-guided cyst drainage; FACT, Functional Assessment of Cancer Therapy; FWB, functional well-being; GI, gastrointestinal; HRQoL, health-related quality of life; LBT, intraluminal brachytherapy; KPS, Karnosfsky Performance Score; MBO, malignant
biliary obstruction; OS, overall survival; pcSEMS, partially covered self-expanding metallic stent; PAN26, pancreatic cancer module-26; PBD, percutaneous biliary drainage; PS, performance score; PWB, physical well-being; SEMS, self-expanding metallic stent; SF-36,
36-item Short Form; STO22, EORTC gastric specific QoL Module; T-bili, total bilirubin; WHOQoL–BREF-22, World Health Organization Quality-of-Life assessment (abbreviated version, 22 items); VASP, visual analogue scale for pruritus assessment.
507
complications or survival were observed between groups. The either PBD or ES. At 1 month follow-up, no significant change
equivalence in morbidity and mortality between procedures in overall HRQoL was observed between the groups; however,
with improved HRQoL suggests ES may be superior to surgical at 3 months a trend toward improvement was noted in both.
bypass in select patients. Compared with ES, PBD was associated with significantly
Techniques of ES and PBD for MBO have undergone greater improvement in physical and psychological domains.
significant evolution and advancement during the past decade. Both groups experienced improvement in symptoms, but PBD
Similarly, studies evaluating not only safety/efficacy of these was associated with significantly greater improvements in
interventions, but also HRQoL, have increased substantially. fatigue compared with ES. Conversely, evaluation of patients
ES for MBO is associated with reasonable clinical and technical randomly assigned to PBD or endoscopic ultrasound-guided
success. Reduction in total bilirubin levels by at least one-third choledochoduodenostomy (EUS-CD) revealed no significant
of preprocedural values with ES has been associated with change in HRQoL in the 30 days following intervention (Artifon
improved social and mental functioning as measured by SF-36; et al, 2012). Furthermore, there were no significant differences
however, these HRQoL improvements appear to be mitigated between the PBD and EUS-CD groups. It appears that, regard-
by the absolute preprocedure bilirubin (preprocedure total bili- less of procedure performed, relief of obstruction per se pro-
rubin ≥ 14 mg/dL; no change in HRQoL parameters) (Abraham vides symptom relief and at best stabilizes HRQoL, which
et al, 2002). Furthermore, in an assessment of 40 patients among this fragile cohort likely represents a positive outcome.
undergoing ES with self-expanding metal stents (SEMS),
global HRQoL on the EORTC QLQ-C30 was improved, as CHALLENGES IN HEALTH-RELATED QUALITY-
were symptoms of nausea/vomiting, appetite, and pruritus at 2 OF-LIFE RESEARCH AND FUTURE DIRECTIONS
weeks postprocedure (Larssen et al, 2011). Using the FACT-G
questionnaire, Barkay and colleagues (2013) also evaluated ES Despite the widespread implementation of HRQoL assessment,
for MBO and found improvement in global HRQoL and physi- interpretation of findings and implementation at the clinical
cal, emotional, and functional well-being at both 1 and 6 practice level is a significant challenge. Merely stating a “P”
months of follow-up. Likewise, a recent study comparing par- value, indicating a statistical difference, is the standard; however,
tially covered self-expandable metal stents with plastic stents, determining how this is to be interpreted clinically is a challenge
revealed no difference between treatments in terms of HRQoL and is one of the major criticisms of HRQoL research. To date,
but a general trend in both arms toward an improvement in this has significantly limited the clinical applicability of HRQoL
global HRQoL and subscales (Moses et al, 2013). Taken in research outcomes. As of 2005, only 25% of HRQoL studies
concert, these findings suggest that ES for MBO improves reported on the clinical significance, and thus applicability, of
short-term HRQoL and symptoms; however, conclusions their findings (Efficace et al, 2007). Several methods have been
regarding longer-term HRQoL outcomes are limited by the suggested to rectify this issue, including establishment of a
substantial decrements in survey completion with prolonged minimal important difference (MID)—the smallest difference
follow-up. in a score in the domain of interest that a patient perceived as
In a prospective evaluation of 109 patients with MBO pre- beneficial and that would mandate a change in patient manage-
senting for PBD, significant improvements in pruritus symp- ment (Jaeschke et al, 1989). Two primary methods for deter-
toms were observed, but overall HRQoL, as assessed by mining a MID for HRQoL outcomes exist: distribution based
FACT-Hep, significantly declined during the 4 week period (estimate based on observed scores in a sample) and anchor
following PBD. The mortality rate was extremely high—28% based (compare, or anchor, HRQoL differences/changes to
at 8 weeks post-PBD, highlighting the degree of illness and other clinical variables or patient’s subjective assessment).
frailty among this patient population (Robson et al, 2010). Although imperfect, a MID provides a means by which to place
PBD followed by ILBT has recently been touted to reduce HRQoL into a clinical context. Proponents for both forms of
localized tumor burden and result in improved biliary drainage measurement of MID exist; however, a combination of both
and symptom relief. In a study of high-dose ILBT 18 patients methods, where feasible, likely yields the most clinically useful
with advanced pancreaticobiliary or metastatic cancer under- and accurate information (Revicki et al, 2008).
went PBD, followed by two ILBT sessions 1 week apart. In Longitudinal studies of HRQoL are essential in the current
conjunction with significant improvement in symptoms (100% era, where disease treatments are becoming more and more
resolution of pruritus), global HRQoL as well as physical and complex and survival “with” disease is common. However, this
social functional scores were improved at all assessment points. assessment is subject to several biases, including “response
It appears that PBD alone significantly improves biliary obstruc- shift,” whereby patients recalibrate their internal standards
tive symptoms; however, in the absence of further treatments, regarding HRQoL during the course of treatment/follow-up
HRQoL seems to decline in a fashion expected of end-stage and “healthy survivor”; while time passes, sicker patients are less
pancreaticobiliary cancers. likely to complete surveys, and therefore longer-term results are
ES was initially preferred to PBD because drainage was heavily biased toward patients who are fairing the best. Although
internal and thought to be more readily achievable. However, methods exist in to attempt combating these issues, their use
advances in interventional radiology and the development of is rare and the approaches lack validation. Furthermore, long-
internal stents for percutaneous delivery suggest that safe can- term studies are often limited by the logistics of survey admin-
nulation of the biliary tree and adequate internal drainage can istration and practicality.
be achieved with both PBD and ES. The growing number of Missing data is a common and critical concern when inter-
techniques and methods for ES and PBD has led to multiple preting HRQoL data. The consequences of missing HRQoL
comparative evaluations of the two procedures. Recently, Saluja information depends on both the amount and the cause of
and colleagues (2008) randomly assigned patients with unre- missing data. It is to be expected that some information will be
sectable gallbladder carcinoma and hilar obstruction to undergo missing at random; the real concern comes when “missingness”
is not a random event. There are multiple mechanisms by personalized medicine, HRQoL researchers have developed
which to assess the importance of missing data, including com- GENEQoL, an initiative assessing potential links between
parison of statistical means from patients with complete HRQoL outcomes and genetics. Given that HRQoL may be
follow-up to those with incomplete follow-up, to see if there is predictive of clinical outcomes, the ultimate goal of such inves-
a difference between groups in HRQoL indices. If this is indeed tigations would be the development of a HRQoL genetic sig-
the fact, data are nonignorable missing data and must be nature to be used as a predictor of patient outcomes (Sloan &
accounted for or explained. Given the significant impact of Zhao, 2006; Sprangers et al, 2009).
missing data in HRQoL assessment, especially in studies during HRQoL research in surgery is in its infancy and, as such,
time, the best method to deal with it is in fact preventive. Trials faces many challenges, from study design and implementation
assessing HRQoL should be designed in a fashion such that to clinical application. However, it is encouraging that enthu-
implementation strategies and rigorous follow-up are in place siasm surrounding development of this field continues to rise
at the outset. among surgeons and institutions alike. In the future, collabora-
Future directions in HRQoL are based on item response tion between surgeons, statisticians, qualitative researchers, and
theory and include computerized adaptive testing in which most importantly, patients will be essential to further the devel-
patients’ responses to a given question will prompt the subse- opment and clinical application of HRQoL outcomes in surgery.
quent line of questions, thus individualizing surveys. Further- Furthermore, as the landscape of medicine continues to evolve,
more, the use of electronic surveys is increasingly become become more complex, and offer more life-saving/prolonging
part of standard clinical care such that questionnaires are treatments, there is no doubt that HRQoL measurement and
completed prior to each clinic visit, immediate feedback understanding will become increasingly valuable and essential
is generated for the patient and treating physician, and impor- for optimal patient care.
tant patient-centered concerns may be addressed and decisions
made. Likewise, in response to increasing fervor surrounding References are available at expertconsult.com.
Chapter 28 Quality of life and hepatobiliary tumors 509.e1
REFERENCES Detmar SB, et al: Health-related quality-of-life assessments and

patient-physician communication: a randomized controlled trial,
Aaronson NK: Quality of life: what is it? How should it be measured?, JAMA 288(23):3027–3034, 2002.
Oncology (Williston Park) 2(5):69–76, 64, 1988. Diouf M, et al: Could baseline health-related quality of life (QoL)
Abraham NS, et al: Palliation of malignant biliary obstruction: a pro- predict overall survival in metastatic colorectal cancer? The results
spective trial examining impact on quality of life, Gastrointest Endosc of the GERCOR OPTIMOX 1 study, Health Qual Life Outcomes
56(6):835–841, 2002. 12:69, 2014.
Aggarwal R, et al: Evaluation of high-dose-rate intraluminal brachy- Dolz C, et al: Antral localization worsens the efficacy of enteral stents
therapy by percutaneous transhepatic biliary drainage in the pallia- in malignant digestive tumors, Gastroenterol Hepatol 34(2):63–68,
tive management of malignant biliary obstruction—a pilot study, 2011.
Brachytherapy 12(2):162–170, 2013. Earlam S, et al: Relation between tumor size, quality of life, and sur-
Artifon EL, et al: Surgery or endoscopy for palliation of biliary obstruc- vival in patients with colorectal liver metastases, J Clin Oncol
tion due to metastatic pancreatic cancer, Am J Gastroenterol 14(1):171–175, 1996.
101(9):2031–2037, 2006. Efficace F, et al: Does a patient’s self-reported health-related quality of
Artifon EL, et al: Biliary drainage in patients with unresectable, malig- life predict survival beyond key biomedical data in advanced colorec-
nant obstruction where ERCP fails: endoscopic ultrasonography- tal cancer?, Eur J Cancer 42(1):42–49, 2006.
guided choledochoduodenostomy versus percutaneous drainage, Efficace F, et al: Has the quality of health-related quality of life report-
J Clin Gastroenterol 46(9):768–774, 2012. ing in cancer clinical trials improved over time? Towards bridging
Avery K, Blazeby JM: Quality of life assessment in surgical oncology the gap with clinical decision making, Ann Oncol 18(4):775–781,
trials, World J Surg 30(7):1163–1172, 2006. 2007.
Banz VM, et al: Long-term quality of life after hepatic resection: health Eid S, et al: Assessment of symptom experience in patients undergoing
is not simply the absence of disease, World J Surg 33(7):1473–1480, hepatic resection or ablation, Cancer 107(11):2715–2722, 2006.
2009. Eltawil KM, et al: Quality of life and survival analysis of patients
Barkay O, et al: Effect of endoscopic stenting of malignant bile duct undergoing transarterial chemoembolization for primary hepatic
obstruction on quality of life, J Clin Gastroenterol 47(6):526–531, malignancies: a prospective cohort study, HPB (Oxford) 14(5):341–
2013. 350, 2012.
Belyaev O, et al: Early and late postoperative changes in the quality of Epelboym I, et al: Quality of life in patients after total pancreatectomy
life after pancreatic surgery, Langenbecks Arch Surg 398(4):547–555, is comparable with quality of life in patients who undergo a partial
2013. pancreatic resection, J Surg Res 187(1):189–196, 2014.
Bernhard J, et al: Missing quality of life data in cancer clinical trials: U.S. Food and Drug Administration: Guidance for industry: patient-
serious problems and challenges, Stat Med 17(5–7):517–532, reported outcome measures: use in medical product development to
1998. support labeling claims: draft guidance, Health Qual Life Outcomes
Billings BJ, et al: Quality-of-life after total pancreatectomy: is it really 4:79, 2006.
that bad on long-term follow-up?, J Gastrointest Surg 9(8):1059– Feinstein AR: Clinimetric perspectives, J Chronic Dis 40(6):635–640,
1066, discussion 1066-1057, 2005. 1987.
Blazeby JM, et al: Health-related quality of life measurement in ran- Fraser SC: Quality-of-life measurement in surgical practice, Br J Surg
domized clinical trials in surgical oncology, J Clin Oncol 24(19):3178– 80(2):163–169, 1993.
3186, 2006. Gotay CC, et al: The prognostic significance of patient-reported
Bonnetain F, et al: Quality of life as a prognostic factor of overall sur- outcomes in cancer clinical trials, J Clin Oncol 26(8):1355–1363,
vival in patients with advanced hepatocellular carcinoma: results 2008.
from two French clinical trials, Qual Life Res 17(6):831–843, 2008. Guyatt GH, et al: Users’ guides to the medical literature. XII. How
Bruner DW, et al: Outcomes research in cancer clinical trial coopera- to use articles about health-related quality of life. Evidence-
tive groups: the RTOG model, Qual Life Res 13(6):1025–1041, Based Medicine Working Group, JAMA 277(15):1232–1237,
2004. 1997.
Bruns H, et al: Quality of life after curative liver resection: a single Hahn EA, et al: Missing data in quality of life research in Eastern
center analysis, World J Gastroenterol 16(19):2388–2395, 2010. Cooperative Oncology Group (ECOG) clinical trials: problems and
Byrne C, et al: Health-related quality of life as a valid outcome in the solutions, Stat Med 17(5–7):547–559, 1998.
treatment of advanced colorectal cancer, Eur J Surg Oncol 33(Suppl Halloran CM, et al: Partial pancreatic resection for pancreatic malig-
2):S95–S104, 2007. nancy is associated with sustained pancreatic exocrine failure and
Chan C, et al: Impact on quality of life after pancreatoduodenectomy: reduced quality of life: a prospective study, Pancreatology 11(6):535–
a prospective study comparing preoperative and postoperative 545, 2011.
scores, J Gastrointest Surg 16(7):1341–1346, 2012. Hamdy H, et al: Impact of hepatocellular carcinoma on health related
Chan G, et al: The role of ciprofloxacin in prolonging polyethylene quality of life in Egyptian patients: a single centre study, J Egypt Soc
biliary stent patency: a multicenter, double-blinded effectiveness Parasitol 43(1):183–194, 2013.
study, J Gastrointest Surg 9(4):481–488, 2005. Han SS, et al: A comparison of the long-term functional outcomes of
Chen L, et al: Quality of life in patients with liver cancer after opera- standard pancreatoduodenectomy and pylorus-preserving pancre-
tion: a 2-year follow-up study, Hepatobiliary Pancreat Dis Int 3(4): atoduodenectomy, Hepatogastroenterology 54(78):1831–1835, 2007.
530–533, 2004. Hartwig W, et al: Total pancreatectomy for primary pancreatic neo-
Chen J, et al: A systematic review of the impact of routine collection plasms: renaissance of an unpopular operation, Ann Surg 261(3):537–
of patient reported outcome measures on patients, providers and 546, 2015.
health organisations in an oncologic setting, BMC Health Serv Res Hofmann B, et al: Ethics of palliative surgery in patients with cancer,
13:211, 2013. Br J Surg 92(7):802–809, 2005.
Choiniere M, et al: Comparisons between patients’ and nurses’ assess- Huang G, et al: Quality of life after surgical resection compared with
ment of pain and medication efficacy in severe burn injuries, Pain radiofrequency ablation for small hepatocellular carcinomas, Br J
40(2):143–152, 1990. Surg 101(8):1006–1015, 2014.
Constitution of the World Health Organization, Am J Public Health Izsak FC, Medalie JH: Comprehensive follow-up of carcinoma patients,
Nations Health 36(11):1315–1323, 1946. J Chronic Dis 24(2):179–191, 1971.
Crippa S, et al: Quality of life in pancreatic cancer: analysis by stage Jaeschke R, et al: Measurement of health status. Ascertaining the
and treatment, J Gastrointest Surg 12(5):783–793, discussion 793- minimal clinically important difference, Control Clin Trials 10(4):407–
784, 2008. 415, 1989.
D’Angelica M, et al: Surgeon-patient communication in the treatment Jarnagin WR, et al: Improvement in perioperative outcome after
of pancreatic cancer, Arch Surg 133(9):962–966, 1998. hepatic resection: analysis of 1,803 consecutive cases over the past
Dasgupta D, et al: Quality of life after liver resection for hepatobiliary decade, Ann Surg 236(4):397–406, discussion 406-397, 2002.
malignancies, Br J Surg 95(7):845–854, 2008. Kahn SB, et al: Quality of life and patients with cancer: a comparative
de Haes JC, van Knippenberg FC: The quality of life of cancer patients: study of patient versus physician perceptions and its implications for
a review of the literature, Soc Sci Med 20(8):809–817, 1985. cancer education, J Cancer Educ 7(3):241–249, 1992.
Kalinowski M, et al: Selective internal radiotherapy with yttrium-90 Quinten C, et al: Baseline quality of life as a prognostic indicator of
microspheres for hepatic metastatic neuroendocrine tumors: a pro- survival: a meta-analysis of individual patient data from EORTC
spective single center study, Digestion 79(3):137–142, 2009. clinical trials, Lancet Oncol 10(9):865–871, 2009.
Karnofsky DA, Burchnel JH: The clinical evaluation of chemothera- Quinten C, et al: Patient self-reports of symptoms and clinician ratings
peutic agents in cancer. In Symposium held at New York academy of as predictors of overall cancer survival, J Natl Cancer Inst 103(24):
medicine, evaluation of chemotherapeutic agents, New York, 1948, 1851–1858, 2011.
Columbia University Press, pp 191–205. Quinten CF: A global analysis of multitrial data investigating quality
Kemmler G, et al: Comparison of two quality-of-life instruments for of life and symptoms as prognostic factors for survival in different
cancer patients: the functional assessment of cancer therapy-general tumor sites, Cancer 120(2):302–311, 2014.
and the European Organization for Research and Treatment of Rees JR, et al: Patient-reported outcomes after hepatic resection of
Cancer Quality of Life Questionnaire-C30, J Clin Oncol 17(9):2932– colorectal cancer metastases, J Clin Oncol 30(12):1364–1370,
2940, 1999. 2012.
Kolligs FT, et al: Pilot randomized trial of selective internal radiation Rees JR, et al: Patient-reported outcomes in long-term survivors of
therapy vs. chemoembolization in unresectable hepatocellular carci- metastatic colorectal cancer needing liver resection, Br J Surg
noma, Liver Int 35(6):1715–1721, 2015. 101(11):1468–1474, 2014.
Kostro J, Sledziński Z: Quality of life after surgical treatment of pan- Revicki D, et al: Recommended methods for determining responsive-
creatic cancer, Acta Chir Belg 108(6):679–684, 2008. ness and minimally important differences for patient-reported out-
Kuroda H, et al: Effects of branched-chain amino acid-enriched nutri- comes, J Clin Epidemiol 61(2):102–109, 2008.
ent for patients with hepatocellular carcinoma following radiofre- Roberts KJ, et al: How severe is diabetes after total pancreatectomy?
quency ablation: a one-year prospective trial, J Gastroenterol Hepatol A case-matched analysis, HPB (Oxford) 16(9):814–821, 2014.
25(9):1550–1555, 2010. Robson PC, et al: Prospective study of outcomes after percutaneous
Langenhoff BS, et al: Quality of life as an outcome measure in surgical biliary drainage for malignant biliary obstruction, Ann Surg Oncol
oncology, Br J Surg 88(5):643–652, 2001. 17(9):2303–2311, 2010.
Langenhoff BS, et al: Quality of life after surgical treatment of colorec- Roila F, Cortesi E: Quality of life as a primary end point in oncology,
tal liver metastases, Br J Surg 93(8):1007–1014, 2006. Ann Oncol 12(Suppl 3):S3–S6, 2001.
Larssen L, et al: Patient-reported outcomes in palliative gastrointestinal Ruers TJ, et al: Comparison between local ablative therapy and che-
stenting: a Norwegian multicenter study, Surg Endosc 25(10):3162– motherapy for non-resectable colorectal liver metastases: a prospec-
3169, 2011. tive study, Ann Surg Oncol 14(3):1161–1169, 2006.
Lee LJ, et al: Quality of life in patients with hepatocellular carcinoma Sajid MS, et al: Use of health-related quality of life tools in hepatobili-
received surgical resection, J Surg Oncol 95(1):34–39, 2007. ary surgery, Hepatobiliary Pancreat Dis Int 7(2):135–137, 2008.
Ljungman D, et al: Cost-utility estimation of surgical treatment of Salem R, et al: Increased quality of life among hepatocellular carci-
pancreatic carcinoma aimed at cure, World J Surg 35(3):662–670, noma patients treated with radioembolization, compared with che-
2011. moembolization, Clin Gastroenterol Hepatol 11(10):1358–1365,
Maisey NR, et al: Baseline quality of life predicts survival in patients e1351, 2013.
with advanced colorectal cancer, Eur J Cancer 38(10):1351–1357, Saluja SS, et al: Endoscopic or percutaneous biliary drainage for gall-
2002. bladder cancer: a randomized trial and quality of life assessment,
Martin RC, et al: Health-related quality of life: return to baseline after Clin Gastroenterol Hepatol 6(8):944–950, e943, 2008.
major and minor liver resection, Surgery 142(5):676–684, 2007. Schag CC, et al: Karnofsky performance status revisited: reliability,
Mbah N, et al: Impact of post-operative complications on quality of validity, and guidelines, J Clin Oncol 2(3):187–193, 1984.
life after pancreatectomy, JOP 13(4):387–393, 2012. Schmidt C, et al: A prospective observational study examining quality
Mehta S, et al: Prospective randomized trial of laparoscopic gastroje- of life in patients with malignant gastric outlet obstruction, Am J
junostomy versus duodenal stenting for malignant gastric outflow Surg 198(1):92–99, 2009.
obstruction, Surg Endosc 20(2):239–242, 2006. Schmidt U, et al: Quality of life and functional long-term outcome after
Mihalache F, et al: Survival and quality of life of cholangiocarcinoma partial pancreatoduodenectomy: pancreatogastrostomy versus pan-
patients: a prospective study over a 4 year period, J Gastrointestin creatojejunostomy, Ann Surg Oncol 12(6):467–472, 2005.
Liver Dis 19(3):285–290, 2010. Schniewind B, et al: Quality of life after pancreaticoduodenectomy for
Miller AR, et al: Health-related quality of life changes following major ductal adenocarcinoma of the pancreatic head, Br J Surg 93(9):1099–
and minor hepatic resection: the impact of complications and post- 1107, 2006.
operative anemia, Am J Surg 206(4):443–450, 2013. Serrano PE, et al: Quality of life in a prospective, multicenter phase 2
Mise Y, et al: Impact of surgery on quality of life in patients with trial of neoadjuvant full-dose gemcitabine, oxaliplatin, and radiation
hepatocellular carcinoma, World J Surg 38(4):958–967, 2014. in patients with resectable or borderline resectable pancreatic adeno-
Moses PL, et al: Randomized trial in malignant biliary obstruction: carcinoma, Int J Radiat Oncol Biol Phys 90(2):270–277, 2014.
plastic vs partially covered metal stents, World J Gastroenterol Shaw CM, et al: Long-term quality of life following pancreaticoduode-
19(46):8638–8646, 2013. nectomy, Hepatogastroenterology 52(63):927–932, 2005.
Muller-Nordhorn J, et al: Health-related quality of life in patients with Shun SC, et al: Quality of life and its associated factors in patients with
pancreatic cancer, Digestion 74(2):118–125, 2006. hepatocellular carcinoma receiving one course of transarterial che-
Nguyen TC, et al: Standard vs. radical pancreaticoduodenectomy for moembolization treatment: a longitudinal study, Oncologist 17(5):
periampullary adenocarcinoma: a prospective, randomized trial eval- 732–739, 2012.
uating quality of life in pancreaticoduodenectomy survivors, J Gas- Slevin ML, et al: Who should measure quality of life, the doctor or the
trointest Surg 7(1):1–9, discussion 9-11, 2003. patient?, Br J Cancer 57(1):109–112, 1988.
Nieveen van Dijkum EJ, et al: Quality of life after curative or palliative Sloan JA, Dueck A: Issues for statisticians in conducting analyses and
surgical treatment of pancreatic and periampullary carcinoma, Br J translating results for quality of life end points in clinical trials,
Surg 92(4):471–477, 2005. J Biopharm Stat 14(1):73–96, 2004.
Oken MM, et al: Toxicity and response criteria of the Eastern Coopera- Sloan JA, Zhao CX: Genetics and quality of life, Curr Probl Cancer
tive Oncology Group, Am J Clin Oncol 5(6):649–655, 1982. 30(6):255–260, 2006.
Osoba D: Lessons learned from measuring health-related quality of life Sprangers MA, Aaronson NK: The role of health care providers and
in oncology, J Clin Oncol 12(3):608–616, 1994. significant others in evaluating the quality of life of patients with
Passik SD, Kirsh KL: The importance of quality-of-life endpoints in chronic disease: a review, J Clin Epidemiol 45(7):743–760, 1992.
clinical trials to the practicing oncologist, Hematol Oncol Clin North Sprangers MA, et al: The establishment of the GENEQOL consortium
Am 14(4):877–886, 2000. to investigate the genetic disposition of patient-reported quality-of-
Poon RT, et al: A prospective longitudinal study of quality of life after life outcomes, Twin Res Hum Genet 12(3):301–311, 2009.
resection of hepatocellular carcinoma, Arch Surg 136(6):693–699, Steel J, et al: Quality of life in patients diagnosed with primary hepa-
2001. tocellular carcinoma: hepatic arterial infusion of cisplatin versus
Presant CA: Quality of life in cancer patients. Who measures what?, 90-yttrium microspheres (Therasphere), Psychooncology 13(2):73–
Am J Clin Oncol 7(5):571–573, 1984. 79, 2004.
Chapter 28 Quality of life and hepatobiliary tumors 509.e3
Stephens RJ, et al: Randomized trials with quality of life endpoints: are Walter J, et al: Palliative partial pancreaticoduodenectomy impairs
doctors’ ratings of patients’ physical symptoms interchangeable with quality of life compared to bypass surgery in patients with advanced
patients’ self-ratings?, Qual Life Res 6(3):225–236, 1997. adenocarcinoma of the pancreatic head, Eur J Surg Oncol 37(9):798–
Tanabe G, et al: Favorable quality of life after repeat hepatic resection 804, 2011.
for recurrent hepatocellular carcinoma, Hepatogastroenterology 48(38): Wang YB, et al: Quality of life after radiofrequency ablation combined
506–510, 2001. with transcatheter arterial chemoembolization for hepatocellular car-
Toro A, et al: Health-related quality of life in patients with hepatocel- cinoma: comparison with transcatheter arterial chemoembolization
lular carcinoma after hepatic resection, transcatheter arterial chemo- alone, Qual Life Res 16(3):389–397, 2007.
embolization, radiofrequency ablation or no treatment, Surg Oncol Wehrmann T, et al: Endoscopic pancreatic duct stenting for relief
21(1):e23–e30, 2012. of pancreatic cancer pain, Eur J Gastroenterol Hepatol 17(12):1395–
van den Berg MW, et al: First data on the evolution duodenal stent for 1400, 2005.
palliation of malignant gastric outlet obstruction (DUOLUTION Whalen GF, Ferrans CE: Quality of life as an outcome in clinical trials
study): a prospective multicenter study, Endoscopy 45(3):174–181, and cancer care: a primer for surgeons, J Surg Oncol 77(4):270–276,
2013. 2001.
Van Heek NT, et al: The need for a prophylactic gastrojejunostomy for World Health Organization: Study protocol for the World Health Orga-
unresectable periampullary cancer: a prospective randomized multi- nization project to develop a Quality of Life assessment instrument
center trial with special focus on assessment of quality of life, Ann (WHOQOL), Qual Life Res 2(2):153–159, 1993.
Surg 238(6):894–902, discussion 902-895, 2003. Wible BC, et al: Longitudinal quality of life assessment of patients with
van Hooft JE, et al: Efficacy and safety of the new WallFlex enteral hepatocellular carcinoma after primary transarterial chemoemboliza-
stent in palliative treatment of malignant gastric outlet obstruction tion, J Vasc Interv Radiol 21(7):1024–1030, 2010.
(DUOFLEX study): a prospective multicenter study, Gastrointest Wiering B, et al: Long-term global quality of life in patients treated for
Endosc 69(6):1059–1066, 2009. colorectal liver metastases, Br J Surg 98(4):565–571, discussion 571-
van Hooft JE, et al: Endoscopic magnetic gastroenteric anastomosis for 562, 2011.
palliation of malignant gastric outlet obstruction: a prospective mul- Wilson IB, Cleary PD: Linking clinical variables with health-related
ticenter study, Gastrointest Endosc 72(3):530–535, 2010. quality of life. A conceptual model of patient outcomes, JAMA
Velanovich V: The quality of quality of life studies in general surgical 273(1):59–65, 1995.
journals, J Am Coll Surg 193(3):288–296, 2001.
Velikova G, et al: Measuring quality of life in routine oncology practice
improves communication and patient well-being: a randomized con-
trolled trial, J Clin Oncol 22(4):714–724, 2004.
CHAPTER 29
Interventional endoscopy: technical aspects
Dennis Yang and Christopher John DiMaio
OVERVIEW
itself into the papilla. Alternatively, the cannula or sphinctero
The field of diagnostic and therapeutic biliary endoscopy has tome can engage the ampullary orifice first before using the
evolved rapidly since the first reports of endoscopic sphincter guidewire to effectively seek out the path to the desired duct
otomy in 1974. A variety of interventional endoscopic pro via fluoroscopy.
cedures have gained widespread acceptance as therapeutic
alternatives to operative management for a spectrum of biliary Endoscopic Retrograde Cholangopancreatography
disorders. Although endoscopic retrograde cholangopancrea
tography (ERCP) remains a common endoscopic technique
Pancreatic Duct Wire or Stent Placement to Facilitate
for the evaluation and management of biliary diseases, there Biliary Access
has been a trend toward alternative modalities, including There are various alternative endoscopic techniques that can
an expanding role for endoscopic ultrasound (EUS)-guided be used when attempts at biliary cannulation result in repeated
therapy (see Chapter 16). This chapter focuses on the current pancreatic duct entry. In the “double-wire technique,” a guide
technical aspects of endoscopic biliary intervention. wire is purposely placed in the pancreatic duct while the biliary
cannulation device is preloaded with a second guidewire to
reattempt biliary cannulation. Theoretically, the pancreatic
TECHNIQUES FOR BILIARY ACCESS guidewire assists by straightening the common channel and
common bile duct (CBD), thus facilitating biliary cannulation.
Endoscopic Retrograde Cholangopancreatography Furthermore, the direction of the pancreatic wire exiting the
Standard Cannulation Technique papilla on endoscopy may provide anatomic cues regarding the
Selective cannulation is a prerequisite for biliary endoscopic optimal axis for biliary cannulation. Although this technique
intervention. There are multiple devices that can be used for has been associated with high rates of cannulation in difficult
successful cannulation during ERCP, including standard can situations, it has not been conclusively shown to be superior to
nulation catheters and sphincterotomes. Standard biliary can standard cannulation (Herreros de Tejada et al, 2009; Maeda
nulation catheters can have one or more lumens, allowing the et al, 2003). Alternatively, instead of a wire, a pancreatic duct
injection of contrast material and passage of a guidewire to stent can also be placed during difficult biliary cannulation. The
achieve cannulation. In contrast to the standard catheter, a pancreatic duct stent may reduce the risk of PEP by ensuring
sphincterotome also has an electrosurgical cutting wire at the pancreatic duct drainage and by minimizing further inadvertent
distal end of the catheter that enables immediate sphincterot entry into the pancreatic duct on repeated attempts at biliary
omy following cannulation when indicated. Furthermore, pro cannulation (Ito et al, 2010; Lee et al, 2012).
viding tension on the cutting wire permits upward bowing of
the sphincterotome, which can facilitate alignment of the tip in Access “Precut” Sphincterotomy for Biliary Access
the proper axis for biliary cannulation. Access “precut” sphincterotomy refers to the technique of
There are various techniques that can be used to gain access incising the papilla prior to obtaining biliary access. Precutting
to the biliary system. The traditional approach entails the use can be a useful technique to achieve selective bile duct cannula
of a catheter to engage the papillary orifice, followed by injec tion when free or wire-guided approaches fail. The needle-knife
tion of contrast to delineate the trajectory and pathway of the (NK) catheter is typically used for this procedure. The NK
bile duct. Following this, the catheter and/or a guidewire can catheters have a retractable bare electrosurgical cutting wire
be inserted directly into the bile duct. Over the past decade, that extends from the tip of the catheter. The exposed needle
guidewire-assisted cannulation has become the favored can then be inserted into the papillary orifice, and the cut is
approach over a contrast-assisted free cannulation technique directed upward in the axis of the bile duct, generally in the 11
because it increases the primary cannulation success rate and to 12 o’clock position. Another variation of this technique
has been associated with lower rates of post-ERCP pancreatitis involves using the NK to begin an incision above the ampullary
(PEP), as it virtually eliminates the possibility of inadvertent orifice to directly access the CBD by creating a biliary fistula
contrast injection of the pancreatic duct (Cheung et al, 2009; (“fistulotomy”).
Lee et al, 2009; Tse et al, 2013). The small-diameter guidewire Precut sphincterotomy has been associated with biliary
with a hydrophilic tip can be advantageous for bile duct can cannulation rates exceeding 90% (Navaneethan et al, 2014).
nulation over the larger-diameter cannula or sphincterotome by However, precutting is not without risk, as the incision is per
minimizing trauma at the papilla. Wire-assisted cannulation can formed without the guidance of a wire within the duct. Previous
be performed by directly probing and advancing the guidewire large multicenter studies have identified precutting as a risk
511
512 PART 4 TECHNIQUES OF BILIARY TRACT INTERVENTION: RADIOLOGIC, ENDOSCOPIC, AND SURGICAL
factor for adverse events, such as acute pancreatitis, hemor The incision is made by upward lifting of the sphincterotome
rhage, and perforation, independent of the endoscopist’s expe with pressure against the papillary roof, but not excessively, to
rience (Freeman et al, 2001; Masci et al, 2001). More recently, avoid a rapid large incision (“zipper”) (Fig. 29.1B). The com
several studies have suggested that precut sphincterotomy may bination of a high cutting current blended with a low coagula
be a surrogate marker of difficult cannulation and not an inde tion current is frequently used, as this is felt to decrease the risk
pendent predictor of PEP (Bailey et al, 2010; Manes et al, of thermal transmission to the adjacent pancreatic tissue and
2009). Indeed, subsequent studies, including multiple meta- hence minimize the risk of pancreatitis. The size of the sphinc
analyses of both randomized and nonrandomized trials indicate terotomy varies on a case-by-case basis and can be limited by
that early precut implementation may be associated with a the length of the intraduodenal portion of the CBD. In general,
lower risk of PEP compared with persistent attempts at the sphincterotomy should be of adequate size to allow the
regular cannulation followed by precutting (Cennamo et al, passage of the stone in the CBD. The size of sphincterotomy
2010; Choudhary et al, 2014). Furthermore, overall complica can be gauged by the ability to move the bowed sphincterotome
tion rates are similar between an early precut sphincterotomy across the opening, by passing an inflated balloon catheter
approach versus persistent attempts at biliary cannulation through the site, and/or by elimination of the tapering or
(Navaneethan et al, 2014). Overall, it is important to empha “pinch” of the intraampullary bile duct seen on fluoroscopy.
size that precut sphincterotomy should be reserved for patients
with a strong indication for biliary access in whom standard Biliary Sphincteroplasty
techniques have failed. This approach should be performed by Endoscopic balloon dilation (sphincteroplasty) of the biliary
an experienced biliary endoscopist familiar with the nuances sphincter muscle was initially proposed as an alternative to
and technical aspects of this approach. endoscopic sphincterotomy. In this procedure, following selec
tive biliary cannulation and placement of a wire in the bile duct,
Transpancreatic Precut Sphincterotomy a balloon-tipped catheter (i.e., CRE balloon or Hurricane RX
(Goff Technique) dilation balloon; Boston Scientific, MA) is advanced over the
Transpancreatic precut (transeptal) sphincterotomy for biliary guidewire. The deflated balloon is positioned across the papilla
access was first described by Goff et al (1995). In this technique, and inflated with radiopaque contrast medium under both endo
following selective cannulation of the pancreatic duct, precut scopic and fluoroscopic visualization (Fig. 29.1C). The inflated
sphincterotomy is performed by cutting the septum between balloon is maintained until the “waist” corresponding with the
the pancreatic and bile duct with the standard sphincterotome biliary sphincter disappears, usually for 15 to 30 seconds. The
directed cephalad toward the bile duct. Since its introduction, optimal balloon size for sphincterotomy has not been estab
several studies have demonstrated high rates of bile duct can lished; however, one should generally avoid inflating the balloon
nulation (82% to 100%) without increased complication rates to a size larger than the diameter of the distal CBD to avoid bile
compared with standard NK precut sphincterotomy (Catalano duct injury (Binmoeller & Schafer, 2001). The main advantage
et al, 2004; Kapetanos et al, 2007). Additional advanced tech of sphincteroplasty is that it results in transient widening of the
niques for biliary access, including those in patients with surgi biliary sphincter such that the biliary sphincter will remain intact
cally altered anatomy, will be covered later in this chapter. and functional postprocedure. This may be advantageous in
children, as an intact biliary sphincter will presumably decrease
TECHNIQUES FOR THE MANAGEMENT the risk of recurrent choledocholithiasis. Furthermore, because
OF CHOLEDOCHOLITHIASIS there is no cutting involved with sphincteroplasty, there is a
lower risk of procedure-related bleeding. Thus use of this tech
Gallstone disease is an endemic condition in developed coun nique may be preferred to sphincterotomy in patients who are
tries, representing a major health burden with annual costs at increased risk of bleeding, such as those on antiplatelet or
of $6.5 billion in the United States (Shaffer et al, 2006). anticoagulation therapy that cannot be held for the procedure.
Choledocholithiasis is concomitantly present in up to 20% Primary biliary sphincteroplasty can also be helpful in facilitating
of patients with cholelithiasis at the time of cholecystectomy biliary cannulation in patients with surgically altered anatomy
(Menezes et al, 2000). Currently, ERCP with biliary sphincter in whom standard sphincterotomy cannot be performed safely
otomy and stone extraction is considered the first-line manage or is technically difficult. The main drawback of performing
ment for choledocholithiasis (Fig. 29.1A) (see Chapters 36 sphincteroplasty alone is its association with a higher risk
and 37). of pancreatitis and lower rates of stone clearance compared
with sphincterotomy (Baron & Harewood, 2004; DiSario et al,
Biliary Sphincterotomy 2004). However, when performed in conjunction with sphinc
Endoscopic sphincterotomy aims at opening the terminal part terotomy, sphincteroplasty has been shown to reduce the need
of the CBD by cutting the papilla and sphincter muscles. The for additional interventions for CBD stone clearance, particu
basic technique of sphincterotomy has not changed significantly larly in the setting of large biliary calculi (≥15 mm) (Madhoun
since its initial description. The standard sphincterotome, the et al, 2014; Teoh et al, 2013). Furthermore, balloon sphinctero
Erlangen “pull-type” model, consists of a catheter containing plasty following sphincterotomy has been shown to be safe, with
an electrosurgical cutting wire exposed 20 to 25 mm near the comparable complication rates compared with sphincterotomy
tip of the sphincterotome. The leading tip distal to the wire, alone (Maydeo & Bhandari, 2007; Weinberg et al, 2006).
the “nose,” is 5 to 10 mm in diameter. Once deep biliary can
nulation has been achieved, the sphincterotome is retracted Stone Extraction
slowly, until one fourth to one half of the wire length is exposed Several series have shown that 82% to 100% of bile duct stones
outside the papilla. The sphincterotome is slightly bowed so can be removed effectively following endoscopic sphincterot
that the cutting wire is in contact with the roof of the papilla. omy with a combination of balloon catheters or wire baskets
Chapter 29 Interventional endoscopy: technical aspects 513
A B
C D
FIGURE 29.1. Endoscopic management of choledocholithiasis. A, Cholangiogram showing diffusely dilated biliary system with stone in the
common bile duct (arrow). Biliary sphincterotomy (B), followed by sphincteroplasty (C). D, Extraction of large stone.
(Bergman et al, 1997; Cotton et al, 1980; Natsui et al, 2013) Lithotripsy
(Fig. 29.1D). Standard stone extraction techniques may fail when a stone
Extraction balloon catheters are available in different sizes is large, impacted, proximal to a stricture, or when stones
(8.5to 20 mm) and in most centers are the standard first-line are multiple. A variety of modalities are currently available to
approach for stone extraction, given their ease of use and lack fragment these difficult stones before extraction, including
of risk of becoming entrapped within the duct. The extraction mechanical lithotripsy, endoscopic intraductal lithotripsy, and
balloon is inflated (to the diameter of the bile duct) above the extracorporeal shock-wave lithotripsy.
stone and pulled back gently to the level of the papilla. In the
setting of multiple stones, it is important to remove the stones Mechanical Lithotripsy
individually starting with the most distal one, to avoid stone Mechanical lithotripsy has been the most frequently used litho
impaction. tripsy approach, given its ease of use and availability, with
Similarly, there are also a variety of wire baskets in different success rates of 90% and higher (Chang et al, 2005; Stefanidis
sizes and configurations. The stone is entrapped between the et al, 2011). There are two variations to the technique of
wires when the basket is closed, and subsequent removal is mechanical lithotripsy: an external-type lithotriptor method
achieved by traction removal of the basket in the axis of the bile and an integrated through-the-endoscope method. Using
duct. The effective traction of the wire basket often allows effec the external-type lithotriptor, the stone is captured within a
tive removal of medium- to large-size stones within the CBD standard Dormia basket, the basket handle is cut off, and
or stones that are “floating” within a dilated bile duct and thus the endoscope is removed. A coiled metal sheath is inserted
easily slip around a balloon. Conversely, the extraction balloon over the wire until its tip is in contact with the stone, and
may be more suitable for the removal of small stones/fragments mechanical lithotripsy is performed by turning the crank handle,
that are difficult to entrap between the wires or when opening crushing the stone between the basket wires and the metal tip
of the basket is constrained by duct caliber. of the sheath. In the integrated through-the-endoscope system,
a special lithotripsy basket contained within a metal sheath is TECHNIQUES FOR THE MANAGEMENT OF
inserted through the accessory channel of the endoscope. Once BILIARY STRICTURES
the stone is captured within the basket, forceful traction on the
wires against the metal sheath results in stone fragmentation. ERCP with stenting is a well-standardized technique com
The most common reasons for failure include large stones monly used to relieve biliary obstruction secondary to both
exceeding 2 cm in size and stone impaction. Basket impaction benign and malignant disease (see Chapters 42, 51, and 52).
within the duct or rupture of the traction wires has been The goal is to relieve the biliary obstruction that can potentially
reported in up to 4% of the cases, which may require external lead to complications, such as jaundice, pruritus, cholangitis,
salvage lithotriptor or surgical retrieval of the retained basket chronic liver disease, and liver failure.
and stone (Garg et al, 2004). Biliary stricture characterization can be a diagnostic chal
lenge that requires a multidisciplinary approach with the inte
Endoscopic Intraductal Lithotripsy gration of laboratory testing, noninvasive and invasive imaging,
Failure of mechanical lithotripsy may be secondary to the and tissue sampling methods. This section focuses on the tech
inability to ensnare the stone within the wires of the basket, nical aspects and outcomes associated with endoscopic man
which can be due to the presence of a stone occupying the agement of benign and malignant biliary strictures. Advances
entire diameter of the duct lumen. This effectively prohibits in endoscopic imaging and tissue sampling for the diagnosis of
adequate opening and manipulation of the basket. Further biliary strictures will be covered later in this chapter.
more, stones located above a stricture, or stones larger than the
biliary sphincterotomy and/or sphincteroplasty may not be Types of Biliary Stents
amenable to standard extraction if the bile duct exit path is Plastic Stents
narrower than the stone. In such instances, it is often necessary Plastic biliary stents are composed of polyethylene, polyure
to fragment the stones using alternative means. thane, or Teflon. Stent diameter and length range from 5 Fr to
Intraductal shock-wave lithotripsy under direct visualization 12 Fr and 5 to 18 cm, respectively (Pfau et al, 2013). Plastic
through a cholangioscope is an alternative modality for the stents are available in a variety of configurations: straight,
fragmentation of these refractory calculi. In electrohydraulic angled, curved, with flaps (flanged), or coiled at one or both
lithotripsy (EHL), the creation of a spark within fluid medium ends (single or double pigtail) for anchorage. All plastic stents
results in the generation of a propagating shock wave capable are radiopaque, some with additional markers at the proximal
of fragmenting intraductal stones. Laser lithotripsy (LL) is and distal end of the stents to facilitate visualization under fluo
based on the principle of transforming optical energy into roscopy. Insertion is via a push catheter over a guidewire.
mechanical energy. In LL, focusing the high-power density Duration of stent patency is largely dependent on the size of
laser light on the surface of the calculi results in the transforma the inner diameter, with 10-Fr and larger stents remaining
tion of matter into a plasma state, a gaseous collection of ions patent, on average, for approximately 3 months, and stent
and free electrons. Subsequently, the plasma expands, inducing occlusion developing secondary to bacterial colonization,
an oscillation wave with tensile/compressive forces directed at sludge, or tissue debris (Donelli et al, 2007).
the stone, leading to fragmentation (DiSario et al, 2007). Once
the stone is fragmented, then standard extraction techniques Self-Expandable Metal Stents
with balloons and baskets can be used. The effectiveness of Self-expandable metal stents (SEMSs) are composed of stain
EHL and LL for stone removal is similar, with clearance rates less steel or a variety of metal alloys, such as nitinol (nickel
ranging from 75% to 98% (Arya et al 2004; Swahn et al, 2010). and titanium combination) or platinol (platinum core with
Adverse events with EHL and LL have been reported in 3% to nitinol encasement). This material is malleable, which allows
14% of cases, with the most common being cholangitis and the SEMSs to adapt to many configurations without compro
bleeding, and rarely, bile duct injury. Prophylactic antibiotics mising radial expansile force. SEMSs are configured into a
are recommended for patients undergoing cholangioscopy cylinder by interwoven wires and are deployed from a con
given the increased risk of cholangitis/sepsis (Piraka et al, strained position within a delivery catheter. Stent diameter and
2007). length vary from 6 mm to 10 mm and 4 to 12 cm, respectively.
The larger stent diameter compared with plastic stents results
Extracorporeal Shock-Wave Lithotripsy in increased duration of stent patency (on average, 6 to 12
Extracorporeal shock-wave lithotripsy (ESWL) is an adjunctive months). SEMSs can be covered, partially covered, or uncov
safe and effective method of delivering shock waves. The tech ered. The covering consists of a silicone, polycaprolactone,
nique can be used for stone fragmentation under ultrasound or polyether polyurethane, polyurethane, or expanded polytetra
fluoroscopic guidance, after contrast is instilled via a nasobiliary fluoroethylene fluorinated ethylene propylene lining (Webb
catheter to aid in stone visualization (Tandan et al, 2009). et al, 2013). Covering reduces tumor or hyperplastic tissue
Multiple ESWL sessions are generally required. Adverse events, ingrowth but is associated with higher migration rate compared
including hemobilia, cholangitis, pancreatitis, and cardiac with uncovered metal stents (Lee et al, 2013). Partial or fully
arrhythmias, have been reported in 10% to 35% of cases covered SEMSs can typically be removed, whereas uncovered
(Muratori et al, 2010). The need for ESWL has decreased as stents are considered permanent.
the effectiveness of intraductal lithotripsy has improved, with
several randomized studies suggesting superior ductal clearance Endoscopic Management of Benign Biliary Strictures
and fewer treatment sessions with LL compared with ESWL Most cases of benign biliary strictures are secondary to post
(Jakobs et al, 1997; Nehaus et al, 1998). Thus ESWL is not operative iatrogenic injury, following cholecystectomy or at
routinely considered the first-line treatment of refractory bile the site of a biliary anastomosis after biliary surgery or liver
duct calculi. transplantation (Fig. 29.2A) (see Chapter 42). Benign biliary
A B C
FIGURE 29.2. A, Cholangiogram showing a benign biliary anastomotic stricture (arrow) after orthotopic liver transplantation. Balloon dilation across
the stricture (B), followed by placement of three plastic biliary stents (C).
in diameter) when the stricture is not amenable to mechanical

TABLE 29.1 Etiologies of Biliary Strictures
dilation. The balloon is inflated with diluted contrast material
Benign Etiologies Malignant Etiologies to the maximum atmospheric pressure allowable and is kept
Iatrogenic (postoperative) Pancreas carcinoma inflated until the stricture “waist” is obliterated, generally after
Cholecystectomy Cholangiocarcinoma 30 to 60 seconds (Fig. 29.2B). Dilation can be safely performed
Biliary anastomosis Ampullary carcinoma to a diameter 1 to 2 mm larger than the downstream bile duct
Chronic pancreatitis Gallbladder carcinoma diameter. Depending on the etiology, stricture dilation alone is
Autoimmune pancreatitis Hepatocellular carcinoma
associated with high restricturing rates, and thus stent place
Primary sclerosing cholangitis Metastasis
Autoimmune sclerosing ment is often pivotal in maintaining patency (Draganov et al,
cholangiopathy 2002). Typically, multiple plastic stents (number and caliber of
Ischemic stents limited by size of the bile duct and stricture) or a single
Vasculitis covered SEMS can be placed and exchanged periodically until
Infectious (viral, parasitic, HIV
cholangiopathy, tuberculosis) stricture resolution is achieved (Fig. 29.2C). It is important to
Radiation therapy highlight that placement of fully covered SEMSs may poten
Postsphincterotomy tially block drainage of a neighboring duct (including the cystic
Portal biliopathy duct); thus these stents are preferably restricted to distal bile
Abdominal trauma
duct strictures and must be used cautiously in patients with an
HIV, Human immunodeficiency virus. in situ gallbladder, as mechanical obstruction of the cystic duct
and resultant cholecystitis may ensue (Fumex et al, 2006).
strictures may also result from ischemic injury and/or inflam Efficacy and Outcomes of Endoscopic Stenting of Benign
matory processes (Table 29.1). Comprehensive evaluation of a Biliary Strictures
biliary stricture and its etiology should always be sought prior Balloon dilation and plastic stenting has traditionally been the
to any therapeutic endoscopic intervention. Magnetic reso mainstay of endoscopic therapy for benign biliary strictures
nance cholangiopancreatography (MRCP) is a noninvasive (Van Boeckel et al, 2009). The long-term results of endoscopic
imaging modality that can accurately delineate the biliary therapy are dependent on the underlying etiology, with postop
anatomy, site, and length of the biliary stricture and thus erative benign biliary strictures (i.e., postcholecystectomy,
provide useful information for ERCP planning (Costamagna & biliary anastomotic strictures following liver transplantation)
Boskoski, 2013). more responsive to treatment compared with biliary strictures
associated with chronic pancreatitis. In postoperative benign
Endoscopic Technique biliary strictures, Bergman and colleagues (2001) reported
ERCP is performed following detailed assessment of the type restenosis rate of 20% after a median follow-up of 9.1 years in
and site of the biliary stricture. The key technical step is to patients who had two 10-Fr stents exchanged every 3 months
negotiate the stricture and achieve biliary access with the use for a total of 12 months. In a separate study of 45 patients with
of a guidewire. This process can be challenging, and the use of postoperative biliary strictures, Costamagna and colleagues
different sizes and types of hydrophilic guidewires in addition (2001) inserted the maximum number of plastic biliary stents
to guidewire manipulation with steerable catheters or sphinc possible (mean, 3.2; range, 1 to 6) per ERCP session, with stent
terotomes is often necessary. Although not required in all cases, exchange performed every 3 months (mean duration of treat
biliary sphincterotomy may need to be performed to facilitate ment, 12 months). This more aggressive biliary stent approach
subsequent therapy, including stricture dilation and stent was associated with no recurrence of symptoms due to biliary
placement. stricture at a mean follow-up of 49 months (range, 2 to 11
Stricture dilation can be performed using push-type dilation years). In contrast, long-term outcomes of endoscopic therapy
catheters (bougies) or hydrostatic balloons (range, 4 to 10 mm for bile duct strictures in the setting of chronic pancreatitis are
significantly less favorable, with one study reporting successful surgical resection, and in those with borderline resectable
endoscopic treatment in only 38% of the cases and persistence disease who may benefit from neoadjuvant chemotherapy. In
of strictures even when endoscopic stenting was performed patients with unknown resectability, placement of the shortest-
beyond a 12-month period (Cahen et al, 2005). length SEMS to bridge a distal stricture should not interfere
Fully-covered SEMSs (FCSEMSs) have been proposed as with subsequent pancreaticoduodenectomy if indicated (Cavell
an alternative to plastic stents for the management of benign et al, 2013). Conversely, if the biliary stricture is proximal and
biliary strictures (Kaffes et al, 2013). Multiple studies have definitive diagnosis and/or staging is not certain, plastic biliary
shown high success rates, with a large prospective multicenter stents can be placed for biliary decompression while workup is
study reporting an overall stricture resolution rate of 90% completed.
(Costamagna et al, 2001). However, similar to prior studies
with plastic stents, long-term stricture resolution rates with Palliative Biliary Drainage of Distal Bile Duct Obstruction
FCSEMs are significantly lower (46% to 75%) for patients with Distal malignant biliary obstructions are typically defined as
chronic pancreatitis (Poley et al, 2012; Tarantino I et al, 2012), those distal to the cystic duct insertion, although this definition
reiterating the refractory nature of this type of strictures to is suboptimal, given the substantial anatomic variability. Thera
endoscopic therapy. peutic options include surgical bypass, percutaneous decom
Overall, the clinical success rates between plastic stents and pression, and endoscopic stenting. A large meta-analysis of
FCSEMSs are comparable, albeit the latter have been associ 2436 patients demonstrated that endoscopic stenting was asso
ated with less endoscopic sessions for stricture resolution ciated with a lower risk of procedural complications than tra
(Kaffes et al, 2012). Further prospective, randomized studies ditional surgical bypass. Overall, SEMSs were not superior to
are needed to compare the long-term efficacy, safety, and cost- plastic stents in terms of technical success, therapeutic success,
effectiveness between metal and plastic stents for the manage or complications but were associated with improved patency
ment of benign biliary strictures. To date, the ideal number of rates (Moss et al, 2007). The shorter patency rate of plastic
stents, the type of stent, and the duration of stent placement stents has been associated with increased repeat endoscopic
for benign stricture resolution remain highly debated and often interventions to reestablish biliary drainage and more hospital
vary from case-by-case basis. ization days compared with SEMSs (Kaassis et al, 2003).
Several studies have not demonstrated differences in stent
Endoscopic Management of Malignant Biliary Strictures patency rates between covered and uncovered SEMSs for distal
Malignant biliary obstruction is most frequently seen in the malignant biliary obstruction, albeit a subsequent meta-analysis
setting of pancreaticobiliary malignancy, but it can also be due suggested possible prolonged patency with covered SEMSs but
to many other etiologies (see Table 29.1). When indicated, a trend toward increased stent migration and tumor overgrowth
endoscopic therapy is considered the mainstay therapy for (Park do H et al, 2006; Saleem et al, 2011; Yoon et al, 2006).
biliary decompression. A careful multidisciplinary review of the Ultimately, the optimal stent choice depends on various factors,
indication and appropriateness of any endoscopic intervention including establishment of diagnosis, need for reinterventions,
should be sought prior to any procedure. operator’s expertise, cost analysis, and the patient’s life
expectancy.
Preoperative Biliary Drainage
The routine need for preoperative biliary decompression in Palliative Biliary Drainage of Proximal Bile Duct
patients with resectable pancreaticobiliary disease remains con Obstruction (Hilar)
troversial. From a technical standpoint, the placement of a Malignant biliary obstructions at the biliary confluence can be
short plastic or metal biliary stent does not appear to interfere technically challenging (see Chapter 51). The extent of the
with subsequent pancreaticoduodenectomy. However, routine biliary obstruction is commonly classified based on the modi
preoperative biliary drainage may be associated with increased fied Bismuth-Corlette classification and can be summarized as
complications and not necessarily with improved postoperative follows: Bismuth type I strictures involve the proximal common
outcomes compared with patients who proceed directly to hepatic duct but not the confluence of the left and right ductal
surgery. In a multicenter randomized, controlled trial of 202 systems, type II involves the confluence but spares the segmen
patients with resectable pancreatic cancer, the rates of compli tal hepatic ducts, type IIIa and IIIb involve either the right or
cations were significantly higher in patients who underwent left segmental hepatic ducts, respectively, whereas both seg
preoperative biliary drainage (74%) compared with those who mental hepatic ducts are involved in type IV (Larghi et al,
underwent surgery alone within 1 week of diagnosis (39%) 2008). Biliary drainage with stent placement can be particularly
(Van der Gaag et al, 2010). Limitations of this study include a difficult in those with advanced complex strictures (Bismuth
25% ERCP failure rate and the use of plastic stents. In a sepa type II and above). Evaluation with noninvasive imaging (i.e.,
rate study, Aadam and colleagues (2012) reported SEMS MRCP) to delineate the anatomy is mandatory for preproce
patency rate of 88% in 55 patients who underwent neoadjuvant dural planning and helps limits contrast injection during the
therapy for a median time of 104 days. A multicenter study of ERCP and the risk of contaminating undrainable segments.
241 patients with resectable and borderline-resectable carci Furthermore, evaluation of the side of the liver (i.e., lack of
noma of the head of the pancreas demonstrated successful atrophy, patent portal vein branch, and lack of extensive seg
placement of SEMSs in all patients as they underwent neoad mental biliary involvement) that will provide the best biliary
juvant therapy with a stent occlusion rate of 5.8% at a mean drainage is critical. It is of utmost importance that the best
time of 6.6 months (Siddiqui et al, 2013). Thus, although approach (including percutaneous approaches) be evaluated in
routine endoscopic intervention prior to curative surgery may a multidisciplinary review.
not be necessary in all patients, there is evidence to suggest a In patients with strictures that do not involve the confluence
role for preoperative biliary drainage in the case of delayed of the right and left hepatic ducts (Bismuth type I), adequate
A B C
FIGURE 29.3. A, Cholangiogram showing a type IIIb malignant hilar obstruction with upstream dilation of the right and left hepatic ducts. Balloon
dilation across the stricture (B) and placement of bilateral uncovered metal biliary stents to drain both lobes of the liver (C).
biliary drainage is often achieved by the placement of a single obstruct drainage of adjacent ducts. For stent implantation, a
biliary stent. Controversy exists as to whether both lobes of the guidewire is advanced across the malignant stricture into the
liver need to be drained when a bifurcation lesion obstructs duct preselected for drainage (Fig. 29.3A). After wire place
both lobes. The main determining factor associated with effec ment, if necessary, dilation of a tight stricture can be performed
tive drainage is the liver volume to be drained. In principle, with a balloon catheter or bougie (Fig. 29.3B). A sphincterot
jaundice can often be relieved when approximately 25% to 30% omy is not necessary when the distal ends of the stents are
of the liver is adequately drained (Dowsett et al, 1989). In a positioned within the duct, which may reduce the risk of post
more recent study, the volume of liver drained (>50%) as stenting cholangitis. On the other hand, stent revision is techni
assessed by computed tomography (CT) imaging was the main cally less demanding and more accessible when the distal end
predictive factor of effective drainage following endoscopic of the stents protrude out of the papilla (Fig. 29.3C). Regard
stenting for malignant hilar biliary strictures. Indeed, drainage less of stent positioning, all patients who undergo endoscopic
of greater than 50% was also associated with a longer median therapy for these complex strictures should receive prophylactic
survival (Vienne et al, 2010). Thus the goal of endoscopic antibiotics.
therapy should be directed at achieving drainage volume greater In summary, endoscopic stenting of proximal biliary obstruc
than 50%, irrespective of whether unilateral or multisegmental tion is challenging. Preprocedural cross-sectional imaging and
stenting is performed. multidisciplinary review is essential in the selection of the target
Several studies have suggested that unilateral stent place parenchyma for drainage and the optimal approach (percutane
ment and drainage is adequate in most patients with nonresect ous vs. endoscopic). If endoscopic approaches are favored, this
able malignant hilar lesions. From a technical standpoint, planning will in turn help maximize biliary drainage by target
unilateral stents have been shown to be associated with a higher ing the dominant biliary systems, limit the use of contrast
rate of successful endoscopic placement and lower rate of com during the procedure, and avoid intubating atrophic segments
plications compared with bilateral stents (De Palma et al, 2001; or areas that cannot be effectively drained.
Mukai et al, 2013). Iwano and colleagues (2011) reported that
unilateral drainage was associated with lower incidence of liver Photodynamic Therapy
abscess compared with bilateral drainage. Nonetheless, endo Photodynamic therapy (PDT) is based on the ability of photo
scopic insertion of multiple stents may be required when effec sensitizers to generate cytotoxic oxygen species in the target
tive drainage of greater than 50% of the liver volume involves tissue upon exposure to light of an appropriate wavelength.
several hepatic segments or to ensure drainage of all opacified Photosensitizing agents (sodium porfimer or aminolaevulinic
ducts. Chang and colleagues (1998) reported a higher survival acid) are injected intravenously preprocedurally, and ERCP is
rate in patients with bilateral drainage compared with those subsequently performed 2 to 4 days thereafter. A catheter with
with persistent duct opacification following unilateral drainage. a quartz fiber coupled with a diode laser emitting a wavelength
Despite the increased technical difficulty and higher costs, of 630 nm is inserted into the bile duct through the accessory
several recent studies have also reported higher stent patency channel of the endoscope. The catheter is directed against the
and decreased reintervention for stent occlusion with bilateral photosensitized malignant cells, causing tumor cell death by the
compared with unilateral stent placement (Liberato & Canena, generation of oxygen free radicals.
2012; Naitoh et al, 2009). In a randomized controlled trial, Ortner and colleagues
Several studies have suggested that SEMSs are preferable to (2003) compared PDT with stenting versus stenting alone in
plastic stents based on higher rates of successful drainage, pro patients with nonresectable cholangiocarcinoma. The authors
longed patency, prolonged survival, and lower rates of compli demonstrated that PDT resulted in prolongation of survival
cations (Naitoh et al, 2009; Vienne et al, 2010). The use of (493 days with PDT plus stenting compared with 98 days in
uncovered SEMSs is preferred for proximal malignant biliary stent-alone group) and improved biliary drainage and quality of
strictures as placement of a covered stent can potentially life. Subsequent nonrandomized studies have also demonstrated
A B
FIGURE 29.4. A, Cholangiogram showing contrast extravasation (arrow) at the biliary anastomosis in a patient after liver transplantation, consistent
with bile leak. B, Bile leak treated by placing a plastic stent across the biliary anastomosis.
reduction in serum bilirubin and increased survival with PDT intraductal filling has occurred (Fig. 29.4A). The aim of endo
(Kahaleh et al, 2008; Prasad et al, 2007). Main limitations of scopic therapy is to decrease the transpapillary pressure gradi
PDT include limited availability, high costs, risk of infection ent, thus favoring transpapillary bile flow rather than
(i.e., cholangitis, liver abscess), and risk of photosensitivity extravasation at the site of the leak (Sandha et al, 2004). This
reaction. can be achieved by performing a biliary sphincterotomy, place
ment of a transpapillary biliary stent, or both. In most instances,
Radiofrequency Ablation placement of a plastic biliary stent (7 Fr or 10 Fr) is sufficient
Radiofrequency ablation (RFA) relies on the generation of without the need of a sphincterotomy and its potential associ
high-frequency alternating electromagnetic energy resulting in ated risks (Katsinelos et al, 2008) (Fig. 29.4B). It is not neces
thermal injury to the target tissue (see Chapter 98B). Intra sary to place the proximal end of the stent beyond the site of
ductal RFA for malignant strictures has become feasible with the leak as reduction of the pressure inside the duct alone
the introduction of a novel RFA probe. The Habib Endo HBP is generally sufficient. Typically, the stent is left in place for
probe (EMcision, London, England) is a bipolar 8-Fr catheter approximately 4 to 6 weeks. Various studies have reported
with two stainless steel electrodes, 8 mm apart, located at the endoscopic success rates for the management of bile leaks
distal tip. The probe can be inserted into the working channel between 90% to 100% (Kaffes et al, 2005; Kim et al, 2014;
of the duodenoscope and advanced into the bile duct over a Ryan et al, 1998). In the minority of cases in which the bile
guidewire. The probe is subsequently activated at 10 W of leak is refractory to endoscopic therapy with plastic stent place
energy for 90 seconds at a time, resulting in local coagulation ment and/or sphincterotomy, upsizing the stent or placing mul
necrosis (Webb & Saunders, 2013). tiple plastic stents can be performed in subsequent sessions
Recent studies have confirmed the safety and efficacy of until resolution is documented. A small pilot study suggested
RFA as an adjunct to SEMSs in patients with a malignant a potential role for FCSEMSs in patients with bile leak refrac
biliary stricture. Figueroa-Barojas and colleagues (2013) dem tory to plastic stent placement; however, these preliminary find
onstrated improvement in stricture diameter following RFA as ings have not been corroborated by larger, multicenter studies
well as 30-day stent patency. The same group subsequently (Kahaleh et al, 2007).
compared metal stenting with RFA versus stenting alone. In It should be noted that, in the presence of a perihepatic bile
their retrospective study, RFA was an independent predictor of collection, endoscopic stenting alone does not result in the
survival, with comparable stent patency rates between both reabsorption of the established biloma. Thus symptomatic
groups (Sharaiha et al, 2014). Further studies including ran bilomas will need to be drained percutaneously (see Chapters
domized controlled trials over a longer period are needed to 27 and 30). An output of less than 10 mL per day through a
validate these preliminary findings. percutaneous drain is associated with bile leak resolution and
can be used as a surrogate indicator for stent removal.
TECHNIQUES FOR THE MANAGEMENT
OF BILE LEAK ENDOSCOPIC MANAGEMENT OF
AMPULLARY ADENOMAS
Bile leak is a well-known complication from injury to the biliary
tree, either secondary to trauma (see Chapter 122) or iatro Ampullary adenomas are dysplastic glandular lesions that arise
genic following laparoscopic cholecystectomy (see Chapters 35 from either the major or minor duodenal papilla. These lesions
and 38), liver resection (see Chapter 103), or transplantation can occur sporadically or arise in the context of genetic syn
(see Chapter 120). Bile leaks can be classified as either high dromes, such as familial adenomatous polyposis. If not removed,
grade or low grade. High-grade leaks demonstrate rapid extrav ampullary adenomas can undergo malignant transformation to
asation of contrast during cholangiogram, whereas low-grade ampullary cancer, with a reported incidence from 25% to 85%
leaks exhibit contrast extravasation only after near-complete (Hirota et al, 2006; Seifert et al, 1991; Takashima et al, 2000).
With advances in therapeutic endoscopy, endoscopic ampul lesion is generally tethered down by the bile and pancreatic
lectomy has become an acceptable alternative therapy to surgery duct. Thus submucosal injection may actually raise the sur
for ampullary adenomas (see Chapter 59). rounding mucosa, create a depressed center (“valley effect”),
and interfere with en bloc excision and subsequent attempts at
Diagnosis and Local Staging bile and pancreatic duct access (Harewood et al, 2005; Irani
Prior to endoscopic ampullectomy, preoperative assessment et al, 2009).
with both a forward- and side-viewing endoscope is routinely For most ampullary lesions, the tip of the snare is positioned
performed to further characterize the lesion. Endoscopic find against the wall of the duodenum at the superior aspect of
ings, including spontaneous bleeding, friability, ulceration, the mass. The snare is then slowly opened and the snare cath
and induration, are often associated with malignant lesions. eter advanced slowly to allow the open snare to encircle the
Biopsies obtained during endoscopy can assess for dysplasia or lesion. Once achieved, the snare is slowly closed while simul
unsuspected carcinoma, although malignancy may be missed taneously advancing the snare catheter toward the base of
in up to 30% of tumors when forceps biopsy specimens are the lesion, followed by polypectomy. Thermal therapy (i.e.,
obtained (Elek et al, 2003). Hence other advanced imaging argon plasma coagulation, monopolar and bipolar coagulation,
modalities, including magnifying endoscopy and narrow-band neodymium:yttrium-aluminum-garnet laser) can be used to
imaging, have been proposed as complementary techniques to fulgurate any residual tissue following piecemeal or incomplete
help predict histologic characteristics of ampullary lesions resection, with caution to avoid excessive tissue destruction
(Uchiyama et al, 2006). around the biliary and pancreatic duct orifices.
EUS has been shown to be superior to CT, angiography, Numerous studies have demonstrated decreased com
and magnetic resonance imaging for local tumor staging, plications with ampullectomy when prophylactic pancreatic
including assessment of the size of lesion and involvement/ stenting is performed (Martin et al, 2003; Yamao et al, 2010).
infiltration of the periampullary wall layers, the CBD, and the Indeed, placement of a prophylactic pancreatic stent is strongly
pancreatic duct (Chen et al, 2009; Manta et al, 2010) (see recommended as it decreases the risk of post-ERCP pancreati
Chapter 16). Many experts agree that smaller lesions (<1 cm) tis following ampullectomy (Harewood et al, 2005; Singh
without suspicious signs of malignancy (ulceration, spontane et al, 2004). Whether ERCP with pancreatic and biliary
ous bleeding, biopsies positive for high-grade dysplasia or car sphincterotomy/stent placement is performed pre-resection or
cinoma) may not require EUS evaluation prior to endoscopic postresection is often dependent upon the endoscopist’s prefer
resection (Baillie, 2005). When performed, EUS can help strat ence (Fig. 29.5C). Because identification of the pancreatic
ify which lesions are amenable to endoscopic ampullectomy. orifice following ampullectomy can be challenging, some
ERCP and intraductal ultrasound (IDUS) can also aid in the authors favor performing pancreatography with iodinated con
detection of tumor extension into either ductal system (De trast diluted with methylene blue or indigo carmine prior to
Palma et al, 2014). resection. The blue-stained pancreatic orifice can theoretically
be more readily identified adjacent to the bile-stained biliary
Endoscopic Therapy orifice and thus facilitate postresection cannulation (El Hajj
Endoscopic ampullectomy (papillectomy) can be considered et al, 2013). Prophylactic biliary sphincterotomy and/or stent
once malignancy has been reasonably excluded. This procedure ing is not widely performed nor uniformly recommended unless
is performed with the standard monopolar diathermic snare there is concern for inadequate biliary drainage following
used for colon polypectomy. The aim of endoscopic excision is ampullectomy (Adler et al, 2006).
to obtain complete removal of the ampullary lesion, preferably Patients who have undergone endoscopic ampullectomy
en bloc (Fig. 29.5A and B). When performing this technique, should undergo routine surveillance. A systematic review esti
it is common for the intraampullary portions of the CBD and mated recurrence rates of 0% to 33% with larger size and
pancreatic duct to be removed as well. Submucosal injection is intraductal extension as recognized risk factors (Han et al,
not routinely recommended, as the center of the ampullary 2006). Endoscopic surveillance should be performed with a
A B C
FIGURE 29.5. Endoscopic ampullectomy. A, Ampullary adenoma. Ampullectomy with en bloc resection (B), followed by biliary and pancreatic
duct stenting (C).
side-viewing duodenoscope; the timing interval is dependent In addition to the potential risks associated with conven
on several factors, including histology and margin status of tional ERCP, Bilroth II anatomy increases the risk of adverse
resected specimen, history of familial adenomatous polyposis, events, including perforations at the gastrojejunal anastomosis
and the patient’s age and comorbidities. Furthermore, any or within the afferent limb itself (Faylona et al, 1999; Kato
specimen with unexpected malignancy should be referred for et al, 2014). Thus the endoscopist should be familiar with
surgical consultation. multiple techniques and be prepared to change strategies on a
case-by-case basis depending on the intraprocedural findings.
ADVANCED TECHNIQUES FOR Endoscopic Retrograde Cholangopancreatography in Patients

BILIARY ACCESS With a Roux-en-Y Anatomy
In patients with a Roux-en-Y anatomy, the segment of small
Techniques for Biliary Access in Patients With bowel upstream to the gastrojejunal anastomosis, which con
Surgically Altered Anatomy tains the biliary orifice, is known as the afferent or biliopancre
ERCP in patients with surgically altered anatomy can be tech atic limb (see Chapter 42). The segment of small bowel from
nically difficult. There are two main challenges that need to be the hepaticojejunostomy to the point where the afferent limb is
overcome to successfully complete the procedure. The first anastomosed (jejunojejunostomy) is known as the efferent or
challenge is to reach the papilla or bilioenteric anastomosis in Roux limb. This type of surgical configuration is commonly
altered luminal anatomy. Once in position, the second chal seen in patients with bariatric surgery (Roux-en-Y gastric
lenge is to be able to cannulate and perform the intended bypass [RYGB]) or Roux-en-Y reconstructions after gastric
intervention from an altered position with the available endo resection, liver transplant, and biliary tract operations. Roux-
scopes and accessories. en-Y anatomy poses a major challenge for ERCP, given the
length of bowel that must be traversed, which is often prohibi
Endoscopic Retrograde Cholangopancreatography in Patients tive for duodenoscopes. Success rates of reaching the biliary
With a Gastrojejunostomy (Bilroth II) anastomosis are as low as 33% (Lee and Shah, 2013). Prior to
The Billroth II anastomosis is created after the distal stomach is an attempt at ERCP, a comprehensive review of the patient’s
resected and an end-to-side gastrojejunostomy is created. From operative report, imaging, and prior procedures are necessary
the gastrojejunal anastomosis, an afferent limb leads toward the to develop a strategic approach.
proximal duodenum, whereas the efferent limb leads to the distal Push enteroscopy (using a forward-viewing enteroscope or
small bowel. ERCP is performed by intubating the afferent limb push colonoscope) can be an alternative when the biliopancre
and cannulating the papilla from a caudal angle. Although the atic limb cannot be reached with the duodenoscope. A case
afferent limb is generally of short length, identification of the series of 18 patients revealed a high success rate (82% to 86%)
limb and navigating through the sharp angulation of this limb in reaching the papilla or ductal anastomosis with either the
can be challenging with the conventional side-viewing duodeno push enteroscope or pediatric colonoscope (Elton et al, 1998).
scope. If this initially fails, using a forward-viewing gastroscope Deep enteroscopy platforms, including single balloon, double
can be help identify the correct limb, which can then be marked balloon, and spiral enteroscopy, were developed to allow access
with submucosal ink tattoo or by the placement of a guidewire to the distal small bowel. All of these technologies are similar
to facilitate subsequent duodenoscopy (Garcia-Cano, 2008). in that they rely on an overtube system to allow deeper endo
Alternatively, the entire procedure can be performed with a cap- scope insertion. Although the longer enteroscopes may facili
fitted forward-viewing gastroscope or pediatric colonoscope. tate navigation through the surgically altered anatomy compared
The transparent cap potentially facilitates navigation through with the duodenoscope, this advantage comes with several limi
the tortuous afferent limb and stabilizes the scope position for tations. First, the lack of a side-viewing perspective can poten
selective biliary cannulation. Recent studies have reported tially make cannulation more difficult. Second, there are limited
rates of access and selective biliary cannulation exceeding 95% accessories specifically designed to use with the longer endo
(Anastassiades et al, 2013; Ki et al, 2015). scopes to perform diagnostic and therapeutic interventions.
When using a duodenoscope, the papilla is usually visible The use of larger-diameter biliary stents can be limited by the
en face upon reaching the second portion of the duodenum. size of the working channel of the endoscope, and even smaller-
From this position, the papilla appears rotated by 180 degrees, caliber accessories may be difficult to advance through the
and biliary cannulation generally progresses toward the 5 channel when the longer endoscope is torqued or looped in the
o’clock position compared with the 11 o’clock position for surgically altered bowel. Last, these procedures can be lengthy
standard ERCP. Hence standard straight cannulas may be pref (90 to 120 minutes), with the increased risk of prolonged
erable for selective bile duct cannulation compared with the general anesthesia (Choi et al, 2013). A multicenter study
upward-curved papillotomes. Alternatively, other devices, such evaluating deep enteroscopy in patients with long-limb surgical
as the Bilroth II sphincterotome with a downward “reversed bypass revealed that the papilla or ductal anastomosis was only
bow” curved wire or rotatable sphincterotomes can be used to successfully reached in 71% of the cases (Shah et al, 2013). In
facilitate bile duct access in these patients (Maluf-Filho, et al, patients in whom enteroscopy was successful, intended ERCP
2008). In a large single-center series of 713 patients, the success was completed in 85% to 90% of the cases. These results
rate for afferent limb intubation and biliary or pancreatic duct suggest that despite the limited ERCP accessories compatible
cannulation using the duodenoscope was 87% and 94%, with these enteroscopes and potential cannulation difficulties
respectively (Bove et al, 2013). Similarly, ERCP success rate with the forward view, the main obstacle for ERCP success
with overtube-assisted enteroscopy (single balloon, double resides in reaching the papilla or ductal anastomosis.
balloon, spiral enteroscopy) exceeds 90% in patients with Transoral ERCP in patients with RYGBcan be even more
Bilroth II anatomy (Skinner et al, 2014). challenging given the relatively longer Roux and biliopancreatic
limbs that must be traversed. Hence alternative access routes the duodenum (Gupta et al, 2014). Bile is aspirated, and con
through the remnant stomach directly to the native papilla trast is injected to confirm position inside the bile system. A
have been explored. Previous studies have reported successful hydrophilic guidewire is then advanced antegrade through the
ERCP through a surgical or percutaneous gastrostomy, with EUS needle and into the bile duct and manipulated across the
the main limitation being the need for a 2-week period for tract papilla. The echoendoscope and needle are carefully removed
maturation, and thus this technique is not applicable in cases while maintaining the guidewire in place. The duodenoscope
requiring urgent ERCP (Gutierrez et al, 2009; Tekola et al, is then inserted and advanced to the duodenum with visualiza
2011). Conversely, to obviate the need for a two-step process, tion of the wire traversing the papilla. The distal end of the
laparoscopy has been used to create a point of access to the indwelling guidewire can be grasped with forceps or snare,
remnant stomach, allowing immediate ERCP by inserting the withdrawn through the accessory channel, and a cannulation
duodenoscope through the trocar. In a retrospective study com catheter back-loaded over the guidewire and readvanced to the
paring laparoscopic-assisted ERCP (LA-ERCP) to balloon- papilla. Alternatively, biliary cannulation can be accomplished
enteroscopy–assisted ERCP (BEA-ERCP), LA-ERCP was in the standard retrograde fashion adjacent to the indwelling
significantly superior to BEA-ERCP in both cannulation rate wire. Overall success and complication rates of the EUS-guided
and therapeutic success (100% vs. 59%, P < .001) (Schreiner rendezvous technique are 81% and 10%, respectively (Iwashita
et al, 2012). More recently, a technique using percutaneous- et al, 2014).
assisted transprosthetic endoscopic therapy has been described
(Law et al, 2013). In this technique, an enteroscope is advanced Endoscopic Ultrasound—Guided Antegrade Biliary Drainage
transorally into the excluded stomach, followed by the creation Antegrade placement of a biliary stent can be accomplished in
of a percutaneous endoscopic gastrostomy. The gastrostomy cases when ERCP cannot be performed due to luminal obstruc
tract is then dilated to allow the placement of a FCSEMS. A tion proximal to the papilla or in cases where the papilla cannot
duodenoscope can then be advanced through the stent to be reached (i.e., RYGB). After obtaining biliary access under
perform antegrade ERCP. Other techniques using EUS to gain EUS guidance, the transmural tract is dilated with either a
gastrostomy access have also been reported. Following identi balloon catheter or bougie to allow advancement of the stent
fication of the excluded stomach under EUS, access can be into the bile system with subsequent deployment across the
obtained by using a lumen-apposing metal stent to create a obstruction. In aggregate, success and complication rates
gastrogastric fistula through which a duodenoscope is inserted reported from small case series are 77% and 5%, respectively
to perform ERCP at the same session (Kedia et al 2014). The (Iwashita et al, 2014).
main concerns with this technique include risk of weight regain
with a patent gastrogastric fistula and stent migration. Further Endoscopic Ultrasound—Guided Transluminal
larger prospective studies are needed before these innovative Biliary Drainage
techniques can be endorsed. EUS-guided transluminal drainage can be considered after
In summary, selection of the appropriate technique for failed ERCP (i.e., luminal obstruction, surgically altered
biliary access in patients with surgically altered anatomy should anatomy, unsuccessful biliary cannulation) and as an alternative
be individualized and would likely involve a combination of to percutaneous transhepatic biliary drainage or surgery. A
methods and endoscopic tools based on patient factors and EUS-guided transluminal approach also permits biliary decom
operator’s expertise. pression in cases when EUS-guided antegrade biliary access is
not feasible because the wire cannot be advanced across the
Endoscopic Ultrasound—Guided Biliary Drainage papilla (i.e., impacted biliary calculi, papillary stenosis, and
ERCP success rates for biliary and pancreatic duct decompres tumor infiltration). Biliary access can be obtained either via an
sion are reported to be around 90% to 95% of the cases in the intrahepatic (hepaticogastrostomy) or extrahepatic choledocho
hands of experienced endoscopists (Kedia et al, 2013). Follow duodenostomy) approach, with studies demonstrating similar
ing initial failed ERCP, the recommended next step is to reat success rates (91% to 96%) between both routes (Artifon et al,
tempt the ERCP. If repeat ERCP is unsuccessful, alternative 2015) (Fig. 29.6A). With either approach, bile duct access is
methods of biliary decompression have traditionally included obtained under EUS and fluoroscopic guidance, followed by
percutaneous transhepatic drainage or surgery. With the dilation of the fistula and stent placement (Fig. 29.6B and D).
advancement of curvilinear-array echoendoscopes and periph The use of both plastic (straight or pigtail) and SEMSs have
eral devices, EUS-guided biliary drainage has become increas been reported, with overall success rates ranging from 72% to
ingly reported (Kahaleh et al, 2006). 100% (Tarantino et al, 2012). There are no comparative studies
between plastic and SEMSs, and the optimal stent has not been
Endoscopic Ultrasound Rendezvous Technique clearly established. In general, plastic stents are less expensive
EUS-guided wire placement into the bile duct in an antegrade but have a shorter patency rate compared with SEMSs. Among
fashion has been used to facilitate subsequent retrograde biliary the types of SEMSs, uncovered SEMSs may have a lower risk
cannulation when standard ERCP has failed. The point of of displacement compared with the partially or fully covered
biliary duct entry (intrahepatic vs. extrahepatic) depends on SEMSs but may be associated with a higher risk of bile leak and
accessibility and which route facilitates wire manipulation. peritonitis if they become displaced (Itoi et al, 2011). More
Whether it is through an intrahepatic or extrahepatic approach, recently, there have been reports on EUS-guided choledocho
a therapeutic linear echoendoscope is used to visualize the bile duodenostomy with a novel lumen-apposing fully covered metal
duct from the stomach or small intestine. Once an avascular stent (AXIOS stent; Xlumena, Mountain View, CA) for malig
plane has been identified using Doppler ultrasonography, a nant biliary obstruction (Glessing et al, 2015; Itoi et al, 2013).
EUS needle is advanced into the bile duct. Effort should be This novel fully covered, wide-lumen, low-profile stent allows
placed in directing the needle angle caudally in the direction of close apposition of the walls of the gastrointestinal lumen and
A B
C D
FIGURE 29.6. Endoscopic ultrasound (EUS)-guided biliary drainage. A, Cholangiogram showing complete distal bile duct obstruction with
diffuse upstream dilation. B, The common bile duct is accessed by using an EUS needle. C, A guidewire is left across the choledochoduodenostomy.
D, After dilating the tract, a stent is successfully deployed.
bile duct, theoretically minimizing the risk of migration or yield (Barkun et al, 2006). Once completed, the brush is
leakage. Further studies are needed to evaluate outcomes associ resheathed and the catheter withdrawn.
ated with these stents and their role in EUS-guided biliary The diagnostic yield of brush cytology for biliary strictures
drainage. is low, with most studies reporting poor sensitivity of 27% to
56% (Victor et al, 2012). Neither longer cytology brush designs
ENDOSCOPIC BILIARY TISSUE ACQUISITION nor prebrushing stricture dilation have conclusively shown
AND ADVANCED IMAGING TECHNIQUES improvement in sensitivity (de Bellis et al, 2003; Fogel et al,
2006). The poor sensitivity of brush cytology has often been
The role of ERCP as a diagnostic tool has diminished with the attributed to sampling error and low cellular yield due to
advent of multiple noninvasive imaging tests of the biliary the scirrous nature of cholangiocarcinoma, and due to the fact
system, including high-resolution CT and MRCP. One notable that pancreatic adenocarcinomas frequently cause only extrin
exception to this trend is the need for ERCP with tissue sam sic compression of the distal bile duct, rather than frank
pling in suspected malignant biliary obstruction. invasion.
Biliary Tissue Acquisition and Analysis Biliary Intraductal Biopsies

Intraductal Brushings Endobiliary forceps biopsy during ERCP is an alternate tech
Intraductal brushing during ERCP is the standard first-line nique routinely used for tissue sampling of biliary strictures. A
approach for tissue acquisition of biliary strictures, primarily variety of flexible forceps are available in adult (7 Fr) and pedi
due to wide availability and technical feasibility. The technique atric (5 to 6 Fr) calibers. Similar to other accessories, the
involves advancing the sheathed cytology brush over a guide forceps can be advanced into the bile duct. A prior sphincter
wire into the bile duct. The brush is advanced beyond the end otomy may facilitate this process, although it is not essential in
of the sheath and multiple to and fro movements are performed the presence of an indwelling guidewire in the biliary system
with the brush across the stricture. Sampling should involve at (Lin et al, 2003). Although previous studies have suggested
least five brush passes across the stricture to maximize cellular that tissue sampling with forceps provides the highest yield for
detection of malignancy (de Bellis et al, 2003), a recent meta- increased the sensitivity from 47% to 84% without a change in
analysis indicates that both brushings and biopsy are compa specificity.
rable (pooled sensitivity of 45% and 48%, respectively) Overall, ERCP with brush cytology and/or intraductal biop
(Navaneethan et al, 2015). ERCP with combined tissue pro sies should be performed in the initial evaluation of indetermi
curement with brush cytology and intraductal biopsies have nate biliary strictures. Advanced cytologic methods, such as
demonstrated a mild improvement, with sensitivity ranging DIA and FISH, can potentially improve sensitivity, especially
from 60% to 75% (Kitajima et al, 2007; Rösch et al, 2004) (see in the setting of negative cytology and histology. It should be
Chapter 30). stressed that a multidisciplinary review of the indication for
intrabiliary tissue sampling is of critical importance. Some stric
Molecular Analysis of Tissue Samples tures do not require biopsy if surgery is indicated on clinical
Chromosomal abnormalities are typically seen in malignant and radiologic grounds, and other biopsy approaches may be
biliary strictures (see Chapter 9C). Flow cytometry has been more appropriate.
commonly used to identify aneuploidy in tissue specimens.
Several studies have shown that flow cytometry for DNA Advanced Endoscopic Biliary Imaging
evaluation yields improved sensitivity (42%) but at the expense Peroral Cholangioscopy
of lower specificity of 70% to 77% (Ryan et al, 1994) for Peroral cholangioscopy is a technique that permits direct endo
the diagnosis of malignant biliary strictures. Furthermore, scopic visualization of the bile ducts by using miniature endo
flow cytometric analysis requires large cellular samples, which scopes and catheters inserted through the accessory port of a
can be challenging with current endoscopic tissue sampling duodenoscope. In the endoscope-based (“mother-daughter”)
techniques. system, a small, thin endoscope (daughter) is inserted through
New ancillary cytologic techniques, including digital image the accessory channel of the duodenoscope (mother). The main
analysis (DIA) and fluorescent in-situ hybridization (FISH) limitation of this system is the requirement of two separate
have been introduced to increase the diagnostic yield of routine endoscopists to operate each scope during the procedure.
cytology. DIA uses a computer assessment to quantify DNA Conversely, the catheter-based system requires only a single
content, chromatin distribution, and nuclear morphology to endoscopist. In this system, an independent instrument with a
assess for aneuploidy. In a single-center prospective study, DIA four-lumen catheter and an operator interface that allows four-
had a higher sensitivity (39%) compared with routine cytology way tip deflection is attached at the shaft of the duodenoscope.
(18%), albeit with a significantly lower specificity (77% vs. A 0.77-mm fiber optic probe can then be inserted through the
98%, respectively) (Baron et al, 2004). FISH is a cytogenetic flexible 10-Fr catheter, which is then advanced into the biliary
technique that uses fluorescent probes that selectively bind to tract through the accessory channel of the duodenoscope. A
specific portions of selected chromosomes, allowing assessment biliary sphincterotomy is often required to allow passage of the
of polysomy via fluoroscopic microscopy. A potential advantage catheter into the duct. In addition to the channel for the optical
of FISH is that it requires fewer cells for analysis compared probe, the catheter includes a 1.2-mm accessory channel and
with routine cytology or flow cytometry. Although the addition two 0.6-mm irrigation channels (Chen et al, 2007).
of FISH has been shown to improve sensitivity (14% to 24%) Peroral cholangioscopy has been primarily used for the man
when cytology is negative for malignancy, and abnormal FISH agement of refractory choledocholithiasis (discussed earlier
results are more likely in the presence of carcinoma, the overall in this chapter) and for the evaluation of indeterminate biliary
accuracy rates are still suboptimal (Fritcher et al, 2009; Levy strictures. Cholangioscopy allows direct visualization and
et al, 2008). More recently, Gonda and colleagues (2012) dem inspection of mucosal abnormalities of the biliary epithelium
onstrated that including the deletion of the 9p21 locus (p16) (Fig. 29.7). Findings of irregularly tortuous and dilated blood
in the diagnostic criteria of FISH for malignant biliary strictures vessels, intraductal lesions, ulcerations, or papillary mucosal
A B C
FIGURE 29.7. Evaluation of a dominant common bile duct stricture (CBD) in primary sclerosing cholangitis (PSC). A, Magnetic resonance
cholangiopancreatography revealing a long-segment distal CBD stricture (arrow) with upstream biliary dilation. B, Cholangiogram showing long CBD
stricture, upstream extrahepatic dilation, and intrahepatic ducts with beading consistent with PSC. C, Direct visualization of the bile duct with peroral
cholangioscopy reveals a smooth, scarred stricture.
projections on cholangioscopy are highly associated with under et al, 2001). Conversely, the restricted depth of penetration
lying malignancy (Kim et al, 2000). In a prospective multi with IDUS and the inability to perform fine needle aspiration
center study, cholangioscopy was able to distinguish malignant significantly limits its utility for assessing advanced tumor
from benign biliary lesions in 92% of cases by visualization extension and nodal and metastatic staging.
alone (Osanai et al, 2013). Furthermore, direct visualization IDUS is a promising advanced endoscopic imaging modality
during cholangioscopy also permits selective targeted tissue that permits high-resolution images of the bile system. This
sampling, with studies showing adequate biopsies in 72% to advantage is hindered by the limited depth of penetration
97% of cases (Chen et al, 2011; Draganov et al, 2011) and and ability to examine more distal sites. Further studies are
an accuracy for malignant lesions as high as 87% (Kalaitzakis needed to validate its place as an adjunct imaging tool to ERCP
et al, 2012). Peroral cholangioscopy has also been used for the and EUS.
diagnosis of malignancy in primary sclerosing cholangitis
(Tischendorf et al, 2006) and evaluation of biliary complica Confocal Laser Endomicroscopy
tions following liver transplantation (Balderramo et al, 2013). Confocal endomicroscopy (CLE) allows real-time high-
Overall, cholangioscopy represents an evolving novel technol resolution evaluation of gastrointestinal mucosal histology in
ogy for the evaluation of undifferentiated biliary strictures, vivo. Imaging is achieved by the projection of a low-power laser
exclusion of occult malignancy, and management of biliary light passed through a confocal aperture. The focused beam
stones. The introduction of a new high-definition cholangio targeted on a specific layer of tissue is then captured by a pho
scope may potentially facilitate improved bile duct examination todetection device and transformed into electrical signals pro
and define its adjunctive role to conventional ERCP within the cessed into grayscale images (Nakai et al, 2014). Because CLE
spectrum of biliary diseases (Parsi et al, 2014). relies upon tissue fluorescence, intravenous fluorescein dye (5
to 10 mL of 10% fluorescein) is administered to highlight tissue
Endoscopic Intraductal Ultrasound structures (individual cell structures, vasculature) prior to
The evolution of EUS has led to the development of small- imaging. The lack of contrast uptake by neoplastic tissue results
caliber ultrasound probes (2.9 mm or less in diameter) for in a contrasted dark appearance compared with adjacent normal
biliary endosonography. These IDUS miniprobes are intro structures.
duced through the accessory channel of the duodenoscope over There are two currently available CLE systems: endoscope-
a guidewire and advanced into the bile duct. IDUS operates at based CLE and probe-based CLE (pCLE). The former is too
higher frequencies (12 to 30 Mhz), with penetration of 2 cm large, as the CLE is integrated in the tip of the endoscope, and
at higher image resolution (0.07 to 0.18 mm) compared with thus biliary examination is generally performed with CLE
standard EUS (Gabbert et al, 2013). The bile duct appears as probes that can be inserted through the accessory channel of
three layers on IDUS. The innermost hyperechoic layer cor the duodenoscope. The laser (488 nm, blue light) is transmit
responds to the mucosa and bile interface. The middle ted through thousands of optimal fibers within the probe
hypoechoic layer corresponds to the discontinuous fibromus (cholangioflex-probe is 9 mm in diameter), and subsequent
cular layer, whereas the outermost hyperechoic layer represents confocal image data are collected at a frame rate of 12 frames/
the subserosal fat plane (Tantau et al, 2008). second with a limited field of vision of 325 µm. pCLE can be
IDUS has been used for the evaluation of suspected cho challenging, as optimal imaging requires significant probe and
ledocholithiasis. Several studies have reported increased patient stability.
sensitivity of IDUS (97% to 100%) over both ERCP and trans The indications for pCLE in biliary disease have not been
abdominal ultrasonography (81% and 45%) for the detection established, but studies have suggested a potential role in the
of bile duct stones in patients with normal cholangiography evaluation of biliary strictures suspicious for malignancy. In a
(Das et al, 2001; Ueno et al, 1997). However, with the high large prospective multicenter study of pCLE during ERCP for
diagnostic accuracy of EUS (sensitivity of 89% and specificity pancreatobiliary strictures, pCLE was associated with an accu
of 94%) for detecting choledocholithiasis in patients with sus racy rate of 81%, sensitivity of 98%, and specificity of 67% in
pected choledocholithiasis, the role of IDUS in the diagnostic the detection of malignancy. pCLE increased accuracy to 90%
algorithm of biliary stone disease remains to be determined compared with 73% by ERCP with tissue sampling alone
(Garrow et al, 2007; Petrov et al, 2009). (Meining et al, 2011). Despite these promising preliminary
Several studies have reported IDUS findings concerning for results, subsequent studies have suggested a wide discrepancy
biliary malignancy. These criteria include the presence of a in the interpretation of pCLE findings between endoscopists
hypoechoic mass, disruption of the normal bile duct structure, for both benign and malignant pancreaticobiliary lesions
abnormal stricture borders (asymmetry, notching), or the pres (Bakhru et al, 2013; Talreja et al, 2014). The inconsistent
ence of suspicious lymph nodes (>1 cm, hypoechoic, round interpretation among practitioners emphasizes the need for
with smooth borders). In a prospective study, ERCP/IDUS was both formal structured training in pCLE (Dunbar et al, 2007)
associated with a correct diagnosis for malignancy in 88% and validation of the currently established diagnostic criteria
patients compared with 76% with ERCP alone (Domagk et al, (Meining et al, 2012).
2004). Furthermore, IDUS has been shown to improve local
tumor staging for cholangiocarcinomas in various studies, even References are available at expertconsult.com.
compared with standard EUS (Menzel et al, 1999; Tamada,
Chapter 29 Interventional endoscopy: technical aspects 524.e1
REFERENCES Cheung J, et al: Guidewire versus conventional contrast cannulation of

the common bile duct for the prevention of post-ERCP pancreatitis:
Aadam AA, et al: Efficacy and safety of self-expandable metal stents a systematic review and meta-analysis, Gastrointest Endosc 70(6):
for biliary decompression in patients receiving neoadjuvant therapy 1211–1219, 2009.
for pancreatic cancer: a prospective study, Gastrointest Endosc Choi EK, et al: ERCP via gastrostomy vs. double balloon enteroscopy
76(1):67–75, 2012. in patients with prior bariatric Roux-en-Y gastric bypass surgery,
Anastassiades CP, et al: Cap-assisted ERCP with a forward-viewing Surg Endosc 27:2894–2899, 2013.
gastroscope as a rescue endoscopic intervention in patients with Choudhary A, et al: Effect of precut sphincterotomy on post-endoscopic
Bilroth II anatomy, Surg Endosc 27(6):2237, 2013. retrograde cholangiopancreatography pancreatitis: a systematic
Artifon EL, et al: Hepaticogastrostomy or choledochoduodenostomy review and meta-analysis, World J Gastroenterol 20(14):4093–4101,
for distal malignant biliary obstruction after failed ERCP: Is there 2014.
any difference?, Gastrointest Endosc 81(4):950–959, 2015. Costamagna G, Boskoski I: Current treatment of benign biliary stric
Arya N, et al: Electrohydraulic lithotripsy in 111 patients: a safe and tures, Ann Gastroenterol 26(1):37–40, 2013.
effective therapy for difficult bile duct stones, Am J Gastroenterol Costamagna G, et al: Long-term results of endoscopic management of
99:2330–2334, 2004. postoperative bile duct strictures with increasing numbers of stents,
ASGE Technology Assessment Committee, et al: Pancreatic and Gastrointest Endosc 54(2):162–168, 2001.
biliary stents, Gastrointest Endosc 77(3):319–327, 2013. Cotton PB: Non-operative removal of bile duct stones by duodeno
Bailey AA, et al: Needle-knife sphincterotomy: factors predicting its use scopic sphincterotomy, Br J Surg 6791:1–5, 1980.
and the relation with post-ERCP pancreatitis (with video), Gastro- Das A, et al: Wire-guided intraductal US: an adjunct to ERCP in the
intest Endosc 71(2):266–271, 2010. management of bile duct stones, Gastrointest Endosc 54(1):31–36,
Baillie J: Endoscopic ampullectomy, Am J Gastroenterol 100:2379– 2001.
2381, 2005. de Bellis M, et al: Tissue sampling at ERCP in suspected malignant
Bakhru MR, et al: Interobserver agreement for confocal imaging of biliary strictures (part 1), Gastrointest Endosc 56:552–561, 2002.
ampullary lesions: a multicenter single-blinded study, J Clin Gastro- de Bellis M, et al: Tissue sampling at ERCP in suspected malignant
enterol 47(5):440–442, 2013. biliary strictures (part 2), Gastrointest Endosc 56:720–730, 2002.
Balderramo D, et al: Prospective evaluation of single-operator peroral de Bellis M, et al: Influence of stricture dilation and repeat brushing
cholangioscopy in liver transplant recpients requiring an evaluation on the cancer detection rate of brush cytology in the evaluation
of the biliary tract, Liver Transpl 19(2):199–206, 2013. of malignant biliary obstruction, Gastrointest Endosc 58:176–182,
Barkun A, et al: Update on endoscopic tissue sampling devices, Gas- 2003.
trointest Endosc 63:71–77, 2006. De Palma GD, et al: Unilateral versus bilateral endoscopic hepatic duct
Baron TH, Harewood GC: Endoscopic balloon dilation of the biliary drainage in patients with malignant hilar biliary obstruction: results
sphincter compared to endoscopic biliary sphincterotomy for removal of a prospective, randomized, and controlled study, Gastrointest
of common bile duct stones during ERCP: a metaanalysis of ran Endosc 53(6):547–553, 2001.
domized controlled trials, Am J Gastroenterol 99:1455–1460, 2004. De Palma GD, et al: Endoscopic papillectomy: Indications, tech
Baron TH, et al: A prospective comparison of digital image analysis niques, and results, World J Gastroenterol 20(6):1537–1543, 2014.
and routine cytology for the identification of malignancy in biliary DiSario JA, et al: Endoscopic balloon dilation compared with sphic
tract strictures, Clin Gastroenterol Hepatol 2(3):214–219, 2004. nterotomy for extraction of bile duct stones, Gastroenterology 127:
Bergman JJ, et al: Randomised trial of endoscopic balloon dilation 1291–1299, 2004.
versus endoscopic sphincterotomy for removal of bile duct stones, DiSario JA, et al: Biliary and pancreatic lithotripsy devices, Gastrointest
Lancet 349:1114–1115, 1997. Endosc 65(6):750–756, 2007.
Bergmann JJ, et al: Long-term follow-up after biliary stent placement Domagk D, et al: Endoscopic retrograde cholangiopancreatography in
for postoperative bile duct stenosis, Gastrointest Endosc 54(2):154– bile duct strictures: a prospective comparison of imaging diagnostics
161, 2001. with histopathological correlation, Am J Gastroenterol 99(9):1684–
Binmoeller KF, Schafer TW: Endoscopic management of bile duct 1689, 2004.
stones, J Clin Gastroenterol 32(2):106–118, 2001. Donelli G, et al: Plastic biliary stent occlusion: factors involved and
Bove B, et al: ERCP in patients with Billroth II gastrectomy: report of possible preventive approaches, Clin Med Res 5:53–60, 2007.
a 30-year experience, Dig Liver Dis 2013;45:S58, 2013. Dowsett JF, et al: Endoscopic biliary therapy using the combined per
Cahen D, et al: Long-term results of endoscopic drainage of common cutaneous and endoscopic technique, Gastroenterology 96:1180–
bile duct strictures in chronic pancreatitis, Eur J Gastroenterol Hepatol 1186, 1989.
52:55–63, 2005. Draganov P, et al: Long-term outcome in patients with benign biliary
Catalano MF, et al: Endoscopic transpancreatic papillary septotomy strictures treated endoscopically with multiple stents, Gastrointest
for inaccessible obstructed bile ducts: comparison with standard Endosc 55(6):680–686, 2002.
pre-cut papillotomy, Gastrointest Endosc 60(4):557–561, 2004. Draganov PV, et al: Prospective evaluation of the clinical utility of
Cavell LK, et al: Biliary self-expandable metal stents do not adversely ERCP-guided cholangiopancreatoscopy with a new direct visualiza
affect pancreaticoduodenectomy, Am J Gastroenterol 108(7):1168– tion system, Gastrointest Endosc 73(5):971–979, 2011.
1173, 2013. Dunbar KB, et al: Confocal laser endomicroscopy image interpreta
Cennamo V, et al: Can early precut implementation reduce endoscopic tion: interobserver agreement among gastroenterologists and pathol
retrograde cholangiopancreatography-related complication risk? ogists, Gastrointest Endosc 65:AB348, 2007.
Meta-analysis of randomized controlled trials, Endoscopy 42(5):381– El Hajj I, Cote GA: Endoscopic diagnosis and management of ampul
388, 2010. lary lesions, Gastrointest Endoscopy Clin N Am 23:95–109, 2013.
Chang WH, et al: Outcome in patients with bifurcation tumors who Elek G, et al: Histological evaluation of preoperative biopsies from
undergo unilateral versus bilateral hepatic duct drainage, Gastrointest ampulla vateri, Pathol Oncol Res 9:32–41, 2003.
Endosc 47:354–362, 1998. Elton E, et al: Diagnostic and therapeutic ERCP using an enteroscope
Chang WH, et al: Outcome of simple use of mechanical lithotripsy of and a pediatric colonoscope in long-limb surgical bypass patients,
difficult common bile duct stones, World J Gastroenterol 11:593–596, Gastrointest Endosc 47:62–67, 1998.
2005. Faylona JM, et al: Small-bowel perforations related to endoscopic ret
Chen YK, et al: SpyGlass single-operator peroral cholangiopancreatos rograde cholangiopancreatography (ERCP) in patients with Billroth
copy system for the diagnosis and therapy of bile-duct disorders: a II gastrectomy, Endoscopy 31:546–549, 1999.
clinical feasibility study (with video), Gastrointest Endosc 65(6):832– Figueroa-Barojas P, et al: Safety and efficacy of radiofrequency ablation
841, 2007. in the management of unresectable bile duct and pancreatic cancer:
Chen CH, et al: Reappraisal of endosonography of ampullary tumors: a novel palliation technique, J Oncol 2013:1–5, 2013.
correlation with transabdominal sonography, CT, and MRI, J Clin Fogel EL, et al: Effectiveness of a new long cytology brush in the evalu
Ultrasound 37:18–25, 2009. ation of malignant biliary obstruction: a prospective study, Gastroin-
Chen YK, et al: Single-operator cholangioscopy in patients requiring test Endosc 63:71–77, 2006.
evaluation of bile duct disease or therapy of biliary stones (with Freeman ML, et al: Risk factors for post-ERCP pancreatitis: a prospec
videos), Gastrointest Endosc 74(4):805–814, 2011. tive, multicenter study, Gastrointest Endosc 54:425–434, 2001.
524.e2 PART 4 TECHNIQUES OF BILIARY TRACT INTERVENTION: RADIOLOGIC, ENDOSCOPIC, AND SURGICAL
Fritcher EG, et al: A multivariable model using advanced cytologic Kaffes AJ, et al: Impact of endoscopic intervention in 100 patients with
methods for the evaluation of indeterminate pancreatobiliary stric suspected postcholecystectomy bile leak, Gastrointest Endosc 61:269–
tures, Gastroenterology 136:2180–2186, 2009. 275, 2005.
Fumex F, et al: Similar performance but higher cholecystitis rate with Kaffes A, et al: Metal stent versus plastic stent for post liver transplant
covered biliary stents: results from a prospective multicenter evalu anastomotic biliary strictures [abstract], Endoscopy 44(Suppl):A97,
ation, Endoscopy 38(8):787–792, 2006. 2012.
Gabbert C, et al: Advanced techniques for endoscopic biliary imaging: Kahaleh M, et al: Interventional EUS-guided cholangiography:
cholangioscopy, endoscopic ultrasonography, confocal, and beyond, evaluation of a technique in evolution, Gastrointest Endosc 64:52–59,
Gastrointest Endosc Clin N Am 23(3):625–646, 2013. 2006.
Garcia-Cano J: A simple technique to aid intubation of the duodeno Kahaleh M, et al: Temporary placement of covered self-expandable
scope in the afferent limb of Billroth II gastrectomies for endoscopic metallic stents in patients with biliary leak: midterm evaluation of a
retrograde cholangiopancreatography, Endoscopy 40:21–22, 2008. pilot study, Gastrointest Endosc 66:52–59, 2007.
Garg PK, et al: Predictors of unsuccessful mechanical lithotripsy and Kahaleh M, et al: Unresectable cholangiocarcinoma: comparison of
endoscopic clearance of large bile duct stones, Gastrointest Endosc survival in biliary stenting alone versus stenting with photodynamic
59:601–605, 2004. therapy, Clin Gastroenterol Hepatol 6(3):290–297, 2008.
Garrow D, et al: Endoscopic ultrasound: a meta-analysis of test per Kalaitzakis E, et al: Diagnostic and therapeutic utility of single-
formance in suspected biliary obstruction, Clin Gastroenterol Hepatol operator peroral cholangioscopy for indeterminate biliary lesions and
5(5):616–623, 2007. bile duct stones, Eur J Gastroenterol Hepatol 24(6):656–664, 2012.
Glessing BR, et al: EUS-guided choledochoduodenostomy with a Kapetanos D, et al: Case series of transpancreatic septotomy as precut
lumen-apposing metal stent before duodenal stent placement for ting technique for difficult bile duct cannulation, Endoscopy 39(9):
malignant biliary and duodenal obstruction, Gastrointest Endosc 802–806, 2007.
81(4):1019–1020, 2015. Kato H, et al: Short double-balloon enteroscopy is feasible and effective
Goff JS, et al: Common bile duct pre-cut sphincterotomy: transpancre for endoscopic retrograde cholangiopancreatography in patients
atic sphincter approach, Gastrointest Endosc 41(5):502–505, 1995. with surgically altered gastrointestinal anatomy, Dig Endosc 26
Gonda TA, et al: Polysomy and p16 deletion by fluorescence in situ (Suppl 2):130–135, 2014.
hybridization in the diagnosis of indeterminate biliary strictures, Katsinelos P, et al: A comparative study of 10-Fr vs. 7-Fr straight
Gastrointest Endosc 75:74–79, 2012. plastic stents in the treatment of postcholecystectomy bile leak, Surg
Gupta K, et al: Endoscopic ultrasound-assisted bile duct access and Endosc 22(1):101–106, 2008.
drainage: multicenter, long-term analysis of approach, outcomes, Kedia P, et al: Endoscopic-guided biliary drainage: how can we achieve
and complications of a technique in evolution, J Clin Gastroenterol efficient biliary drainage?, Clin Endosc 46:543–551, 2013.
48(1):80–87, 2014. Kedia P, et al: Internal EUS-directed transgastric ERCP (EDGE):
Gutierrez JM, et al: Surgical gastrostomy for pancreatobiliary and duo game over, Gastroenterology 147:566–568, 2014.
denal access following Roux-en-Y gastric bypass, J Gastrointest Surg Ki HS, et al: Feasibility of cap-assisted endoscopic retrograde cholan
12(12):2170–2175, 2009. giopancreatography in patients with altered gastrointestinal anatomy,
Han J, Kim MH: Endoscopic papillectomy for adenomas of the Gut Liver 9(1):109–112, 2015.
major duodenal papilla (with video), Gastrointest Endosc 63:292–301, Kim HJ, et al: Tumor vessel: a valuable cholangioscopic clue of malig
2006. nant biliary stricture, Gastrointest Endosc 52(5):630–634, 2000.
Harewood GC, et al: Prospective, randomized, controlled trial of Kim KH, et al: Endoscopic management of bile leakage after cholecys
prophylactic pancreatic stent placement for endoscopic snare tectomy: a single-center experience for 12 years, Clin Endosc 47(3):
excision of the duodenal ampulla, Gastrointest Endosc 62:367–370, 248–253, 2014.
2005. Kitajima Y, et al: Usefulness of transpapillary bile duct brushing cytol
Herreros de Tejada A, et al: Double-guidewire technique for difficult ogy and forceps biopsy for improved diagnosis in patients with biliary
bile duct cannulation: a multicenter randomized, controlled trial, strictures, J Gastroenterol Hepatol 22:1615–1620, 2007.
Gastrointest Endosc 70(4):700–709, 2009. Larghi A, et al: Management of hilar biliary strictures, Am J Gastroen-
Hirota WK, et al: Standards of Practice Committee, American Society terol 103(2):458–473, 2008.
for Gastrointestinal Endoscopy (2006) ASGE guideline: the role of Law R, et al: Single-session ERCP in patients with previous Roux-en-Y
endoscopy in the surveillance of premalignant conditions of the gastric bypass using percutaneous-assisted transprosthetic endo
upper GI tract, Gastrointest Endosc 63:570–580, 2006. scopic therapy: a case series, Endoscopy 45(08):671–675, 2013.
Irani S, et al: Papillectomy for ampullary neoplasm: results of a single Lee A, Shah JN: Endoscopic approach to the bile duct in the patient
referral center over a 10-year period, Gastrointest Endosc 70:923–932, with surgically altered anatomy, Gastrointest Endosc Clin N Am 23(2):
2009. 483–504, 2013.
Ito K, et al: Can pancreatic duct stenting prevent post-ERCP pancre Lee TH, et al: Can wire-guided cannulation prevent post-ERCP pan
atitis in patients who undergo pancreatic duct guidewire placement creatitis? A prospective randomized trial, Gastrointest Endosc 69:444–
for achieving selective biliary cannulation? A prospective randomized 449, 2009.
controlled trial, J Gastroenterol 45(11):1183–1191, 2010. Lee TH, et al: Prophylactic temporary 3F pancreatic duct stent to
Itoi T, et al: Stent selection and tips on placement technique of EUS- prevent post-ERCP pancreatitis in patients with a difficult biliary
guided biliary drainage: transduodenal and transgastric stenting, cannulation: a multicenter, prospective, randomized study, Gastro-
J Hepatobiliary Pancreat Sci 18(5):664–672, 2011. intest Endosc 76(3):578–585, 2012.
Itoi T, et al: Endoscopic ultrasonography-guided cholecystogastros Lee JH, et al: Comparison of the utility of covered metal stents
tomy using a lumen-apposing metal stent as an alternative to extra versus uncovered metal stents in the management of malignant
hepatic bile duct drainage in pancreatic cancer with duodenal biliary strictures in 749 patients, Gastrointest Endosc 78(2):312–324,
invasion, Dig Endosc 25(Suppl 2):137–141, 2013. 2013.
Iwano H, et al: Unilateral versus bilateral drainage using self-expandable Levy MJ, et al: Prospective evaluation of advanced molecular markers
metallic stent for unresectable hilar biliary obstruction, Dig Endosc and imaging techniques in patients with indeterminate bile duct
23:43–48, 2011. strictures, Am J Gastroenterol 103:1263–1273, 2008.
Iwashita T, et al: Endoscopic ultrasound-guided biliary drainage: a Liberato MJ, Canena JM: Endoscopic stenting for hilar cholangiocar
review, Clin J Gastroenterol 7:94–102, 2014. cinoma: efficacy of unilateral and bilateral placement of plastic and
Jakobs R, et al: Fluoroscopically guided laser lithotripsy versus extra metal stents in a retrospective review of 480 patients, BMC Gastro-
corporeal shock wave lithotripsy for retained bile duct stones: a enterol 12:103, 2012.
prospective randomized study, Gut 40:678–682, 1997. Lin LF, et al: Guidewire technique for endoscopic transpapillary pro
Kaassis M, et al: Plastic or metal stents for malignant stricture of the curement of bile duct biopsy specimens without endoscopic sphinc
common bile duct? Results of a randomized prospective study, Gas- terotomy, Gastrointest Endosc 58(2):272–274, 2003.
trointest Endosc 57:178–182, 2003. Madhoun MF, et al: Endoscopic papillary large balloon dilation
Kaffes AJ, Liu K: Fully covered self-expandable metal stents for treat reduces the need for mechanical lithotripsy in patients with large bile
ment of benign biliary strictures, Gastrointest Endosc 78(1):13–21, duct stones: a systematic review and meta-analysis, Diagn Ther
2013. Endosc 2014:309618, 2014.
Chapter 29 Interventional endoscopy: technical aspects 524.e3
Maeda S, et al: Prospective randomized pilot trial of selective biliary Piraka C, et al: Transpapillary cholangioscopy directed lithotripsy in
cannulation using pancreatic guidewire placement, Endoscopy 35(9): patients with difficult bile duct stones, Clin Gastroenterol Hepatol
721–724, 2003. 5:1333–1338, 2007.
Manes G, et al: An analysis of the factors associated with the develop Poley J, et al: A prospective group sequential study evaluating a new
ment of complications in patients undergoing precut sphincterot type of fully covered self-expandable metal stent for the treatment
omy: a prospective, controlled, randomized, multicenter study, Am of benign biliary strictures, Gastrointest Endosc 75:783–789, 2012.
J Gastroenterol 104(10):2412–2417, 2009. Prasad GA, et al: Factors predicting survival in patients with cholan
Manta R, et al: Linear endoscopic ultrasonography vs magnetic reso giocarcinoma treated with photodynamic therapy, Clin Gastroenterol
nance imaging in ampullary tumors, World J Gastroenterol 16:5592– Hepatol 5:743–748, 2007.
5597, 2010. Rösch T, et al: ERCP or EUS for tissue diagnosis of biliary strictures?
Maluf-Filho F, et al: Rotatable sphincterotome facilitates bile duct A prospective comparative study, Gastrointest Endosc 60(3):390–396,
cannulation in patients with altered ampullary anatomy, Gastroenterol 2004.
Hepatol 4(1):59–62, 2008. Ryan ME, et al: Comparison of flow cytometry for DNA content and
Martin JA, Haber GB: Ampullary adenoma: clinical manifestations, brush cytology for detection of malignancy in pancreaticobiliary
diagnosis, and treatment, Gastrointest Endosc Clin N Am 13:649–669, strictures, Gastrointest Endosc 40:133–139, 1994.
2003. Ryan ME, et al: Endoscopic intervention for biliary leaks after laparo
Masci E, et al: Complication of diagnostic and therapeutic ERCP: a scopic cholecystectomy: a multicenter review, Gastrointest Endosc
prospective multicenter study, Am J Gastroenterol 96:417–423, 2001. 47:261–266, 1998.
Maydeo A, Bhandari S: Balloon sphincteroplasty for removing difficult Saleem A, et al: Meta-analysis of randomized trails comparing the
bile duct stones, Endoscopy 39(11):958–961, 2007. patency of covered and uncovered self-expandable metal stents for
Meining A, et al: Direct visualization of indeterminate pancreaticobili palliation of distal malignant bile duct obstruction, Gastrointest
ary strictures with probe-based confocal laser endomicroscopy: a Endosc 74:321–327, 2011.
multicenter experience, Gastrointest Endosc 74:961–968, 2011. Sandha GS, et al: Endoscopic therapy for bile leak based on a new
Meining A, et al: Classification of probe-based confocal laser endomi classification: results in 207 patients, Gastrointest Endosc 60:567–574,
croscopy findings in pancreaticobiliary strictures, Endoscopy 44(3): 2004.
251–257, 2012. Schreiner MA, et al: Laparoscopy-assisted versus balloon enteroscopy-
Menezes N, et al: Prospective analysis of a scoring system to predict assisted ERCP in bariatric post-Roux-en-Y gastric bypass patients,
choledocholithiasis, Br J Surg 87:1176–1181, 2000. Gastrointest Endosc 75(4):748–756, 2012.
Menzel J, et al: Intraductal ultrasonography (IDUS) of the pancreato- Seifert E, et al: Adenoma of the ampulla of Vater: putative precancer
biliary duct system. Personal experience and review of the literature, ous lesion, Gut 32:1558–1561, 1991.
Eur J Ultrasound 10(2–3):105–115, 1999. Shaffer EA: Gallstone disease: Epidemiology of gallbladder stone
Moss AC, et al: Malignant distal biliary obstruction: a systematic disease, Best Pract Res Clin Gastroenterol 20(6):981–996, 2006.
review and meta-analysis of endoscopic and surgical bypass results, Shah RJ, et al: A multicenter, U.S. experience of single-balloon,
Cancer Treat Rev 33(2):213–221, 2007. double-balloon, and rotational overtube-assisted enteroscopy ERCP
Mukai I, et al: Metallic stents are more efficacious than plastic stents in patients with surgically altered pancreaticobiliary anatomy (with
in unresectable malignant hilar biliary strictures: a randomized con video), Gastrointest Endosc 77:593–600, 2013.
trolled trial, J Hepatobiliary Pancreat Sci 20:214–222, 2013. Sharaiha RZ, et al: Comparison of metal stenting with radiofrequency
Muratori R, et al: ESWL for difficult bile duct stones: a 15-year single ablation versus stenting alone for treating malignant biliary stric
centre experience, World J Gastroenterol 16:4159–4163, 2010. tures: is there an added benefit?, Dig Dis Sci 50:3099–3102, 2014.
Naitoh I, et al: Unilateral versus bilateral endoscopic metal stenting for Siddiqui AA, et al: Self-expanding metal stents (SEMS) for preopera
malignant hilar biliary obstruction, J Gastroenterol Hepatol 24:552– tive biliary decompression in patients with resectable and borderline-
557, 2009. resectable pancreatic cancer: outcomes in 241 patients, Dig Dis Sci
Nakai Y, et al: Confocal laser endomicroscopy in gastrointestinal and 58(6):1744–1750, 2013.
pancreatobiliary diseases, Dig Endosc 26(1):86–94, 2014. Singh P, et al: Does prophylactic pancreatic stent placement reduce the
Natsui M, et al: Long-term outcomes of endoscopic papillary balloon risk of post-ERCP acute pancreatitis? A meta-analysis of controlled
dilation and endoscopic sphincterotomy for bile duct stones, Dig trials, Gastrointest Endosc 60(4):544–550, 2004.
Endosc 25(3):313–321, 2013. Skinner M, et al: ERCP with the overtube-assisted enteroscopy tech
Navaneethan U, et al: Early precut sphincterotomy and the risk of nique: a systematic review, Endoscopy 46(7):560–572, 2014.
endoscopic retrograde cholangiopancreatography related complica Standards of Practice Committee, et al: The role of endoscopy in
tions: an updated meta-analysis, World J Gastrointest Endosc ampullary and duodenal adenomas, Gastrointest Endosc 64(6):849–
6(5):200–208, 2014. 854, 2006.
Navaneethan U, et al: Precut sphincterotomy: efficacy for ductal access Stefanidis G, et al: Large balloon dilation vs. mechanical lithotripsy for
and the risk of adverse events, Gastrointest Endosc 81(4):924–931, the management of large bile duct stones: a prospective randomized
2015. study, Am J Gastroenterol 106(2):278–285, 2011.
Navaneethan U, et al: Comparative effectiveness of biliary brush cytol Swahn F, et al: Ten years of Swedish experience with intraductal elec
ogy and intraductal biopsy for detection of malignant biliary stric trohydraulic lithotripsy and laser lithotripsy for the treatment of
tures: a systematic review and meta-analysis, Gastrointest Endosc difficult bile duct stones: an effective and safe option for octogene
81(1):168–176, 2015. rians, Surg Endosc 24:1011–1016, 2010.
Nehaus H, et al: Randomized study of intracorporeal laser lithotripsy Takashima M, et al: Carcinoma of the ampulla of Vater associated with
versus extracorporeal shock-wave lithotripsy for difficult bile duct or without adenoma: a clinicopathology analysis of 198 cases with
stones, Gastrointes Endosc 47:327–334, 1998. reference to p53 and Ki-67 immunohistochemical expressions, Mod
Ortner ME, et al: Successful photodynamic therapy for nonresectable Pathol 13:1300–1307, 2000.
cholangiocarcinoma: a randomized prospective study, Gastroenterol- Talreja JP, et al: Pre- and post-training session evaluation for interob
ogy 125(5):1355–1363, 2003. server agreement and diagnostic accurarcy of probe-based confocal
Osanai M, et al: Peroral video cholangioscopy to evaluate indetermi laser endomicroscopy for biliary strictures, Dig Endosc 26(4):577–
nate bile duct lesions and preoperative mucosal cancerous extension: 580, 2014.
a prospective multicenter study, Endoscopy 45(8):635–642, 2013. Tamada K, et al: Transpapillary intraductal US prior to biliary drainage
Park do H, et al: Covered versus uncovered wallstent for malignant in the assessment of longitudinal spread of extrahepatic bile duct
extrahepatic biliary obstruction: a cohort comparative analysis, Clin carcinoma, Gastrointest Endosc 53(3):300–307, 2001.
Gastroenterol Hepatol 4(6):790–796, 2006. Tandan M, et al: Extracorporeal shock wave lithotripsy of large difficult
Parsi MA, et al: Diagnostic and therapeutic cholangiopancreatoscopy: common bile duct stones: efficacy and analysis of factors that
performance of a new digital cholangioscope, Gastrointest Endosc favor stone fragmentation, J Gastroenterol Hepatol 8:1370–1374,
79(6):936–942, 2014. 2009.
Petrov MS, et al: Systematic review of endoscopic ultrasonography Tantau M, et al: Intraductal ultrasonography for the assessment of
versus endoscopic retrograde cholangiopancreatography for sus preoperative biliary and pancreatic strictures, J Gastrointest Liver Dis
pected choledocholithiasis, Br J Surg 96(9):967–974, 2009. 17(2):217–222, 2008.
524.e4 PART 4 TECHNIQUES OF BILIARY TRACT INTERVENTION: RADIOLOGIC, ENDOSCOPIC, AND SURGICAL
Tarantino I, et al: Endoscopic ultrasound guided biliary drainage, Van Boeckel P, et al: Plastic or metal stents for benign extrahepatic
World J Gastrointest Endosc 4(7):306–311, 2012. biliary strictures: a systematic review, BMC Gastroenterol 9:96, 2009.
Tarantino I, et al: Fully covered self-expandable metallic stents in Van der Gaag NA, et al: Preoperative biliary drainage for cancer of the
benign biliary strictures: a multicenter study on efficacy and safety, head of the pancreas, N Engl J Med 362(2):129–137, 2010.
Endoscopy 44:923–927, 2012. Victor DW, et al: Current endoscopic approach to indeterminate biliary
Tekola B, et al: Percutaneous gastrostomy tube placement to perform strictures, World J Gastroenterol 18(43):6197–6205, 2012.
transgastrostomy endoscopic retrograde cholangiopancreatography Vienne A, et al: Prediction of drainage effectiveness during endoscopic
in patients with Roux-en-Y anatomy, Dig Dis Sci 56(11):3364–3369, stenting of malignant hilar strictures: the role of liver volume assess
2011. ment, Gastrointest Endosc 72(4):728–735, 2010.
Teoh AY, et al: Randomized trial of endoscopic sphincterotomy with Webb K, Saunders M: Endoscopic management of malignant bile duct
balloon dilation versus endoscopic sphincterotomy alone for removal strictures, Gastrointest Endoscopy Clin N Am 23:313–331, 2013.
of bile duct stones, Gastroenterology 144(2):341–345, 2013. Weinberg BM, et al: Endoscopic balloon sphincter dilation (sphinc
Tischendorf JJ, et al: Cholangioscopic characterization of dominant teroplasty) versus sphincterotomy for common bile duct stones,
bile duct stenosis in patients with primary sclerosing cholangitis, Cochrane Database Syst Rev (4):CD004890, 2006.
Endoscopy 38(7):665–669, 2006. Yamao T, et al: Endoscopic snare papillectomy with biliary and pan
Tse F, et al: Guide wire-assisted cannulation for the prevention of creatic stent placement for tumors of the major duodenal papilla,
post-ERCP pancreatitis: a systematic review and meta-analysis, Surg Endosc 24:119–124, 2010.
Endoscopy 45(8):605–618, 2013. Yoon WJ, et al: A comparison of covered and uncovered Wall-stents
Uchiyama Y, et al: New approach to diagnosing ampullary tumors by for the management of distal malignant biliary obstruction, Gastro-
magnifying endoscopy combined with a narrow-band imaging intest Endosc 63(7):996–1000, 2006.
system, J Gastroenterol 41(5):483–490, 2006.
Ueno A, et al: Diagnosing extrahepatic bile duct stones using intra
ductal ultrasonography: a case series, Endoscopy 29(5):356–360,
1997.
CHAPTER 30
Radiologic hepatobiliary interventions
Karen T. Brown and Anne M. Covey
RADIOLOGIC HEPATOBILIARY INTERVENTIONS In patients with hepatic encephalopathy and portal hyper-
tension, or patients with “left-sided portal hypertension”
Minimally invasive hepatic intervention is indicated in a wide (gastric varices due to splenic vein occlusion), balloon-occluded
range of pathologic conditions and can be generally divided into retrograde transvenous obliteration (BRTO) or balloon-
vascular, biliary, and hepatic parenchymal procedures. The occluded antegrade transvenous obliteration (BATO) may be
objective of this chapter is to provide a broad overview of the preferable to TIPSS. BRTO and BATO refer to procedures in
spectrum of interventions that can be performed in a percuta- which a high-risk or bleeding gastric varix is catheterized
neous fashion by using imaging guidance. Greater detail will be and sclerosed typically with a mixture of Ethiodol, Sotradechol,
found in individual chapters devoted to each topic. and air agitated through a three-way stopcock (Saad et al,
2012).
Vascular Procedures In some cases, portal vein narrowing or occlusion due to
The liver is an extremely vascular organ with nutrient supply extrinsic compression by tumor may cause symptoms similar
from both the portal vein and hepatic artery and drainage via to those seen in cirrhotic portal hypertension. In such cases,
the hepatic veins. These vessels are common targets for the placement of a self-expanding stent can relieve varices and
interventional radiologist. ascites. This is most commonly seen in patients with locally
advanced pancreaticobiliary cancer where portal vein stenting
Portal Vein may also improve thrombocytopenia, broadening chemother-
The majority of the nutrient blood flow to the liver is via the apy options (Fig. 30.1).
portal vein, which drains the splanchnic circulation and spleen. Another procedure that has become relatively common is
The most common abnormality involving the portal vein is portal vein embolization (PVE), as an adjunctive procedure
portal hypertension, typically as a sequela of cirrhosis (see prior to hepatic resection. In patients with a suboptimal future
Chapters 76, 81, 83, and 87). Clinical manifestations of portal liver remnant (FLR) volume, which can be assessed volumetri-
hypertension include splenomegaly, thrombocytopenia, varices, cally (i.e., with computed tomography [CT] or magnetic reso-
ascites, and liver failure. In 1969, Rosch and colleagues reported nance imaging [MRI]) or functionally (e.g., indocyanine green
the first case of transjugular intrahepatic portosystemic shunt clearance), contralateral PVE can be performed to induce pre-
(TIPSS) in dogs. Thirteen year later, Colapinto and colleagues operative hypertrophy of the FLR (May et al, 2013) (see
(1982) reported the first human application of TIPSS. In this Chapter 108C). In patients with cirrhosis, most surgeons
procedure, a path is created from the hepatic vein to the portal believe that a FLR of greater than 40% of the total liver volume
vein through the liver parenchyma, thereby decreasing portal (TLV) is optimal. In patients without underlying liver disease,
pressure and relieving patients from intractable ascites or acute a FLR of greater than 25% may be acceptable. Other risk
variceal bleeding. Initially the tract was formed with serial dila- factors for impaired liver function include diabetes, prior che-
tors or balloon dilation with limited success. When the Palmaz motherapy, and steatosis, and so the desired volume of the FLR
metallic balloon expandable stents became available in the mid- is estimated on a case-by-case basis.
1980s (Palmaz et al, 1986), procedural success improved, and PVE to improve the safety of hepatic resection was first
the technique gained widespread acceptance. Further refine- proposed by Makuuchi and colleagues in 1989. Initially, this
ment to the use of covered self-expanding stents has improved procedure was performed via a transileocolic approach that
long-term patency, making this a viable option for not only required laparotomy and general anesthesia. Although ligation
patients with life-threatening hemorrhage but also as a means of a portal vein branch can be carried out during a laparotomy,
to control intractable ascites (Boyer & Haskal, 2005; Saad, today this procedure is most commonly performed percutane-
2014). ously, typically as an outpatient. Access to the ipsilateral portal
The most significant complication of TIPSS is hepatic vein is made with a 21-gauge needle, and typical angiographic
encephalopathy due to the volume of blood shunted past the catheters are used to perform venography and embolization. A
liver parenchyma. As a result, the presence of hepatic encepha- wide range of agents have been used to perform the procedure,
lopathy is a relative contraindication to the procedure. Other including ethanol, Gelfoam, thrombin, polyvinyl alcohol, glue,
contraindications include right heart failure, hepatic vein occlu- spheric embolic agents, coils, and sclerosing agents. No agent
sion, and sepsis. has proven superior; each is expected to increase the absolute
525
A B
C D
FIGURE 30.1 A, Transhepatic portal venogram in a patient with cholangiocarcinoma shows irregular narrowing of the portal vein. Varices are seen
as a manifestation of portal hypertension. B, After placement of a self-expanding metallic stent across the stenosis, there is good flow across the
stent. The varices are no longer evident. Coronal computed tomography image before (C) and after (D) portal vein stent shows decrease in spleno-
megaly. The platelet count rose from 70 to 180 K/mcL within days of stent placement.
FLR/TLV in the range of 8% to 10%. Complications are Subsequently, transarterial treatments have been applied to a
uncommon, the most significant being nontarget embolization wide variety of hypervascular tumors, including sarcoma
to the main portal vein or portal vein supplying the FLR, which and breast cancers, as well some tumors that are not particu-
could preclude operation. This occurs in less than 1% of larly hypervascular by imaging, such as colon cancer or
patients (van Lienden et al, 2013). cholangiocarcinoma.
Different forms of treatment have been administered via
Hepatic Artery the hepatic artery to treat such tumors, including chemother-
Unlike portal vein interventions, which are most commonly apy infusion, bland (particle) embolization (transcatheter arte-
undertaken to treat the sequela of portal hypertension or, in the rial embolization [TAE]), transarterial chemoembolization
case of PVE, as adjunct to hepatic resection, most transarterial (TACE), embolization with drug-eluting beads (DEB-TACE),
interventions in the liver are done to effect treatment of unre- and radioembolization (RAE) (see Chapter 96). Two random-
sectable malignancy or for control of bleeding in the setting of ized trials have demonstrated improvement in overall survival
trauma (see Chapters 21, 96, 122, 124, and 125). in patients with HCC treated with TACE compared with
Both primary and metastatic liver tumors derive most of patients who received best supportive care (Llovet et al, 2002;
their blood supply from the hepatic artery, unlike the non– Lo et al, 2002). To date, there has been no study conclusively
tumor-bearing parenchyma, which receives the majority of demonstrating a significant difference in overall survival
nutrient flow from the portal vein. Therefore administering a among any method of TACE, TAE, or RAE (Brown et al,
treatment to the artery can effect tumor regression without 2012).
causing collateral damage to the underlying parenchyma. Indications for arterially directed therapy include control of
In the mid-1970s, it was recognized that the unusual dual symptoms (pain or hormonal-related symptoms due to neuro-
vascular supply to the liver might allow effective transarterial endocrine liver metastases), progression of disease on systemic
treatment for hypervascular metastases from neuroendocrine treatment, prolongation of survival, and local tumor control to
tumors, as well as primary hepatocellular carcinoma (HCC). maintain eligibility for liver transplant (in the case of selected
Chapter 30 Radiologic hepatobiliary interventions 527
patients with HCC). Transarterial therapies are rarely, if ever, Early posttransplant screening Doppler ultrasound (US) can
curative and instead are intended to be repeated upon disease be used to detect abnormal flow in the hepatic artery. If this
progression. In cases of minimal disease burden, ablation may test is abnormal, a contrast study (US, CT, or angiography)
be performed in conjunction with embolization as a potentially should be considered. To salvage the organ, a precious resource,
curative therapy (see Chapter 98). In this instance, performing revascularization is often attempted after documentation of
the embolization immediately prior to ablation has the advan- abnormal flow, even in asymptomatic patients (Pareja et al,
tage of depositing contrast-laden particles within the tumor to 2010).
assist in targeting with the ablation device. In addition, occlud- Hepatic artery complications after transplant include occlu-
ing arterial blood flow may increase the zone of ablation sion, stenosis, and pseudoaneurysm. In the case of stenosis or
(Tanaka et al, 2013). In some cases, the angiogram may iden- occlusion, revascularization with catheter-directed thromboly-
tify additional sites of disease undetected on preprocedure sis, angioplasty, and/or stent placement has been effective in the
imaging, changing the treatment plan. majority of cases. Pseudoaneurysm is a rare, but potentially
Selection criteria differ slightly with each treatment option. fatal, complication that may be treated with a covered stent
Broadly speaking, patients with unresectable disease involving graft (see Chapters 120 and 124).
less than 50% of the liver without underlying liver disease or Following blunt trauma, the liver is the second most com-
with well-compensated (Childs-Pugh score A or B7) cirrhosis monly injured abdominal organ (the spleen is first) (see Chapter
may be candidates. In the past, portal vein occlusion was con- 122). The American Association for the Surgery of Trauma
sidered an absolute contraindication because of the reported Injury Scoring Scale was developed to help guide management
higher complication rate and risk of death. More recently, series of these patients (Tinkoff et al, 2008). Injuries to the hepatic
of patients with portal vein occlusion treated with TAE, TACE, artery include pseudoaneurysms, which can be unifocal or mul-
and RAE have shown efficacy without a significant increase in tiple resulting in a “starry sky” appearance of multiple sites of
complications, thus supporting its use in this group of patients extravasation/injury on angiography. Focal extravasation or
with limited treatment options (Kim et al, 2009; Lau et al, pseudoaneurysm is usually treated with coil embolization of the
2013; Maluccio et al, 2008). affected vessel just distal and proximal to the injury, or with a
The complication profiles differ slightly between the various covered stent. In the case of multifocal injury, particle emboli-
transarterial therapies. TACE is infrequently associated with zation of the hepatic artery may be performed. Because of the
bone marrow suppression and alopecia. Radiation-induced dual blood supply to the liver discussed earlier, embolization of
liver failure occurs in 1% to 2% of patients who undergo RAE. the hepatic artery in the presence of a patent portal vein is rarely
With varying frequency, each intraarterial therapy is associated of clinical consequence to the patient. Hepatic artery injury
with arterial sclerosis and arterial occlusion, which can make may occur following iatrogenic hepatic interventions, either
future intervention more difficult (Miyayama et al, 2006). This surgical or percutaneous, such as biliary drainage or TIPSS,
is more commonly seen with TACE and drug-eluting beads and is treated similarly with coil embolization or covered stent
than with TAE (Erinjeri et al, 2010; Geschwind et al, 2003) placement.
but may be seen with any modality. The clinical relevance of
this angiographic finding is that over time tumors can derive Hepatic Vein
arterial supply from nonhepatic collateral vessels, making The least common vascular target of endovascular intervention
treatment more challenging and creating a higher risk of in the liver is the hepatic vein. Budd-Chiari is a potentially life-
nontarget embolization. Branches that commonly give rise to threatening disease of heterogeneous etiology resulting in
extrahepatic tumor supply include the right phrenic, internal obstruction of hepatic venous outflow, which occurs in less than
mammary, gastroduodenal, intercostal, and renal capsular 1 per million persons (see Chapter 88). Acutely, patients are
arteries. symptomatic with abdominal pain and ascites, and over time,
Complications include nontarget embolization, liver failure, centrilobular fibrosis and cirrhosis may develop. Initial therapy
vessel injury, and postembolization syndrome. Postemboliza- includes systemic anticoagulation, but the benefit of anticoagu-
tion syndrome occurs in the majority of patients and consists lation alone is debatable. Patients with ongoing symptoms
of some degree of pain, fever, and/or nausea. This can last for may benefit from thrombolysis, venoplasty, and/or stent place-
several days. Prolonged pain may suggest nontarget emboliza- ment and, in some cases, TIPSS (Copelan et al, 2015) (see
tion to the pancreas, resulting in pancreatitis, or to the gallblad- Chapter 87).
der or upper gastrointestinal (GI) tract, resulting in cholecystitis, Stenosis of the intrahepatic or suprahepatic inferior vena
or gastric or duodenal ulceration. cava may occur as a complication following orthotopic liver
The hepatic artery is also a vessel that may require interven- transplantation, and symptomatology mimicking Budd-Chiari
tion after liver transplant. Following primary graft malfunction, may ensue. Elevated velocities by Doppler US suggest the diag-
hepatic artery thrombosis is the second leading cause of graft nosis, and a pressure gradient of greater than 6 mm Hg across
failure after liver transplant and is a major cause of transplant- the stenosis at venography is diagnostic (Kubo et al, 2006).
related mortality. This complication can result from technical Venoplasty or, in select cases, stent placement can alleviate
issues with the anastomosis, including disparate diameters of symptoms and preserve graft function.
donor and recipient vessels, and tension on, or kinking of, the
anastomosis. In most cases, HAT occurs within the first 100 Biliary Intervention
days and manifests as fulminant hepatic necrosis and/or biliary Noninvasive imaging of the biliary tree with contrast-enhanced
tract ischemia and necrosis, resulting in sepsis. Because these CT and MRI have virtually eliminated the need for invasive
patients are immunosuppressed to prevent graft rejection, the percutaneous transhepatic cholangiography to diagnose biliary
gram-negative sepsis resulting from biliary necrosis can be very tract disorders (see Chapter 20). Therefore percutaneous tran-
difficult to treat (see Chapter 120). shepatic cholangiography is rarely performed for diagnostic
A B C
FIGURE 30.2 A, Cholangiogram in a patient with metastatic colon cancer shows a short-segment occlusion of the central right hepatic duct. After
stent placmeent extending from the right posterior hepatic duct to the common hepatic duct (B), there is good flow into the duodenum across the
preserved sphincter of Oddi (C).
purposes alone, but rather at the time of planned biliary inter- a longer median patency (3 to 9.1 months vs. 1.8 to 5.5
vention. Percutaneous transhepatic biliary drainage or stent months), but there is no difference in overall patient survival
placement can be performed to relive symptoms caused by (Levy et al, 2004). When feasible, primary stent placement is
obstructive jaundice, including pruritus, anorexia, and cholan- preferred because it does not require an exteriorized device.
gitis. Bliiary drainage can also be performed to lower the bili- Additionally, if there are incompletely drained ducts, having a
rubin preoperatively or to allow for chemotherapy or, in the catheter may put patients at risk for developing cholangitis
setting of bile leak, to divert bile (Migita et al, 2009). Less because bile colonization occurs within 48 hours, and when a
common indications include access to treat biliary stone disease primary stent is placed, there is less opportunity for contami-
or to facilitate intraductal therapies, such as brachytherapy. nating incompletely drained ducts. Finally, in high bile duct
In general, bile duct obstruction below the common hepatic obstruction, stents can often be placed above the papilla.
duct (i.e., “low” bile duct obstruction) (Bismuth et al, 1992) is Without reflux of bowel contents or an exteriorized device, the
best treated endoscopically because placement of a plastic or sterility of the biliary tree is maintained so that if or when
self-expanding metal stent (SEMS) can drain the entire biliary patients present with occluded stents the likelihood of present-
tree through the normal orifice of the sphincter of Oddi. In ing with cholangitis is greatly diminished (Fig. 30.2).
cases of high bile duct obstruction, (at or above the cystic duct Metal stents also may be covered with polyurethane, with
insertion) percutaneous drainage is preferred because this the goal of preventing tumor ingrowth, resulting in improved
allows a specific duct to be targeted for drainage without con- patency. Unfortunately, this has not been shown to be the case,
tamination of nontarget ducts. and increased complications of acute cholecystitis (Isayama
Ideally, the need for drainage and optimal plan for a given et al, 2004) and stent migration have been seen (Yoon et al,
patient is made by multidisciplinary consensus involving hepa- 2006). Covered stents are not generally indicated in high bile
tobiliary surgeons, gastroenterologists, and interventional radi- duct obstruction because of the risk of occluding segmental bile
ologists. Preprocedure planning must include high-quality ducts. Further studies are needed to establish the indications
imaging of the liver and biliary tree. Contrast-enhanced CT for covered versus bare metal stents.
and MRI are extremely useful to show biliary anatomy and
pathology (see Chapters 18 and 19). In some cases, US may Bile Duct Biopsy
add additional information about the level of obstruction and In some cases, the etiology of biliary obstruction may not be
patency of portal vein branches, but US alone is not sufficient evident at the time of drainage. Bile obtained at the time of
for preprocedure planning (see Chapter 15). A dose of prophy- drainage may be sent for cytology but diagnostic sensitivity is
lactic broad-spectrum antibiotic to cover enterococci, strepto- relatively low. Bile duct biopsy may be helpful in establishing
cocci, and aerobic and anaerobic gram-negative bacilli is a diagnosis of intraductal pathology (e.g., cholangiocarcinoma)
recommended prior to biliary intervention. This is particularly and in differentiating recurrent tumor from ischemic/
important when there has been bile duct reconstruction or postoperative stricture (see Chapters 22 and 51). At the time
biliary instrumentation because, in this setting, the incidence of biliary drainage, endoluminal brush or forceps biopsy of the
of colonized bile is high. stricture depicted by cholangiography can be used to obtain a
With malignant biliary obstruction in the absence of signs sample. These techniques are most useful for intraductal
or symptoms of cholangitis, placement of a primary SEMS may pathology (e.g., cholangiocarcinoma) in contradistinction to
be performed. Compared with plastic stents, metal stents have extrinsic masses that may cause biliary obstruction (e.g., hilar
A C
B
FIGURE 30.3 A, Ultrasound demonstrates multiple echogenic stones (arrow) with posterior acoustic shadowing. B, Coronal computed tomography
demonstrates gallbladder wall thickening and small amount of pericholecystic fluid, as well as endoscopic stent placed the day before that extends
to common hepatic duct, covering cystic duct origin. C, Image obtained after placement of cholecystostomy catheter demonstrates multiple stones
in the gallbladder with no opacification of the cystic duct or contrast entering the stent.
adenopathy or parenchymal liver mass.) In cases in which gallbladder or biliary tree for stone removal or to provide biliary
establishing a diagnosis is difficult, the biliary tree can be opaci- drainage in patients with common bile duct obstruction distal
fied through an indwelling drainage catheter and the stricture to the insertion of the cystic duct. When performed in the
targeted with a percutaneous fine needle. Cholangiographic setting of gallstones, it is often used as a temporizing measure,
guided-needle biopsy can be used to diagnose both intraductal to allow the resolution of sepsis and optimization of the patient’s
and extraductal causes of obstruction. medical conditions for subsequent cholecystectomy.
In cancer patients with bile duct obstruction, it is not
Percutaneous Cholecystostomy uncommon to see cholecystitis complicating SEMS placement
Percutaneous cholecystostomy is most commonly indicated for when gallstones are present and the stent covers the cystic duct
the treatment of acute acalculous cholecystitis in severely ill orifice, particularly when contrast is seen within the cystic duct
hospitalized patients but may also be used to treat calculous or gallbladder (Fig. 30.3). Some of these patients undergo
cholecystitis in patients who are too sick for, or in whom comor- subsequent cholecystectomy, but in palliative situations, the
bidities preclude, definitive cholecystectomy (see Chapters 33 patients may live with a cholecystostomy catheter. Cholecystos-
and 34). It may also be performed to provide access to the tomy has been reported effective as definitive therapy in
very-high-risk patients with calculous cholecystitis who have not seen when the patient returns for a repeat study after their
remained asymptomatic after catheter placement (Nasim et al, first dilation, the stricture can be dilated with a larger balloon.
2011). After cholecystostomy, clinical resolution of infection Particularly resistant strictures dilated have been treated by
occurs within 24 to 48 hours in the vast majority of cases. When some operators using cutting balloons with good, durable
performed for acalculous cholecystitis, the catheter can be results (Murkund et al, 2015).
capped once it begins to drain bile, a clear sign that the cystic
duct is patent, but the catheter should not be removed until Hepatic Cysts
tract maturation has occurred. In general, this occurs within 2 Liver cysts occur in up to 5% of the population and are inci-
to 3 weeks (Hatjidakis et al, 1998). Tract maturation may dentally seen on cross-sectional imaging (see Chapters 75 and
happen earlier when a transhepatic rather than transperitoneal 90B). They can slowly enlarge and rarely become symptomatic,
route to the gallbladder has been employed. If stone removal although symptoms from mass effect may develop if they
from the gallbladder is to be undertaken this should also be impinge on adjacent structures. In addition, bleeding may occur
performed after tract maturation. into the cyst, causing pain. A simple liver cyst should be dif-
ferentiated from a cystic tumor. The most common mimic of
Biliary Stone Disease the simple cyst is a cystadenoma that has the potential to
Stones can be removed from the gallbladder or biliary tree using become a cystadenocarcinoma. Unfortunately, cross-sectional
a variety of approaches and methods (see Chapters 36 and 37). imaging studies do not reliably distinguish between these
When there is a retained common bile duct stone after chole- two entities. Metastases that become necrotic can appear
cystectomy and a T-tube has been left in place, this is accom- cystic, particularly treated metastases from GI stromal tumor;
plished via the T-tube tract following tract maturation. When however, history and previous imaging will usually make the
choledocholithiasis occurs in a patient with remote cholecys- distinction.
tectomy, it is performed through a transhepatic approach. Cho- Treatment of simple liver cysts is not warranted unless the
ledocholithiasis may present with cholangitis or be asymptomatic. patient is symptomatic and the symptoms are clearly related to
The first step in percutaneous stone removal is placement of a the cyst. Simple aspiration can provide temporary relief, and,
percutaneous biliary drainage catheter. After a period of 2 to 3 when the symptoms are relieved, serves as proof of association,
weeks of drainage, during which a mature tract forms, the but recurrence after aspiration is 100%. Drainage and sclerosis
biliary catheter is exchanged for a large diameter sheath. The has been reported but has generally been replaced by laparo-
first order of business then is balloon sphincterotomy, followed scopic unroofing of the cyst, which is typically well tolerated by
by pushing smaller calculi into the duodenum using a balloon. the patient and avoids problems associated with an indwelling
If the stones are too large to pass through the dilated ampulla, catheter (see Chapter 75). Cystadenomas, because of the risk
they can be broken up using baskets, snares, or even lithotripsy. of malignancy, require surgical resection.
Once the duct is thought to be clear, the internal external drain-
age catheter is replaced. The patient returns in 1 to 2 weeks, Hepatic Abscess
and sheath cholangiography is performed to look for retained Most liver abscesses in the United States are pyogenic, caused
stones. If the ducts are clear, the internal external catheter is by bacteria (see Chapter 72). Amebic abscesses caused by Ent-
replaced with an external drain above the ampulla and then amoeba histolytica and fungal abscesses each account for about
capped. If the patient does well with the catheter capped, it can 10% of liver abscesses (see Chapter 73). Pyogenic liver abscesses
be removed without further imaging. are usually polymicrobial, and an etiology can often be discov-
ered. Ascending hematogenous infection from the GI tract is a
Bile Duct Injury common etiology, and patients with diverticulitis or appendicitis
The bile duct can be injured from blunt or penetrating trauma may present with a liver abscess. Cross-sectional imaging studies
but is probably most commonly injured at the time of surgery are fairly typical, demonstrating a complex collection that
(see Chapters 42 and 122). There was a fairly high rate of bile appears to have many septa (Fig. 30.5), resembling a cauli-
duct injury when laparoscopic cholecystectomy was initially flower. Despite this appearance, these can be successfully treated
introduced, at least in part related to lack of operator experi- percutaneously in combination with antibiotics, although cath-
ence, but now mostly related to unrecognized bile duct anoma- eter drainage may be prolonged (Mezhir et al, 2010). Imaging
lies (see Chapters 35 and 38). When the bile duct is clipped or should include the pelvis, with careful evaluation for a potential
transected, there is little that can be done percutaneously other source. Patients who develop bile duct obstruction, particularly
than draining the obstructed duct or diverting the transected in the setting of bactibilia related to previous biliary-enteric
one. When postoperative strictures occur, patients may present bypass, transampullary stent placement, or a preexisting percu-
with cholangitis, pruritus, jaundice, or any combination. Some- taneous drain, also may present with a liver abscess, although,
times, the biliary tree does not appear dilated or is only mini- in this situation it is better classified as an infected biloma (Fig.
mally dilated. Percutaneous access to the biliary tree is 30.6) and appears less complex than a true pyogenic liver
established first, and then the stricture is crossed (Fig. 30.4). abscess. Causes of infected biloma include stent obstruction,
An internal external catheter is left across the stricture for a recurrent tumor, or anastamotic stricture, and a catheter placed
couple of weeks as tract maturation occurs. The patient returns, for biloma drainage will continue to drain bile unless the cause
and the stricture is dilated with a balloon (Lee et al, 2012). of obstruction is eliminated. Liver abscess can also occur as a
When applied to bilioenteric anastomoses, success is likely in complication following embolization of liver tumors in the same
the majority of patients, with stricture recurrence in as few as patient group with compromised sphincter of Oddi and bactibilia
5% of patients on long-term follow-up (Janssen et al, 2014). (Mezhir et al, 2011) (see Chapter 96). This should always
Dilation of the common bile duct is not likely to be successful be considered in the preprocedure evaluation of patients under-
using a balloon of less than 10 mm. If an acceptable result is going hepatic artery embolization that have had a previous
A B
C D
FIGURE 30.4 A, Coronal computed tomography demonstrating bile duct dilatation to level of bilioenteric anastomosis. B, Stricture at anastomosis
shown on cholangiogram at time of initial drainage. Filling defects are air bubbles. C, Stricture is dilated with a 10-mm balloon. D, Cholangiogram
obtained 3 weeks after dilation with widely patent bilioenteric anastomosis.
pancreatoduodenectomy (Fig. 30.7) or colonized bile for any fever and abdominal pain. Hepatomegaly occurs often, par-
reason. True liver abscesses should be distinguished from “peri- ticularly with large cysts. When cysts are small, aspiration may
hepatic” collections that occur postoperatively, are not in the be necessary to differentiate amebic from pyogenic abscess
parenchyma of the liver, and are usually much easier to treat because amebic antibodies may not be detected at presenta-
(Fig. 30.8) (see Chapters 27 and 103). tion, although they typically appear later. Symptoms resolve
quickly with administration of metronidazole, and interven-
Amebic Abscess tion beyond that is rarely needed unless there has been
Unlike pyogenic hepatic abscess, amebic abscesses are typi- rupture. A recent randomized trial of 57 patients with
cally associated with travel history to an endemic region (see abscesses 5 to 10 cm in size found that although fever and
Chapter 73). At presentation, patients almost universally have pain resolved sooner in the group treated with aspiration and
A B
FIGURE 30.5 A, Ultrasound of abscess demonstrating complex echotexture with multiple septa and enhanced through-transmission. B, Coronal
computed tomography scan image demonstrating typical “cauliflower” appearance of pyogenic liver abscess with multiple septa.
metronidazole, compared with metronidazole alone, the dif- cancer will develop liver metastases during the course of their
ference was not statistically significant, and there was no dif- disease (see Chapter 92). In almost half of these patients,
ference in morbidity, mortality, treatment failure, days to disease is limited to the liver, and up to 25% of these patients
normalization of leukocytosis, or duration of hospital stay have resectable disease. More effective systemic chemotherapy
between the two groups (Bammigatti et al, 2013). (see Chapters 99 and 100) and advances in techniques
of hepatic resection (see Chapter 103) have combined to
Echinococcal Cysts improve survival and increase rates of hepatic resection. Con-
Hydatid cystic disease is a parasitic disease cause by Echinococ- comitantly, there have been advances in interventional radio-
cus granulosa (see Chapter 74). The larvae of this parasite cause logic techniques of percutaneous thermal ablation, including
the disease, which is endemic in Mediterranean, Middle radiofrequency, microwave, laser, and cryoablation, as well as
Eastern, and South American countries and in New Zealand irreversible electroporation (see Chapter 908). All of these tech-
and Turkey, where people are in close contact with sheep and niques are less costly, safer, and result in shorter hospital stays
dogs. The most common site of disease is in the liver (50% to than hepatic resection and could be applied in place of surgical
80%), followed by the lung (5% to 30%). Surgery is the primary resection in well-selected cases, although the risk of local recur-
method of treatment; however, percutaneous approaches have rence is higher. With increased screening of patients with hepa-
been investigated in recent years. In a study from 2005, Paksoy titis C and earlier detection of smaller HCCs, percutaneous
and colleagues reported on 59 patients with 109 hydatid cysts ablative techniques also have application in patients with HCC
that were treated percutaneously, injecting either hypertonic and underlying cirrhosis who may not be candidates for resec-
saline or albendazole sodium as the scolicidal agent. All patients tion because of their underlying liver disease, or as an adjunc-
were given 10 mg/kg/day of albendazole beginning 48 hours tive method of maintaining them on a transplant list. There
before their procedure, and this was continued for 2 months are other tumors that might be treated with ablation, assuming
postprocedure. Directly before the procedure, they received they meet number, size, and location criteria for successful
diphenhydramine and hydrocortisone to prevent anaphylactic treatment.
reactions. Treatment was safe and effective in both groups, with
only one recurrence in the group treated with hypertonic saline. Criteria for Treatment
This compares well with a reported 4% incidence of recurrence Even the most advanced ablative techniques are limited with
following surgery. Although all cysts returned to their initial size regard to the tumor size that can be successfully treated. It has
directly following aspiration and injection, successful treatment recently been shown that an ablative “margin” is critical to a
was associated with decrease in size over time. successful ablation, with a margin of at least 5 mm on CT 4 to
8 weeks following ablation associated with the best local tumor
Hepatic Ablation control (Wang et al, 2013). Most commercially available abla-
Tumors tion systems result in an elliptic volume of coagulative necrosis,
Colorectal cancer is the third most common malignancy in the with a maximum long axis of 4 cm, limited by properties of the
United States, and most disease-related deaths are secondary local tumor/tissue environment. For this reason, successful abla-
to metastatic disease. Up to 50% of patients with colorectal tion, with a low rate of local tumor recurrence is seen most often
A B
C
FIGURE 30.6 A, Computed tomography (CT) of a patient with pancreatic cancer and distal bile duct obstruction who has had a metallic stent
placed; note gas in bile ducts. B, Three months later, the patient presents with fever and elevated white blood cells, and CT at the same level now
demonstrates a gas/fluid collection in right liver (arrow). C, Catheter study 1 week later outlines communication to biliary tree with an obstructing
filling defect at the confluence of the bile ducts (arrow).
in tumors less than 3 cm in size, and the best results are obtained Subcapsular tumors can be difficult to treat effectively. Struc-
in tumors 2.5 cm or less. Larger tumors, or those with complex tures that may be injured by heat or cold, such as bile ducts and
geometry, are unlikely to be effectively treated by attempting to other adjacent organs, might preclude safe treatment, although
extend the thermal effect by using multiple overlapping applica- irreversible electroporation has been developed, in part, to deal
tions. The location of the tumor may also limit effectiveness of with that issue. It is fair to say that ideal patients for percutane-
ablation or ability to use the technique safely. Tumors that are ous ablation should have oligometastatic disease with three or
adjacent to high-flow blood vessels (portal vein, hepatic vein, fewer tumors, all less than 3 cm in size, that are not adjacent to
inferior vena cava) will be cooled where they are in contact a major bile duct or bile duct branch or in contact with any
with the blood vessel so that the application of heat will be less high-flow vessel, especially one greater than 5 mm in size. Local
effective at that margin and the risk of recurrence higher. tumor recurrence and survival rates are both related to the
A B
FIGURE 30.7 A, Patient with neuroendocrine tumor metastatic to liver, with several lesions evident in right liver. B, Computed tomography 3 weeks
postembolization when patient presented with fever and elevated white blood cells, demonstrating thick-walled abscess at location of treated tumor.
A B
FIGURE 30.8 Axial (A) and coronal (B) images from a patient with perihepatic postoperative fluid collection. Note how it “pushes” on hepatic
parenchyma.
number of tumors treated, and the best outcome is found in the 1- to 2-cm range because they retain dense contrast. Finally,
patients with a solitary tumor. it is possible to target areas of the tumor that may show less
Size criteria can probably be extended in HCC or other deposition of embolic material and in which recurrence is likely
vascular tumors that can be embolized first (Fig. 30.9). In this (Wang et al, 2013).
instance, addition of ablation to embolization provides several
theoretical advantages. In the first instance, it provides a Results
method of “double kill,” whereby the tumor is exposed to two Unfortunately, existing evidence for ablation of hepatic colorec-
tumoricidal events: ischemia and lethal temperature (Elnekave tal metastases is primarily from single-arm retrospective and
et al, 2013). Because embolization results in cessation of prospective studies with no published randomized trial (Wong
flowing arterial blood, the primary blood supply to hepatic et al, 2010). Overall survival, ranging from 14% to 55%, and
tumors, heat resulting from ablation does not need to overcome local recurrence rates of 4% to 60% are reported following
the cooling effects of flowing blood in the tumor or immediate treatment of colorectal metastases. Optimal conditions—small
environment, theoretically increasing efficiency of the process, tumors limited to the liver, not adjacent to blood vessels, and
resulting in more effective heating of the tissue. Second, when in the hands of experienced interventionalists—should result in
ablation is performed directly following embolization, it local control in 98% of lesions (Gervais et al, 2009). For small
becomes much easier to radiologically target small tumors in HCC, three randomized trials do exist that compare surgical
A B
C D
E
FIGURE 30.9 Arterial (A) and (B) venous phase computed tomography (CT) in a patient with 5-cm solitary hepatocellular carcinoma in a background
of nonalcoholic steatohepatitis. Immediately following embolization (C) there is retained contrast in the embolized tumor, and ablation is performed
with a single-needle radiofrequency electrode. After ablation, the track of the needle electrode can be seen with tiny “gas bubbles” at the periphery
of the tumor (D). Complete radiologic response by imaging on follow-up CT 6 weeks after treatment (E).
resection to radiofrequency ablation (Chen et al, 2006; Feng 2013), this may have been due to a difference in tumor size. In
et al, 2012; Huang et al, 2010 ); two of these reported equiva- the resection group, 98% of solitary lesions were less than 3 cm
lence with 3-year survivals of about 70% and recurrence-free in size; in the ablation group, only 47% of solitary tumors were
survival of 50% to 60%. Only the study by Huang demon- less than 3 cm. It is likely that, with careful selection, small
strated a difference with surgical resection being superior to solitary HCC can be treated as effectively with ablation as with
ablation, providing both a 5-year survival (75.7% vs. 54.8%) surgical resection.
and recurrence-free survival (51.3 vs. 28.7%) advantage. As
was alluded to in a review article from 2013 (Cucchetti et al, References are available at expertconsult.com.
Chapter 30 Radiologic hepatobiliary interventions 536.e1
REFERENCES carcinoma: a randomised controlled trial, Lancet 359(9319):1734–

1739, 2002.
Bammigatti C, et al: Percutaneous needle aspiration in uncomplicated Lo CM, et al: Randomized controlled trail of transarterial lipiodol
amebic liver abscess: a randomized trial, Trop Doct 43(1):19–22, chemoembolization for unresectable hepatocellular carcinoma, Hep-
2013. atology 35(5):1164–1171, 2002.
Bismuth H, et al: Management strategies in resection for hilar cholan- Maluccio MA, et al: Transcatheter arterial embolization with only par-
giocarcinoma, Ann Surg 215(1):31–38, 1992. ticles for the treatment of unresectable hepatocellular carcinoma,
Boyer TD, Haskal ZJ: The role of TIPS in the management of portal J Vasc Interv Radiol 19(6):862–869, 2008.
hypertension, Hepatology 41(2):386–400, 2005. May BJ, et al: Update on portal vein embolisation: evidence based
Brown DB, et al: Quality improvement guidelines for transhepatic arte- outcomes, controversies and novel strategies, J Vasc Interv Radiol
rial chemoembolization, embolization, and chemotherapeutic infu- 24(2):241–254, 2013.
sion for hepatic malignancy, J Vasc Interv Radiol 23(3):287–294, Mezhir JJ, et al: Current management of pyogenic liver abscess: surgery
2012. is now second line treatment, J Am Coll Surg 2010(6):975–983,
Chen MS, et al: A prospective randomized trial comparing percutane- 2010.
ous local ablative therapy and partial hepatectomy for small hepato- Mezhir JJ, et al: Pyogenic liver abscess after hepatic arery embolisation:
cellular carcinoma, Ann Surg 243(3):321–328, 2006. a rare but potentially lethal complication, J Vasc Interv Radiol
Colapinto RF, et al: Creation of an intrahepatic portosystemic shunt 22(2):177–182, 2011.
with a Grüntzig balloon catheter, Can Med Assoc J 126(3):267–268, Migita K, et al: Clinical outcome of malignant biliary obstruction
1982. caused by metastatic gastric cancer, World J Surg 33(11):2396–2402,
Copelan A, et al: Diagnosis and management of Budd-Chiari syn- 2009.
drome: an update, Cardiovasc Intervent Radiol 38(1):1–12, 2015. Miyayama S, et al: Extrahepatic blood supply to hepatocellular carci-
Cucchetti A, et al: Systematic review of surgical resection vs. radiofre- noma: angiographic demonstration and transcatheter arterial chemo-
quency ablation for hepatocellular carcinoma, World J Gastroenterol embolization, Cardiovasc Intervent Radiol 29(1):39–48, 2006.
19(26):4106–4118, 2013. Murkund A, et al: Percutaneous management of resistant biliary-
Elnekave E, et al: Long-term outcomes comparing surgery to enteric anastomotic strictures with use of a combined cutting and
embolization-ablation for treatment of solitary HCC <7cm, Ann Surg conventional ballon cholangioplasty protocol: a single-center experi-
Oncol 20(9):2881–2886, 2013. ence, J Vasc Interv Radiol 26:560–565, 2015.
Erinjeri JP, et al: Arterial patency after repeated hepatic artery bland Nasim S, et al: Emerging indications for percutaneous cholecystostomy
particle embolization, J Vasc Interv Radiol 21(4):522–526, 2010. for the management of acute cholecystitis—a retrospective review,
Feng K, et al: A randomized controlled trial of radiofrequency ablation Int J Surg 9:456–459, 2011.
and surgical resection in the treatment of small hepatocellular carci- Paksoy Y, et al: Percutaneous treatment of liver hydatid cysts: compari-
noma, J Hepatol 57(4):794–802, 2012. son of direct injection of albendazole and hypertonic saline solution,
Gervais DA, et al: Society of Interventional Radiology position state- AJR Am J Roentgenol 185(3):727–734, 2005.
ment on percutaneous radiofrequency ablation for the treatment of Palmaz JC, et al: Expandable intrahepatic portacaval shunt stents in
liver tumors, J Vasc Interv Radiol 20:S342–S347, 2009. dogs with chronic portal hypertension, AJR Am J Roentgenol 147(6):
Geschwind JF, et al: Transcatheter arterial chemoembolization of liver 1251–1254, 1986.
tumors: effects of embolization protocol on injectable volume of Pareja E, et al: Vascular complications after orthotopic liver transplan-
chemotherapy and subsequent arterial patency, Cardiovasc Intervent tation: hepatic artery thrombosis, Transplant Proc 42(8):2970–2972,
Radiol 26(2):111–117, 2003. 2010.
Hatjidakis AA, et al: Maturation of the tract after percutaneous chole- Rosch J, et al: Transjugular portal venography and radiologic portoca-
cystostomy with regard to access route, Cardiovasc Intervent Radiol val shunt: an experimental study, Radiology 92(5):1112–1114, 1969.
21:36–40, 1998. Saad WE: The history and future of TIPS: food for thought, Semin
Huang J, et al: A randomized trial comparing radiofrequency ablation Intervent Radiol 31(3):258–261, 2014.
and surgical resection for HCC conforming to the Milan criteria, Saad WE, et al: BATO with or without BRTO for the management of
Ann Surg 252(6):903–912, 2010. gastric varices: concept and technical applications, Tech Vasc Interv
Isayama H, et al: A prospective randomised study of “covered” versus Radiol 15(3):203–225, 2012.
“uncovered” diamond stents for the management of distal malignant Tanaka T, et al: Superselective particle embolisation enhances efficacy
biliary obstruction, Gut May 53(5):729–734, 2004. of radiofrequency ablation: effects of particle size and sequence of
Janssen JJ, et al: Percutaneous balloon dilatation and long-term drain- action, Cardiovasc Intervent Radiol 36(3):773–782, 2013.
age as treatment of anastomotic and nonanastomotic benign biliary Tinkoff G, et al: American Association for the Surgery of Trauma
strictures, Cardiovasc Intervent Radiol 37(6):1559–1567, 2014. Organ Injury Scale I: spleen, liver, and kidney, validation based on
Kim KM, et al: Reappraisal of repeated transarterial chemoemboliza- the National Trauma Data Bank, J Am Coll Surg 207(5):646–655,
tion in the treatment of hepatocellular carcinoma with portal vein 2008.
invasion, J Gastroenterol Hepatol 24(5):806–814, 2009. van Lienden KP, et al: Portal vein embolisation before liver resection:
Kubo T, et al: Outcome of percutaneous transhepatic venoplasty for a systematic review, Cardiovasc Intervent Radiol 36(1):25–34, 2013.
hepatic venous outflow obstruction after living donor liver transplan- Yoon WJ, et al: A comparison of covered and uncovered Wallstents for
tation, Radiology 239(1):285–290, 2006. the management of distal malignant biliary obstruction, Gastrointest
Lau WY, et al: Treatment for hepatocellular carcinoma with portal vein Endosc 63(7):996–1000, 2006.
tumor thrombosis: the emerging role for radioembolization using Wang X, et al: Margin size is an independent predictor of local tumor
yttrium-90, Oncology 84(5):311–318, 2013. progression after ablation of colon cancer liver metastatses, Cardio-
Lee AY, et al: Percutaneous transhepatic balloon dilation of biliary vasc Intervent Radiol 36(1):166–175, 2013.
enteric anastomotic strictures after surgical repair of iatrogenic bile Wang X, et al: Patterns of retained contrast on immediate postproce-
duct injuries, PLoS ONE 7(10):e46478, 2012. dure computed tomography (CT) after particle embolisation of liver
Levy MJ, et al: Palliation of malignant extrahepatic biliary obstruction tumors predicts subsequent treatment response, Cardiovasc Intervent
with plastic versus expandable metal stents: An evidence based Radiol 36(4):1030–1038, 2013.
approach, Clin Gatroenterol Hepatol 2(4):273–285, 2004. Wong SL, et al: American Society of Clinical Oncology 2009 Clinical
Llovet JM, et al: Arterial embolisation or chaemoembolisation versus evidence review on radiofrequency ablation of hepatic metastases
symptomatic treatment in patients with unresectable hepatocellular from colorectal cancer, JCO 28(3):493–508, 2010.
CHAPTER 31
Bile duct exploration and biliary-enteric anastomosis
Sean M. Ronnekleiv-Kelly and Clifford S. Cho
OVERVIEW of patients (Babel et al, 2009) (see Chapter 2). There are a
number of ductal anomalies related to the convergence of the
Minimally invasive techniques to manage biliary pathology left and right hepatic ducts and the insertion of cystic duct.
have reduced the need for open operative intervention. Although Although the left biliary system is fairly consistent, the right
open exploration for biliary disease has become less common, biliary system is prone to anatomic variation; the most common
specific situations, such as obstructive common duct stones not variants include the right anterior or posterior sector ducts
amenable to endoscopic therapy or restoration of biliary-enteric traversing a longer extrahepatic course before joining the left
continuity following resection of bile duct tumors, remain indi- biliary system (Blumgart et al, 1984).
cations for more invasive approaches. It is important to have a
thorough understanding of biliary anatomy and familiarity with BILE DUCT EXPLORATION
various options for operative exposure and management. Oper-
ative techniques encompassing bile duct exploration and Overview
biliary-enteric bypass will be the focus of this chapter. It is estimated that approximately 10% of the adult U.S. popula-
tion is affected by cholelithiasis, with nearly one third of these
ANATOMY patients requiring cholecystectomy over their lifetime (see
Chapter 32). Approximately 10% to 15% of patients undergo-
Knowledge of extrahepatic biliary anatomy and an appreciation ing cholecystectomy have choledocholithiasis; thus common
of anatomic variants are essential for safe operative conduct (see bile duct (CBD) stones remain the most common etiology for
Chapter 2). The intrahepatic bile ducts draining the various biliary exploration (Costi et al, 2014) (see Chapters 36 and 37).
sectors ultimately coalesce into the right hepatic duct draining There are a number of techniques available for evaluation and
the right hemiliver and the left hepatic duct draining the left clearance of the common bile duct, including percutaneous,
hemiliver, which then converge at the liver hilum to form the endoscopic, laparoscopic, and open methods. Open common
common hepatic duct, the most anterior structure of the portal bile duct exploration (CBDE) is only necessary in a small
triad at this location. In approximately 80% to 90% of cases, subset of patients. This subset includes patients undergoing
the right hepatic artery courses posterior to the common an open cholecystectomy (or a laparoscopic cholecystectomy
hepatic duct toward the right liver, whereas in the minority of converted to open) in which choledocholithiasis is suspected,
cases, it can be found anterior to the duct. The right hepatic patients with large or multiple stones, and patients requiring
duct has a short extrahepatic segment (approximately 1.5 to transduodenal sphincteroplasty. Although it is more invasive,
2 cm) compared with the left, which traverses beneath segment open CBDE is highly effective and safe. A recent analysis dem-
IVB for approximately 3 to 4 cm after exiting the liver paren- onstrated equivalent duct clearance, morbidity, and mortality
chyma. When segment IVB is broad, the left hepatic duct between endoscopic measures and open surgical techniques
assumes a longer extrahepatic course with a transverse orienta- (Clayton et al, 2006). Because percutaneous, endoscopic, and
tion, as opposed to a shorter and more oblique orientation laparoscopic modalities are discussed in other chapters, open
when segment IVB is narrow (Fig. 31.1). The left hepatic duct bile duct exploration will be the focus of the next section.
and left portal vein travel within a peritoneal reflection of the
gastrohepatic ligament; exposure of these structures is facili- Incision and Exposure
tated by lowering the hilar plate at the base of segment IVB A right subcostal incision affords satisfactory exposure of the
(Fig. 31.2). As the left hepatic duct and left portal vein enter gallbladder, portal structures, and duodenum; alternatively, an
the umbilical fissure, they are joined by the left hepatic artery, upper midline incision may be equally effective, particularly in
and vascular branches to segments II, III, and IV travel with thin patients. Division of the lateral peritoneal attachments of
biliary drainage from these segments. To maximize access to the right colon, followed by mobilization of transverse colon
these structures, a bridge of hepatic tissue at the base of the mesentery off of the duodenum, provides visualization of the
umbilical fissure (bridging segment III and IV) must often be duodenum. A generous Kocher maneuver facilitates access to
divided (Fig. 31.3). The ligamentum teres (obliterated umbili- the CBD located in the lateral border of the hepatoduodenal
cal vein) at the base of the falciform ligament runs across the ligament. Cephalad retraction of the undersurface of the liver
umbilical fissure, separating segment IV from segment II and at the base of segment IVB will optimize visualization of the
III. Segment I hepatic ducts flow into both the right and left extrahepatic biliary system (Fig. 31.4). Additionally, cholecys-
biliary systems, with the majority of drainage entering the left tectomy can enhance exposure of the hepatoduodenal ligament
hepatic duct just proximal to common hepatic duct. and can facilitate intraoperative transcystic cholangiography,
Safe operative conduct requires a close familiarity with ana- which can help to delineate biliary anatomy. Dissection of the
tomic variations of biliary drainage, as they occur in up to 25% gallbladder and cystic duct also identifies the confluence of
537
VIII II
I
VII
III
Segment III pedicle

VI V
IV
Umbilical fissure
Left hepatic pedicle

Right hepatic pedicle
FIGURE 31.1. Diagrammatic expanded view of the liver showing FIGURE 31.3. The bridge of liver tissue frequently presents between
its segmental structure. Elements of the portal triad are distributed to the base of segment IV, and the left lobe of the liver can be divided. This
the right and left liver on a segmental basis. The left hepatic duct always is conveniently done by passing a curved director beneath it and cutting
pursues an extrahepatic course beneath the base of segment IV in the it with diathermy. Such division can be useful in aiding an approach to
groove, separating segment IV from the caudate lobe (segment I; see the left hepatic duct, particularly if the course of the duct is vertical and
Fig. 31.2). The ligamentum teres marks the umbilical fissure and runs to the base of segment IV is short. If the bridge of tissue is present, the
join the umbilical portion of the left branch of the portal vein. Each portal maneuver is always necessary to allow dissection of the segment III duct
triad is composed of hepatic artery, portal vein, and biliary duct. Note (see Fig. 31.8).
the distribution of the left portal triad in the umbilical fissure; major
branches recur to segment IV medially, and two major branches pursue
a lateral course to segments II and III of the left lobe.
Segment IV
Anterior
CL
A B C
FIGURE 31.2. Sagittal section showing the relationship of segment IV
and caudate lobe (CL) to the left portal triad, which is encased within a
reflection of the lesser omentum that fuses with the Glisson capsule at
the base of segment IV. The arrow indicates the point of incision for the
dissection to lower the hilar plate (see also Fig. 31.5). A, Left hepatic
duct; B, left branch, portal vein. Note the left hepatic artery (C) joins the
FIGURE 31.4. To optimize exposure of the extrahepatic biliary system,
left duct and left branch of the portal vein at the umbilical fissure.
the undersurface of the liver at segment IVB is retracted cephalad. Also
depicted here is the initial line of incision for an approach to the left
hepatic duct by lowering of the hilar plate. The incision is made at pre-
cystic duct and CBD. Once the CBD has been successfully cisely the point at which the Glisson capsule reflects to the lesser
identified, the overlying peritoneal tissue can be dissected free omentum (see Fig. 31.2).
to allow access to the anterior portion of the duct.
Supraduodenal Exploration caution should be taken to avoid injuring the posterior wall of
The anterior aspect of the CBD is exposed, and two stay the duct. The length of the choledochotomy will depend upon
sutures are placed on either side of the midpoint of the planned the diameter of the duct and size of stones present with the
longitudinal incision. The choledochotomy should be located lumen but is generally 1 to 2 cm. Another consideration regard-
anteriorly to avoid compromising blood vessels that typically ing the location of this incision is proximity to duodenum; in
run along the medial and lateral aspects of the CBD and the event that a choledochoduodenostomy should be required
common hepatic duct. To avoid injury to the cystic duct, the for appropriate drainage, every effort should be made to posi-
site of cystic duct insertion should be identified, as this may tion the incision distally enough to permit a tension-free anas-
occur in a medial or posterior location. During the incision, tomosis to the proximal duodenum.
Chapter 31 Bile duct exploration and biliary-enteric anastomosis 539
After choledochotomy, flushing of the duct (distally toward traverse into the left or right hepatic duct proximally or into the
the ampulla) with saline irrigation can result in expulsion of duodenum distally, and by excising the back wall of the hori-
stones. Following this, a deflated Fogarty balloon catheter is zontal portion of the T (to minimize the risk of tube occlusion
passed distally through the ampulla into the duodenum. Once and to facilitate tube removal). The tube is inserted through
the tip of the catheter is palpated within the duodenal lumen, the choledochotomy, and the remainder of the duct is closed
the balloon is inflated and the catheter is withdrawn until resis- with fine absorbable sutures around the tube. Care should be
tance is encountered, indicating that the balloon is positioned taken to leave enough redundancy in the intraperitoneal portion
against the sphincter of Oddi. The balloon is slowly and par- of the tube to avoid tension (and possible tube dislodgement)
tially deflated while applying continuous tension, allowing in the event of significant postoperative abdominal distension.
passage into the distal CBD. At this point, the balloon is care- Postoperatively, the T-tube is placed to dependent drainage
fully reinflated, and the catheter is swept proximally, sweeping until resolution of postoperative papillary edema allows physi-
retained stones up toward the cholodochotomy. This process ologic flow of bile into the duodenum. If persistently elevated
is repeated until no stones return. The catheter is then passed output or drainage around the tube occurs, investigation via
proximally to retrieve any stones within the common hepatic cholangiography can identify malfunction, dislodgement, or
duct and intrahepatic biliary tree. Rigid instruments such as distal obstruction secondary to retained stone. If a repeat chol-
clamps or stone forceps can cause injury to the bile duct and angiogram at approximately 2 to 3 weeks is normal, the T-tube
are therefore generally best avoided. Rather, a choledochoscope may be removed. If choledocholithiasis persists, the T-tube can
can be utilized to visualize remnant stones. These can be cap- be clamped to promote stone passage. If signs or symptoms of
tured with basket retrieval, in which the basket is passed beyond cholangitis occur, the tube can be unclamped and repeat
the stone, opened, and pulled back to ensnare the stone. The imaging is obtained. Residual obstruction may be amenable to
choledochoscope and associated basket with stone are then stone extraction via T-tube or endoscopic or percutaneous
withdrawn. At the conclusion of duct exploration, confirmation access.
of proximal and distal duct clearance is performed either using
choledochoscopy or completion cholangiography (Verbesey Outcomes
et al, 2008). Open CBDE has been shown to be a safe procedure, with
favorable mortality and morbidity rates. A recent analysis of
Transduodenal Exploration multiple trials comparing open CBDE with endoscopic retro-
When an impacted stone at the distal CBD cannot be cleared grade cholangiopancreatography (ERCP) identified similar
via choledochotomy, a transduodenal approach can be used. A rates of mortality (1% vs. 3%) and morbidity (20% vs. 19%)
2- to 4-cm longitudinal incision on the lateral aspect of the in the open surgery versus endoscopy groups, respectively
second portion of the duodenum can allow visualization of the (Dasari et al, 2013). Even in high-risk patients, open surgical
ampulla. Stay sutures are placed on either side of the incision intervention has comparable mortality (4% vs. 6%) and mor-
to maximize exposure. If there is difficulty visualizing the bidity (23% vs. 16%) versus ERCP (Targarona et al, 1996).
ampulla, a small catheter can be passed through the choledo- The rate of retained stones was historically lower after open
chotomy into the duodenum. A sphincterotomy is performed CBD exploration than after ERCP (6% vs. 20%) (Suc et al,
at the 10 to 11 o’clock position, and the sphincter is incised to 1998). However, more recently, studies have demonstrated
the level of the impacted stone or probe (catheter). Placing rates of retained stones after ERCP to be approximately 2% to
the incision at this location in the ampulla minimizes the chance 10%, although higher rates are reported for difficult scenarios
of pancreatic duct injury, which is generally located opposite (i.e., larger stones, cholangitis) (Rogers et al, 2010; Wan et al,
the planned sphincterotomy site. The stone is then extracted, 2011) (see Chapter 29). For instance, a recent study of retained
and the CBD mucosa is approximated to the duodenal mucosa stones after ERCP in a high-volume center demonstrated clear-
with absorbable sutures (sphincteroplasty) to avoid postopera- ance of the CBD in 91.7% of patients with stones less than
tive papillary stenosis. Once again, a catheter is passed to 2 cm in diameter versus 77.8% successful clearance in patients
ensure resolution of obstruction, and choledochoscopy or with stones greater than 2 cm in diameter (Wan et al, 2011).
cholangiography is used to confirm the absence of residual There is a relative paucity of recent data encompassing open
stones. The duodenotomy is typically closed in one layer CBDE due to options of ERCP and laparoscopic CBDE.
(Verbesey et al, 2008). If sphincteroplasty is not successful, the However, given the safety profile and necessity when less inva-
obstruction can be bypassed with a choledochoduodenostomy sive alternatives fail, open CBDE remains an important option
(discussed in Choledochoduodenostomy later). for CBD stones.
T-Tubes BILIARY-ENTERIC ANASTOMOSIS

Although T-tube insertion for choledochotomy closure is not
mandatory, advantages include drainage of bile in the setting Overview
of CBD or papillary edema and access for postoperative stone There are three key aspects to consider when planning a biliary-
retrieval or cholangiography or choledochoscopy. Disadvan- enteric anastomosis (see Chapter 42). These include identifica-
tages include tube migration, obstruction, and bile leak follow- tion of a healthy segment of bile duct tissue proximal to the site
ing removal. Although closure over a T-tube is typically the of obstruction; preparation of a segment of alimentary tract such
preferred approach, a recent analysis suggested that the duct as duodenum or, more commonly, Roux-en-Y jejunal limb; and
can be closed primarily without increased morbidity or mortal- construction of a mucosa-to-mucosa anastomosis. It is therefore
ity (Gurusamy et al, 2013). If a T-tube is used, size 14 Fr or important to use preoperative imaging to clearly delineate the
larger will permit cholangiography and choledochoscopy. The biliary anatomy prior to undertaking operative intervention
T-tube is prepared by cutting two limbs at lengths that will not for biliary decompression. It should be noted that invasive
cholangiography is no longer necessary to delineate biliary Kocher maneuver, will further enhance exposure for biliary-
anatomy in the great majority of cases. Cross-sectional imaging enteric bypass.
with magnetic resonance imaging/magnetic resonance cholan- Caution must be exercised in the setting of long-standing
giopancreatography, or even thin-cut computed tomography biliary obstruction or conditions associated with ipsilateral
scans, can accurately characterize the anatomy of the biliary tree hepatic atrophy and contralateral hypertrophy. In the scenario
and underlying pathology (see Chapter 19). Invasive imaging of of marked right hemiliver atrophy, the liver hilum and portal
the biliary tree with endoscopic or percutaneous cholangiogra- structures will become rotated in a counterclockwise manner.
phy allows stent placement, which can facilitate intraoperative Consequently, the portal vein will assume a more anterior loca-
identification of right and left hepatic ducts (see Chapters 20, tion, and the CBD or common hepatic duct will be posteriorly
29, and 30). However, it should be stressed that utilization of displaced. Hypertrophy of segment IV protrudes over the porta
an intrabiliary catheter should not be necessary in most cases hepatis and may provide additional complexity for access to the
for duct identification in the hands of an experienced biliary hilum and left biliary system (Pottakkat et al, 2009). In cases
surgeon. It is also important to recognize that instrumentation of very profound right liver atrophy, access to the biliary conflu-
of the biliary tree introduces bacterial contamination that, in a ence may require a thoracoabdominal incision.
setting of biliary stasis, can result in cholangitis, periductal
inflammation, and a higher risk of postoperative infections. Hepaticojejunostomy
Caution must also be exercised to avoid percutaneous drainage Despite the need for an additional anastomosis (at the jeju-
if it is unlikely that the stent can be passed across the obstructing nojejunostomy), Roux-en-Y hepaticojejunostomy is the most
lesion. Insertion of a percutaneous drain into an excluded biliary common surgical reconstruction for biliary obstruction. The
segment will result in bacterial colonization of static bile; if that jejunum is typically anastomosed to the common hepatic
hepatic segment cannot be decompressed by a subsequent oper- duct just distal to the confluence of the right and left hepatic
ative intervention, it will not be possible to remove that external ducts. If this approach is not feasible due to tumor infiltra-
drain without risking refractory cholangitis. These complexities tion or a high stricture, drainage can be obtained via the right
underscore the critical importance of an experienced multidis- hepatic duct or left hepatic duct. Moreover, ducts draining
ciplinary team reviewing and treating complex biliary obstruc- segment III can be utilized when the left hepatic duct is not
tion, particularly at the biliary confluence. accessible (Blumgart et  al, 1984). Disadvantages include
Depending on the underlying pathology, there are a number necessity for two anastomoses and exclusion of bile from the
of options for restoration of biliary continuity with the duodenum.
alimentary tract. For instance, choledocholithiasis refractory
to local exploration may require choledochoduodenostomy. Approach to Right Hepatic Duct
Other benign etiologies, such as iatrogenic bile duct injury, As the portal pedicles enter the liver parenchyma, they remain
strictures from previous biliary-enteric operations, choledochal enclosed within a fibrous sheath derived from Glisson’s capsule.
cysts, or inflammatory strictures, may require restoration with Access to the right portal pedicle containing the right hepatic
Roux-en-Y choledochojejunostomy or hepaticojejunostomy. duct can be achieved by isolating the pedicle in an extrahepatic
Additionally, benign proximal biliary strictures as well as malig- location or by exposing the pedicle via intrahepatic dissection.
nancy (cholangiocarcinoma) may require anastomosis between The extrahepatic approach begins by lowering the hilar plate;
intrahepatic ducts and jejunum. Finally, the gallbladder may the peritoneum along the posterior aspect of segment IV is
also be utilized to facilitate drainage (cholecystoduodenostomy divided, allowing separation of Glisson’s capsule from the peri-
and cholecystojejunostomy). Although nonoperative measures toneal reflection enveloping the porta hepatis. By reflecting the
can be utilized in most situations, familiarity with the various base of segment IV in a cephalad direction, this dissection
surgical techniques can enable appropriate restoration of effectively exposes the confluence of the right and left hepatic
biliary-enteric continuity when the situation demands (see duct. By continuing this plane of dissection to the right
Chapter 42). (onto the cystic plate), the right hepatic duct may be exposed
(Strasberg et al, 2008). If the extrahepatic portion of the right
Incision and Exposure hepatic duct is too short to be visualized in this way (as is often
A right subcostal incision with or without an upper midline the case), the intrahepatic approach may be used. This requires
extension or a left subcostal extension followed by upward hepatotomies in the caudate process just posterior to the porta
elevation and cephalad retraction of the costal margin provides hepatis and along the base of the gallbladder fossa. A blunt
adequate exposure for construction of any biliary-enteric anas- right angle clamp may be passed between these hepatotomies
tomosis. The ligamentum teres is ligated and divided, and the to encircle the right pedicle, which can then be delivered for
falciform ligament is divided to its most cephalad extent on the further dissection to identify the right hepatic duct, or the
diaphragm. Retraction of the ligamentum teres is helpful for anterior or posterior sectional ducts (Fig. 31.5) (Jamieson &
optimal visualization of the vascular inflow and biliary drainage Launois, 1992).
of segments II, III, and IV. If direct decompression of the gall-
bladder is not to be undertaken, cholecystectomy can be advan- Approach to Left Hepatic Duct
tageous for identification of the cystic duct, which can be The left hepatic duct traverses an extrahepatic course below
dissected to its point of insertion onto the common hepatic segment IVB from the umbilical fissure to the porta hepatis.
duct. Cholecystectomy will also expose the cystic plate, which Because of its longer extrahepatic course, the left hepatic duct
runs in continuity with the hilar plate. By lowering the hilar is the preferred target. After ligation of the ligamentum teres
plate, the left hepatic duct may be exposed as it runs against with a firm tie, retraction is applied to elevate the left hemiliver.
the base of segment IVB (Blumgart, 1987). Mobilization of The bridge of tissue at the base of the umbilical fissure contains
the right colon with caudal retraction, followed by generous no large vessels and can be divided to provide mobility of the
A B
FIGURE 31.5. A, Exposure of the right hepatic pedicle in the setting of a cholangiocarcinoma, extending into the left hepatic duct. Intrahepatic
exposure of the right hepatic duct is accomplished initially by controlling the right portal pedicle. After performing hepatotomies in the caudate process
just posterior to the porta hepatis and along the base of the gallbladder fossa, a blunt right angle can be used to facilitate encircling and delivery of
the right pedicle. The overlying hepatic parenchyma is further dissected, which allows identification of the right hepatic duct, as well as the anterior
or posterior sectional ducts. B, The relevant duct, usually the anterior sectoral duct, is opened, and anastomosis is carried out.
base of segment IVB and to enhance exposure of the left portal

pedicle (see Fig. 31.3). The left hepatic duct runs with the left
portal vein and is exposed by lowering the hilar plate (Fig.
31.6). Dissection toward the hepatic confluence to expose an
adequate segment of the left hepatic duct is facilitated by
retracting segment IVB anteriorly (Blumgart et al, 1984).
Approach to Segment III Duct

The presence of a bulky unresectable tumor at the hepatic hilus
may require construction of a more proximal anastomosis.
This can be achieved by exposing the biliary drainage of the
left hemiliver within the umbilical fissure (Blumgart, 1987;
Jarnagin et al, 1998; Voyles et al, 1983). If the left hemiliver
has not atrophied from long-standing biliary obstruction, uni-
lateral left-sided biliary decompression will effectively relieve
obstructive jaundice and restore hepatic function even when
the biliary drainage of the excluded right hemiliver remains
obstructed. In a series evaluating segment III bypass for malig-
nant biliary obstruction, patients experienced durable relief of
jaundice and pruritus; there was an 80% patency rate at 1 year
(Jarnagin et al, 1998). Segment III bypass is best utilized in the
setting of a dilated left biliary system, which facilitates identifi-
cation of the segment III duct within the liver parenchyma; the
procedure is much more challenging in the face of normal-
caliber ducts. Also, it must be remembered that, if the right
biliary system has been instrumented or is otherwise colonized
and is isolated from the left hepatic duct, the segment III bypass FIGURE 31.6. The hilar plate is lowered, and the left hepatic duct is
approach will not provide adequate drainage. exposed for dissection. The exposure is carried medially and to the right
After ligation and retraction of ligamentum teres, the band to expose the confluence and the right hepatic duct.
of liver parenchyma at the base of the ligamentum teres joining
segment III and IVB (if present) is divided to enhance exposure
of the segment III duct, which, if dilated, may be more readily small-gauge needle. On occasion, a wedge resection of a portion
apparent. This exposure can be facilitated by fashioning a of segment III can also be performed to provide exposure of
superficial hepatotomy along the left of the ligamentum teres, the segment III duct (Fig. 31.8) (Blumgart, 1987). To prepare
through which the segment III duct may be exposed and for biliary-enteric anastomosis, the duct should be dissected
opened without risk of injury to the vascular pedicle to segment free for 1.5 cm but should not be cleared circumferentially. A
III (Fig. 31.7). In circumstances where identification is diffi- defunctionalized jejunal loop is then brought up in a retrocolic
cult, localization can be confirmed by aspiration with a fashion and prepared for anastomosis.
A B
C D
FIGURE 31.7. A, The liver is held up so that its inferior surface is seen. The bridge of liver between segment IV and the left lobe of the liver has
been divided. The base of the ligamentum teres is seen. B, The ligamentum teres is then pulled downward. The peritoneum of its upper surface on
the left side is incised, and the extensions passing into the liver are exposed. The left sides of these extensions are divided between ligatures, which
must be passed carefully using aneurysm needles. This part of the dissection is tedious and should be carried out meticulously because hemorrhage
within the recess adjacent to the segment III duct can be difficult to control. C, The segment III duct is exposed. D, The duct is opened longitudinally
for anastomosis, which is carried out by the technique illustrated in Figure 31.13.
A B
FIGURE 31.8. A, The liver is split to the left of the ligamentum teres in the umbilical fissure, and it may be necessary to remove a small wedge of
liver tissue. B, The segment III duct is exposed at the base of the liver split above and behind its accompanying vein and is ready for
anastomosis.
FIGURE 31.10. Adjacent ductal orifices may be approximated before

anastomosis.
FIGURE 31.9. Manner of fixation of transanastomotic tubes. Note

the introduction of the absorbable suture in a mattress fashion across
the ductal wall proximal to the future site of anastomosis. This secures
the tube conveniently during anastomosis.
Construction of Anastomosis
For the purposes of Roux-en-Y reconstruction, the most proxi-
mal loop of jejunum that can be brought to lie against the
planned site of anastomosis without tension is selected. The
jejunum is transected, and a Roux limb of 50 to 70 cm is passed
in a retrocolic fashion through the avascular portion of the
transverse mesocolon to the right of the middle colic artery. A
The jejunojejunostomy may be fashioned in a sutured or stapled
manner. We do not typically use transanastomotic stent;
however, if a stent is to be used, it is preferable to pass the stent
through the cut hepatic duct prior to construction of the anas-
tomosis. The stent may be affixed against the duct wall with a
single 4-0 or 5-0 absorbable catgut suture that is tied on the
outside of the duct wall (Fig. 31.9); this maneuver helps to
avoid inadvertent dislodgment of the tube during placement of
the anastomotic sutures. When more than one biliary duct
orifice is present, it is preferable to create a single-duct orifice
by approximating the two ducts with a single row of 4-0 or 5-0
absorbable sutures (Fig. 31.10).
The hepaticojejunostomy may be performed in an end-to-
side or side-to-side anastomosis (see Chapter 42). For an end-
to-side hepaticojejunstomy, the bile duct segment is transected,
and an adjacent jejunotomy is fashioned at a safe distance from
the mesenteric margin of the bowel approximately 2 cm from
the staple line. The jejunotomy length should be shorter than
the ductotomy, as the bowel is more pliable than the duct. An
anterior row of full-thickness, single interrupted 4-0 or 5-0
absorbable sutures is placed in the bile duct, passed from inside B
to outside, and working from the patient’s left to right. This
FIGURE 31.11. A, The initial step in the creation of hepaticojejunos-
row of sutures, with needles intact, is retracted and elevated;
tomy Roux-en-Y. The anterior layer of suture (3-0 Vicryl) on the bile duct
this maneuver effectively pulls the anterior aspect of the duct is inserted first, and the sutures are passed from the inside out, starting
away, facilitating exposure and placement of the posterior row from the patient’s left and working toward the right. The needles are
of sutures (Fig. 31.11). A single full-thickness posterior row of retained, and the sutures are kept in order. B, The anterior layer of
interrupted 4-0 or 5-0 absorbable sutures is used to approxi- sutures is elevated. This displays the posterior ductal wall, and the
mate the inferior edge of the biliary duct to the superior edge posterior row of sutures is now placed, again from left to right.
of jejunum, also working from patient’s left to right. These generally completed with a single layer to avoid narrowing.
sutures are tied and cut short except for the two corner sutures, Placement of a sponge circumferentially around the anastomo-
which are secured with clamps. The previously placed row of sis allows an intraoperative test to confirm absence of a large
anterior sutures is then used to complete the anastomosis. Each bile leak.
needle is passed through the jejunum, tied, and cut short; the The alternative approach of constructing a side-to-side anas-
corner stay sutures are then cut. tomosis offers the advantages of less devascularization, with the
A side-to-side anastomosis may be appropriate in the setting use of a nontransected duct, and preservation of biliary-
of left hepatic duct or segment III duct hepaticojejunostomy duodenal continuity, in case future ERCP should be desirable
(see Fig. 31.6). Additionally, benign strictures (e.g., iatrogenic) (Winslow et al, 2009). To perform this anastomosis, the ante-
can be approached in this fashion. When decompression is rior surface of the duct is exposed and opened longitudinally
undertaken at the level of the proximal hepatic duct, increased for a distance of 2.5 to 3.0 cm, avoiding the medial and lateral
length may be achieved by extending the incision onto the left locations of periductal vasculature. The biliary-enteric anasto-
hepatic duct. A longitudinal ductomy of approximately 2.5 to mosis can then be completed in a similar manner as described
3.0 cm is performed, with a corresponding jejunotomy on the for side-to-side hepaticojejunostomy (see earlier). A single
antimesenteric border, 2 cm from the staple line. An anterior interrupted anterior (or right) row of sutures allows exposure
row of full-thickness, single interrupted 4-0 or 5-0 absorbable for placement of the posterior (or left) row, approximating duct
sutures, passed from outside to inside, are retracted to allow to jejunum, followed by completion of the anterior row
exposure for placement of the posterior row. Full-thickness, (Winslow et al, 2009).
single interrupted 4-0 or 5-0 absorbable sutures approximate
the inferior edge of the duct to the superior edge of jejunum. Choledochoduodenostomy
Following placement, the posterior row of sutures is tied with Although choledochoduodenostomy has the physiologic advan-
knots on the inside. Subsequently, the preplaced anterior row tage of enabling bile flow into the duodenum, this procedure is
of sutures are used to complete the anastomosis; they are effective only for distal strictures or impacted stone(s) in the
passed from inside to outside on the jejunum, and the knots distal CBD. After mobilization of hepatic flexure of the colon,
are tied on the outside (Winslow et al, 2009). a generous Kocher maneuver is performed to allow sufficient
mobility of the duodenum to enable construction of a tension-
Choledochojejunostomy free anastomosis. The gallbladder is removed and the cystic
Decompression of the biliary tree utilizing the common duct ligated. A 2.5- to 3.0-cm longitudinal incision is made
hepatic or CBD is typically undertaken in the setting of distal on the anterior surface of the supraduodenal CBD. Corre-
obstruction. Impacted ampullary stone(s), iatrogenic injuries, spondingly, the duodenum adjacent to the CBD is incised
stricture formation in which a more durable drainage procedure longitudinally along its superior-lateral border (Fig. 31.13).
is required, and irresectable periampullary tumors in which This produces incisions that are perpendicular to one another
biliary stenting is not effective are clinical scenarios where cho- (unlike the parallel configuration used during hepaticojejunos-
ledochojejunostomy may be utilized. Advantages of a surgical tomy). As with the hepaticojejunostomy, the duodenotomy is
approach versus nonoperative decompressive modalities are generally shorter in length than the ductotomy. The anastomo-
long-term patency and durability without the uncommon need sis is then constructed in a manner that anastomoses the bile
for repeat stent placement or revisions. duct transversely to the longitudinally-oriented duodenotomy.
Choledochojejunostomy can be performed as an end-to-side Three corner 4-0 or 5-0 absorbable sutures are placed: The
or side-to-side anastomosis. For the end-to-side technique, a first two are positioned between the midpoints of the ductot-
cholecystectomy with cystic duct ligation is performed, fol- omy and proximal and distal aspects of the duodenotomy, and
lowed by ligation of the CBD or common hepatic duct (hepati- the third is placed between the distal aspect of the CBD incision
cojejunostomy) at the level of planned transection. The duct is and midpoint of the superior lip of the duodenotomy. Traction
opened at the level of the planned anastomosis, and the endo- on the three corner sutures reorients the distal portion of the
biliary stent, if present, is removed. If desired, bile cultures may ductotomy transversely, facilitating placement of the posterior
also be collected at this time. The remainder of the CBD is row of absorbable sutures between the superior edge of the
transected, and the distal stump is oversewn with 3-0 absorb- duodenotomy and the distal half of the CBD. Full-thickness
able suture. It is important to identify healthy, well-vascularized sutures are made to approximate the ductal mucosa to duode-
duct proximal to the level of injury or pathology to avoid isch- nal mucosa. A fourth corner suture can then be placed between
emic stricture. Similarly, care should be taken to avoid excessive the proximal end of the ductotomy and the midpoint of the
circumferential dissection of the duct, as this can compromise anterior duodenal wall. Retraction of this fourth corner suture
its blood supply. A 50- to 70-cm Roux-en-Y limb of jejunum tents the remaining anterior wall of the anastomosis forward,
is passed retrocolic and to the right of the middle colic vessels facilitating placement of the remaining anterior row of sutures
and positioned to reside adjacent to the proximal bile duct in (Fig. 31.14). As before, use of a single row of sutures minimizes
a tension-free manner. The posterior wall of the duct is sutured the risk of anastomotic narrowing.
to the jejunum with a running 3-0 or 4-0 absorbable suture.
The tail of the suture and needle are left intact. A jejunotomy Cholecystoduodenostomy and Cholecystojejunostomy
is fashioned along the duct, and single interrupted 3-0 or 4-0 Loss common approaches to biliary bypass are cholecystoduo-
absorbable sutures are used to approximate the jejunal mucosa denostomy or cholecystojejunostomy. The cholecystoenteric
to the duct mucosa with the knots tied on the inside of the bypass is relatively easy to construct, but long-term patency
lumen. The anterior portion of the anastomosis is then com- rates are suboptimal compared with maneuvers that directly
pleted using the running suture used to construct the posterior decompress the extrahepatic biliary ducts. As a result, use of
row (Fig. 31.12). Due to the small lumen, the anastomosis is cholecystoduodenostomy or cholecystojejunostomy is limited
II
I
III IV
FIGURE 31.12. A, Technique for end-to-side anastomosis of the bile duct below the hilus to jejunum. I, A 3-0 Vicryl suture is used, and the serosa
of the jejunum is sutured to the full thickness of the bile duct. II, This suture is developed as the posterior wall of the bile duct is attached to the
jejunal serosa. The dotted line marks the point of incision in the jejunum, which is made after the posterior layer is attached. III, The suture is now
developed either as a continuous or an interrupted suture on the anterior layer. The posterior layer of the jejunal mucosa is not sutured directly to
the bile duct mucosa. Several interrupted sutures may be inserted before completing the anterior layer, however, to approximate the mucosa (inset).
IV, The anastomosis is completed. The inset shows the posterior layer with mucosal apposition. B, Alternatively, the jejunum may be opened, and
a mucosa-to-mucosa anastomosis may be performed with a continuous polydioxanone suture as illustrated.
FIGURE 31.13. After the gallbladder is removed, the common bile duct is
removed through a conventional longitudinal incision following a Kocher inci-
sion (1), freeing the lateral duodenum around the common duct. Routine
common duct exploration is carried out. If indications for a choledochoduo-
denostomy exist, the anastomosis is performed. The incision in the common
duct (2) is extended to 2.5 cm by direct measurement. In almost all instances,
the incision in the duct carries into the common hepatic duct. The incision in
the postbulbar duodenum (3) is slightly smaller because the stoma in the
duodenum stretches to approximate the choledochal incision.
FIGURE 31.14. A, Each side of the choledochoduodenostomy

D is bisected by suture (A and B) of absorbable material (chromic
C
catgut or polyglycolic acid) that passes from the end of the duo-
denal incision through the midpoint of the choledochal incision.
B
Likewise, the duodenal incision is bisected by a suture through the
posterior wall of the duodenal incision and the lower apex of the
A B
choledochal incision (C). These stitches convert part of the longi-
A tudinal choledochotomy incision into a transverse ostium. The lax
approximation of the duodenal and choledochal incisions occurs,
with the duodenum mobilized, by placing tension on a lateral stay
suture (A or B) and the middle stay suture (C). B, Sutures may be
placed to complete the posterior suture line, approximating the
common bile duct to the posterior duodenal incision. After place-
ment of the sutures, they are tied so that the knots are within the
lumen. The anterior wall is similarly approximated using a suture
bisecting the anterior duodenal incision (D) and through the original
apex of the bile duct incision. C, With this bisecting suture
(D) tented forward, each of the segments between the tied lateral
stay suture and this anterior suture is similarly approximated using
A B interrupted sutures with the knots tied on the outside. The anasto-
mosis is completed by completing the third segment of this triangle
with sutures placed (E) between the remaining lateral stay suture
and the bisecting, anterior suture (D). It is important in the place-
D ment of these last sutures that they do not catch the posterior
suture line. The benefit of placing all the sutures in one line of the
triangular closure and tying them all after placement is that it allows
an internal inspection before the lumen of the choledochoduode-
nostomy is obscured. A single row of sutures is all that is used. A
A second row would do nothing but decrease the choledochoduo-
denostomy orifice size and should be avoided. The sutures
should be placed close enough for a bile-tight approximation.
E Digital pressure on the duodenum or the common duct should
B give no evidence of leakage. D, The completed anastomosis
allows a thumb-sized defect to be palpated through the duodenal
tongue that has been brought on to the common bile duct and
common hepatic duct. The anastamosis may be drained or not,
according to preference (the leak rate is 1%). The presence of a
closed-suction drain (Jackson-Pratt type) obviates the need for a
subsequent percutaneous drainage catheter if this uncommon
complication occurs.
C D
A B
C
FIGURE 31.15. Technique of cholecystojejunostomy. A, The cystic duct should join the common bile duct above the tumor. If it enters at the
level of the tumor (dashed line), the procedure is contraindicated. B, The posterior layer of the anastomosis is performed with a running suture
between the openings at the fundus of the gallbladder and the jejunum. C, The anterior layer of the side-to-side anastomosis is completed. (We do
not use an associated enteroenterostomy.)
to circumstances of advanced malignancy that require simple performed and the gallbladder evacuated of stones and bile; a
operative interventions and only short-term palliation (Dayton, bile specimen can be sent for analysis. Continuity with the
et al, 1980). Cholecystoenterostomy may be suitable in situa- common hepatic duct is confirmed, and a corresponding
tions where major tumoral obstruction obscures access to the enterotomy mirroring the cholecystotomy is fashioned. An
porta hepatis; however, the obstruction must not extend to the anterior row of sutures is then placed to complete the anasto-
level of the cystic duct insertion. The presence of cholelithiasis mosis (Fig. 31.15).
is another consideration, as significant stone burden within the
gallbladder makes this operative strategy less attractive. OUTCOMES
Operatively, the gallbladder and cystic duct are evaluated
to ensure their suitability for biliary decompression; specifi- A recent single institution retrospective analysis of 45 patients
cally, a patent cystic duct is required for this technique to undergoing reconstruction after biliary injury measured a
provide effective drainage. To construct a cholecystoduodenos- postoperative biliary fistula rate of 3% and a biliary stricture
tomy, a Kocher maneuver is used to provide enough duodenal rate of less than 5% over 4 years (Winslow et al, 2009) (see
mobility for a tension-free anastomosis. If cholecystojejunos- Chapter 42). Other analyses have also confirmed the safety
tomy is performed, the most proximal loop of jejunum that will and longevity of biliary decompression, with low rates of
easily lie adjacent to the gallbladder is selected; a Roux-en-Y fistula and stricture formation necessitating subsequent opera-
is not routinely performed. The gallbladder fundus is secured tive intervention (Chapman et al, 1995; Jarnagin et al, 1998;
to the antimesenteric border of duodenum or jejunum with Murr et al, 1999; Tocchi et al, 1996). Although the limited
interrupted 3-0 absorbable sutures. A cholecystotomy is number of patients requiring biliary-enteric bypass prohibits
comparative analysis of the various techniques, the larger decompression, cholangiocarcinoma developed in 5.8% of
series demonstrate low perioperative morbidity and mortality patients after transduodenal sphincteroplasty, 7.6% of patients
and adequate long-term patency. In patients undergoing after choledochoduodenostomy, and 1.9% of patients after
bypass for benign disease, consideration should be given to hepaticojejunostomy after an interval of 132 to 218 months
prolonged clinical monitoring, as there appears to be both a (Tocchi et al, 2001).
risk of delayed stricture and an elevated risk of cholangiocar-
cinoma. In a review of 1003 patients undergoing biliary References are available at expertconsult.com.
Chapter 31 Bile duct exploration and biliary-enteric anastomosis 548.e1
REFERENCES Pottakkat B, et al: Surgical management of patients with post-

cholecystectomy benign biliary stricture complicated by atrophy-
Babel N, et al: Iatrogenic bile duct injury associated with anomalies of hypertrophy complex of the liver, HPB (Oxford) 11:125–129, 2009.
the right hepatic sectoral ducts: a misunderstood and underappreci- Rogers SJ, et al: Prospective randomized trial of LC + LCBDE vs
ated problem, HPB Surg 2009:153269, 2009. ERCP/S + LC for common bile duct stone disease, Arch Surg 145:
Blumgart LH: Surgical approaches to the left hepatic duct, Langenbecks 28–33, 2010.
Arch Chir 370:235–249, 1987. Strasberg SM, et al: Isolation of right main and right sectional portal
Blumgart LH, Kelley CJ: Hepaticojejunostomy in benign and malig- pedicles for liver resection without hepatotomy or inflow occlusion,
nant high bile duct stricture: approaches to the left hepatic ducts, J Am Coll Surg 206:390–396, 2008.
Br J Surg 71:257–261, 1984. Suc B, et al: Surgery vs endoscopy as primary treatment in symptom-
Chapman WC, et al: Postcholecystectomy bile duct strictures: manage- atic patients with suspected common bile duct stones, Arch Surg
ment and outcome in 130 patients, Arch Surg 130:597–604, 1995. 133:702–708, 1998.
Clayton ESJ, et al: Meta-analysis of endoscopy and surgery versus Targarona EM, et al: Randomised trial of endoscopic sphincterotomy
surgery alone for common bile duct stones with the gallbladder in with gallbladder left in situ versus open surgery for common bile
situ, Br J Surg 93:1185–1191, 2006. duct calculi in high-risk patients, Lancet 347:926–929, 1996.
Costi R, et al: Diagnosis and management of choledocholithiasis in the Tocchi A, et al: The long-term outcome of hepaticojejunostomy in the
golden age of imaging, endoscopy and laparoscopy, World J Gastro- treatment of benign bile duct strictures, Ann Surg 224:162–167,
enterol 20:13382–13401, 2014. 1996.
Dayton MT, et al: Efficacy of the gallbladder for drainage in biliary Tocchi A, et al: Late development of bile duct cancer in patients who
obstruction, Arch Surg 115:1086–1089, 1980. had biliary-enteric drainage for benign disease: a follow-up study of
Dasari BVM, et al: Surgical versus endoscopic treatment of bile duct more than 1000 patients, Ann Surg 234:210–214, 2001.
stones (review), Cochrane Database Syst Rev (12):CD003327, 2013. Verbesey JE, Birkett DH: Common bile duct exploration for choledo-
Gurusamy KS, et al: T-tube drainage versus primary closure after cholithiasis, Surg Clin North Am 88:1315–1328, 2008.
open common bile duct exploration, Cochrane Database Syst Rev (6): Voyles CR, et al: Carcinoma of the proximal extrahepatic biliary tree
CD005640, 2013. radiologic assessment and therapeutic alternatives, Ann Surg 197:
Jarnagin WR, et al: Intrahepatic biliary enteric bypass provides effective 188–194, 1983.
palliation in selected patients with malignant obstruction at the Wan XJ, et al: Success rate and complications of endoscopic extraction
hepatic duct confluence, Am J Surg 175:453–460, 1998. of common bile duct stones over 2 cm in diameter, Hepatobiliary
Launois B, Jamieson GG: The posterior intrahepatic approach for Pancreat Dis Int 10:403–407, 2011.
hepatectomy or removal of segments of the liver, Surg Gynecol Obstet Winslow ER, et al: “Sideways”: results of repair of biliary injuries using
174:155–158, 1992. a policy of side-to-side hepatico-jejunostomy, Ann Surg 249:426–
Murr MM, et al: Long-term results of biliary reconstruction after lapa- 434, 2009.
roscopic bile duct injuries, Arch Surg 134:604–610, 1999.
PART 5 Biliary Tract Disease
SECTION I Inflammatory, Infective, and Congenital

A. Gallstones and Gallbladder
CHAPTER 32
Natural history of gallstones and
asymptomatic gallstones
Pierre F. Saldinger and Omar Bellorin-Marin
HISTORY corresponding to 20 to 25 million Americans who have or will

have cholelithiasis (Stinton et al, 2012).
Gallstones have been recognized since antiquity and have even With 6.3 million affected men and 14.2 million affected
been found during autopsies of Egyptian mummies. The Greek women between the ages of 20 and 74 years, and at an annual
physician Trallianus described calculi within radicles of a cost of $6.5 billion, gallstones is the most expensive digestive
human liver (Glenn & Grafe, 1966). During the sixteenth tract disorder today in the United States, having increased more
century, Vesalius and Fallopius described gallstones in the gall- than 20% over the last 3 decades (Shaffer, 2006; Stinton,
bladders of dissected human bodies (Schwartz, 1981). Ferne- 2012). The mortality rate for gallstones disease is relatively low
lius is among the first authors who took note of them when at 0.6%, with an estimated 1092 gallstone-related deaths in the
examining stool samples (Coe, 1757). Attempts to treat gall- United States in 2004. This represents a marked decrease in
stones started in the eighteenth century, based on in vitro mortality from more than 5000 deaths in 1950, with a fall of
experiments by Percival dissolving gallstones by using water more than 50% between 1979 and 2004 (Everhart, 2009).
impregnated with fixed air (Percival, 1775). In 1788 in his book Geography plays an important role in the prevalence of chole-
“Considerations on Bilious Diseases: and Some Particular Affections lithiasis and also the type of stone. Overall, cholesterol stones
of the Liver and the Gallbladder,”Andree recommended a light are more common in the Western World (>85%), and brown-
infusion of ginger, orange peel, and warm water on an empty pigmented stones are predominant in Asia, primarily as a result
stomach for those suffering with symptomatic gallstones. The of bacterial infections, biliary parasites, and stasis from partial
first attempt at surgical treatment of gallstones is ascribed to biliary obstruction. This likely reflects combined differences in
John Stough Boobs, considered the father of gallbladder environmental, dietary, and genetic factors (Stinton, 2012).
surgery, who successfully performed a cholecystotomy and The highest rates of gallstones are seen in Native American
extraction of gallstones in 1867 (Ellis, 2009). In 1878, James populations. North American Indians have rates of gallstones
Marion Sims performed the first planned cholecystostomy for at 64.1% in women and 29.5% in men. Among Mapuche
biliary drainage as a treatment option for cholecystitis (Fowler, Indians of Chile, there is also a high prevalence of gallstones,
1900). In 1882, after animal experimentation and studies on at 49.4% in women and 12.6% in men (Stinton, 2012). In the
the cadaver, Carl Langenbuch became convinced that surgical United States after the Native Americans, the highest incidence
removal of the gallbladder did not represent a technically dif- of gallstones among women is in Hispanics (27%), non-
ficult operation and should lead to a permanent cure. He per- Hispanic whites (17%), and non-Hispanic blacks (14%). Men
formed the first successful cholecystectomy, setting the path for have a lower risk in all ethnic groups, with similar prevalence
therapeutic surgical intervention for symptomatic cholelithiasis (Everhart, 2001). More detailed information about geographic
(Halpert, 1932). One century later, in 1985, Eric Muhe per- disparities in the prevalence of gallstone disease can be found
formed the first laparoscopic cholecystectomy by using a in Table 32.1.
custom-made laparoscope called the “galloscope,” a hemoclip
and a pistol-grip scissors. This represented one significant Risk Factors for Gallstones Development
advance in surgery that opened the modern era in the surgical Gallstone disease results from a complex interaction of genetic
treatment of gallstones (Walker, 2001). and environmental factors. This is confirmed by an altered
incidence of gallstone detection when people from diverse
EPIDEMIOLOGY geographic areas travel and settle in areas with a different
dietary regimen, level of physical activity, and hygiene. The
Gallstones represent a significant public health problem in common mechanisms of gallstone formation include choles-
modern societies, affecting 10% to 15% of the adult population, terol hypersecretion, alteration in intestinal bile salt/cholesterol
551
552 PART 5 BILIARY TRACT DISEASE Section I Inflammatory, Infective, and Congenital
TABLE 32.1 Geographic Prevalence of Gallstones as Assessed by Ultrasound

Study Geographic Area Male Population (%) Female Population (%)
North America
Everhart et al, 1999, 2001, 2002 American Indians 64.1 29.5
Hispanic 8.9 27
Shaffer, 2005 Non-Hispanic White 8.6 16.6
Non-Hispanic Black 5.3 13.9
South America
Miguel et al, 1998; Covarrubias Chilean Indians 12.6 49.4
et al, 1995; Moro et al, 2000
Palermo et al, 2013 Argentina 18.2 25
Europe
Glambek et al, 1987 Norway 17 21
Loria et al, 1994 Italy 2.7 8.4
Mellström et al, 1988 Sweden NR 27
Kratzer et al, 1998 Germany 5.8 6.3
Martinez et al, 1997 Spain 7.8 11.5
Asia
Khuroo et al, 1989 Kashmir 3.07 9.6
Unisa et al, 2011 India 2 5.6
Chen et al, 1998 Taiwan 5.3 2.4
Shen et al, 2014 China 9.3 9.8
Kono et al,1992 Japan 3.6 NR
Chapman et al, 2000 New Zealand 18.1 23.1
Africa
Walker et al, 1989; South Africa NR 10
Bagi Abdel, 1991 Sudan 5.6 5.1
NR, Not reported.
absorption, and gallbladder hypokinesia, which leads to bile can be in part explained by a hormonal effect; estrogen decreases
cholesterol supersaturation and nucleation (see Chapter 8). bile salts secretion and increases cholesterol, whereas progestins
Bacterial infection and an increase in bilirubin load play a role act by impairing gallbladder emptying, causing stasis. Preg-
in the development of black and brown gallstones. The follow- nancy is associated with up to 30% risk of developing biliary
ing is an overview of the most common risk factors for the sludge. Oral contraceptive use and low-dose estrogen therapy
development of gallstones. in postmenopausal women also increase the risk of gallstone
formation (Hulley et al, 1998; Maringhini, et al, 1993).
Age
The incidence of gallstones increases with age across all ethnic Obesity, Diet, Physical Activity, and Weight Loss
groups, becoming 4 to 10 times more likely in individuals older Obesity is a well-established major risk factor for the develop-
than 40 years, with a very low rate among infants and children ment of gallstones, with an incidence of 25% in morbidly obese
(Chen et al, 1998; Einarsson et al, 1985). In a population- patients (Li et al, 2009).The association with simple obesity is
based study, the prevalence of gallstones in children was 1.9% important in females, whereas in males it is mostly associated
(Wesdorp et al, 2000). As previously mentioned, some specific with intraabdominal (central) obesity and metabolic syndrome
populations, such as American Indians (Pima), have an (De Santis et al, 1997; Heaton et al, 1991; Maclure et al, 1989;
increased incidence of gallstones (up to 70% by 30 years of Tsai et al, 2004).
age), implicating hereditary metabolic factors (Sampliner et al, The frequency of detection of gallstones has almost doubled
1970). Pediatric populations with chronic hemolytic condi- in the last 60 years, in parallel with a changed composition of
tions, such as sickle cell anemia, also represent another distinct stones from pigment to cholesterol. In humans, most choles-
group in whom gallstones develop early. In a study of 82 chil- terol found in gallstones is from the diet because de novo
dren (0 to 18 years of age) with cholelithiasis, gallstone disease hepatic synthesis is only about 20% (Paigen & Carey, 2002).
was associated to a hemolytic disease in 39% of cases (Wesdorp This emphasizes the importance of high caloric and carbohy-
et al, 2000). drate intake as risk factors for the development of gallstones in
modern society.
Gender Bariatric patients are at a high risk for the development of
Female gender is a risk factor for developing gallstones, surpass- cholelithiasis after weight-reduction surgery. Sludge and gall-
ing males in the incidence of gallstones and the chance of having stones form in 30% to 71% of patients following bariatric
surgery by 2 : 1 or 3 : 1 in most studies (see Table 32.1). This procedures, with weight loss more than 1.5 kg/week being a risk
A. Gallstones and Gallbladder Chapter 32 Natural history of gallstones and asymptomatic gallstones 553
TABLE 32.2 Genes Involved in Gallstone Pathogenesis

Genes and Encoded Proteins Mechanism Significance in Humans
ABCB4, ATP-binding cassette transporter B4 Decreased biliary phospholipid secretion +

ABCB11, ATP-binding cassette transporter B11 Decreased bile salts secretion +
ABCG5/ABCG8/ABCA1/ABCB1, ATP-binding cassette Increased cholesterol intestinal efflux ?
G5/G8/A1/B1
APOA1, apolipoprotein A1 Increased biliary cholesterol and bile acid +/?
output
APOB, apolipoprotein B Decreased hepatic VLDL secretion +/?
APOE, apolipoprotein E Increased intestinal cholesterol absorption, +/?
decreased bile acid secretion
MUC1/MUC2, membrane bound mucin 1/2 Promotes gallbladder cholesterol absorption +/?
and impairs gallbladder motility
CCKAR, cholecystokinin A receptor Gallbladder hypomotility ?
ADRB3, β3-adrenergic receptor Gallbladder hypomotility ?
CETP, cholesteryl ester transfer protein, plasma Increased HDL metabolism +
CFTR, cystic fibrosis transmembrane conductance regulator Increased biliary bilirubin secretion and bile +
salt excretion in feces, decreased bile pH
ANK1, EPB42, SPTA1, SPTB, SLC4A1, associated with Increased bilirubin load +
hereditary spherocytosis
HBB, hemoglobin-β, associated with sickle cell anemia and Increased bilirubin load +
thalassemia
AK1, G6PD, GPI, GSR, PGK1, PKLR, associated with Increased bilirubin load +
erythrocyte enzyme deficiencies
CYP39A1, oxysterol 7-α-hydroxylase Induced hypercholesterolemia +
UGT1A1, Gilbert syndrome–associated protein Increased bilirubin load +
NPC1L1, NPC1 (Niemann-Pick disease, type C1)–like Reduced cholesterol intestinal absorption +
protein 1
+, Clear evidence; ?, potentially related but no definitive evidence; ATP, adenosine triphosphate; HDL, high-density lipoprotein; VLDL, very-low-density lipoprotein.
factor (Shiffman et al, 1991; Weinsier et al, 1995). The stones control subjects (van der Linden et al, 1973), suggesting a
are likely to become apparent 6 weeks after surgery and symp- familial component for gallstone formation. Concordance rates
tomatic in 7% to 16% of patients. This presentation is associ- of gallstones in monozygotic twins of both genders are signifi-
ated with weight loss that exceeds more than 25% of initial cantly higher than for dizygotic twins, representing conclusive
body weight (Li et al, 2009). evidence for a genetic predisposition in developing gallstone
Patients receiving total parenteral nutrition (TPN) are disease (Katsika et al, 2005). Hepatic cholesterol hypersecre-
another group with a high incidence of gallstone and sludge tion, resulting in bile supersaturation and cholesterol stone
development. Biliary sludge appears as early as 5 to 10 days formation, may be caused by increased hepatic uptake and
after fasting, with all patients developing sludge after 6 weeks synthesis of cholesterol, decreased hepatic synthesis of bile
of TPN therapy. Sludge usually resolves within 4 weeks of ces- salts, decreased hepatic synthesis of cholesterol esters for inclu-
sation of TPN when the patient has resumed oral intake. After sion in very-low-density lipoprotein (VLDL), and derange-
3 months of TPN therapy, approximately 45% of patients will ments in the intestinal absorption of cholesterol and bile
develop gallstones (Messing et al, 1983; Murray et al, 1992; salts. Genes controlling hepatic cholesterol metabolism have
Shaffer, 2001). Low levels of physical activity are associated been implicated in the pathogenesis of gallstone formation in
with development of gallstones in large observational and popu- transgenic/knockout mouse models and in humans (Table
lation studies. These patients also typically have related obesity 32.2). Excess cholesterol load in the smooth muscle of the
and typical dietary risk factors (Leitzmann et al, 1999). Finally, gallbladder can affect signaling via the cholecystokinin A recep-
patients receiving long-term somatostatin analogue therapy tor and contribute to the development of gallstones by inducing
for various disorders (acromegaly, neuroendocrine tumors) bile stasis. Intestinal cholesterol/bile salt absorption is tightly
are at increased risk for gallstone formation as a result of controlled, and genes responsible for cholesterol absorption
reduced gallbladder motility and reduced bile flow (Bornschein (NPC1L1, Niemann-Pick C1–like protein) or cholesterol efflux
et al, 2009). (ABCG5/G8) have been implicated in the development of gall-
stones in certain populations because of intestinal bile salt loss.
Genetics Single-gene mutations resulting in hemolysis increase the inci-
Population studies indicate a 30% genetic component in the dence of black stones because of an excess of unconjugated bile
development of gallstone disease. Family and twin studies have acids forming polymerized calcium bilirubinate stones. A
contributed to the understanding of gallstone formation. Cho- summary of genes found to be involved in gallstone pathogen-
lesterol gallstones are found at a ratio of 3 : 1 in family members esis can be found in Table 32.2 (Chuang et al, 2013; Grünhage
of the affected person, compared with spouses or unrelated & Lammert, 2006).
colleagues (1985) followed 135 asymptomatic men and women

Other Risk Factors with gallstones who were subscribers to the Health Insurance
Certain medical conditions and medications can cause gall- Plan of Greater New York. Of these subjects, 10% developed
stones by multiple mechanisms. Cirrhosis is a risk factor for symptoms, and 7% required cholecystectomy over a median
gallstones formation (black stones) with an overall prevalence follow-up of 46 months. A placebo group of 193 asymptomatic
of 25% to 30%. Cirrhosis can induce decreased bile salt syn- patients who were part of a chemical dissolution trial were fol-
thesis and malabsorption, chronic hemolysis, and a hyperestro- lowed for 24 months (Thistle et al, 1984). In this study, 31%
genic state, all of which contribute to gallstone formation of these patients eventually developed biliary pain, a higher
(Conte et al, 1999). Inflammatory bowel disease, ileal resec- than previously reported incidence. This high rate of symptoms
tion, and cystic fibrosis can also interfere with bile salt absorp- may have been related to the intense surveillance or require-
tion, resulting in bile that is supersaturated with cholesterol. ment that participants have no symptoms only for the 12
Estrogen and thiazide diuretics can increase biliary cholesterol months preceding the trial. Cucchiaro and colleagues (1990)
secretion, and progesterone can impair gallbladder motility. followed 125 asymptomatic patients for 5 years. Fifteen patients
In addition to causing gallbladder hypomotility, octreotide developed symptoms during that time, and 2 required emergent
increases the amount of reabsorbed deoxycholic acid and surgery for gallstone complications; 54 patients died during that
results in cholesterol supersaturation. Ceftriaxone is excreted period because of malignancies, cardiovascular disease, or renal
in bile in significant amounts and can cause biliary pseudoli- insufficiency, but none died of gallstone disease. Friedman and
thiasis and sludge, which usually disappear after antibiotic coworkers (1989) followed 123 asymptomatic patients for 20
therapy is discontinued (Lambou-Gianoukos & Heller, 2008). years and found that 6% of patients developed severe gallstone-
related symptoms during the first 5 years after diagnosis. Hall-
NATURAL HISTORY OF GALLSTONES destamm and colleagues (2004) followed 120 asymptomatic
patients, 14 of whom required treatment for symptoms or
Gallstones do not cause any symptoms in up to 80% of carriers. gallstone-related complications over a median observation
The risk of becoming symptomatic is 2% to 3% per year with period of 87 months. The authors calculated that the cumula-
a cumulative risk of 10% at 5 years. A smaller proportion (1% tive risk of requiring treatment during the first 5 years after
to 2% per year) will develop other complications related to detection of gallstones was 7.6%. Davide and colleagues (2010)
gallstones, such as acute pancreatitis or choledocholithiasis. followed 580 asymptomatic patients with gallstones for a period
Given the low risk of symptoms and gallstone-related complica- of 8.7 years. At the end of the period, 453 (78%) of patients
tions, nonoperative management is recommended for patients remained asymptomatic, whereas 10.5% developed mild symp-
with asymptomatic gallstones (Attili et al, 1995; Friedman, toms (abdominal pain not requiring rest), and 11.4% devel-
1993; Gibney, 1990; Gracie & Ransohoff, 1982; McSherry oped severe symptoms (abdominal pain requiring rest), with an
et al, 1985; Sakorafas et al, 2007; Thistle et al, 1984). Half of overall complication rate of 3% (presence of jaundice or fever).
the patients with symptoms develop a second attack within a Symptoms disappeared in 58% of the mildly symptomatic
year, but in 30% of all cases, there is only one symptomatic patients, and only 17.4% of patients underwent cholecystec-
attack. The analysis of the natural history of gallstones started tomy. Of the severe symptomatic patients, symptoms dis
with a landmark study by Gracie and Ransohoff (1982), which appeared in 52%, with 41.3% of patients undergoing
followed 123 University of Michigan faculty members—110 cholecystectomy. The authors concluded that patients with
men and 13 women—with asymptomatic gallstones at screen- symptomatic gallstones can have a relatively benign natural
ing for 15 years. At 5, 10, and 15 years of follow-up, 10%, 15%, history. However, given the risk of symptoms and/or complica-
and 18%, respectively, had become symptomatic. None of the tions, cholecystectomy was recommended in this group. A
subjects developed complications before the onset of typical summary of the natural history of gallstones as reported by
symptoms. The approximate rate at which the subjects devel- various authors is shown in Table 32.3.
oped biliary pain was 2% per year for the first 5 years, with a Patients with diabetes may have a higher incidence of gall-
subsequent decrease over time. Three patients in this study stones from the indirect effects of the metabolic syndrome,
developed biliary complications, all of which were preceded by obesity, and a family history of gallstones. No data show higher
biliary pain. rates of symptoms or complications among diabetics with
Attili and colleagues (1995) also followed 151 subjects with asymptomatic cholelithiasis. Therefore prophylactic cholecys-
gallstones over 10 years. Of those subjects, 33 had symptoms, tectomy in asymptomatic gallstones carriers with diabetes is not
and 118 were asymptomatic at the beginning of the study. The recommended (Sleisenger & Fordtran, 1993).
cumulative probability of developing biliary colic was 12% at 2 There are several instances where prophylactic cholecystec-
years, 17% at 4 years, and 26% at 10 years. The cumulative tomy may be considered. The size of gallstones is thought to
probability of developing complications by 10 years was 3% in be linked to a higher risk for the development of gallbladder
the initial asymptomatic group and 7% in the symptomatic cancer, but this is controversial (see Chapter 49). Cholecystec-
group. The authors concluded that the natural history of gall- tomy is possibly indicated in some populations carrying a high
stone disease might not be as benign as previously thought. risk of gallbladder cancer (Chilean Hispanics, Mapuche Indians,
The results of several other natural history studies have been and Maoris from Easter Island), even when associated with
reported. In a Japanese study, Wada and Imamura (1993) asymptomatic cholelithiasis (Grimaldi et al, 1993; Randi et al,
found that one third of 1850 patients with cholelithiasis were 2006). Prophylactic cholecystectomy also could be indicated in
symptomatic. Of the remaining 680 asymptomatic patients, patients with active hemolytic anemias, such as sickle cell
20% became symptomatic over a median follow-up of 13 years. disease, where symptomatic gallstones can be difficult to
Patients older than 70 years were more likely to become symp- distinguish from complications such as infarction of intraab-
tomatic than patients younger than 70 years. McSherry and dominal viscera. Another relative indication for prophylactic
A. Gallstones and Gallbladder Chapter 32 Natural history of gallstones and asymptomatic gallstones 555
TABLE 32.3 Natural History of Gallstones

Study N No. Years Biliary Pain (%) Biliary Complications (%)
Gracie & Ransohoff, 1982 123 15 18 2

Attili et al, 1995 118 10 26 3
Juhasz et al, 1994 110 6 22* —
Wada & Imamura, 1993 680 10-17 20 —
McSherry et al, 1985 135 4 17 0
Friedman et al, 1989 123 5 — 6
Thistle et al, 1984 193† 2 31 —
Cucchiaro et al, 1990 125 5 12 2
Angelico et al, 1997 426 10 38 —
Halldestamm et al, 2004 120 7 12‡ —
Davide et al, 2010 580 8.7 26.9 3
*Patients requiring cholecystectomy.
†
Asymptomatic for 12 months preceding the study.
‡
Symptoms or complications.
cholecystectomy is before or during renal transplantation or 10-year period. As a result, cholecystectomy to prevent the
during bariatric surgery (Bonatsos et al, 2001; Graham et al, development of gallbladder cancer in patients with gallstones is
1995; Shiffman et al, 1993). not recommended. Some populations, such as Pima Indians,
Chileans, and Pakistani women with gallstones, have an unusu-
ally high incidence of gallbladder cancer (12 to 21 per 100,000)
GALLSTONES, PORCELAIN GALLBLADDER, (Goldin & Roa, 2009). In these unique populations, prophy-
AND GALLBLADDER CARCINOMA lactic cholecystectomy to prevent the development of gallblad-
(SEE CHAPTER 49) der cancer is probably justified.
Calcification of the gallbladder is encountered in up to
Gallstones are closely associated with gallbladder cancer, with 0.8% of all gallbladder pathology specimens. Porcelain gall-
a pooled relative risk (compared with patients without gallbladder or diffuse gallbladder calcification was long considered
stones) of 4.9 and a calculated risk of 0.02% per year (Randi to be a significant risk factor for the presence of gallbladder
et al, 2006). Chemical exposure, tobacco, family history of cancer, with a risk of cancer ranging from 12% to 62%
gallbladder cancer, and duration of cholelithiasis are all poten- (Cunningham & Alexander, 2007; Stephen & Berger, 2001;
tial risk factors for the development of gallbladder cancer (Kajal Towfigh et al, 2001). Two studies from the United States ana-
et al, 2012; Miyazaki et al, 2008). Gallstones are present in lyzed large databases of cholecystectomy specimens and
60% to 90% of gallbladder cancer cases (Goldin & Roa, 2009), reported a much lower incidence of cancer in porcelain gall-
although no direct causal relationship between gallstones and bladder than in previous studies (Stephen & Berger, 2001;
gallbladder cancer has been proven. Gallstones larger than Towfigh et al, 2001). In a more recent study, calcification-
3 cm have been associated with a higher risk of gallbladder associated gallbladder cancer was found in only 6% compared
cancer; however, the presence of large gallstones may simply with 1% in a matched cohort of patients without gallbladder
reflect the duration of cholelithiasis and inflammation rather wall calcification (Schnelldorfer, 2013). Based on these data, at
than a direct causal relationship with cancer. Also, a significant least in Western populations, the incidence of cancer in porce-
difference in prevalence of gallbladder cancer worldwide points lain gallbladder should be considered to be increased but not
to the influence of other factors. The risk of developing carci- as high as reported previously. Given the small but significantly
noma has been estimated to be 1% of calculous gallbladders higher risk of gallbladder cancer associated with porcelain gall-
20 years after the initial diagnosis of gallstones, with the risk bladder, compared with the general population, and the lethal-
increased mainly in men (Maringhini et al, 1987); however, ity of this cancer, cholecystectomy is generally indicated in
given the high incidence of gallstones and the rarity of gallblad- patients with a calcified gallbladder.
der cancer, this estimate would seem to be high. In some series,
no patients with gallstones developed gallbladder cancer over a References are available at expertconsult.com.
A. Congenital Disorders Chapter 32 Natural history of gallstones and asymptomatic gallstones 555.e1
REFERENCES Grimaldi CH, et al: Increased mortality with gallstone disease: results
of a 20-year population-based survey in Pima Indians, Ann Intern
Andree J: Considerations on bilious diseases: and some particular affections Med 118(3):185–190, 1993.
of the liver and the gallbladder, London, 1788, J Murray and W Grünhage F, Lammert F: Gallstone disease: pathogenesis of gallstones—
Lowndes. a genetic perspective, Best Pract Res Clin Gastroenterol 20(6):997–
Angelico F, et al: Ten-year incidence and natural history of gallstone 1015, 2006.
disease in a rural population of women in central Italy. The Rome Halldestamm I, et al: Development of symptoms and complications in
Group for the Epidemiology and Prevention of Cholelithiasis individuals with asymptomatic gallstones, Br J Surg 91:734–738,
(GREPCO), Ital J Gastroenterol Hepatol 29:249–254, 1997. 2004.
Attili AF, et al: The natural history of gallstones: the GREPCO experi- Halpert B: Fiftieth anniversary of the removal of the gallbladder, Arch
ence. The GREPCO Group, Hepatology 21(3):655–660, 1995. Surg 25:178–182, 1932.
Bagi Abdel M, et al: Prevalence of gallbladder disease in Sudan: first Heaton KW, et al: Why do men get gallstones? Role of abdominal
sonographic study in adult population [abstract], Gastroenterology fat and hyperinsulinaemia, Eur J Gastroenterol Hepatol 3:745–751,
100:A307, 1991. 1991.
Bonatsos G, et al: Laparoscopic cholecystectomy in adults with sickle Hulley S, et al: Randomized trial of estrogen plus progestin for second-
cell disease, Surg Endosc 15(8):816–819, 2001. ary prevention of coronary heart disease in postmenopausal women.
Bornschein J, et al: Octreotide LAR: safety and tolerability issues, Heart and Estrogen/Progestin Replacement Study (HERS) Research
Expert Opin Drug Saf 8(6):755–768, 2009. Group, JAMA 280(7):605–613, 1998.
Chapman BA, et al: Gallstone prevalence in Christchurch: risk factors Jain K, et al: Risk factors for gallbladder cancer: a case-control study,
and clinical significance, N Z Med J 113:46–48, 2000. Int J Cancer 132:1660–1666, 2012.
Chen CY, et al: Age is one of the risk factors in developing gallstone Juhasz ES, et al: Incidental cholecystectomy during colorectal surgery,
disease in Taiwan, Age Ageing 27(4):437–441, 1998. Ann Surg 219:467–472, 1994.
Chuang S, et al: Genetics of gallstone disease, Adv Clin Chem 60:143– Katsika D, et al: Genetic and environmental influences on symptomatic
185, 2013. gallstone disease: a Swedish study of 43,141 twin pairs, Hepatology
Coe T: A treatise on biliary concretions: or, stones in the gall-bladder and 41:1138–1143, 2005.
ducts, London, 1757, D Wilson and T Dorham. Khuroo MS, et al: Prevalence of biliary tract disease in India: a sono-
Conte D, et al: Close relation between cirrhosis and gallstones: cross- graphic study in adult population in Kashmir, Gut 30(2):201–205,
sectional and longitudinal survey, Arch Intern Med 159:49–52, 1999. 1989.
Cucchiaro G, et al: Clinical significance of ultrasonographically Kono S, et al: Prevalence of gallstone disease in relation to smoking,
detected coincidental gallstones, Dig Dis Sci 35:417–421, 1990. alcohol use, obesity, and glucose tolerance: a study of self-defense
Cunningham SC, Alexander HR: Porcelain gallbladder and cancer: officials in Japan, Am J Epidemiol 136:787–794, 1992.
ethnicity explains a discrepant literature? Am J Med 120(4):e17–e18, Kratzer W, et al: Gallstone prevalence in Germany: the Ulm gallblad-
2007. der stone study, Dig Dis Sci 43:1285–1291, 1998.
Covarrubias C, et al: The role of ethnicity and other risk factors for Lambou-Gianoukos S, Heller SJ: Lithogenesis and bile metabolism,
cholelithiasis in a highly prevalent area: cross-sectional and cohort Surg Clin North Am 88(6):1175–1194, vii, 2008.
studies in Chilean and Amerindian Araucanos, Gastroenterology Leitzmann MF, et al: Recreational physical activity and the risk of
108:A1053, 1995. cholecystectomy in women, N Engl J Med 341:777–784, 1999.
Davide F, et al: Natural history of gallstone disease: expectant manage- Li VK, et al: Predictors of gallstone formation after bariatric surgery:
ment or active treatment? Results from a population-based cohort a multivariate analysis of risk factors comparing gastric bypass,
study, J Gastro Hepatol 25:719–724, 2010. gastric banding, and sleeve gastrectomy, Surg Endosc 23:1640–1644,
De Santis A, et al: Gallstones and diabetes: a case-control study in a 2009.
free-living population sample, Hepatology 25(4):787–790, 1997. Loria P, et al: Prevalence rates of gallstone disease in Italy. The Chi-
Einarsson K, et al: Influence of age on secretion of cholesteroland anciano population study, Eur J Epidemiol 10:143–150, 1994.
synthesis of bile acids by the liver, N Engl J Med 313:277–282, 1985. Maclure KM, et al: Weight, diet, and the risk of symptomatic gallstones
Ellis H: John Stough Bobbs: father of gallbladder surgery, Br J Hosp in middle-aged women, N Engl J Med 321(9):563–569, 1989.
Med (Lond) 70:650, 2009. Maringhini A, et al: Gallstones, gallbladder cancer, and other gastro-
Everhart JE, et al: Prevalence and ethnic differences in gallbladder intestinal malignancies: an epidemiologic study in Rochester,
disease in the United States, Gastroenterology 117(3):632–639, 1999. Minnesota, Ann Intern Med 107:30–35, 1987.
Everhart JE: Gallstones and ethnicity in the Americas, J Assoc Acad Maringhini A, et al: Biliary sludge and gallstones in pregnancy: inci-
Minor Phys 12:137–143, 2001. dence, risk factors, and natural history, Ann Intern Med 119(2):116–
Everhart JE, et al: Prevalence of gallbladder disease in American Indian 120, 1993.
populations: findings from the Strong Heart Study, Hepatology 35(6): Martinez C, et al: Prevalence and associated factors for gallstone
1507–1512, 2002. disease: results of a population survey in Spain, J Clin Epidemiol
Everhart JE: Burden of digestive diseases in the United States part 1: 50:1347–1355, 1997.
overall and upper gastrointestinal diseases, Gastroenterology 136:376– McSherry CK, et al: The natural history of diagnosed gallstone disease
386, 2009. in symptomatic and asymptomatic patients, Ann Surg 202(1):59–63,
Fowler J: Surgery of the liver and biliary passages, Brooklyn Med J 1985.
14:929–930, 1900. Mellström D, et al: Gallstones and previous cholecystectomy in 77- to
Friedman GD: Natural history of asymptomatic and symptomatic gall- 78-year-old women in an urban population in Sweden, Scand J
stones, Am J Surg 165(4):399–404, 1993. Gastroenterol 23(10):1241–1244, 1988.
Friedman GD, et al: Prognosis of gallstones with mild or no symptoms: Messing B, et al: Does total parenteral nutrition induce gallbladder
25 years of follow-up in a health maintenance organization, J Clin sludge formation and lithiasis? Gastroenterology 84:1012–1019,
Epidemiol 42:127–136, 1989. 1983.
Gibney EJ: Asymptomatic gallstones, Br J Surg 77(4):368–372, 1990. Miguel JF, et al: Genetic epidemiology of cholesterol cholelithiasis
Glambek I, et al: Prevalence of gallstones in a Norwegian population, among Chilean Hispanics, Amerindians, and Maoris, Gastroenterol-
Scand J Gastroenterol 22(9):1089–1094, 1987. ogy 115(4):937–946, 1998.
Glenn F, Grafe WR Jr: Historical events in biliary tract surgery, Arch Miyazaki M, et al, Japanese Association of Biliary Surgery, Japanese
Surg 93:848–852, 1966. Society of Hepato-Pancreatic Surgery, Japan Society of Clinical
Goldin RD, Roa JC: Gallbladder cancer: a morphological and molecu- Oncology: Risk factors for biliary tract and ampullary carcinomas
lar update, Histopathology 55(2):218–229, 2009. and prophylactic surgery for these factors, J Hepatobiliary Pancreat
Gracie WA, Ransohoff DF: The natural history of silent gallstones: the Surg 15(1):15–24, 2008.
innocent gallstone is not a myth, N Engl J Med 307(13):798–800, Moro PL, et al: Gallstone disease in Peruvian coastal natives and high-
1982. land migrants, Gut 46(4):569–573, 2000.
Graham SM, et al: The utility of prophylactic laparoscopic cholecys- Murray FE, et al: Development of biliary sludge in patients on intensive
tectomy in transplant candidates, Am J Surg 169(1):44–48, discus- care unit: results of a prospective ultrasonographic study, Gut 33(8):
sion 48–49, 1995. 1123–1125, 1992.
555.e2 PART 5 BILIARY TRACT DISEASE Section I Inflammatory, Infective, and Congenital
Paigen B, Carey MC: Gallstones. In King RA, et al, editors: The genetic Sleisenger MH, Fordtran JS, editors: Gastrointestinal disease,
basis for common diseases, London, 2002, Oxford University Press, Philadelphia, 1993, Saunders, pp 973–984.
pp 298–335. Stephen AE, Berger DL: Carcinoma in the porcelain gallbladder: a
Palermo M, et al: Prevalence of cholelithiasis in Buenos Aires, Argen- relationship revisited, Surgery 129(6):699–703, 2001.
tina, Acta Gastroenterol Latinoam 43:98–105, 2013. Stinton LM, Shaffer E: Epidemiology of gallbladder disease: choleli-
Percival T: On the solution of stones of the urinary, and of the gallbladder, thiasis and cancer, Gut Liver 6:172–187, 2012.
by water impregnated with fixed air, London, 1775. Thistle JL, et al: The natural history of cholelithiasis: the National
Randi G, et al: Gallbladder cancer worldwide: geographical distribu- Cooperative Gallstone Study, Ann Intern Med 101(2):171–175,
tion and risk factors, Int J Cancer 118(7):1591–1602, 2006. 1984.
Sakorafas GH, et al: Asymptomatic cholelithiasis: is cholecystectomy Towfigh S, et al: Porcelain gallbladder is not associated with gallblad-
really needed? A critical reappraisal 15 years after the introduction der carcinoma, Am Surg 67(1):7–10, 2001.
of laparoscopic cholecystectomy, Dig Dis Sci 52:1313–1325, 2007. Tsai CJ, et al: Prospective study of abdominal adiposity and gallstone
Sampliner RE, et al: Gallbladder disease in Pima Indians: demonstra- disease in U.S. men, Am J Clin Nutr 80(1):38–44, 2004.
tion of high prevalence and early onset by cholecystography, N Engl Unisa S, et al: Population-based study to estimate prevalence and
J Med 283(25):1358–1364, 1970. determine risk factors of gallbladder diseases in the rural gangetic
Schnelldorfer T: Porcelain gallbladder: a benign process or concern of basin of north India, HPB 13:117–125, 2011.
malignancy? J Gastrointest Surg 17(6):1161–1168, 2013. van der Linden W, Simonson N: Familial occurrence of gallstone
Schwartz SI: Sequence of stones, Contemp Surg 18:9, 1981. disease. Incidence in parents of young patients, Hum Hered 23:123–
Shaffer EA: Gallbladder sludge: what is its clinical significance? Curr 127, 1973.
Gastroenterol Rep 3(2):166–173, 2001. Wada K, Imamura T: Natural course of asymptomatic gallstone disease,
Shaffer EA: Epidemiology and risk factors for gallstone disease: has the Nippon Rinsho 51:1737–1743, 1993.
paradigm changed in the 21st century? Curr Gastroenterol Rep 7:132– Walker AR, et al: Prevalence of gallstones in elderly black women in
140, 2005. Soweto, Johannesburg, as assessed by ultrasound, Am J Gastroenterol
Shaffer EA: Gallstone disease: epidemiology of gallbladder stone 84(11):1383–1385, 1989.
disease, Best Pract Res Clin Gastroenterol 20(6):981–996, 2006. Walker R: The first laparoscopic cholecystectomy, JSLS 5:89–94,
Shen C, et al: Association of cholecystectomy with metabolic syndrome 2001.
in a Chinese population, PLoS ONE 9(2):e88189, 2014. Weinsier RL, Wilson LJ, et al: Medically safe rate of weight loss for the
Shiffman ML, et al: Gallstone formation after rapid weight loss: a treatment of obesity: a guideline based on risk of gallstone formation,
prospective study in patients undergoing gastric bypass surgery for Am J Med 98(2):115–117, 1995.
treatment of morbid obesity, Am J Gastroenterol 86(8):1000–1005, Wesdorp I, et al: Clinical presentations and predisposing factors of
1991. cholelithiasis and sludge in children, J Pediatr Gastroenterol Nutr
Shiffman ML, et al: Gallstones in patients with morbid obesity: rela- 31:411–417, 2000.
tionship to body weight, weight loss and gallbladder bile cholesterol
solubility, Int J Obes Relat Metab Disord 17(3):153–158, 1993.
CHAPTER 33
Cholecystitis
Kaitlyn J. Kelly and Sharon Marie Weber
OVERVIEW of the right upper quadrant of the abdomen. Other presenting

symptoms include nausea, vomiting, and anorexia.
Cholecystitis, a common condition usually resulting from com-
plications of cholelithiasis, occurs in two forms: acute and Differential Diagnosis
chronic. Acute cholecystitis requires urgent intervention, typi- Several disease processes can present similarly to cholecystitis
cally with antibiotics and cholecystectomy. In the setting of and should be considered in the differential diagnosis. These
acute cholecystitis, cholecystectomy is optimally performed include peptic ulcer disease, gastritis and gastroenteritis, irri-
soon after the diagnosis is made. If urgent cholecystectomy is table bowel syndrome, inflammatory bowel disease, right lower
not feasible, operation can be delayed until after the acute lobe pneumonia, and biliary dyskinesia. An initial chest
episode has resolved, and then the operation may be performed radiograph is generally sufficient to assess for a right lower
electively, provided that the acute process can be controlled and lobe infiltrate. The other diagnoses should be entertained
the symptoms resolve. Chronic cholecystitis is the manifesta- and worked up appropriately in symptomatic patients without
tion of ongoing, intermittent inflammation and biliary colic. gallstones on ultrasound (US).
Patients with this condition benefit from elective cholecystec-
tomy. A less common version of cholecystitis is acute acalculous Diagnostic Evaluation and Imaging
cholecystitis, which occurs most often in critically ill patients. Abdominal US (see Chapter 15) is useful for assessing patients
Although gallstones are, by definition, absent in this condition, suspected to have acute cholecystitis. Typical findings include
cholecystectomy specimens in patients with acute acalculous gallstones, gallbladder wall thickening (>4 mm), and pericho-
cholecystitis often reveal biliary sludge. lecystic fluid (Fig. 33.1). In addition, the sonographer can
assess for pain and inspiratory arrest when the gallbladder is
ACUTE CHOLECYSTITIS directly compressed by the US probe (sonographic Murphy
sign). Typically, conventional grayscale imaging is used,
Pathogenesis which, together with clinical picture and sonographic Murphy
The cause of acute cholecystitis is an impacted gallstone in the sign, is sensitive and specific for diagnosing acute cholecysti-
outlet of the gallbladder, either in the infundibulum or in the tis, with an overall accuracy of greater than 90% (Pinto et al,
cystic duct (Sjodahl et al, 1978). The impacted gallstone results 2013). Other techniques of imaging that assess blood flow,
in gallbladder distension and edema with acute inflammation, such as color velocity imaging, may improve accuracy in
which eventually can result in venous stasis and obstruction, selected cases.
followed by thrombosis of the cystic artery. Ultimately, isch- Hepatobiliary scintigraphy (see Chapter 17) is a useful study
emia and necrosis of the gallbladder can occur. Because the in selected patients when the diagnosis is uncertain. This
fundus of the gallbladder is the greatest distance from the cystic nuclear medicine study is performed with derivatives of
arterial blood supply, it is more sensitive to ischemia and is the aminodiacetic acid (hepatoiminodiacetic acid, isopropylacet-
most common location for necrosis of the gallbladder. The anilido imidodiacetic acid, or diisopropylacetanilido iminodi-
acute inflammation of cholecystitis may be complicated by acetic acid), which are taken up by hepatocytes and secreted in
secondary biliary infection. Positive bile cultures are found in the bile. When the tracer is labeled with technetium, it is pos-
approximately 20% of patients with acute cholecystitis (den sible to visualize biliary function with scintigraphy. A normal
Hoed et al, 1998), the most common of which are gram- scan delineates the biliary tree, including the gallbladder, and
negative bacteria of gastrointestinal origin, such as Klebsiella shows prompt emptying of the agent into the duodenum. Non-
spp. and Escherichia coli. The incidence of bactobilia is even visualization of the gallbladder on hepatobiliary scan implies
higher in patients who have had endoscopic sphincterotomy or obstruction of the cystic duct and is consistent with acute cho-
other biliary instrumentation and has been reported to in as lecystitis (Fig. 33.2).
many as 60% in this population (Reinders et al, 2011). An added utility of this test is that, in addition to gallbadder
filling or nonfilling, it also evaluates gallbladder emptying.
Clinical Manifestations Abnormal emptying is defined as an ejection fraction (EF) of
Most patients with acute cholecystitis are seen with severe, 35% or less on scintigraphy, a finding suggestive of biliary
constant, right upper quadrant abdominal or epigastric pain, pathology. Middleton and Williams (2001) reported on 141
sometimes with radiation to the subscapular area. This pain patients with normal US and scintigraphy with normal gallblad-
may be preceded by intermittent, self-limited bouts of abdomi- der filling and an EF of 35% or less who underwent cholecys-
nal pain from episodes of biliary colic. Acute cholecystitis is tectomy. Of these patients, 95% had significant improvement
frequently associated with fever and leukocytosis, findings that or complete resolution of symptoms after surgery, and 41% of
are not present in cases of uncomplicated biliary colic. Patients these patients had cholecystitis on histologic evaluation of the
also may have a Murphy sign, or inspiratory arrest on palpation gallbladder.
556
A. Gallstones and Gallbladder Chapter 33 Cholecystitis 557
Hepatobiliary scintigraphy is not useful in patients with poor cholecystitis are the same as those seen on US and include wall
hepatic function because it requires hepatic excretion of bile, thickening, pericholecystic stranding, distension of the gallblad-
but it is accurate in approximately 90% of patients and may be der, high-attenuation bile, pericholecystic fluid, and subserosal
more accurate than US alone. Hepatobiliary scintigraphy is edema. CT is generally less sensitive than US for diagnosing
more involved, more expensive, and requires a longer time than acute cholecystitis, particularly early in the course, when the
US; however, it should be reserved for selected cases (Chatzi- imaging findings may be subtle (Fidler et al, 1996; Harvey &
ioannou et al, 2000; Kalimi et al, 2001). Mahid and colleagues Miller, 1999).
(2009) have suggested a diagnostic approach that starts with
US for patients with biliary symptoms. If no gallstones are Treatment
definitively identified, this should be followed by esophagogas- Initial treatment with antibiotics active against enteric bacteria
troduodenoscopy to exclude alternative causes of symptoms, should begin as soon as the patient is diagnosed with acute
such as peptic ulcer disease or gastritis. If this test result is cholecystitis. In addition, oral intake should be held, and appro-
negative, hepatobiliary scintigraphy should follow. priate intravenous (IV) fluid resuscitation should be started in
Computed tomography (CT; see Chapter 18) can also help preparation for surgery with parenteral analgesics administered
diagnose acute cholecystitis and provides more detailed ana- as needed.
tomic information than US. CT is particularly useful in patients The definitive treatment for acute cholecystitis is cholecys-
whose symptoms suggest a complication such as pericholecystic tectomy (see Chapter 35). From the time this operation was
abscess or an alternative diagnosis. The CT findings of acute first performed in 1882 by Langenbuch, open cholecystec-
tomy has been the standard of care for patients with acute
cholecystitis. With the advent of laparoscopic cholecystectomy
in the 1980s, the standard approach has changed such that
cholecystectomy is now routinely performed laparoscopically.
The benefits of laparoscopic cholecystectomy are discussed in
depth elsewhere (see Chapter 35), but they include a shorter
TRANS RT GB
postoperative stay and decreased analgesia requirements (Cox
et al, 1993). Although the laparoscopic approach is now stan-
dard for most cases, it is interesting to note that two prospec-
tive, randomized studies suggested little or no difference in
several outcome measures between laparoscopic versus small-
incision open cholecystectomy (Keus et al, 2008; Majeed
et al, 1996).
Early analysis of the results of laparoscopic cholecystectomy
in patients with acute versus chronic cholecystitis showed
increased morbidity and mortality rates for patients with simple
or complicated acute cholecystitis. Because of the increased
morbidity and mortality, acute cholecystitis initially was consid-
ered a relative contraindication to laparoscopic cholecystectomy
(Flowers et al, 1991). Subsequent reports, however, have shown
FIGURE 33.1. Transverse view of the gallbladder on ultrasound in a the improved safety of this technique in the acute setting (Chan-
patient with calculous cholecystitis, revealing gallstones and gallbladder dler et al, 2000; Johansson et al, 2003; Lai et al, 1998; Lo et al,
wall thickening. 1998). The conversion rate to an open procedure is higher for
A B
FIGURE 33.2. A, Normal hepatoiminodiacetic acid (HIDA) scan demonstrating contrast-filled gallbladder (arrow). B, Abnormal HIDA scan demon-
strating nonfilling of the gallbladder consistent with cystic duct obstruction. (Courtesy Dr. Scott Perlman, University of Wisconsin Hospital and Clinics.)
patients with acute cholecystitis compared with patients under- compared with those treated conservatively with IV antibiotics
going elective operations for simple biliary colic (Schirmer et al, and bowel rest, followed by delayed cholecystectomy. In high-
1991), but most patients with acute cholecystitis (>80%) can risk patients in whom general anesthesia is contraindicated,
undergo laparoscopic cholecystectomy successfully (Papi et al, percutaneous stone extraction has been used successfully
2004). Retrospective series have reported that risk factors for (Gibney et al, 1987; Wong et al, 1999). More recently, single
conversion to open cholecystectomy include obesity (Rosen and repeated percutaneous transhepatic gallbladder aspiration
et al, 2002), elevated white blood cell count (Alponat et al, has been shown to be successful in high-risk patients with acute
1997; Kanaan et al, 2002), and male gender (Kanaan et al, cholecystitis who do not respond to conservative therapies
2002). (Tsutsui et al, 2007). This treatment modality is not widely
Other novel surgical approaches to cholecystectomy have practiced, however.
been proposed to treat patients’ symptomatic gallstones,
including so-called mini-laparoscopic cholecystectomy (see Chapter Timing of Surgery
35), which uses 2- to 3-mm ports (Hsieh, 2003); mini- The optimal interval of time between the diagnosis of acute
cholecystectomy (Assalia et al, 1997), in which a small (mean, cholecystitis and definitive treatment with cholecystectomy has
5.5 cm) incision is used to remove the gallbladder; single-incision been the subject of many prospective randomized trials, nine
laparoscopic cholecystectomy; and natural orifice transluminal evaluating open cholecystectomy and five evaluating laparo-
endoscopic (NOTES) cholecystectomy with transvaginal extrac- scopic cholecystectomy (Papi et al, 2004; Siddiqui et al,
tion. Prospective, randomized studies evaluating the safety of 2008). The concern in operating on patients with early chole-
these techniques are lacking, but existing data suggest decreased cystitis (typically defined as <3 days) is the potential for
postoperative pain and improved cosmesis at the expense of increased postoperative complications, including common bile
slightly longer operating times with these techniques. duct injury. The risk of performing cholecystectomy late
Laparoscopic subtotal cholecystectomy (LSC) has also (weeks after the diagnosis of cholecystitis) is that a subset of
been evaluated as a means of decreasing the conversion rate patients has recurrent symptoms during the period between
to open procedure in patients with acute cholecystitis (Horiu- diagnosis and surgical treatment, which leads to recurrent hos-
chi et al, 2008; Singhal et al, 2009) (see Chapter 35). This pital admissions and urgent surgery (Papi et al, 2004). In mul-
procedure involves leaving the posterior wall of the gallbladder tiple randomized prospective trials evaluating the timing of
when it is significantly difficult to dissect from the liver bed, open cholecystectomy, patients undergoing early operation
with cauterization of the remnant mucosa and suturing or sta- experienced no increased perioperative morbidity or mortality
pling of the gallbladder neck in cases of severe inflammation and had a shorter length of hospital stay compared with
of the triangle of Calot. Indeed, in the face of severe inflam- patients undergoing delayed operation (Norrby et al, 1983;
matory change, such an approach is safer than risking injury Van der Linden & Edlund, 1981). In addition, a meta-analysis
to the common hepatic duct by pursuing dissection within the of these trials demonstrated that more than 20% of patients
porta hepatis. Horiuchi and colleagues (2008) demonstrated a did not respond to medical management while awaiting defini-
significant decrease in conversion rate to open procedure with tive treatment, and approximately half of these patients
the use of this technique, with no increase in postoperative required urgent surgical treatment as a result (Papi et al,
complications. In this study, all patients undergoing LSC had 2004). In this same analysis, no increase in morbidity was seen
temporary subhepatic drains placed. This technique, although in patients undergoing early definitive treatment with laparo-
innovative, does not have wide clinical applicability because of scopic (p = 0.6) or open (p = 0.2) cholecystectomy compared
limited availability of instruments, specialist surgeons trained with delayed treatment, but a clear difference was seen in
in their use, or both. Laparoscopic cholecystectomy remains length of hospital stay, with patients undergoing delayed treat-
the standard therapy for definitive treatment of patients with ment requiring a more prolonged hospitalization (Papi et al,
acute cholecystitis, with conversion to an open procedure if 2004). Although it is safe to perform cholecystectomy either
necessary. early or late after an episode of acute cholecystitis, patients
In patients with a high perioperative risk related to sepsis or benefit when surgery is performed early.
other underlying medical comorbidities, initial treatment of Performing laparoscopic cholecystectomy in patients with
acute cholecystitis with percutaneous cholecystostomy tube acute cholecystitis poses unique challenges to surgeons because
placement is preferred (see Chapter 34). These tubes can be of the potential for increased morbidity, including the most
placed percutaneously using either US or CT guidance (Hat- feared complication—common bile duct injury (see Chapter
zidakis et al, 2002). This procedure effectively decompresses 42). The concern is that the acute inflammation may obscure
the gallbladder by evacuating the infected bile and relieving the the anatomy, leading to an increase in postoperative complica-
pain associated with gallbladder distension from outlet obstructions. Multiple prospective randomized trials have evaluated the
tion, and it is associated with a low complication rate (Akyürek outcome of patients with acute cholecystitis undergoing early
et al, 2005; Byrne et al, 2003; Spira et al, 2002; Werbel et al, versus late laparoscopic cholecystectomy (Table 33.1). Although
1989). In addition, most patients (>80%) improve clinically a significant increase in operation time was experienced for
within a short time (Byrne et al, 2003; Hatzidakis et al, 2002; those undergoing early compared with delayed cholecystecomy
Vauthey et al, 1993). After stabilization of the patient, and if (p = 0.002), the results of these trials uniformly showed no
the clinical situation otherwise warrants, a delayed interval significant difference in postoperative morbidity or mortality,
cholecystectomy should be performed, which often can be including common bile duct injury, when surgery is performed
accomplished laparoscopically (Spira et al, 2002). Akyürek early (Siddiqui et al, 2008). Additionally, no significant differ-
and colleagues (2005) demonstrated decreased hospital stay ence was found in the conversion rate to open cholecystectomy,
and cost in high-risk patients undergoing percutaneous chole- although it was clearly higher (20% to 30%) in patients with
cystostomy, followed by early laparoscopic cholecystectomy cholecystitis compared with prior studies evaluating patients
TABLE 33.1 Results of Prospective, Randomized Trials Comparing Early versus Delayed Laparoscopic Cholecystectomy for Acute
Cholecystitis
Length of Conversion to Open
Reference N Definition Morbidity Rate Stay (days) Cholecystectomy
Ozkardes et al, 2014 60 Early: <24 hours 27% 5 13%

Late: 6-8 weeks 0 8 0
Saber & Hokkam, 2014 120 Early: <72 hours NR 2 5%
Late: 6-8 weeks NR 6 2%
Gutt et al, 2013 618 Early: <24 hours 12% 5 10%
Late: 7-45 days 34% 10 12%
Macafee et al, 2009 72 Early: <72 hours 22% 6 3%
Late: 3 months 11% 6 3%
Kolla et al, 2004 40 Early: <4 days 20% 4 25%
Late: 6-12 weeks 15% 10 25%
Johansson et al, 2003 145 Early: <7 days 18% 5 31%
Late: 6-8 weeks 10% 8 21%
Lai et al, 1998 104 Early: <24 hours 9% 8 24%
Late: 6-8 weeks 8% 12 8%
Lo et al, 1998 99 Early: <3 days 29% 6 11%
Late: 6-8 weeks 13% 11 23%
undergoing elective laparoscopic cholecystectomy in the non- surgery was more cost effective because of reduced overall
acute setting. Perhaps the most important finding was that in length of hospital stay and avoidance of readmissions for recur-
all but one study, patients randomly assigned to late cholecys- rent cholecystitis or biliary colic.
tectomy did not respond to conservative management in 15%
to 30% of cases. Although patients in the early group generally CHRONIC CHOLECYSTITIS
experienced a longer postoperative hospitalization (p = 0.004),
most of these trials demonstrated a decrease in overall length Pathogenesis and Clinical Manifestations
of hospital stay in the early group versus the delayed group Chronic cholecystitis may result after one or more episodes of
(cumulative p < 0.001) (Chandler et al, 2000; Johansson et al, acute cholecystitis, or it may evolve, initially without symptoms,
2003; Kolla et al, 2004; Lai et al, 1998; Lo et al, 1998). Early merely from the presence of gallstones. In most cases, patients
laparoscopic cholecystectomy is therefore the preferred surgical describe one or more episodes of abdominal pain that is clini-
technique for patients with acute cholecystitis. Catena and col- cally consistent with biliary colic. The term colic is a misnomer
leagues (2009) have proposed the use of a harmonic scalpel for because the pain from chronic cholecystitis is usually constant
improved hemostasis and biliostasis in laparoscopic cholecys- in nature and is similar to that seen initially with acute chole-
tectomy, and preliminary data suggested it may decrease the cystitis, although it is self-limited and often less severe. The
conversion rate to open procedure in patients undergoing lapa- pain associated with chronic cholecystitis seems to be the result
roscopic cholecystectomy for acute cholecystitis. A prospective, of intermittent obstruction of the gallbladder outflow.
randomized controlled trial subsequently confirmed these find- There are numerous well-described risk factors for the
ings (Catena et al, 2009). development of gallstones and subsequent chronic cholecystitis.
The majority of trials of early versus delayed laparoscopic Patients at particularly high risk include obese women, in whom
cholecystectomy define “early” as within 72 hours of symptom pathologic changes of chronic inflammation are found even in
onset. A nonrandomized, prospective study by Tzovaras and the absence of gallstones (Calhoun & Willbanks, 1987; Csendes
colleagues (2006) assessed 129 patients undergoing laparo- et al, 2003); this may be due to an increase in the cholesterol
scopic cholecystectomy for acute cholecystitis during the index saturation of bile in obese patients (St. George & Shaffer, 1993;
admission. The patients were divided into three groups regard- see Chapter 8). These patients are often asymptomatic. Xan-
ing the timing of their surgery from symptom onset: within 3 thogranulomatous cholecystitis is a subtype of chronic chole-
days, between 4 and 7 days, and after 7 days. They found no cystitis that can mimic gallbladder cancer (Agarwal, et al. 2013)
significant difference in conversion rate, morbidity, or postop- (see Chapter 49). It is characterized by the presence of destruc-
erative hospital stay among these groups and thus suggest that tive inflammation of the gallbladder wall, often accompanied
the benefits of early cholecystectomy are not limited to patients by proliferative fibrosis. The histologic appearance is that of
who are seen within 72 hours of symptom onset. foamy histiocytes in a background of acute and chronic inflam-
Another factor to consider is whether early or delayed lapa- matory cells. This process can result in the appearance of asym-
roscopic cholecystectomy for acute cholecystitis is more cost metric gallbladder wall thickening and/or mass formation in the
effective. A meta-analysis of randomized and other studies per- gallbladder wall and can be difficult or impossible to distinguish
formed by Lau and colleagues (2006) concluded that early from gallbladder cancer (Fig. 33.3). Further confusing this
A B
FIGURE 33.3. A, Coronal image of a patient seen with chronic right upper quadrant pain and a distended gallbladder with asymmetric wall thick-
ening concerning for malignancy. B, Gross image of the resected gallbladder, which was firm and markedly abnormal. Final pathology demonstrated
xanthogranulomatous cholecystitis.
picture, serum cancer antigen 19-9 levels can be elevated in burns, or in those undergoing a prolonged recovery from major
xanthogranulomatous cholecystitis (Clarke et al, 2007). operations who are unable to tolerate oral intake (Gu et al,
2014). It has been speculated that in such situations, there is no
Diagnostic Imaging stimulus for gallbladder contraction, the bile becomes inspis-
US examination of the gallbladder most often reveals circum- sated, and biliary sludge forms. The exact pathogenesis is
ferential thickening of the gallbladder wall with cholelithiasis unknown but likely involves some combination of ischemia,
(see Chapter 15). In advanced cases, a small, shrunken gall- biliary stasis, and sepsis (Hakala et al, 1997; Warren, 1992), and
bladder with a thickened wall and multiple gallstones are seen. inspissated bile and sludge seem to play some causative role.
Discomfort may be reproduced with direct pressure on the Although this condition traditionally has been described in the
gallbladder with the US probe. Commonly, particularly in patient groups mentioned earlier, several reports suggest an
obese patients and in those with mild symptoms, the US exami- increase in the de novo presentation of acalculous cholecystitis
nation shows no particular gallbladder wall abnormalities in the outpatient population, including patients with atheroscle-
(Calhoun & Willbanks, 1987). rotic vascular disease, such as seen in hypertension and diabetes
(Parithivel et al, 1999; Ryu et al, 2003; Savoca et al, 1990).
Treatment Overall, acalculous cholecystitis represents approximately 5%
Elective cholecystectomy is the treatment of choice for patients to 15% of all cases of acute cholecystitis. A male predominance
with symptoms of chronic cholecystitis (see Chapter 35). In is seen in cases of acalculous cholecystitis, in contrast to acute
most (≥ 90%) cases, cholecystectomy can be accomplished calculous cholecystitis, which occurs more commonly in women
laparoscopically. Patients occasionally are seen with atypical (Kalliafas et al, 1998; Ryu et al, 2003; Wang et al, 2003).
pain (left hypochondrium) or with minimal or no pain but, A prospective study evaluating trauma patients with serial
rather, intermittent nausea or bloating. In such cases, evalua- US examinations found that the incidence of acalculous chole-
tion for other possible causes of symptoms should be under- cystitis in severely injured patients (injury severity score ≥ 12,
taken, particularly if US shows gallstones but no sequelae of requiring intensive care for > 4 days) was 11% (Pelinka et al,
chronic cholecystitis are seen. In cases where there is asym- 2003), which is similar to other reports (Imhof et al, 1992). In
metric gallbladder wall thickening or mass formation concern- addition, three factors were correlated with an increased risk
ing for malignancy, a laparoscopic exploration is reasonable but for acalculous cholecystitis in this high-risk population: (1) high
with a low threshold for conversion to an open operation. The injury-severity score, (2) increased heart rate, and (3) transfu-
surgeon should be prepared to perform a definitive gallbladder sion requirement at the time of admission. This study suggests
cancer operation, including liver resection and regional lymph- that more acutely injured patients, who are expected to require
adenectomy. In these cases, it is critical to avoid gallbladder prolonged ventilatory and nutritional support, are at higher risk
perforation or spillage intraoperatively. Frozen-section analysis for acalculous cholecystitis (Pelinka et al, 2003).
of any grossly abnormal tissue can be considered, but it is often
appropriate to proceed with a cancer operation if gross findings Clinical Manifestations
are suspicious for malignancy (see Chapter 49). Part of the difficulty in making the diagnosis of acalculous
cholecystitis is that many patients seen with this condition are
ACUTE ACALCULOUS CHOLECYSTITIS critically ill and require ventilatory support and sedation. The
symptoms and signs are often masked by the patient’s underly-
Pathogenesis ing condition or the interventions used to treat it (Kalliafas
Acalculous cholecystitis usually occurs in patients with coexist- et al, 1998). In the outpatient population seen with acalculous
ing major illnesses, such as generalized sepsis, major trauma, or cholecystitis, the diagnosis is more straightforward, mimicking
the signs and symptoms of acute calculous cholecystitis (Ryu assessed 255 severely injured trauma patients with serial US
et al, 2003). examinations. In this trial, all patients with US findings consis-
The most frequent physical and laboratory findings are tent with acalculous cholecystitis and major clinical symptoms—
fever, right upper quadrant pain, leukocytosis, and hyperbiliru- abdominal pain or distension or both, hemodynamic instability,
binemia. These findings are often nonspecific, however, in the or organ failure—underwent cholecystectomy, and all had acal-
setting of sepsis and critical illness (Kalliafas et al, 1998). The culous cholecystitis on final pathology. A few patients had posi-
incidence of gangrene and perforation seems to be increased in tive US findings, but no major clinical symptoms were observed.
patients with acalculous cholecystitis compared with acute cal- In this group, all US findings returned to normal within 3 weeks.
culous cholecystitis, likely because of the delay in diagnosis that In addition, 15% of patients experienced hydrops of the gall-
is common with this disease. Severe gallbladder complications bladder without clinical symptoms, and subsequent normaliza-
such as gangrene, perforation, and empyema occur more com- tion of the US was observed in all (Pelinka et al, 2003). These
monly in older patients with an elevated white blood cell count findings suggest that although many critically ill patients may
(Ryu et al, 2003; Wang et al, 2003). In many series, the risk develop US abnormalities consistent with acalculous cholecys-
of severe gallbladder complications was found to be 50% to titis, the combination of clinical symptoms and positive imaging
60% (Kalliafas et al, 1998; Swayne, 1986; Wang et al, 2003). findings will identify most patients who require intervention.
This high risk may be the result of the disturbance in capillary Because of the difficulty establishing a diagnosis in these
microcirculation that has been shown in pathologic studies on critically ill patients, CT has been used as an additional imaging
gallbladder specimens after cholecystectomy for acalculous test. The advantage of CT is that imaging of the entire chest,
cholecystitis (Hakala et al, 1997; Warren, 1992). In addition, abdomen, and pelvis can be obtained; this is particularly impor-
partly as a result of the severity of the patient’s underlying tant in this patient cohort, in whom clinical signs and symptoms
condition, the mortality rates are high—15% in some series may be misleading or whose signs and symptoms may result
(Wang et al, 2003). from other causes. CT may be more sensitive and specific than
US (Blankenberg et al, 1991; Mirvis et al, 1986), but it has the
Diagnostic Evaluation and Imaging disadvantage of requiring transport of the patient outside of the
Imaging algorithms for patients with suspected acalculous cho- intensive care unit.
lecystitis are similar to algorithms for patients with acute cho- When a patient’s diagnosis is questionable based on physical
lecystitis. The initial imaging test is usually US, which classically findings or US evaluation or both, hepatobiliary scintigraphy
reveals gallbladder distension, a thickened gallbladder wall, and may be used. In past reports, a high rate of false-positive results
biliary sludge without stones (Fig. 33.4; Wang et al, 2003) (see was seen in patients with acalculous cholecystitis (Mirvis et al,
Chapter 15). The difficulty with interpreting these findings is 1986), but more recent evaluation of this modality has shown
that many critically ill, parenteral nutrition–dependent patients improved accuracy in diagnosing acalculous cholecystitis (Kal-
have these findings. Partly because of this difficulty, the accu- liafas et al, 1998; Ryu et al, 2003). In some series, scintigraphy
racy of US to diagnose acalculous cholecystitis has been highly has been found to be more sensitive than CT or US in diagnos-
institution dependent, with some series showing it to be highly ing acalculous cholecystitis (Kalliafas et al, 1998; Prevot et al,
sensitive and specific (Pelinka et al, 2003) and others showing 1999; Puc et al, 2002; Ryu et al, 2003). The specificity of
it to be less accurate (Kalliafas et al, 1998). US is widely avail- hepatobiliary scintigraphy can be improved by administering
able and easy to use, even in acutely ill patients, because it can morphine to cause constriction of the sphincter of Oddi and
be performed at the bedside, and it is inexpensive. US should improve gallbladder filling (see Chapter 17). This maneuver
be performed as the initial imaging modality for suspected decreases the incidence of false-positive studies and improves
acalculous cholecystitis. specificity, but it does not result in improvements in sensitivity
Evaluating the natural history of abnormalities visualized on compared with conventional scintigraphy (Cabana et al, 1995;
US in acutely injured patients is problematic; a prospective trial Flancbaum et al, 1989; Kalliafas et al, 1998).
Treatment
The definitive treatment for acalculous cholecystitis is chole-
cystectomy, which can be performed laparoscopically in most
cases. In patients who are critically ill, placement of a percuta-
neous cholecystostomy tube allows decompression of the gall-
bladder and drainage of contained, infected bile (see Chapter
34); this allows time for the patient to recover from the acute
illness before considering proceeding with cholecystectomy
(Akyürek et al, 2005; McClain et al, 1997). Percutaneous cho-
lecystostomy may be the definitive treatment for acalculous
cholecystitis because there is no chronic obstruction of the
gallbladder outlet as in acute cholecystitis (Davis et al, 1999;
Vauthey et al, 1993) (see Chapters 28 and 32).
COMPLICATIONS OF CHOLECYSTITIS
Gangrenous Cholecystitis
FIGURE 33.4. Ultrasound of a patient with acalculous cholecysti- Gangrenous cholecystitis is a more common finding in diabetic
tis reveals marked thickening of the gallbladder wall (arrows). patients with acute cholecystitis who present with a leukocytosis
(Fagan et al, 2003). In addition, the risk of gangrenous chole-

cystitis is higher in patients with acalculous cholecystitis, likely
owing to the delay in diagnosis that commonly occurs in this
disease (Kalliafas et al, 1998; Swayne, 1986; Wang et al, 2003).
As previously mentioned, the most common site for necrosis to
occur is in the fundus. Full-thickness necrosis is by definition
always present in patients with gangrenous cholecystitis, but
this condition may or may not result in free perforation of the
gallbladder. In patients with free perforation, bile-stained
abdominal fluid is present.
Because these patients are generally ill, imaging with CT
scan is often performed. Findings most specific for acute gan-
grenous cholecystitis on CT scan include air in the wall or
lumen, intraluminal membranes, an irregular wall, or pericho-
lecystic abscess. As expected, a contrast-enhanced CT scan
may show a lack of mural enhancement in patients with gan-
grenous cholecystitis (Bennett et al, 2002).
FIGURE 33.5. Ultrasound of the gallbladder revealing air within
Empyema the gallbladder wall (arrows) consistent with emphysematous
cholecystitis.
In cases of empyema, the gallbladder is filled with purulent bile.
This condition usually is associated with acute cholecystitis and
occurs in the setting of infected bile and an obstructed cystic
duct. Most patients with empyema have calculous cholecystitis,
but empyema also can occur in patients with acalculous disease
(Tseng et al, 2000). The clinical course can mimic that of an
intraabdominal abscess from other causes, and patients are
often seen initially with clinical manifestations of sepsis. Patients
with gallbladder empyema require urgent cholecystectomy or
percutaneous cholecystostomy, depending on the severity of R
2
L
2
illness at the time of presentation (Tseng et al, 2000). Critically 2 7
7
2
ill patients may be best served by a temporary cholecystostomy
tube followed by elective cholecystectomy.
Emphysematous Cholecystitis
Emphysematous cholecystitis is a rare entity that results from
the presence of gas-forming bacteria in the bile. Emphysema-
tous cholecystitis may be seen in association with acute or
gangrenous cholecystitis, and it is more common in men and
patients with diabetes (Tellez et al, 1999). The diagnosis occa-
FIGURE 33.6. Computed tomography scan (noncontrast) in a
sionally can be made by simple abdominal radiographs, but
patient with emphysematous cholecystitis showing gallbladder
more often it is diagnosed on US (Fig. 33.5) or CT scan (Fig.
filled with air.
33.6)(Bennett & Balthazar, 2003; Gill et al, 1997; Konno et al,
2002). Patients should receive IV antibiotics to include coverage
for Clostridium species, followed by emergent cholecystectomy. any stones may be all that is possible and usually resolves the
condition. In the acute setting, the biliary obstruction often
Mirizzi Syndrome resolves after cholecystectomy and resolution of the inflamma-
Mirizzi syndrome is defined as biliary obstruction secondary to tory process. In some cases, however, the chronic inflamma-
cholecystitis. It occurs in 0.3% to 3% of patients undergoing tion leads to fistulation from the gallbladder to the bile duct,
cholecystectomy (Lai & Lau, 2006). An impacted stone in the or a biliary stricture results, both of which will complicate the
gallbladder infundibulum or cystic duct can compress the bile operative procedure and may require biliary reconstruction
duct, usually at the level of the common hepatic duct (type I), (see Chapter 42).
or a stone can erode from the gallbladder or cystic duct into
the common hepatic duct, resulting in a cholecystocholedochal Cholecystoenteric Fistula
fistula (type II). Patients are seen with symptoms of acute Cholecystoenteric fistula, or perforation of the gallbladder into
cholecystitis but with the additional finding of hyperbilirubine- an adjacent hollow organ, is a rare complication of acute cho-
mia and elevated alkaline phosphatase. A laparoscopic approach lecystitis. The duodenum and the transverse colon are the most
to this condition has been shown to result in high conversion common sites of fistulation, which results in decompression of
and complication rates and is generally not recommended the gallbladder and may result in brief symptomatic improve-
(Antoniou et al, 2009; Lai & Lau, 2006). Open cholecystec- ment. This complication occurs most frequently in women in
tomy is the gold standard for treatment when this condition is their sixth to seventh decades and has been shown to be associ-
identified preoperatively. If inflammation has obliterated the ated with Mirizzi syndrome or an additional hepatobiliary
triangle of Calot, a partial cholecystectomy with removal of abnormality such as gallbladder cancer (Beltran et al, 2008;
Chowbey et al, 2006; Costi et al, 2009). Contamination of the concomitant Mirizzi syndrome should be considered. Patients
biliary tree by enteric organisms may result, and patients may with gallstone ileus require removal of the obstructing stone
be seen with cholangitis and pneumobilia. Approximately 10% via enterotomy (Chowbey et al, 2006), and it is important to
to 15% of patients with cholecystoenteric fistulae will pass perform a thorough examination of the bowel for any other
gallstones into the small intestine and be seen with small bowel stones. Fistula takedown and cholecystectomy can be performed
obstruction, termed gallstone ileus. In patients with cholecysto- at the same procedure or at a second, delayed procedure if the
colonic fistula, chronic diarrhea is the most common presenting patient is too unstable to tolerate a prolonged initial operation,
symptom in nonemergent cases (Costi et al, 2009). or if the pericholecystic inflammation is so severe as to make
Patients with cholecystoenteric fistula require cholecy initial cholecystectomy unsafe.
stectomy with takedown and closure of the fistula. When cho-
lecystoenteric fistula is encountered at the time of surgery, References are available at expertconsult.com.
A. Gallstones and Gallbladder Chapter 33 Cholecystitis 563.e1
REFERENCES Gibney RG, et al: Combined surgical and radiologic intervention for
complicated cholelithiasis in high-risk patients, Radiology 165:715–
Agarwal AK, et al: Mass-forming xanthogranulomatous cholecystitis 719, 1987.
masquerading as gallbladder cancer, J Gastrointest Surg 17:1257– Gill KS, et al: The changing face of emphysematous cholecystitis, Br
1264, 2013. J Radiol 70:986–991, 1997.
Akyürek N, et al: Management of acute calculous cholecystitis in high- Gu MG, et al: Risk factors and therapeutic outcomes of acute acalcu-
risk patients: percutaneous cholecystotomy followed by early laparo- lous cholecystitis, Digestion 90:75–80, 2014.
scopic cholecystectomy, Surg Laparosc Endosc Percutan Tech Gutt CN, et al: Acute cholecystitis: early versus delayed cholecystec-
15(6):315–320, 2005. tomy, a multicenter randomized trial (ACDC study NCT00447304),
Alponat A, et al: Predictive factors for conversion of laparoscopic cho- Ann Surg 258(3):353–393, 2013.
lecystectomy, World J Surg 21:629–633, 1997. Hakala T, et al: Microangiopathy in acute acalculous cholecystitis,
Antoniou SA, et al: Laparoscopic treatment of Mirizzi syndrome: a Br J Surg 84:1249–1252, 1997.
systematic review, Surg Endosc 24:33–39, 2010. Harvey RT, Miller WT Jr: Acute biliary disease: initial CT and
Assalia A, et al: Emergency minilaparotomy cholecystectomy for acute follow-up US versus initial US and follow-up CT, Radiology 213:
cholecystitis: prospective randomized trial—implications for the 831–836, 1999.
laparoscopic era, World J Surg 21:534–539, 1997. Hatzidakis AA, et al: Acute cholecystitis in high-risk patients: percuta-
Beltran MA, et al: The relationship of Mirizzi syndrome and cholecys- neous cholecystostomy vs. conservative treatment, Eur Radiol 12:
toenteric fistula: validation of a modified classification, World J Surg 1778–1784, 2002.
32:2237–2243, 2008. Horiuchi A, et al: Delayed laparoscopic subtotal cholecystectomy in
Bennett GL, et al: CT findings in acute gangrenous cholecystitis, AJR acute cholecystitis with severe fibrotic adhesions, Surg Endosc 22(12):
Am J Roentgenol 178:275–281, 2002. 2720–2723, 2008.
Bennett GL, Balthazar EJ: Ultrasound and CT evaluation of emergent Hsieh CH: Early minilaparoscopic cholecystectomy in patients with
gallbladder pathology, Radiol Clin North Am 41:1203–1216, 2003. acute cholecystitis, Am J Surg 185:344–348, 2003.
Blankenberg F, et al: Computed tomography as an adjunct to ultra- Imhof M, et al: Acute acalculous cholecystitis complicating trauma:
sound in the diagnosis of acute acalculous cholecystitis, Gastrointest a prospective sonographic study, World J Surg 16:1160–1165,
Radiol 16:149–153, 1991. 1992.
Byrne MF, et al: Percutaneous cholecystostomy in patients with acute Johansson M, et al: Management of acute cholecystitis in the laparo-
cholecystitis: experience of 45 patients at a US referral center, J Am scopic era: results of a prospective, randomized clinical trial, J Gas-
Coll Surg 197:206–211, 2003. trointest Surg 7:642–645, 2003.
Cabana MD, et al: Morphine-augmented hepatobiliary scintigraphy: a Kalimi R, et al: Diagnosis of acute cholecystitis: sensitivity of sonogra-
meta-analysis, Nucl Med Commun 16:1068–1071, 1995. phy, cholescintigraphy, and combined sonography-cholescintigraphy,
Calhoun R, Willbanks O: Coexistence of gallbladder disease and J Am Coll Surg 193:609–613, 2001.
morbid obesity, Am J Surg 154:655–658, 1987. Kalliafas S, et al: Acute acalculous cholecystitis: incidence, risk factors,
Catena F, et al: The HAC trial (harmonic for acute cholecystitis): a diagnosis, and outcome, Am Surg 64(5):471–475, 1998.
randomized, double-bline, controlled trial comparing the use of har- Kanaan SA, et al: Risk factors for conversion of laparoscopic to open
monic scalpel to monopolar diathermy for laparsocopic cholecystec- cholecystectomy, J Surg Res 106:20–24, 2002.
tomy in cases of acute cholecystitis, World J Surg 20;9(1):53, 2014. Keus F, et al: Randomized clinical trial of small-incision and laparo-
Catena R, et al: Prospective analysis of 101 consecutive cases of lapa- scopic cholecystectomy in patients with symptomatic cholecystitis:
roscopic cholecystectomy for acute cholecystitis operated with har- primary and clinical outcomes, Arch Surg 143:371–377, 2008.
monic scalpel, Surg Laparosc Endosc Percutan Tech 19(4):312–316, Kolla SB, et al: Early versus delayed laparoscopic cholecystectomy
2009. for acute cholecystitis: a prospective randomized trial, Surg Endosc
Chandler CF, et al: Prospective evaluation of early versus delayed lapa- 18(9):1323–1327, 2004.
roscopic cholecystectomy for treatment of acute cholecystitis, Am Konno K, et al: Emphysematous cholecystitis: sonographic findings,
Surg 66:896–900, 2000. Abdom Imaging 27:191–195, 2002.
Chatziioannou SN, et al: Hepatobiliary scintigraphy is superior to Lai PB, et al: Randomized trial of early versus delayed laparoscopic
abdominal ultrasonography in suspected acute cholecystitis, Surgery cholecystectomy for acute cholecystitis, Br J Surg 85:764–767, 1998.
127:609–613, 2000. Lai EC, Lau WY: Mirizzi syndrome: history, present and future devel-
Chowbey PK, et al: Laparoscopic management of cholecystoenteric opment, ANZ J Surg 76(4):251–257, 2006.
fistulas, J Laparoendosc Adv Surg Tech A 16(5):467–472, 2006. Lau H, et al: Early versus delayed-interval laparoscopic cholecystec-
Clarke T, et al: Gallbladder mass with carbohydrate antigen CA 19-9 tomy for acute cholecystitis: a meta-analysis, Surg Endosc 20(1):82–
level in the thousands: malignant or benign pathology?, Surg Today 87, 2006.
37:342–344, 2007. Lo CM, et al: Prospective randomized study of early versus delayed
Costi R, et al: Cholecystocolonic fistula: facts and myths—a review of laparoscopic cholecystectomy for acute cholecystitis, Ann Surg
the 231 published cases, J Hepatobiliary Pancreat Surg 16(1):8–18, 227:461–467, 1998.
2009. Macafee DA, et al: Prospective randomized trial using cost-utility
Cox MR, et al: Laparoscopic cholecystectomy for acute inflammation analysis of early versus delayed laparoscopic cholecystectomy for
of the gallbladder, Ann Surg 218:630–634, 1993. acute gallbladder disease, Br J Surg 96(9):1031–1040, 2009.
Csendes A, et al: Histologic findings of gallbladder mucosa in 87 Majeed AW, et al: Randomised, prospective, single-blind comparison
patients with morbid obesity without gallstones compared to 87 of laparoscopic versus small-incision cholecystectomy, Lancet 347:
control subjects, J Gastrointest Surg 7:547–551, 2003. 989–994, 1996.
Davis CA, et al: Effective use of percutaneous cholecystostomy in high- Mahid SS, et al: Meta-analysis of cholecystectomy in symptomatic
risk surgical patients: techniques, tube management, and results, patients with positive hepatobiliary iminodiacetic acid scan results
Arch Surg 134:727–731, 1999. without gallstones, Arch Surg 144(2):180–187, 2009.
den Hoed PT, et al: Infections and bacteriological data after laparo- McClain T, et al: Laparoscopic cholecystectomy in the treatment of
scopic and open gallbladder surgery, J Hosp Infect 39:27–37, 1998. acalculous cholecystitis in patients after thermal injury, J Burn Care
Fagan SP, et al: Prognostic factors for the development of gangrenous Rehabil 18:141–146, 1997.
cholecystitis, Am J Surg 186:481–485, 2003. Middleton GW, Williams JH: Diagnostic accuracy of 99Tcm-HIDA
Fidler J, et al: CT evaluation of acute cholecystitis: findings and useful- with cholecystokinin and gallbladder ejection fraction in acalculous
ness in diagnosis, AJR Am J Roentgenol 166:1085–1088, 1996. gallbladder disease, Nucl Med Commun 22(6):657–661, 2001.
Flancbaum L, et al: Use of cholescintigraphy with morphine in criti- Mirvis SE, et al: The diagnosis of acute-acalculous cholecystitis—a
cally ill patients with suspected cholecystitis, Surgery 106:668–673, comparison of sonography, scintigraphy, and CT, AJR Am J Roent-
1989. genol 147:1171–1175, 1986.
Flowers JL, et al: The Baltimore experience with laparoscopic manage- Norrby S, et al: Early or delayed cholecystectomy in acute cholecystitis:
ment of acute cholecystitis, Am J Surg 161:388–392, 1991. a clinical trial, Br J Surg 70:163–165, 1983.
Novitsky YW, et al: Advantages of mini-laparoscopic vs conventional Sjodahl R, et al: Pathogenesis of acute cholecystitis, Surg Gynecol Obstet
laparoscopic cholecystectomy: results of a prospective randomized 146:199–202, 1978.
trial, Arch Surg 140(12):1178–1183, 2005. Spira RM, et al: Percutaneous transhepatic cholecystostomy and
Ozkardes AB, et al: Early versus delayed laparoscopic cholecystectomy delayed laparoscopic cholecystectomy in critically ill patients with
for acute cholecystitis: A prospective, randomized study, Int Surg acute calculous cholecystitis, Am J Surg 183:62–66, 2002.
99(1):56–61, 2014. St. George CM, Shaffer EA: Spontaneous obesity and increased bile
Papi C, et al: Timing of cholecystectomy for acute calculous cholecys- saturation in the ground squirrel, J Surg Res 55:314–316, 1993.
titis: a meta-analysis, Am J Gastroenterol 99:147–155, 2004. Swayne LC: Acute acalculous cholecystitis: sensitivity in detection
Parithivel VS, et al: Acute acalculous cholecystitis in young patients using technetium-99m iminodiacetic acid cholescintigraphy, Radiol-
without predisposing factors, Am Surg 65:366–368, 1999. ogy 160:33–38, 1986.
Pelinka LE, et al: Acute acalculous cholecystitis after trauma: a pro- Tellez LGS, et al: Acute emphysematous cholecystitis: report of twenty
spective study, J Trauma 55:323–329, 2003. cases, Hepatogastroenterology 46:2144–2148, 1999.
Pinto A, et al: Accuracy of ultrasonography in the diagnosis of acute Tseng LJ, et al: Palliative percutaneous transhepatic gallbladder drain-
cholecystitis: review of the literature, Crit Ultrasound J 5(Suppl age of gallbladder empyema before laparoscopic cholecystectomy,
1):S11, 2013. Hepatogastroenterology 47:932–936, 2000.
Prevot N, et al: Contribution of cholescintigraphy to the early diagnosis Tsutsui K, et al: Usefulness of single and repetitive percutaneous tran-
of acute acalculous cholecystitis in intensive-care-unit patients, Eur shepatic gallbladder aspiration for the treatment of acute cholecys-
J Nucl Med 26:1317–1325, 1999. titis, J Gastroenterol 42(7):583–588, 2007.
Puc MM, et al: Ultrasound is not a useful screening tool for acute Tzovaras G, et al: Timing of laparoscopic cholecystectomy for acute
acalculous cholecystitis in critically ill trauma patients, Am Surg cholecystitis: a prospective nonrandomized study, World J Gastroen-
68:65–69, 2002. terol 12(34):5528–5531, 2006.
Reinders JS, et al: Incidence of bactobilia increases over time after Van der Linden W, Edlund G: Early versus delayed cholecystectomy:
endoscopic sphincterotomy, Dig Surg 28:289–292, 2011. the effect of a change in management, Br J Surg 68:753–757, 1981.
Rosen M, et al: Predictive factors for conversion of laparoscopic cho- Vauthey JN, et al: Indications and limitations of percutaneous chole-
lecystectomy, Am J Surg 184:254–258, 2002. cystostomy for acute cholecystitis, Surg Gynecol Obstet 176:49–54,
Ryu JK, et al: Clinical features of acute acalculous cholecystitis, J Clin 1993.
Gastroenterol 36:166–169, 2003. Wang AJ, et al: Clinical predictors of severe gallbladder complications
Saber A, Hokkam EN: Operative outcome and patient satisfaction in in acute acalculous cholecystitis, World J Gastroenterol 9:2821–2823,
early and delayed laparoscopic cholecystectomy for acute cholecys- 2003.
titis, Minim Invasive Surg 2014:162643, 2014. Warren BL: Small vessel occlusion in acute acalculous cholecystitis,
Savoca PE, et al: The increasing prevalence of acalculous cholecystitis Surgery 111:163–168, 1992.
in outpatients: results of a 7-year study, Ann Surg 211:433–437, Werbel GB, et al: Percutaneous cholecystostomy in the diagnosis and
1990. treatment of acute cholecystitis in the high-risk patient, Arch Surg
Schirmer BD, et al: Laparoscopic cholecystectomy: treatment of choice 124:782–786, 1989.
for symptomatic cholelithiasis, Ann Surg 213:665–676, 1991. Wong SK, et al: Percutaneous cholecystostomy and endoscopic chole-
Siddiqui T, et al: Early versus delayed laparoscopic cholecystectomy cystolithotripsy in the management of acute cholecystitis, Surg
for acute cholecystitis: a meta-analysis of randomized clinical trials, Endosc 13:48–52, 1999.
Am J Surg 195(1):40–47, 2008.
Singhal T, et al: Laparoscopic subtotal cholecystectomy: initial experi-
ence with laparoscopic management of difficult cholecystitis, Surgeon
7(5):263–268, 2009.
CHAPTER 34
Percutaneous treatment of gallbladder disease
Jad Abou Khalil, George Zogopoulos, and Jeffrey Stewart Barkun
OVERVIEW and illustrate some of its technical aspects and potential com-
plications. We will place a particular emphasis on the quality of
The first reports of an operative cholecystostomy are attribut- the evidence available in the literature and propose guidelines
able to Johannes Fabricius (1618) and Stalpert Von Der Wiel for the management of patients considered for PC.
(1667), who described the procedure as occurring almost by
happenstance (Hardy, 1993). The following two centuries only
revealed further sporadic reports, until Marion Sims, an Ameri- INDICATIONS AND CONTRAINDICATIONS FOR
can surgeon in Paris, performed a clearly intentional cholecys- PERCUTANEOUS CHOLECYSTECTOMY
tostomy in 1878. Bobbs had previously performed an inadvertent
kitchen-table cholecystostomy. Kocher and Tait (Traverso, Acute Calculous Cholecystitis in High-Risk Patients
1976) formalized the procedure in 1878, many months after Despite its prevalent use, there is a dearth of data from high-
Sims but independent of his efforts. quality clinical trials to support the use of PC in patients with
At a time when cholecystectomy (see Chapter 35) became acute calculous cholecystitis (ACC) who also have significant
the gold standard for the management of most acute gallblad- medical comorbidities (see Chapter 33). Despite this lack of
der diseases, operative cholecystostomy remained an attractive supporting data, this procedure is relatively common. In reviews
alternative in situations of significant patient comorbidity or of surgical databases, patients who receive PC are demonstrably
intraoperative risk, often as a bridge to cholecystectomy (Glenn, older and have greater medical comorbidity than those receiv-
1977; Skillings et al, 1980). ing LC (Anderson et al, 2013; Smith et al, 2013).
The first description of an ultrasound-guided percutaneous A systematic review of cohort studies comparing PC with
cholecystostomy (PC) for acute cholecystitis followed the open and/or LC for ACC identified 53 studies examining the
development of percutaneous biliary drainage for the manage- question; however, differences in outcome reporting, biases in
ment of obstructive jaundice and dates back to 1980 (Radder, control selection, and significant heterogeneity of study popula-
1980; van Sonnenberg et al, 1992). Early case series showed tions made it impossible to draw conclusive recommendations
encouraging results in patients who were not candidates for on clear indications in high-risk surgical patients (Winbladh
open cholecystectomy (Eggermont et al, 1985), and ensuing et al, 2009). The review confirmed that mortality rates after PC
cohort studies popularized its use in circumstances in which are high (15.4% vs. 4.5% with cholecystectomy). Aside from
cholecystectomy was not feasible. The popularity of this pro- the obvious selection bias, this rate of mortality is also likely
cedure has persisted into the laparoscopic era. In the absence inflated by the high mortality in early cohorts. The review
of comparative data, PC has supplanted open and laparoscopic reported a combined mortality following PC of 22.1% and
alternatives to cholecystostomy and become widely accepted as 13.3% in cohorts before and after 1995, respectively (P <.001).
a treatment for cholecystitis in situations in which surgical In one of the only matched studies comparing PC and
intervention is not feasible or deemed too risky (Fig. 34.1). A LC, Smith and colleagues (2013) reported a single-institution
review of nationwide medicoadministrative data shows that the retrospective analysis during two time periods: 1989-1998
number of PCs performed in the United States for all indica- and 1998-2009 (143 vs. 286 PC and LC, respectively). Patients
tions has increased sixfold between 1994 and 2004. The vast were matched for time period alone rather than any clinical
majority of PCs, 97% in that time period, have been performed patient factor. The authors demonstrated changing trends
by interventional radiologists (Duszak & Behrman, 2012) (see in the characteristics of patients receiving PC for acute chole-
Chapter 30). An examination of the National Inpatient Sample cystitis: From 1989-1998, 100% of patients were American
database shows that between 1998 and 2010, 1.5% of calculous Society of Anesthesiologists (ASA) 3 and 4, whereas from
and 7.5% of acalculous cholecystitis in the United States were 1998 to 2009, that proportion dropped to 82%. The use of PC
treated with PC (Anderson et al, 2013) rather than by open or in patients with ASA scores of 1 and 2 is therefore increasing
laparoscopic cholecystectomy (LC). with time.
In addition to its use for the management of acute cholecys- A retrospective cohort comparing 51 patients undergoing
titis, the PC technique also provided its early practitioners with PC with 151 undergoing LC for ACC and acute acalculous
the opportunity to investigate the management of gallstone cholecystitis (AAC) within the Veteran’s Affairs Boston Health
disease nonsurgically by chemical dissolution, mechanical System demonstrated again that patients selected for PC were
extraction, or lithotripsy of gallbladder calculi (van Sonnenberg older, had higher Charlson Comorbidity Index (CCI) scores,
et al, 1990, 1992). However, these approaches are not curative ASA classifications, white blood cell counts, and alkaline phos-
and have largely been abandoned due to the high rate of recur- phatase levels. The authors performed a multivariate logistic
rence of cholecystitis and the superiority of LC. In this chapter, regression and demonstrated that high CCI was the only inde-
we will examine the indications and contraindications for PC pendent predictor of receiving a PC.
564
A. Gallstones and Gallbladder Chapter 34 Percutaneous treatment of gallbladder disease 565
A B
FIGURE 34.1. Computer tomography–guided percutaneous cholecystostomy for acute acalculous cholecystitis in a patient with severe arterioscle-
rosis and ischemic heart disease. A, Percutaneously placed drainage catheter. B, Pigtail catheter in the gallbladder. The gallbladder wall is thickened,
and there is hyperdense sludge within the gallbladder.
Only two randomized studies examining the use of PC in the proportion of conversion to open cholecystectomy at opera-
acute cholecystitis have been published; however, they ask dif- tion. The authors did not define inclusion or exclusion criteria,
ferent questions, and both have significant methodologic limi- and no definition was provided for “a high-risk” surgical
tations. Hatzidakis and colleagues (2002) randomly assigned patient. The study was also not adequately powered to detect
123 high-risk patients (Acute Physiology and Chronic Health meaningful differences in the outcomes it compared. Further-
Evaluation [APACHE] score > 12) with ultrasound-proven more, the trial did not state how randomization was performed,
ACC or AAC to ultrasound-guided PC versus conservative lacked blinding or allocation concealment, and analysis was
medical therapy (63 and 60 patients, respectively). In this trial, performed strictly per-protocol despite the high drop-out rate.
ultrasound-guided PC was unsuccessful in 5% of patients who Such methodologic shortcomings make it difficult to make
had to undergo computed tomography (CT)-guided PC. Early recommendations based on this study.
in the series, the investigators used a nephrostomy tube instead A Cochrane Collaboration Systematic Review (Gurusamy
of a more secure pigtail catheter, resulting in a 15% (9/63) rate et al, 2013) on the use of PC in high-risk patients with ACC
of tube dislodgment, with 3 (5%) patients requiring operative identified the aforementioned two randomized trials (Akyürek,
intervention. These 3 patients all had a transperitoneal route 2005; Hatzidakis et al, 2002), grading their quality as very low
of entry. Only 3 (5%) of the PC patients required cholecystec- and their risk of bias as very high. The group was unable to
tomy within 30 days of failure of PC to control symptoms. The generate a recommendation on the usefulness of PC compared
authors were unable to demonstrate a difference in symptom with conservative therapy, calling for higher-quality random-
resolution after 3 days or in 30-day mortality and concluded ized trials. The largest multicenter randomized trial to date, the
that PC was indicated if symptoms failed to resolve after 3 days CHOCOLATE trial, has published its protocol but is currently
of conservative medical management. The study, however, had still recruiting (Kortram et al, 2012).
significant methodologic flaws. There was no power analysis, In conclusion, despite the absence of convincing high-level
no measures taken to conceal allocation, and no attempt to evidence, and based mainly on retrospective unmatched
blind the investigators, the patients, or the care providers. cohorts, PC is frequently recommended for patients with ACC
Moreover, an evolving procedural learning curve within the who do not respond to medical therapy and who are deemed
study period makes PC appear more morbid than it likely is in unfit for surgery due to age or medical comorbidity.
more experienced hands. The second randomized trial by
Akyürek and colleagues (2005) randomly assigned 70
Acute Calculous Cholecystitis With
ultrasound-proven AAC patients to PC, followed by early LC Delayed Presentation
3 to 4 days later (37 patients), compared with conservative PC, followed by interval LC, has been used in the specific
management and delayed LC 8 weeks following recovery (33 context of ACC late in the course of the disease, hoping that
patients). The investigators aimed to compare hospital length conversion to open cholecystectomy would be less than during
of stay, cost, and treatment success. Treatment success was LC at the index presentation (see Chapters 33 and 35). This
defined as resolution of signs and symptoms of cholecystitis, as was the subject of a retrospective cohort study of ASA one half
well as a decrease in white blood cell count. Six of 37 patients patients seen with ACC after more than 72 hours of symptoms
in the PC group did not go on to have early LC because they and not responding to nonoperative treatment for 48 hours. It
persistently had an APACHE score greater than 12 for 120 compared PC with delayed LC 4 weeks later (48 patients) to
hours after PC. Three of 33 patients in the delayed LC group LC (43 patients). The authors found a lower frequency of
were excluded—1 because of death from sepsis and 2 because conversion to open surgery, shorter hospital stay, and fewer
of refusal to undergo LC. In this study, PC followed by early total complications with PC and delayed LC (40% vs. 19%;
LC decreased hospital length of stay and costs compared with P = .029 in the early LC vs. early PC and delayed LC group)
delayed LC after conservative management but did not decrease (Karakayali et al, 2014). The main methodologic limitation of
A B
C
FIGURE 34.2. Percutaneous transhepatic cholecystostomy for acute acalculous hemorrhagic cholecystitis. After an initial gallbladder puncture
under ultrasound guidance with a 22-gauge needle, the procedure was continued under fluoroscopic control. A, Direct cholecystography showed
a large filling defect in the infundibulum of the gallbladder. B, The 22-gauge access was converted to a larger caliber, and a guidewire was introduced
into the gallbladder. A 6.5-Fr self-retaining pigtail catheter was placed in the gallbladder lumen. There was no opacification of the cystic duct at the
time of the initial procedure. C, Repeat cholangiography was done 1 week after emergency cholecystostomy. The cystic duct was patent, and there
was unimpeded flow of contrast material into the duodenum. Most of the debris within the gallbladder had disappeared. The cholecystostomy
catheter was removed 1 week after cholecystostomy; elective cholecystectomy was not required.
the study was selection bias: It is unclear how patients in the trend toward higher 30-day mortality in the PC group (13%).
early cholecystectomy group were selected, and no attempt was The deaths in the PC group were due to advanced cancer,
made at multivariate adjustment for baseline differences illustrating the significant selections bias that went into treat-
between the groups. Zehetner and colleagues (2014) compared ment attribution, a bias that remains uncorrected by matching
23 patients with ultrasound-confirmed ACC presenting more on basic demographic characteristics. It is also difficult to draw
than 72 hours after symptom onset and treated with PC with a statistically sound conclusion from such a small sample size.
controls treated by LC. The authors pair-matched patients for
age, sex, race, body mass index, diabetes, and sepsis. Contrary Acute Acalculous Cholecystitis in High-Risk Patients
to the previous study, the authors demonstrated increased The use of PC in high-risk patients with AAC is supported by
length of hospital stay (LOS) with PC and no difference in many cohort studies demonstrating decreased morbidity and
major morbidity; however, there was a significant proportion of mortality when compared with LC or open cholecystectomy
conversion to open surgery in the no-PC group (17%) and a (OC) (Simorov et al, 2013) (Fig. 34.2). However, when
A. Gallstones and Gallbladder Chapter 34 Percutaneous treatment of gallbladder disease 567
compared with conservative medical therapy, different studies mortality or serious morbidity. These case reports and case
reach different conclusions, with some large studies showing no series suggest that PC is feasible during pregnancy and may be
benefit attributable to PC (see Chapter 33). used as a safe alternative to operative intervention after failure
An examination of 43,341 patients with AAC and severe of medical management in the first or third trimesters.
sepsis or shock within the California Office of Statewide Health
Planning and Development Patient Discharge Database (1995- Sepsis of Unknown Origin in the Intensive Care
2009) demonstrated that patients undergoing PC had the In the critically ill patient with persistent sepsis of unknown
greatest number of comorbidities and were the sickest. However, origin, PC affords a safe and rapid therapeutic intervention in
no improvement in survival was associated with PC versus the event of suspected AAC or ACC, and it can be used to
medical management after extensive multivariate adjustment exclude the gallbladder as the septic focus. Two case series
(Anderson et al, 2014), and this remained true even in the describe such an approach: In 24 (Lee et al, 1991) and 82
subgroup of patients with severe sepsis or shock and those patients (Boland et al, 1994), cholecystostomy resulted in reso-
dependent on mechanical ventilation. In fact, patients seemed lution of sepsis in 58% and 51% of patients, respectively. These
to fare best when they could undergo cholecystectomy alone or case series were, however, not followed by controlled trials.
treated with PC followed by cholecystectomy. The study did Moreover, it is possible that the incidence of AAC was higher
not consider short-term outcomes, length of stay, symptom compared with contemporary practice due to different intensive
control, and quality-of-life measures. This large study of medicare practices, such as prolonged enteric starvation. We postu-
coadministrative data is the strongest evidence that PC might late that “diagnostic PC” has less utility in the modern intensive
not offer a mortality benefit over medical management in care setting.
patients with AAC who are deemed unfit for surgery.
An analysis of 58,518 patients with AAC from another med- Conclusion
icoadministrative database, the National Inpatient Sample PC should be reserved for patients who are not medically fit to
(Anderson et al, 2013), demonstrated that, after multivariate undergo open or LC (Abi-Haidar et al, 2012). It may be used
adjustment (for sex, age, race, Charlson index, and hospital as a bridge to cholecystectomy in patients failing medical
teaching status), PC remained associated with increased odds therapy for ACC and not fit for surgery at the time of presenta-
of death, increased length of stay, and decreased odds of tion, and it can be used as definitive therapy in patients with
gallbladder-specific and total complications, indicating that in AAC.
both ACC and AAC its use was reserved for selected patients.
The study did not compare PC with nonoperative manage- TECHNICAL ASPECTS AND COMPLICATIONS
ment, and there is no doubt that residual statistical confound-
ing is a concern in such large retrospective databases. Insertion Technique
Therefore whether or not PC in critically ill patients with 1. Drainage versus aspiration: Drainage has been demonstrated
ACC offers a survival advantage over the absence of gallbladder- to be superior to simple aspiration of the acutely inflamed
directed interventions remains unclear, as our knowledge is gallbladder in a single randomized trial (Ito et al, 2004) and
derived from retrospective cohort studies with varying degrees is currently most widely used.
of risk adjustment and significant residual confounding 2. Operator: Most commonly, PC is performed by radiologists
(Mansour & Yopp, 2014). (Duszak & Behrman, 2012) (see Chapter 30) under ultra-
sonographic guidance, but a single publication reports on a
Special Populations series performed by surgeons (Silberfein et al, 2007), with
Pregnancy acceptable results and improved access times, although this
Although the safety of LC is well established in the second experience may not be generalizable to most institutions.
trimester (Lu et al, 2004; Swisher et al, 1994), most surgeons Either way, it is advisable that only individuals highly facile
are reluctant to operate in the first and third trimesters due to with ultrasonographic or CT-guided interventional tech-
fear of spontaneous abortion and preterm labor, respectively. niques perform PC and that there be regular institutional
PC may thus provide a safe temporizing measure for pregnant quality monitoring, according to proposed procedure-
patients seen with severe biliary colic or cholecystitis in the first specific quality-control benchmarks (Saad et al, 2010).
and third trimester who fail conservative medical therapy. The 3. Approach: Patient preparation consists of skin disinfection
evidence to support this approach is scarce, and no large cohort and standard sterile precautions, procedural sedation, and
or controlled trials has examined this question. Two small either prophylactic administration or ready availability of
studies have reported the feasibility and safety of PC in preg- atropine, as vagal reactions have been documented (Little
nancy for acute cholecystitis and biliary colic in a limited et al, 2013). PC is performed under direct ultrasound or CT
number of patients. Allmendinger and colleagues (1995) pub- guidance, with equivalent success and complication rates
lished a case report of two pregnant patients who underwent (Loberant et al, 2010). After the planned anatomic course
PC for cholecystitis at 30 and 32 weeks of gestation, respec- is identified, the gallbladder is entered using a finder needle,
tively, followed by elective cholecystectomy in the postpartum and a guidewire is inserted into it, with a size 7- to 10-Fr
phase. Chiappetta Porras and colleagues (2009) described a pigtail catheter inserted over the guidewire using a Seldinger
cohort of 122 (Technical Aspects and Complications) for biliary technique. Other catheters, such as a central venous cathe-
colic in the first trimester, four cholecystostomies in the first ter, can be used in lieu of a pigtail catheter and may be of
trimester for cholecystitis, and four gallbladder aspirations in use in certain institutional settings, but data on catheter
the third trimester (three for biliary colic and one for cholecys- displacement and patency times with these other options
titis.) In this series, patients seen with biliary disease were are limited (Park et al, 2005). A catheter placed by transhe-
treated exclusively with LC, with no reported fetal or maternal patic access is preferred to the transperitoneal route. The
transhepatic tract matures faster than the transperitoneal 22% at 1 year but also a 1 year mortality of 37% following
tract (2 weeks vs. 3 weeks, respectively), allowing faster index admission, mainly from medical causes unrelated to gall-
removal of the catheter as well as less bile leaks and stone disease (Sanjay et al, 2013). Although cholecystectomy
bile peritonitis (Hatjidakis et al, 1998). The transhepatic may be desirable after PC for ACC, a review of a large medi-
approach is safer because it avoids the peritoneum in patients coadministrative database demonstrates that PC is, in fact, only
with significant ascites or bowel interposition (Ginat & Saad, followed by elective cholecystectomy in 40% of patients at 1
2008). However, it is associated with the potential for pneu- year, whereas an additional 18% die without undergoing cho-
mothorax, bleeding, and hemobilia (van Sonnenberg, 1990). lecystectomy (de Mestral et al, 2013). It is therefore prudent
to weigh the risk of interval LC with the patient’s concurrent
Complications medical comorbidities, evolving surgical risks, and life expec-
In experienced hands, PC has both a high technical success rate tancy against the risk of recurrence and ensuing morbidity and
(>98.9%) (Winbladh et al, 2009) and a low complication rate. mortality.
This includes hemorrhage (2.5%), pneumothorax (0.5%), and
postprocedural catheter dislodgement (<1%) (Saad et al, PERCUTANEOUS TREATMENT OF GALLSTONES:
2010). Following catheter withdrawal, 3% of patients experi- TECHNIQUES OF HISTORIC INTEREST
ence severe bile peritonitis, and another 3% experience a mild
symptomatic biliary leak (Wise et al, 2005). Bile leaks are asso- Many techniques have been developed in an attempt to avoid
ciated with transperitoneal access (Sanjay et al, 2013) and may the need for elective cholecystectomy. The 1980s saw the rise
require operative therapy or repeat PC. In a systematic review in popularity of extracorporeal shock-wave lithotripsy (ESWL),
of 35 papers, displacement of the PC catheter was reported in which in association with oral bile salt therapy, promised to
8.57%; however, this may be an underestimation in an outpa- clear the gallbladder of gallstones without resorting to surgery.
tient population (Winbladh et al, 2009). In the cohort described However, its use was limited to patients with less than three
by Smith and colleagues (2013), catheter-related complications cholesterol stones of size 2 cm or less, and it required multi-
occurred in 14.5% of patients, including tube dislodgement in ple treatments over months. The proportion of patients with
10 of 143 (7%), extra-abdominal peritube bile leak in 7 of 143 recurrent cholelithiasis at 6 years was 69%, making ESWL an
(5%), pain in 4 of 143 (3%), and occlusion in 4 of 143 (3%). unacceptable option for most patients (Cesmeli et al, 1999).
These complications were managed mostly with tube reposi- Moreover, it has been proven to not be a cost-effective
tioning or upsizing, and none required operation. approach (Barkun et al, 1997), and its use has been largely
abandoned.
Management of the Percutaneous Cholecystectomy Methyl tert-butyl ether (MTBE), a cholesterol solvent, has
Catheter and the Gallbladder been used to dissolve gallbladder calculi after the placement of
As already stated, the catheter tract usually matures after 2 a PC. However, its use is cumbersome, requires specialized
weeks for transhepatic catheters and after 3 weeks for transperi- equipment and tubing that will not be destroyed by MTBE, as
toneal drains. A cholecystogram can be obtained prior to well as a bothersome, prolonged administration over many days
removal of the catheter 3 to 6 weeks after insertion, but imaging (Thistle et al, 1989). Its use is mainly of historic interest.
is not necessary in cases where a small-bore catheter was placed A publication has suggested the use of PC as an adjunct to
and where the cystic duct was patent at insertion. In AAC, cystic duct stenting in patients deemed unfit for LC (Elmunzer
assuming the resolution of precipitating factors, PC is consid- et al, 2011). Only four patients were described in this study,
ered curative, and LC is usually not necessary. There remains and no information on long-term patency of this technique was
a debate as to whether PC can be considered curative after described. A small series has described the feasibility of endo-
ACC (Chung et al, 2012). Recurrence rates of cholecystitis scopic ultrasound-guided transduodenal drainage of the gall-
after PC in ACC without interval cholecystectomy vary between bladder; however, this experience has not been reproduced to
studies: from 4.1% at 1 year in one study (Li et al, 2013) to date (Lee et al, 2007).
11.7% at 38 months in another (Chang et al, 2014). A cohort
from the United Kingdom shows a recurrence rate as high as References are available at expertconsult.com.
A. Gallstones and Gallbladder Chapter 34 Percutaneous treatment of gallbladder disease 568.e1
REFERENCES Kortram K, et al: Acute cholecystitis in high risk surgical patients:

percutaneous cholecystostomy versus laparoscopic cholecystectomy
Abi-Haidar Y, et al: Revisiting percutaneous cholecystostomy for acute (CHOCOLATE trial): study protocol for a randomized controlled
cholecystitis based on a 10-year experience, Arch Surg 147(5):416– trial, Trials (Online) 13:7, 2012.
422, 2012. Lee SS, et al: EUS-guided transmural cholecystostomy as rescue man-
Akyürek N, et al: Management of acute calculous cholecystitis in high- agement for acute cholecystitis in elderly or high-risk patients: a
risk patients: percutaneous cholecystotomy followed by early laparo- prospective feasibility study, Gastrointest Endosc 66(5):1008–1012,
scopic cholecystectomy, Surg Laparosc Endosc Percutan Tech 15(6): 2007.
315–320, 2005. Lee MJ, et al: Treatment of critically ill patients with sepsis of unknown
Allmendinger N, et al: Percutaneous cholecystostomy treatment of cause: value of percutaneous cholecystostomy, AJR Am J Roentgenol
acute cholecystitis in pregnancy, Obstet Gynecol 86(4 Pt 2):653–654, 156(6):1163–1166, 1991.
1995. Li M, et al: Percutaneous cholecystostomy is a definitive treatment for
Anderson JE, et al: A nationwide examination of outcomes of percuta- acute cholecystitis in elderly high-risk patients, Am Surg 79(5):524–
neous cholecystostomy compared with cholecystectomy for acute 527, 2013.
cholecystitis, 1998-2010, Surg Endosc 27(9):3406–3411, 2013. Little MW, et al: Percutaneous cholecystostomy: the radiologist’s role
Anderson JE, et al: Cholecystostomy offers no survival benefit in in treating acute cholecystitis, Clin Radiol 68(7):654–660, 2013.
patients with acute acalculous cholecystitis and severe sepsis and Loberant N, et al: Comparison of early outcome from transperitoneal
shock, J Surg Res 190(2):517–521, 2014. versus transhepatic percutaneous cholecystostomy, Hepatogastroen-
Barkun AN, et al: Costs and effectiveness of extracorporeal gallbladder terology 57(97):12–17, 2010.
stone shock wave lithotripsy versus laparoscopic cholecystectomy: a Lu EJ, et al: Medical versus surgical management of biliary tract disease
randomized clinical trial, Int J Technol Assess Health Care 13(04):589– in pregnancy, Am J Surg 188(6):755–759, 2004.
601, 1997. Mansour JC, Yopp AC: Percutaneous cholecystostomy: the challenges
Boland GW, et al: Percutaneous cholecystostomy in critically ill of cohort analysis, J Surg Res 190(2):417–418, 2014.
patients: early response and final outcome in 82 patients, AJR Am Park SH, et al: Percutaneous cholecystostomy using a central venous
J Roentgenol 163(2):339–342, 1994. catheter is effective for treating high-risk patients with acute chole-
Cesmeli E, et al: Gallstone recurrence after successful shock wave cystitis, Surg Laparosc Endosc Percutan Tech 15(4):202–208, 2005.
therapy: the magnitude of the problem and the predictive factors, Radder RWR: Ultrasonically guided percutaneous catheter drainage for
Am J Gastroenterol 94(2):474–479, 1999. gallbladder empyema, Diagn Imaging 49(6):330–333, 1980.
Chang YR, et al: Percutaneous cholecystostomy for acute cholecystitis Saad WE, et al: Quality improvement guidelines for percutaneous tran-
in patients with high comorbidity and re-evaluation of treatment shepatic cholangiography, biliary drainage, and percutaneous chole-
efficacy, Surgery 155(4):615–622, 2014. cystostomy, J Vasc Interv Radiol 21(6):789–795, 2010.
Chiappetta Porras LT, et al: Minimally invasive management of acute Sanjay P, et al: Clinical outcomes of a percutaneous cholecystostomy
biliary tract disease during pregnancy, HPB Surg 2009:829020, for acute cholecystitis: a multicentre analysis, HPB (Oxford)
2009. 15(7):511–516, 2013.
Chung YH, et al: Can percutaneous cholecystostomy be a definitive Silberfein EJ, et al: Percutaneous cholecystostomy for acute cholecys-
management for acute acalculous cholecystitis? J Clin Gastroenterol titis in high-risk patients: experience of a surgeon-initiated interven-
46(3):216–219, 2012. tional program, Am J Surg 194(5):672–677, 2007.
de Mestral C, et al: Cholecystostomy: a bridge to hospital discharge Simorov A, et al: Emergent cholecystostomy is superior to open cho-
but not delayed cholecystectomy, J Trauma Acute Care Surg lecystectomy in extremely ill patients with acalculous cholecystitis: a
74(1):175–179, discussion 179-180, 2013. large multicenter outcome study, Am J Surg 206(6):935–940, discus-
Duszak R Jr, Behrman SW: National trends in percutaneous cholecys- sion 940-931, 2013.
tostomy between 1994 and 2009: perspectives from Medicare pro- Skillings JC, et al: Cholecystostomy: a place in modern biliary surgery?
vider claims, J Am Coll Radiol 9(7):474–479, 2012. Am J Surg 139(6):865–869, 1980.
Eggermont AM, et al: Ultrasound-guided percutaneous transhepatic Smith TJ, et al: Changing trends and outcomes in the use of percutane-
cholecystostomy for acute acalculous cholecystitis, Arch Surg ous cholecystostomy tubes for acute cholecystitis, Ann Surg 257(6):
120(12):1354–1356, 1985. 1112–1115, 2013.
Elmunzer BJ, et al: Percutaneous cholecystostomy as a bridge to defini- Swisher SG, et al: Biliary disease during pregnancy, Am J Surg 168(6):
tive endoscopic gallbladder stent placement, Clin Gastroenterol 576–579, discussion 580-571, 1994.
Hepatol 9(1):18–20, 2011. Thistle JL, et al: Dissolution of cholesterol gallbladder stones by methyl
Ginat D, Saad WE: Cholecystostomy and transcholecystic biliary tert-butyl ether administered by percutaneous transhepatic catheter,
access, Tech Vasc Interv Radiol 11(1):2–13, 2008. N Engl J Med 320(10):633–639, 1989.
Glenn F: Cholecystostomy in the high risk patient with biliary tract Traverso LW: Carl Langenbuch and the first cholecystectomy, Am J
disease, Ann Surg 185(2):185–191, 1977. Surg 132(1):81–82, 1976.
Gurusamy KS, et al: Percutaneous cholecystostomy for high-risk surgi- van Sonnenberg E, et al: Percutaneous gallbladder puncture and cho-
cal patients with acute calculous cholecystitis, Cochrane Database Syst lecystostomy: results, complications, and caveats for safety, Radiology
Rev (8):CD007088, 2013. 183(1):167–170, 1992.
Hardy KJ: Carl Langenbuch and the Lazarus Hospital: events and van Sonnenberg E, et al: Interventional gallbladder procedures, Radiol
circumstances surrounding the first cholecystectomy, ANZ J Surg Clin North Am 28(6):1185–1190, 1990.
63(1):56–64, 1993. van Sonnenberg E, et al: Percutaneous gallbladder puncture and cho-
Hatjidakis AA, et al: Maturation of the tract after percutaneous chole- lecystostomy: results, complications, and caveats for safety, Radiology
cystostomy with regard to the access route, Cardiovasc Intervent 183(1):167–170, 1992.
Radiol 21(1):36–40, 1998. Winbladh A, et al: Systematic review of cholecystostomy as a treatment
Hatzidakis AA, et al: Acute cholecystitis in high-risk patients: percuta- option in acute cholecystitis, HPB (Oxford) 11(3):183–193, 2009.
neous cholecystostomy vs conservative treatment, Eur Radiol 12(7): Wise JN, et al: Percutaneous cholecystostomy catheter removal and
1778–1784, 2002. incidence of clinically significant bile leaks: a clinical approach to
Hatzidakis AA, et al: Acute cholecystitis in high-risk patients: percuta- catheter management, AJR Am J Roentgenol 184(5):1647–1651,
neous cholecystostomy vs conservative treatment, Eur Radiol 12(7): 2005.
1778–1784, 2002. Zehetner J, et al: Percutaneous cholecystostomy versus laparoscopic
Ito K, et al: Percutaneous cholecystostomy versus gallbladder aspira- cholecystectomy in patients with acute cholecystitis and failed con-
tion for acute cholecystitis: a prospective randomized controlled trial, servative management: a matched-pair analysis, Surg Laparosc Endosc
AJR Am J Roentgenol 183(1):193–196, 2004. Percutan Tech 24(6):523–527, 2014.
Karakayali FY, et al: Emergency cholecystectomy vs percutaneous cho-
lecystostomy plus delayed cholecystectomy for patients with acute
cholecystitis, Hepatobiliary Pancreat Dis Int 13(3):316–322, 2014.
CHAPTER 35
Technique of cholecystectomy:
open and minimally invasive
Flavio G. Rocha and Jesse Clanton
OVERVIEW back. These attacks are frequently postprandial, related to a

fatty or “spicy” meal, or sometimes they may awaken the
Gallstones are an extremely common condition, occurring in patient from sleep. Although a large proportion of the popula-
approximately 10% to 20% of the adult population (see Chapter tion has gallstones, prophylactic surgery for incidentally found
32). The preferred treatment for symptomatic gallstone disease stones is controversial. An asymptomatic patient generally has
is cholecystectomy. Open cholecystectomy remained the pre- less than a 20% chance of ever having symptoms, and therefore
ferred surgical option for patients with symptomatic gallstone in most patients the risks of a prophylactic operation outweigh
disease until the advent of laparoscopic cholecystectomy in the the potential benefit (Fendrick et al, 1993; Ransohoff & Gracie,
late 1980s. Because early results were so promising regarding 1993; Ransohoff et al, 1983). In certain settings, however, such
decreasing pain, costs, and hospital stay, rapid adoption of the as patients with sickle cell disease, those requiring long-term
laparoscopic technique ensued worldwide (Escarce et al, 1995; total parenteral nutrition, and those who are therapeutically
Legorreta et al, 1993; Nenner et al, 1994; Steiner et al, 1994). immunosuppressed due to solid-organ transplantation, prophy-
Shortly thereafter, there was a rising concern that the new lapa- lactic cholecystectomy may be justified. Because hepatic or
roscopic technique was associated with an increased risk for vasoocclusive disease from sickle cell disease may be difficult
common bile duct (CBD) injuries (see Chapters 38 and 42). to differentiate from acute cholecystitis (Tagge et al, 1994),
After this was recognized, increased efforts in awareness, educa- prophylactic surgery is often considered in these patients. Cho-
tion, and training have decreased the risk. Laparoscopic lecystectomy is generally recommended in the transplant popu-
cholecystectomy is now considered the gold standard for lation due to the concern that chronic immunosuppression can
the treatment of symptomatic gallstone disease (Soper et al, mask signs and symptoms of cholecystitis, leading to a potential
1992). During the last several years, additional techniques for delay in diagnosis and risk of more severe complications
have been developed in an attempt to improve upon the lapa- (Hull et al, 1994). However, the timing of such cholecystec-
roscopic approach. This has taken many forms, including tomy is not universally accepted, with some literature suggest-
smaller and fewer ports and instruments introduced through a ing mandatory screening and treatment of biliary disease before
single mini-incision or via robotic surgery. However, at present transplantation (Girardet et al, 1989) to prophylactic cholecys-
there is no compelling evidence that any of these newer innova- tectomy 6 months after transplantation (Boline et al, 1991) to
tions improve outcome metrics, such as pain, complications, expectant management of all asymptomatic patients (Fendrick
return to function, or cost, although they may have some et al, 1993; Steck et al, 1991). The patient without gallstones
impact on patient satisfaction when compared with standard but whose symptoms are typical for biliary pain should be
four-port laparoscopic cholecystectomy. Because open chole- considered for the procedure if there is evidence of acalculous
cystectomy is performed more infrequently, younger surgeons cholecystitis or biliary dyskinesia (Soper, 1991). Additional
trained in this modern era are less familiar with this operation. pathologies such as gallstone pancreatitis and gallbladder
This chapter discusses the technique of open and laparoscopic polyps greater than 1 cm in size (see Chapter 49) warrant
cholecystectomy, current status of laparoscopic cholecystec- cholecystectomy. At the time of laparotomy or laparoscopy for
tomy, and variations on the technique with special attention to an unrelated pathology, cholecystectomy can be considered for
details for reducing operative complications, including bile duct those with documented gallstones even if currently asymptom-
injury. atic because the incremental risk of the added procedure to the
index operation is generally small. National and worldwide
INDICATIONS trends of increased rates of cholecystectomy rising along with
the popularity of laparoscopy suggest that surgeons have
Indications for cholecystectomy should be the same, regardless become more liberal with performing cholecystectomy.
of technique (i.e., open or laparoscopic procedure) (Box 35.1). Once the indication for cholecystectomy is confirmed,
Because cholelithiasis afflicts 10% to 20% of the population in timing of surgery is the next consideration. For acute cholecys-
the United States, many patients will be treated for symptoms titis this may depend on the severity of inflammation (see
or complications of their gallstones (see Chapter 32). Most Chapter 33). Unless patient comorbidities or hemodynamic
commonly, the pain arising from a stone that temporarily instability do not allow it, the preferred approach to cholecys-
obstructs the gallbladder, known as biliary colic, is severe and titis is urgent surgery. Confirmed in both the open and laparo-
episodic in nature. It is predominantly in the right upper quad- scopic literature, performing cholecystectomy within 72 hours
rant of the abdomen or epigastrium and often radiates to the of symptom onset has been demonstrated in prospective
569
associated with decreased hospital stay and quicker recovery to

BOX 35.1 Indications for Cholecystectomy
normal activities (Ros et al, 2001; Rosenmüller et al, 2013).
Indications for Cholecystectomy An initial concern regarding the laparoscopic technique was
Symptomatic cholelithiasis a higher risk of bile duct injury (see Chapter 42). During the
• Biliary colic
beginning of widespread adoption of laparoscopic cholecystec-
• Acute cholecystitis
tomy, initial reports and small series of patients demonstrated
• Chronic cholecystitis
Asymptomatic cholelithiasis an increased rate of bile duct injuries compared with the
• Sickle cell disease accepted rates of an open procedure (Dunn, 1994; Steele,
• Total parental nutrition 1995). Larger single-center and multicenter studies were then
• Chronic immunosuppression conducted and indicated that bile duct injury rates were actu-
• No immediate access to health care facilities (e.g., ally similar to those during the open era, ranging from less than
missionaries, military personnel, Peace Corps workers, relief 0.01% to 0.3%. However, they have not been completely elimi-
workers) nated (Croce et al, 1994; Cuschieri et al, 1991; Pesce et al,
• Incidental cholecystectomy for patients undergoing 2012; Underwood, 2000). Although bile duct injury is not the
abdominal surgery for other indications
only complication of cholecystectomy, a large population-based
Acalculous cholecystitis (including biliary dyskinesia)
study in Italy recently demonstrated laparoscopic cholecystec-
Gallstone pancreatitis
Gallbladder polyps greater than 1 mm in diameter tomy to be associated with a lower rate of both surgical-related
Porcelain gallbladder and systemic complications compared with an open technique
Proven or concern for gallbladder cancer (Agabiti et al, 2013).
Regarding efficacy, the laparoscopic approach treats gall-
stone disease as effectively as traditional open surgery. The
proportion of patients afflicted with biliary-type pain after lapa-
randomized trials to reduce complications and hospital stay roscopic removal of the gallbladder ranges from 6.4% to 10.6%
compared with interval cholecystectomy (Gurusamy et al, (Kane et al, 1995; Ure et al, 1995), similar to rates reported
2013; Lo et al, 1998). Nonurgent pathologies are generally after open cholecystectomy. Symptom relief is also influenced
performed on the basis of patient and surgeon preference. by the presence of gallstones, with 82% of patients with gall-
stones receiving symptom relief after 3 months compared with
LAPAROSCOPIC VERSUS OPEN TECHNIQUES only 52% of patients without gallstones (Fenster et al, 1995).
Cost-effectiveness has not proven to be superior in either group,
Laparoscopic cholecystectomy remains the standard of care for with some studies favoring the open and others the laparoscopic
acute cholecystitis and symptomatic cholelithiasis. Advantages approach (Bass et al, 1993; Fullarton et al, 1994; McMahon
include less postoperative pain, a lower incidence of incisional et al, 1994). This is likely influenced by a number of factors
hernias and adhesions, smaller scars, a shorter hospital stay, an and the equipment used during laparoscopy, because there
earlier return to full activity, and a decrease in the overall cost exists a wide range of equipment preferences for this common
(Barkun et al, 1992; Bass et al, 1993; McMahon et al, 1994; operation.
Soper et al, 1992). Surgeon volume also affects laparoscopic outcomes, with
A significant body of literature has been dedicated to the decreased rates of conversion, bile duct injury, lower length of
comparison of the open and laparoscopic techniques, including stay, and lower in-hospital mortality demonstrated by surgeons
randomized controlled trials (RCTs) and meta-analyses. Two who perform more than 15 laparoscopic cholecystectomies per
early, large prospective randomized trials showed promising but year (Csikesz et al, 2009). The value of experience and a learn-
mixed results. A large prospective randomized trial compared ing curve effect is also accepted, whereby the incidence of bile
310 patients undergoing laparoscopic cholecystectomy to 155 duct injuries is much higher in a surgeon’s early cases compared
patients undergoing open cholecystectomy via a minilaparot- with later in his or her experience (Moore & Bennett, 1995).
omy and demonstrated a higher rate of complications (9% vs. This is not surprising given the inherent differences and non-
3%) in the laparoscopic group but reduced postoperative anal- transferability of skills between the open and laparoscopic
gesia requirements and earlier return to normal activities and approaches.
work. (McGinn et al, 1995). An additional prospective ran- Finally, familiarity with the anatomy of the extrahepatic
domized trial of 200 patients during the same time period failed biliary tree and adjacent structures within the porta hepatis,
to demonstrate any differences in length of stay, return to work, especially the most common anatomic variations, is essential
or time to full recovery (Majeed et al, 1996). However, both of for any surgeon performing these procedures (see Chapter 2).
these studies were conducted during the relative infancy of
laparoscopic cholecystectomy. The mean length of stay in each LAPAROSCOPIC TECHNIQUE
study was 2 to 3 days for both groups of patients, and outpa-
tient cholecystectomy was not considered the norm that it is Operating Room Setup
today. Currently, laparoscopic cholecystectomy as an outpa- Before beginning the procedure, the surgeon must ensure that
tient procedure is universally accepted and considered safe even all equipment is present and functioning properly. This includes
in the elderly population (Rao et al, 2013). More recent pro- not only usual laparoscopic equipment, including clips, cautery,
spective trials performed after the initial learning curve of irrigation, and suction as needed, but also appropriate equip-
laparoscopic cholecystectomy are more favorable toward the ment for performing cholangiography and having instruments
laparoscopic approach. Large single-institution and multicenter readily available in the event that the procedure is converted to
studies have consistently demonstrated that although laparo- an open laparotomy (Box 35.2). A monitor is positioned over
scopic cholecystectomy has slightly longer operative times, it is each shoulder of the patient and angled toward the contralateral
A. Gallstones and Gallbladder Chapter 35 Technique of cholecystectomy: open and minimally invasive 571
cavity is performed, including not only the gallbladder but also

BOX 35.2 Equipment for Laparoscopic Cholecystectomy
the stomach, liver, omentum, intestines, and pelvic organs.
Equipment for Laparoscopic Cholecystectomy Evaluation is performed to identify any additional pathology,
Laparoscopic viewing equipment assess for adhesions, plan additional port placement, and
• Camera (5-mm and 10-mm sizes, 0 and 30 degrees each)
ensure that no iatrogenic injury was created upon initial entry
• Light source
into the abdomen. Two additional 5 mm ports are then placed
• Monitor (one over each shoulder)
Insufflator and tubing in the right upper quadrant under direct visualization. The first
Trocars (multiple 5-mm trocars and at least one 10-mm; long trocar is placed along the right anterior axillary line 1 to 2
trocars available for obese patients) centimeters below the costal margin. The second port is then
Instruments for dissection and removal of gallbladder inserted in a subcostal position along midclavicular line on the
• Clips (5-mm and 10-mm variety) right. A final 5 or 10 mm port is placed in the subxiphoid
• Electrocautery region, usually 2 to 4 cm below the xiphoid, depending on the
• Suction relative locations of the gallbladder and falciform ligament, with
• Irrigation the tip angled to the right of the falciform toward the gallblad-
• Endoscopic retrieval bag
der (Fig. 35.1). Placing this trocar too far superiorly can cause
Instruments available in the setting of complications or unusual
difficulty with dissection and induce arm strain. Alternatively,
circumstances
• Endoloop if a Veress needle is used in the left upper quadrant to obtain
• Argon laser pneumoperitoneum, it may be replaced here with a 5 mm-sized
• Cholangiography equipment port, eliminating the need for a subxiphoid port.
• Cholangiocatheter (Mixter, Olsen-Reddick, Whistle-tip, etc.) To facilitate exposure laparoscopically, the patient is placed
• Injectable contrast dye and saline in a reverse Trendelenburg position by 30 degrees and rotated
• C-arm and monitor for fluoroscopy to the left by 15 degrees. This maneuver allows the omentum,
Open cholecystectomy tray and instruments duodenum, and colon to move away from the liver and gall-
bladder. Occasionally, in patients with significant intraabdomi-
nal fat, additional port(s) may be required for retraction. Most
side. The patient is placed in the supine position with both arms commonly, this can be performed with a 10 mm port placed in
out on the operating room table. The patient should be placed the right midabdomen for insertion of a fan or similar instru-
on the table in such a position so that fluoroscopy can be ment to retract the greater omentum inferiorly and out of the
achieved (requiring a free space under the patient’s abdominal field of view.
cavity), even if a cholangiogram is not routinely scheduled for
the procedure. We recommend appropriate straps and a foot- Dissection and Critical View of Safety
board to secure the patient for necessary rotation of the table Once all the ports are placed, the surgeon may choose to work
to facilitate optimal exposure. After general anesthesia is via the subxiphoid port to dissect and clip while his left hand
induced, an orogastric tube is inserted and placed to suction to is used to either drive the camera or manipulate the gallbladder
decompress the stomach. from the right upper quadrant port. This will often depend on
the expertise of the assistant during the procedure as well as
Pneumoperitoneum surgeon preference. The description that follows assumes that
Pneumoperitoneum is essential for creating a proper visualiza- the surgeon is manipulating the camera with his left hand while
tion and a working space in the abdominal cavity. This can be the assistant retracts via both ports placed in the right upper
obtained in any fashion that the surgeon is most comfortable quadrant.
performing, traditionally by either a Veress needle or a Hasson Atraumatic grasping forceps are placed through each right
direct cut-down technique. Specifics of obtaining pneumoperi- upper quadrant trocar to manipulate and retract the gallblad-
toneum is out of the scope of this chapter, but either technique der. The most lateral retractor is first used to grasp the fundus
has been proven to be equally safe and effective in experienced of the gallbladder. Once a firm grasp is secured, the jaws of the
hands (Ahmad et al, 2012). Surgeons should ultimately be grasper are locked into place. The gallbladder is then retracted
familiar and comfortable with multiple techniques. Initial entry cephalad and laterally, raising the liver edge with it to facilitate
and the first port is commonly placed in an infraumbilical or exposure of the gallbladder neck. This maneuver may be chal-
supraumbilical position, depending on patient size and body lenging in the setting of a fixed heavy cirrhotic or fatty liver. In
habitus. Alternatively, a Veress needle may be placed in the left the case of a significantly tense and distended gallbladder, a
upper or right upper quadrant for initial access and insufflation. large-bore needle is used to puncture the gallbladder at the
Once access into the abdominal cavity is confirmed, the fundus and decompress the contents for easier manipulation.
abdomen is insufflated to 12 to 15 mm Hg using carbon Once the gallbladder is sufficiently retracted cephalad to
dioxide. The anesthesia team is alerted during insufflation to expose the neck of the gallbladder, an additional atraumatic
observe for hypercarbia, hypotension, or arrhythmias. If any of grasper is placed through the other right upper quadrant
these are observed, immediate desufflation is initiated. Gradual trocar to grasp the infundibulum. Avascular adhesions seen
reinsufflation may be attempted after the patient’s condition on the body of the gallbladder to the omentum, hepatic
has stabilized. flexure, or duodenum are taken down at this time with a com-
bination of blunt dissection and electrocautery until these adhe-
Port Placement and Exposure sions are satisfactorily stripped down to the level of the
After pneumoperitoneum is obtained, a 5- or 10-mm port is infundibulum (Fig. 35.2). The infundibulum is retracted later-
placed in the periumbilical region. A 0- or 30-degree laparo- ally and caudally to expose the triangle of Calot (Fig. 35.3). A
scope is then inserted, and inspection of the entire abdominal cautery instrument is then placed through the subxiphoid
Gallbladder Laparoscope
Liver
A
A B
FIGURE 35.1. Positions for insertion of trocars during laparoscopic cholecystectomy. The laparoscope (A) is positioned in the periumbilical region,
the graspers for gallbladder retraction (B) and manipulation (C) are positioned in the RUQ along the subcostal region, while the subxyphoid region
(D) is used for the dissector, diathermy, and clip appliers.
A. Triangle of
cholecystectomy
B. Triangle of
FIGURE 35.2. The omentum and adhesions are divided with cautery
Calot
or by carefully pulling them away with atraumatic grasping forceps
(shown here). FIGURE 35.3. A, Triangle of cholecystectomy limited by the common
hepatic duct, right hepatic duct, cystic duct, and inferior liver edge.
B, The triangle of Calot limited by the common hepatic duct, cystic duct,
trocar, and the peritoneum is opened along the gallbladder wall and cystic artery.
at the neck. This dissection is continued superiorly along both
sides of the gallbladder 1 to 2 mm from the liver edge. Medial attachments surrounding the neck of the gallbladder are dis-
retraction of the gallbladder is helpful to facilitate this dissec- sected free until only two structures remain: the cystic duct and
tion on the lateral wall. cystic artery. It may be useful to alternatively retract the infun-
Once the peritoneum is opened on the gallbladder wall, dibulum either medially or laterally to expose both sides of the
attention is then turned again to the triangle of Calot to develop gallbladder and dissect the adhesions free (Fig. 35.4). There is
a “window” of dissection and obtain the critical view of no need to continue dissection to the level of the CBD, because
safety (Strasberg et al, 1995). This technique has been widely this may only increase the risk of injury. Blunt dissection using
adopted and demonstrated to reduce the incidence of iatrogenic a Maryland grasper or a laparoscopic right-angle grasper may
bile duct injuries during cholecystectomy (Strasberg & Brunt, be performed to create a window behind the cystic duct by
2010). Using a combination of blunt dissection and cautery, all gently spreading the avascular tissue. The gallbladder is
A B
FIGURE 35.4. Proper gallbladder retraction for exposure of triangle of Calot (A) and reverse triangle of Calot (B). (Courtesy Quality Medical Publish-
ing, St. Louis, MO.)
FIGURE 35.6. Critical view of safety during laparoscopic

cholecystectomy
first, and additional dissection off the cystic plate may be per-
formed to confirm the location and course of the cystic duct.
FIGURE 35.5. Critical view. Hepatocystic triangle is dissected free of Completion of Cholecystectomy
all tissue except for the cystic duct and artery, and the cystic plate on After anatomy is confirmed, the cystic artery and duct are
the liver bed is exposed. When this view is achieved, the two structures
clipped and divided. We prefer to use clips, either metallic or
entering the gallbladder can only be the cystic duct and artery.
plastic. Before each clip is secured, it is important to properly
visualize the posterior jaw of the clip applier to avoid injury to
dissected off of the liver’s edge, exposing the cystic plate. Once the surrounding structures and ensure proper deployment of the
the cystic artery and duct can clearly be identified entering the clip. The clips should not be placed too distally so that the CBD
gallbladder and only the cystic plate is seen behind these struc- is “tented up” into the clip. Occasionally, a large or inflamed
tures, the critical view of safety is achieved, and it is considered duct may hinder the ability to sufficiently secure clips across the
safe to divide the duct and artery (Figs. 35.5 and 35.6). If there entire lumen, and in these cases, an Endoloop preformed suture
is a question regarding anatomy at this time, a cholangiogram can be used on the distal cystic duct. Although used by some
may be performed. Additionally, the cystic artery may be divided surgeons, we strongly discourage the use of a linear stapler for
FIGURE 35.7. Separation of the gallbladder from its bed by dissection

with a blunt-tipped thermal energy probe. The neck of the gallbladder
is placed on traction in a superior direction, then twisted to the left and
right to place tension on the junction between the gallbladder and the
hepatic fossa.
a large inflamed duct, because any duct large enough to require

a stapler for division should raise the suspicion of an impending
CBD injury. Confirmation of anatomy with cholangiography or
an open procedure should be strongly considered in this situa-
tion. Division between clips is performed using hook or straight
scissors. Cautery is avoided for division due to the risk of necro- FIGURE 35.8. Blind placement of clips or clamps for hemostasis can
sis and subsequent clip slippage. The gallbladder is then freed result in injury to the hepatic artery or bile duct.
off the gallbladder fossa from the infundibulum to the fundus
by using electrocautery, maintaining a careful plane between the
cystic plate and the gallbladder wall. Appropriate retraction may placed in a bag before extraction from the patient to avoid stone
need to be adjusted frequently during this step to maintain spillage into the peritoneal cavity. In the case of multiple or
tension on the inferior border of the gallbladder that is attached large stones, a large Kelly clamp can be used to dilate the
to the liver. The fundus should remain retracted in a cephalad umbilical site for specimen extraction. Once a portion of the
direction, whereas the neck of the gallbladder can be adjusted specimen bag is protruding through the skin, this can be
anterosuperiorly and then alternatively medially and laterally to opened, and stones or bile can be removed to facilitate delivery
best facilitate exposure (Fig. 35.7). Intermittent blunt dissection from the abdomen. Any spilled bile or debris should be copi-
facilitates exposure of the proper plane. ously irrigated and then aspirated dry with suction. Any fascial
In the setting of acute inflammation, this step can be diffi- defect greater than or equal to 10 mm should be closed by
cult, and bleeding from the liver may be a sign that dissection using sutures to avoid subsequent incisional hernia. The skin
is being conducted past the cystic plate into the parenchyma of incisions are then closed with glue or subcuticular sutures.
the liver. Blind placement of clips, clamps, or use of cautery at
this time risks worsening hemorrhage or inadvertent injury to Three-Port and Two-Port Techniques
the hepatic blood supply, superficial biliary pedicles, or termi- Additional laparoscopic methods have been created that alter
nal middle hepatic vein branches and should be avoided (Fig. the number and type of trocar sites used for laparoscopic cho-
35.8). Just before the gallbladder is completely free of the liver, lecystectomy, in an attempt to further improve upon the tradi-
it may be useful to inspect the clips and gallbladder fossa for tional four-port technique. Most series are small but indicate
bleeding at this time, while exposure is still maximized. Gentle that these techniques are both feasible and generally safe. Pro-
irrigation and suction can assist with the identification of any ponents claim that reducing the number of trocars is no more
bleeding or bile leakage in the operative field. The clips are technically challenging, is as safe, and can result in fewer scars
carefully inspected a final time to ensure that they are placed and reduced costs (Sun et al, 2009; Tuveri & Tuveri, 2007).
securely and adequately compressing across the entire lumen This technique is typically performed by omitting one or both
without impinging on adjacent structures. The final attach- of the right upper quadrant trocars and adjusting the technique
ments to the liver are then divided. The camera is placed in the used for retraction. This can be achieved by using sutures in
upper subxiphoid trocar, and the gallbladder is grasped by the the gallbladder and/or needlescopic graspers without trocars
neck by using a large grasper placed through the umbilical (Dan et al, 2013). However, a recent Cochrane Review con-
trocar. The gallbladder is typically delivered through the peri- cluded that there is insufficient evidence to determine whether
umbilical port, because enlarging this port causes less pain and there are any true benefits from performing a cholecystectomy
has a better cosmetic effect than enlarging other ports. If the by using less than four ports, and the safety profile is yet to be
gallbladder is punctured during the procedure, it should be established (Gurusamy et al, 2014). Newer instruments are
being introduced for these purposes and may contribute to an not be justified for routine use in simple cases due to significant
improvement in outcomes. Until that time, larger studies are cost and resource burden, performing robotic surgery on more
needed to demonstrate a true effect. straightforward cases, such as cholecystectomy, has been pro-
posed as an ideal way for surgeons to acquire experience, skill,
Single-Incision Laparoscopic Surgery and confidence (Jayaraman et al, 2009). Cholecystectomy is
Single-incision laparoscopic surgery (SILS) cholecystectomy is often the first procedure that a general surgeon attempts with
an extension of the attempt to reduce the number and size of robotic assistance. This is due to many factors, including rela-
laparoscopic ports. SILS is performed through one small trans- tive ease and frequency with which cholecystectomy is per-
abdominal incision, usually at the umbilicus, rather than the formed in the modern era and familiarity with the operation.
standard multiple trocar sites. The theoretical benefits include To date, few studies have compared robotic cholecystectomy
less pain and a better esthetic result. Similar to three-port and with the laparoscopic approach, and there are no RCTs to
two-port cholecystectomy, SILS is often assisted with needle- compare them. A recent Cochrane Review concluded that
scopic graspers and/or sutures placed on the gallbladder for although robotic cholecystectomy is safe, there are no obvious
retraction and exposure. However, no widely accepted standard benefits over laparoscopic cholecystectomy (Gurusamy et al,
technique currently exists. Because it requires the same basic 2012). This review was significantly limited by the small
skill set and instruments of a traditionally laparoscopic chole- number of trials and the high risk of systemic errors and bias
cystectomy, SILS can be readily implemented. Several specifi- within these trials.
cally designed single-port systems and instruments have now Although outcomes are no different, there is a significantly
been approved by the U.S. Food and Drug Administration. higher cost associated with robotic surgery. This is due not only
There have been several RCTs comparing SILS with either to the significant upfront cost of initially acquiring a robotic
standard- or reduced-port cholecystectomy. However, all of console (2 million US dollars), but also the cost of maintenance
these trials have relatively small numbers of patients and a low and replacement of disposable instruments. Economic analyses
overall rate of complications. The largest such trial randomly have demonstrated significantly higher costs in the robotic
assigned 200 patients to SILS or standard four-port laparo- group, both in single-institution studies as well as large national
scopic cholecystectomy. The total rate of adverse events was samples, ranging from $1730 to $8182 more than laparoscopic
not significantly different in either group, but the incidence of cholecystectomy (Breitenstein et al, 2008; Kamiński et al, 2014;
wound complications, including postoperative hernia, was Rosemurgy et al, 2015). With the recent focus on cost savings
higher in the SILS group (Marks et al, 2013). The hernia rate and responsible spending becoming more pervasive in the
in this study was 1.2% for the laparoscopic group, compared medical system, these costs may be prohibitive for all but when
with 8.4% in the SILS group (P = .03). Operative times are robotic surgery is considered absolutely necessary.
also often longer for a SILS cholecystectomy compared with a A recent trend has been to combine robotics with SILS.
standard laparoscopic approach (Marks et al, 2013; Tsimoyi- Combining the two approaches allows the robotic console’s
annis et al, 2010). Two RCTs evaluating postoperative scars improved dexterity and ergonomics to offset the inherent limi-
found improved cosmetic scores in the SILS group (Lai et al, tations in the single-site technique. The robotic platform
2011; Marks et al, 2013). is thought to improve visualization, reduce instrument colli-
Although SILS appears to be feasible and safe, patient selec- sions, and restore intuitive instrument manipulation that limits
tion is a crucial factor. The primary benefit—that of improved standard single-site surgery. Initial studies suggest that single-
cosmetic results—appears to come at the cost of increased incision robotic surgery is not only safe and effective but
wound complications and hernia rate. This trade-off must be also requires less of a learning curve than traditional SILS
thoroughly addressed with the patient during the consent (Spinoglio et al, 2012). Although initially considered in highly
process. Although other benefits are yet to be determined, the selected patients, recent data suggest that liberally applying
true safety profile is also yet to be established. It is most impor- the technique to a wider population can also be safe and effec-
tant to maintain appropriate surgical principles, including visu- tive (Svoboda et al, 2015; Vidovszky et al, 2014).
alization of the critical view of safety, regardless of the approach
used. Although four-port cholecystectomy still remains the gold CONTRAINDICATIONS
standard treatment, SILS may play a role in selected patients.
Contraindications to cholecystectomy can be considered in two
Robotic Cholecystectomy groups: contraindications to cholecystectomy and contraindica-
Robotic surgery has significantly increased in popularity and tions to laparoscopy. Absolute contraindications to operative
availability during the last decade. While technology has cholecystectomy, open or laparoscopic, include refractory coag-
advanced, both the surgical community and public have become ulopathy, inability to tolerate general anesthesia for any reason,
more aware of its existence, and robotic consoles have become and severe sepsis causing hemodynamic instability or end-
prominent at most major medical centers. Much of this has organ failure. In the latter case, considered grade III acute
arguably been driven by a combination of patient demand cholecystitis based on the 2013 Tokyo Guidelines for the diag-
and company marketing. However, the perceived benefits of nosis and management of acute cholecystitis and cholangitis,
improved visualization, ergonomics, and a gentler learning the patient may be better served by emergent gallbladder
curve have made robotic surgery an attractive alternative to decompression via a percutaneous cholecystostomy tube (see
traditional laparoscopy for the general surgeon. Although there Chapter 34) (Mayumi et al, 2007). Once the patient is stabi-
is a paucity of data to demonstrate objective improvements in lized, an interval cholecystectomy can be considered. However,
measurable outcomes when using a robotic approach in general if the patient has diffuse peritonitis with hemodynamic instabil-
surgery, its rapid adoption is becoming more widespread. Some ity and the diagnosis is unclear, an open procedure is then
groups have recommended that although robotic surgery may recommended. Laparotomy offers the ability to confirm the
diagnosis and rapidly correct the disorder. Relative contraindi- experience, and comfort with both laparoscopic and open tech-
cations to surgery can include cirrhosis with portal hyperten- niques, as well as the patient’s anatomy and pathology. A
sion, severe cardiopulmonary disease, and pregnancy, but are surgeon whose primary experience is with open cholecystec-
ultimately at the surgeon’s discretion. tomy may more liberally convert, whereas many younger sur-
Contraindications to laparoscopy continue to decrease while geons today have only a limited experience with the open
surgeons gain expertise and technology advances. Any contra- approach and feel more comfortable using advanced laparo-
indication to open cholecystectomy would also obviously be scopic techniques to troubleshoot problems or accomplish the
considered a contraindication to laparoscopic cholecystectomy. procedure. No one attitude is correct, as long as patient safety
In the case of hemodynamic instability, pneumoperitoneum is the ultimate goal. Although conversion rates are generally
may reduce venous return and exacerbate hypotension; thus greater with less experienced surgeons, a nationwide study
laparoscopy should not be used in these situations. In addition reported a conversion rate to open procedure of 5% to 10%
to the previous, absolute contraindications unique to laparo- (Livingston & Rege, 2004).
scopic cholecystectomy are dictated primarily by the surgeon’s Primary reasons to consider a conversion include unclear
philosophy and experience and include suspicion for gallblad- anatomy, intraoperative complications, failure to progress in a
der cancer, previous abdominal operations with extensive adhe- timely manner, and pathology, including choledocholithiasis
sions, morbid obesity, and pregnancy. not amenable to laparoscopic or postoperative endoscopic tech-
Previous lower abdominal or pelvic operations rarely inter- niques (see Chapters 36 and 37). Although conversion has been
fere with cholecystectomy. Ideally, an unscarred area around considered a “complication” by some, it should be regarded as
the umbilicus is used to gain entry, but if this is not possible, mature judgement, accepting a potentially longer hospital stay
a Veress needle to the left upper or right upper quadrant can to avoid a disastrous complication. Often the complications
be a safe and effective method to insufflate the abdomen. A requiring conversion to open laparotomy are obvious, including
5 mm camera can then be inserted at this site to better evaluate massive hemorrhage, bowel perforation, or major injury to the
any adhesions that may be present along the midline. If neces- bile duct. An open procedure offers the advantage of palpation
sary, right upper quadrant or subxiphoid ports may be inserted and tactile sensation to the surgeon. This may be beneficial
next to free adhesions from the midline before a trocar and when the anatomy is unclear from inflammation, adhesions, or
camera are inserted in that position. anomalies. Although rare, the discovery of a fistula between the
biliary system and bowel may require laparotomy for definitive
Cirrhosis management.
Cirrhosis is considered by many to be a contraindication to at
least laparoscopic cholecystectomy, and possibly to cholecys- OPEN TECHNIQUE
tectomy in general (see Chapter 77). This is due to concerns
over the ability to retract a heavy, friable liver, and the risk of Incision
coagulopathy. A recent population-based study demonstrated The traditional incision used during an open cholecystectomy
that the laparoscopic technique is well tolerated by patients is a right subcostal or Kocher incision. This incision is made 2
with cirrhosis and associated with significantly lower rates of cm below the costal margin and generally extends from the
infection, bleeding, transfusion, liver failure, and overall mor- midline to the edge of the rectus muscle laterally. The length
tality compared with an open procedure, making it the proce- of the incision should be tailored to the exposure required and
dure of choice for cholecystectomy (Chmielecki et al, 2012). A body habitus of the patient. The incision is carried down to the
meta-analysis also demonstrated that the laparoscopic tech- anterior fascia. The rectus muscle is divided with cautery,
nique is preferred in Child’s A and B cirrhotics, with a shorter ensuring appropriate hemostasis by any perforating vessels in
hospital stay and fewer overall postoperative complications the muscle. The posterior rectus sheath is incised, and the
compared with an open technique (de Goede et al, 2013). To abdomen is entered
successfully perform a cholecystectomy in a patient with cir-
rhosis, any coagulopathy should be reversed before surgery. Exposure and Placement of Retractors
Careful attention should be made to observe for any aberrant Exposure during open cholecystectomy is paramount to per-
portosystemic venous collateralization, either in the liver bed, forming the procedure adequately and safely. The round
porta hepatis, or even in the abdominal wall. Liberal use of ligament is divided between ties or, occasionally, with an
hemostatic agents can be used, and energy sources, including energy-sealing device near the anterior abdominal wall. The
the argon beam cautery should be readily available if needed. falciform ligament is incised with cautery for several centime-
If bleeding is unusual or cannot be controlled laparoscopically, ters superiorly to further facilitate exposure. A circular plastic
prompt conversion to an open procedure should be performed. wound edge protector can be inserted and has been demon-
Once the procedure is complete, attention should be placed to strated in RCTs to prevent wound infections (Mihaljevic et al,
carefully ensure that all incision closures are made watertight 2014). A major retractor such as a Thompson- or Bookwalter-
to minimize postoperative ascitic leak (Poggio et al, 2000). type retractor is then attached to the operating room table and
properly set up.
CONVERSION TO OPEN Two abdominal wall retractors are placed on the superior
edge of the incision, one medially and one laterally, and secured
Although laparoscopy is the preferred approach in the majority to the retractor system. Moist sponges are used to pack the
of patients, there are many situations the surgeon may encoun- bowel out of the operative field inferiorly, and additional packs
ter that necessitate conversion to an open procedure. The placed behind the liver often help elevated the gallbladder into
choice and timing of intraoperative conversion to open depends the field for easier dissection. A large malleable retractor may
on a variety of factors and are often based on the surgeon’s skill, be placed on the sponge to maintain this exposure, but it is not
always needed in a small patient. An additional retractor is then Once all the avascular attachments are dissected in this area
placed on the undersurface of the liver to the left of the gallblad- and only two structures are seen going into the gallbladder with
der to retract the liver edge out of the way and bring the porta the cystic plate behind, the critical view is confirmed. The cystic
hepatis into view. duct is then palpated to evaluate for stones and “milked” back
into the gallbladder (Fig. 35.10). The cystic artery is then
Dissection ligated and transected between sutures or clips close to the
Retrograde Cholecystectomy gallbladder wall (Fig. 35.11).
When performing a retrograde cholecystectomy, the critical If a cholangiogram is necessary, it is performed at this time
view of safety is first identified before removal of the gallbladder by placing a clip at the junction of the gallbladder and cystic
from the liver. This technique is the usual technique used duct and then opening the cystic duct via a small transverse
during laparoscopic cholecystectomy but may be more chal- incision 2 to 3 mm distally for insertion of a cholangiocatheter
lenging for a novice surgeon in the open setting without the (see Chapters 23 and 36). The cystic duct is then ligated and
magnification of laparoscopy and when exposure is difficult transected between sutures or clips. The gallbladder is then
secondary to inflammation or obesity. However, it is important dissected from its fossa by using cautery. Gentle traction is
to maintain the same principles to avoid inadvertent biliary placed to facilitate dissection. This dissection plane is kept close
injury. to the gallbladder wall while the cystic plate is left behind to
The fundus of the gallbladder is identified and grasped with avoid deeper dissection into the liver parenchyma and resultant
a Kelly or similar-type clamp (Fig. 35.9A). The peritoneum injury. In the case of acute cholecystitis, this dissection may be
overlying the infundibulum is incised. The incision is carried best done by using sharp or suction dissection. After the gall-
along anteriorly and posteriorly to exposure the cystic triangle. bladder is removed, the gallbladder fossa is inspected to ensure
Care should be made to keep this dissection close to the gall- hemostasis. Any small bleeding points are controlled with
bladder. Once the cystic duct is identified, a suture ligature can cautery. If available, the argon beam coagulator may be valu-
be passed around it and left untied to place tension on the duct able here. If there is more brisk bleeding from a lacerated area
for better exposure and to prevent any stone migration into the of the liver bed, a gauze pack should be held in place with
CBD. The cystic artery is identified as it usually lies just above constant pressure for 5 to 10 minutes. Hemostatic agents may
the cystic duct (Fig. 35.9B). The artery is dissected toward the also be used at this time. This is often sufficient to stop most
gallbladder, visualizing its termination at the gallbladder wall venous bleeding. If hemorrhage is not controlled after simple
to avoid ligating an aberrant or anterior right hepatic artery. pressure, deep hemostatic sutures can be placed with care taken
The gallbladder should be freed along the cystic plate for not to injury the right portal pedicle, which may lie in close
greater exposure and to visualize the critical view of safety. proximity at the base of the gallbladder fossa.
A B
FIGURE 35.9. A, The gallbladder is grasped with a clamp, and dissection of the cholecystectomy triangle is started. B, The cystic artery is in its
normal position, above the cystic duct.
FIGURE 35.12. While the anterograde dissection progresses toward

the gallbladder neck medially, the cystic artery is identified, ligated, and
divided.
fossa. The anterior and posterior fascial layers are then closed
separately, followed by staples or sutures on the skin.
FIGURE 35.10. Palpation reveals a stone in the cystic duct, which
may be “milked” back into the gallbladder.
Anterograde, or Fundus-Down, Cholecystectomy
In some cases, an anterograde approach may be necessary. This
technique should not be used in the presence of severe inflam-
mation obstructing visualization of the cystic duct, however.
There is still significant risk of biliary injury in these cases unless
proper visualization of the cystic duct and artery are obtained.
First, a large Kelly clamp is used to grasp the fundus of
the gallbladder. Another clamp, such as a right-angle clamp is
then used to grasp the peritoneum on the liver edge. An incision
in the gallbladder serosa is made approximately 0.5 cm away
from the liver edge by using cautery. A plane is developed
between the gallbladder wall and the cystic plate; the surgeon
should ensure not to dissect too deeply into liver parenchyma,
because this may risk injury. This plane is created medially and
laterally toward the fundus while the gallbladder is dissected
away from the fossa. Occasionally, an energy-sealing device may
be used if the peritoneal covering is edematous and excessively
vascular. Dissection is performed in a posterolateral manner to
free the gallbladder and best expose the cystic duct and artery.
The cystic artery is encountered while the dissection is contin-
ued medially toward the cystic triangle. The cystic artery is then
ligated and transected as it enters the gallbladder wall (Fig.
35.12). The infundibulum is next dissected free while it courses
down toward the gallbladder neck. Careful palpation should
identify any stones that may have migrated into the cystic or
CBDs. The cystic duct is dissected free for a length of no more
than 1 cm to avoid injury to the common hepatic or CBD. At
this point, the only attachment to the gallbladder should be the
cystic duct. Once the anatomy is confirmed, the cystic duct is
FIGURE 35.11. The cystic artery is ligated close to the gallbladder
ligated and transected between sutures or clips, and the speci-
wall.
men is removed from the field. The abdomen is closed as
described earlier.
The gallbladder can be opened and inspected on the back
table for the presence of stones, tumors, and to evaluate the PARTIAL OR SUBTOTAL CHOLECYSTECTOMY
cystic duct. A drain is rarely necessary after an uncomplicated
cholecystectomy, but if there is concern for bile leakage, then it A severely inflamed gallbladder may be encountered at the time
may be left postoperatively, positioned close to the gallbladder of operation that limits the ability to safely achieve the critical
view of safety (see Chapter 33). Fibrosis or inflammation may the case or after the gallbladder has been inadvertently injured
cause the surgeon to mistake the CBD for the cystic duct (Fig. or entered. In these cases, a partial or subtotal cholecystectomy,
35.13). In these cases, the surgeon is urged not to proceed in which a portion of the gallbladder is removed and gallstones
boldly but instead to err on the side of caution. Cholecystitis extracted while the posterior wall is left behind, is the procedure
can often be managed safely until inflammation has had a of choice for source control and surgical management.
chance to subside, but bile duct injuries have the potential to This approach can be used during either open or laparo-
cause significant long-term harm or even death. If the difficulty scopic surgery, and the technique is similar. The gallbladder is
of the operation and risk for injury is identified early and the drained and opened with hook diathermy at the fundus or
gallbladder has not been perforated, a cholecystostomy tube occasionally at the infundibulum. Bile, stones, and debris are
may be placed, the patient closed, and surgery may be delayed suctioned out and removed. During laparoscopy, this may be
until a period of medical management has occurred. Unfortu- facilitated by placing an Endobag to collect the contents of
nately, this is often difficult to recognize until further along in the gallbladder. The incision may need to be extended to the
gallbladder neck without dissecting the cystic duct or artery. If
this is performed via laparotomy, a finger may be inserted
into the gallbladder and be useful to guide dissection (Fig
35.14A). The anterior wall is then excised with cautery
with a small strip of the posterior wall left attached to the liver
(Fig. 35.15). The remnant mucosa can then be either removed
FIGURE 35.15. Laparoscopic subtotal cholecystectomy, demonstrat-

FIGURE 35.13. The common hepatic duct can be mistaken for the ing the anterior wall excised and a small strip of the posterior wall left
cystic duct when the region of the infundibulum cannot be delineated attached to the liver. The remnant mucosa can then be either removed
because of fibrosis and inflammation. or coagulated with cautery or argon laser.
A B
FIGURE 35.14. A, The fundus of the gallbladder has been opened, and a finger is introduced into the gallbladder for palpation. B, Partial chole-
cystectomy. The superficial part of the fundus and body of the gallbladder has been excised, leaving in place its attachment to the liver. The remaining
mucosa is removed by curettage and/or fulgurated with electrocoagulation, and a closed-suction drain is placed near the infundibulum.
or coagulated with cautery or argon beam coagulator. Intraop-

erative cholangiography can be performed; however, if acute
cholecystitis has obliterated the cystic duct, this may not be
technically feasible. If bile is found oozing from the gallblad-
der stump, it may be closed with surgical clips, suture liga-
tion, or an Endoloop. A drain should be placed near the
gallbladder stump to drain any potential future bile collection
(see Fig. 35.14B).
A recent systematic review and meta-analysis of 1231 sub-
total cholecystectomies demonstrated a higher rate of post
operative bile leakage but overall comparable morbidity rates,
including bile duct injury when compared with complete cho-
lecystectomy (Elshaer et al, 2015). Additionally, when compar-
ing laparoscopic with open approach of partial cholecystectomy,
this review found less fluid collections, retained stones, wound
infections, and overall mortality when the laparoscopic approach
was used, but a higher rate of bile leaks.
INTRAOPERATIVE PROBLEMS
Anatomic Variations
An intimate knowledge of gallbladder and biliary anatomy and
the common variations is essential to safely performing chole-
cystectomy, laparoscopic or open. Normal anatomy of the gall-
bladder and biliary system is discussed in detail in Chapters 1
and 2. Hepatobiliary anatomy can present with anomalies
and variations. Duplicated, left-sided, bilobed, or congenitally
absent gallbladder is considerably rare, predominantly identi-
fied preoperatively, and not associated with postoperative injury
or complications. Our focus instead will be on relevant biliary
anatomy and variations that can potentially lead to inadvertent FIGURE 35.16. Variations in the confluence of the cystic duct and
injury. common hepatic duct.
Significant variation of the junction between the cystic duct
and common hepatic duct can exist (Fig. 35.16). The common
hepatic duct can range in length from 1 to 7.5 cm. In 22% of
patients, the cystic duct is seen to run parallel to the common
hepatic duct for an average of 17 mm before forming the CBD
(Newman & Northrup, 1963). The cystic duct can be signifi-
cantly shortened or absent secondary to inflammation and scar-
ring from repeated bouts of cholecystitis (Dorrence et al, 1999).
This is usually not predicted or identified until at the time of
the operation.
Large series of cholangiographic studies have demonstrated
anatomic variations and congenital abnormalities of the biliary
tree in nearly half of patients (Puente et al, 1983). The cystic
duct may drain directly into the right hepatic duct or even a
low right sectoral duct (Fig. 35.17). Anomalous extrahepatic
bile ducts can occur in as many as 12% of patients, with the
most common being an anomalous right sectoral duct that
empties into either the common hepatic duct or cystic duct
(Berci, 1981). Fig. 35.18 demonstrates the common anatomic
biliary variations of the right sectoral ducts. When these varia-
tions do occur, they are often replaced ducts and represent the
only biliary drainage from the corresponding segment(s) of the
liver. Therefore if these ducts are occluded, biliary obstruction
will ensue with resultant liver atrophy and/or obstructive chol-
angitis. Ligation of the cystic duct may need to occur distal to
an aberrant duct insertion site, to preserve proper biliary drain-
age (Fig. 35.19). Although it may be suspected when encoun-
tering an anatomic anomaly during surgery, true duplication of
the cystic duct is exceedingly rare. It is essential to be aware FIGURE 35.17. High insertion of the cystic duct demonstrated on
of the common anatomic variations and have a high index of endoscopic retrograde cholangiopancreatography.
complication and subsequent difficulty of repair (Strasberg &

Helton, 2011) (see Chapter 42).
Gallbladder Perforation
Gallbladder perforation is likely the most common intraopera-
tive problem or complication encountered by most surgeons,
especially during the learning curve of cholecystectomy.
Although the leakage of bile or stones can be troublesome, it
rarely requires conversion to an open procedure. Perforation
primarily occurs either secondary to traction of forceps or from
electrocautery thermal injury. Although it is difficult to accu-
rately measure the rates worldwide, perforation may occur in
as many as one third of patients during cholecystectomy (Jones
et al, 1995). When it does occur, the loss of gallstones and their
retention in the abdominal cavity should be noted in the
description of the surgical procedure (Gerlinzani et al, 2000).
It is unclear what consequences, if any, are secondary to
FIGURE 35.18. Variations in the confluence of the extrahepatic bile gallbladder perforation during surgery. For most patients, per-
ducts and cystic duct. foration does not result in any immediate complications or
clinical difference (Jones et al, 1995; Memon et al, 1999). If
gallstones are spilled into the abdomen and not retrieved, prob-
lems may occur. Pigmented stones often harbor bacteria and
may cause postoperative infection. There are several case
reports and series of such complications, including intraab-
dominal (Horton & Florence, 1998) and abdominal wall
(Eisenstat, 1993) abscess formation secondary to gallstone
spillage during cholecystectomy. A recent prospective analysis
of clinical outcomes after accidental gallbladder spillage sug-
gests that it may cause more postoperative pain, ileus, and
infection when compared with uncomplicated cases (Suh et al,
2012).
Biliary Injury
Although fortunately quite rare, biliary injury remains one of
the most dreaded complications of cholecystectomy (see Chap-
ters 38 and 42). Not only can bile duct injury be fatal in the
short or long term, but it can cause disturbances in quality of
life and need for prolonged medical care (Cates et al, 1993;
Landman et al, 2013). Additionally, it commonly results in
litigation (Asbun et al, 1993; Perera et al, 2010). Although they
are infrequent, and designing adequate studies to evaluate
proper management and treatment can be difficult, much has
been learned about the etiology of iatrogenic bile duct injuries.
FIGURE 35.19. An aberrant duct draining directly into the cystic duct Interestingly, the majority of bile duct injuries are caused by
as demonstrated on endoscopic retrograde cholangiopancreatography. attending surgeons and are not recognized at the time of opera-
tion (Bingham et al, 2000). Injuries are more often due to
suspicion during surgery. Intraoperative cholangiogram may be misperceptions of normal anatomy than lack of knowledge,
useful in identifying or confirming abnormal biliary anatomy. skill, or judgment (Way et al, 2003). In experienced hands, the
Arterial anatomic variations are also common and can con- rate of iatrogenic biliary is likely 0.3% or less in both open and
tribute to or be the direct cause of significant morbidity if laparoscopic cholecystectomy.
injured. The cystic artery originates from the right hepatic Routine use of cholangiography was initially proposed to
artery most commonly (76%) but may also come from the left, reduce injury after the widespread and rapid adoption of lapa-
common, or proper hepatic arteries (Chen et al, 2000). A roscopic cholecystectomy was found to result in a higher inci-
double cystic artery can be found in 12% to 25% of patients. dence of biliary injuries (Fletcher et al, 1999; Flum et al,
Additionally, when the cystic artery originates from an aberrant 2003). However, risk of injury to the CBD or a segmental duct
source, it frequently traverses outside the triangle of Calot while performing cholangiography can occur if proper anatomic
(Ding et al, 2007). Although not technically an anatomic varia- identification is not accomplished. Additionally, the ability to
tion, the right hepatic artery may frequently loop onto the properly interpret the images intraoperatively can remain a
infundibulum of the gallbladder and become misidentified as challenge and lead to bile duct injury even with a reported
the cystic artery. Arterial injury may accompany bile duct injury “normal” cholangiogram (Chapman et al, 2003).
(vasculobiliary injury), with the right hepatic artery involved in The critical view of safety, as described earlier and developed
92% of cases, and can significantly add to the morbidity of the by Strasberg, remains the current standard technique to reduce
the incidence of bile duct injury during cholecystectomy. Often, a biliary injury may not be identified until postopera-
Although no level I data exist to prove that this technique defi- tively. In these cases, the surgeon must have a high index of
nitely lowers the rate of bile duct injury, a large body of evi- suspicion and make every effort to avoid a delay in diagnosis.
dence does support it (Strasberg & Brunt, 2010). Due to the Symptoms of worsening abdominal pain, fevers, and abdominal
rare frequency of iatrogenic biliary injury, a RCT proving the distention that persist for several days after surgery warrant
efficacy of this technique may not be feasible. further investigation. Initial diagnostic tests include ultrasonog-
Common circumstances that are thought to risk iatrogenic raphy and computed tomography of the abdomen. If fluid is
injury include misidentification of anatomy, a short cystic duct, seen in the gallbladder fossa, right paracolic gutter, hepatorenal
significant inflammation, previous scar or adhesions, obesity, recess, or subdiaphragmatic space, it should be percutaneously
or bleeding obscuring the field of view (Adams et al, 1993; drained and sampled. Biliary injuries have been well described
Asbun et al, 1993; Soper et al, 1993). The “classic injury” and classified by the Strasberg scheme (Fig. 35.21) (Strasberg
occurs when the CBD is mistaken for the cystic duct and et al, 1995). Evaluation of the extent of injury can be deter-
clipped, thought to occur if the cystic duct is pulled too vigor- mined with magnetic resonance cholangiopancreatography or
ously laterally, causing the cystic duct and CBD to stretch into cholangiogram via either endoscopic retrograde cholangiopan-
alignment and appear as one. If this mistake is not immediately creatography (ERCP) or percutaneous transhepatic cholangi-
identified, dissection often continues along the CBD postero- ography. A majority of simple postoperative bile leaks can be
superiorly until an “aberrant” or “duplicated” duct is encoun- managed with drainage alone or with drainage and ERCP
tered also entering toward the gallbladder. If this is then ligated, placement of a biliary stent (Kozarek et al, 1994), but more
a more proximal bile injury will then occur (Fig 35.20). complex injuries require multidisciplinary management and
If a bile duct injury is identified at the time of surgery, the may require operative repair. Repair of biliary injuries and
surgeon must make an assessment of not only the extent of biliary-enteric anastomoses are discussed in more detail else-
injury and their own ability to potentially repair it, but the where in this text (see Chapter 31).
patient’s current condition, resources of the hospital, and
the availability of specialized hepatobiliary expertise. Often, the Bleeding
most prudent action is to terminate the operation, leave a drain, The incidence of uncontrollable bleeding from laparoscopic
and close the patient with immediate referral to a specialized cholecystectomy is 0.1% to 1.9% and can occur from three
hepatobiliary center. Early referral to a specialist has been dem- distinct sites: (1) the liver, (2) arterial sources, and (3) port-
onstrated to result in better long-term outcomes (Perera et al, insertion sites. Significant bleeding from the liver bed is fairly
2011). As noted previously, if an attempt is made to forge ahead common and is now appreciated to be from the close proximity
and remove the gallbladder, there is significant risk in worsen- of the middle hepatic vein and its terminal branches, which run
ing the bile duct injury and/or creating a vascular injury (Stras- close to the gallbladder fossa in as many as 10% to 15% of
berg & Helton, 2011). patients. Major bleeding usually occurs during the final aspects
of the removal of the gallbladder from the hepatic fossa and
generally requires immediate conversion to open surgery to
control profuse hemorrhage through stitch ligation if initial
attempts at laparoscopic hemostatic control fail. Bleeding from
the cystic artery can be controlled with clips, but only if ana-
tomic landmarks can be safely confirmed. Caution must be
exercised in this setting because of a high association with right
hepatic arterial injury with injudicious placement of clips or
sutures. Significant bleeding may also first be evident postop-
3 eratively as an acute hemodynamic decline requiring resuscita-
tion, transfusions, and often reoperation. The culprit in this
2 scenario is usually a dislodged clip.
When a trocar injury to a major blood vessel is suspected,
the patient must be opened immediately without removing the
1 trocar, and the involved blood vessel must be isolated and
controlled. In contrast, if the small-bore Veress needle enters a
viscus or blood vessel, the operation generally can be com-
pleted, and the patient is monitored closely for signs of com-
plications in the postoperative period. Additionally, bleeding
may occur if a trocar lacerates a blood vessel within the abdomi-
FIGURE 35.20. Pathogenesis of the “classic” injury. 1. The common nal wall, such as the epigastric artery or one of its branches.
bile duct is mistaken as the cystic duct and is clipped and divided. 2. Before removal, each trocar should be visualized from the peri-
The dissection is carried up along the left side of the common hepatic toneal aspect by using the laparoscope. If significant hemor-
duct in the belief that this is the underside of the gallbladder. 3. The rhage is seen, it generally can be controlled with cautery or a
common hepatic duct is transected while the surgeon tries to dissect
through-and-through suture on each side of the trocar insertion
what he or she believes is the gallbladder from the liver bed. If the struc-
ture is recognized as a bile duct at this point, it is often thought to be a
site.
second cystic duct or an accessory duct. While the common hepatic
duct is divided, the right hepatic artery is often injured. (From Strasberg
Bowel Injury
SM, Helton WS: An analytical review of vasculobiliary injury in laparo- Bowel injury is a rare complication but can be potentially fatal
scopic and open cholecystectomy. HPB (Oxford) 13:1-14, 2011.) if unrecognized. It is most likely to occur at two common
A B C
>2 cm <2 cm E3
E1 E2
E4
E5
FIGURE 35.21. Strasberg classification of biliary injuries. Type A injuries are leaks from the cystic duct or small ducts in the liver bed. Type B and
C injuries almost always involve aberrant right hepatic ducts. Type D injuries are lateral injuries to the major bile ducts. Type E injuries are subdivided
according to the Bismuth classification system. (From Strasberg SM, et al: An analysis of the problem of biliary injury during laparosopic cholecys-
tectomy. J Am Coll Surg 180:101-125, 1995.)
points: during laparoscopic entry or during dissection of the indolent course, presenting with an enterocutaneous fistula,
cystic structures near the porta hepatis. Bowel injury has been standard drainage and nutritional support is appropriate.
described in approximately one to four patients in 1000 lapa-
roscopic procedures and is slightly more common with an open POSTOPERATIVE MANAGEMENT
cut-down technique than with a Veress needle, but it can occur
with both. Immediately after insufflation occurs and the The majority of patients who undergo elective laparoscopic
abdomen is entered, a brief but thorough inspection of the cholecystectomy are discharged the same day, both for patient
abdomen should be performed to ensure that no injury occurred preference and to reduce hospital costs. A patient admitted
during initial entry. All additional trocars should be placed from the clinic or emergency department with acute cholecys-
under direct laparoscopic visualization to avoid injury or per- titis is often observed in the hospital overnight postoperatively.
foration to nearby bowel. The trocar should be grasped in such Other patients who may benefit from an overnight stay include
a manner that a hand can be used as a “brake” to prevent elderly patients, those with significant comorbid illnesses, those
inadvertently introducing the trocar too far. Although some requiring substantial analgesia postoperatively, and patients
injuries may be reparable laparoscopically, conversion to open with complications. The patient is allowed a diet shortly after
should not be delayed if necessary for an adequate repair. surgery, usually beginning with clear liquids then advanced as
Another feared postoperative complication of cholecystec- tolerated to a regular diet without any specific restrictions.
tomy is that of a missed enterotomy. Severe abdominal pain Incisional and referred should pain is common. The patient is
and distention, fevers, diarrhea, persistent radiographic free air, often prescribed oral narcotics for postoperative pain control,
and, potentially, even sepsis and subsequent hemodynamic but many patients are able to recover with only over-the-coun-
instability should alert the surgeon to this serious complication. ter pain medications. It is unnecessary to prescribe a course of
These symptoms usually occur within the first 3 to 4 days antibiotics postoperatively except in unusual circumstances.
postoperatively. An immediate laparotomy is usually indicated Minimal activity restrictions are placed on the patient postop-
to diagnose and treat the condition. If the patient has a more eratively if they undergo laparoscopic surgery. Due to a risk of
hernia formation, heavy lifting is discouraged for 4 to 6 weeks operation in an ambulatory setting. Attempts at improving out-
after an open procedure. The patient may return to work once comes from laparoscopic cholecystectomy will continue to
their pain is well controlled, usually by postoperative day 2 or drive the development of novel surgical tools and techniques.
3. Routine follow-up by the surgeon in clinic can occur between Due to the relative safety and low number of complications,
1 and 4 weeks after surgery. further improvement may be difficult and much of the recent
technology has been focused on cosmetic or other ancillary
CONCLUSION improvements. These newer approaches to cholecystectomy,
including single-incision laparoscopy and robotics must be held
Cholecystectomy remains the gold standard for the manage- to the current standard of laparoscopic cholecystectomy in the
ment of gallbladder disease in the developed world, with face of growing patient and industry pressure.
the vast majority now performed minimally invasively with
excellent results. Many patients are now able to undergo this References are available at expertconsult.com.
A. Gallstones and Gallbladder Chapter 35 Technique of cholecystectomy: open and minimally invasive 584.e1
REFERENCES Fullarton G, et al: Evaluation of the cost of laparoscopic and open

cholecystectomy, Br J Surg 81:124–126, 1994.
Adams DB, et al: Bile duct complications after laparoscopic cholecys- Gerlinzani S, et al: Is the loss of gallstones during laparoscopic chole-
tectomy, Surg Endosc 7:79–83, 1993. cystectomy an underestimated complication?, Surg Endosc 14:373–
Agabiti N, et al: Thirty-day complications after laparoscopic or open 374, 2000.
cholecystectomy: a population-based cohort study in Italy, BMJ Girardet R, et al: Significance of asymptomatic biliary tract disease
Open. 3(2):pii:e001943, 2013. in heart transplantation recipients, J Heart Transplant 8:391–399,
Ahmad G, et al: Laparoscopic entry techniques, Cochrane Database Syst 1989.
Rev (2):CD006583, 2012. Gurusamy KS, et al: Robot assistant versus human or another robot
Asbun HJ, et al: Bile duct injury during laparoscopic cholecystectomy: assistant in patients undergoing laparoscopic cholecystectomy,
mechanism of injury, prevention, and management, World J Surg Cochrane Database Syst Rev (9):CD006578, 2012.
17:547–552, 1993. Gurusamy KS, et al: Early versus delayed laparoscopic cholecystec-
Barkun JS, et al: Randomized controlled trial of laparoscopic versus tomy for people with acute cholecystitis, Cochrane Database Syst Rev
mini cholecystectomy. The McGill Gallstone Treatment Group, (6):CD005440, 2013.
Lancet 340(8828):1116–1119, 1992. Gurusamy KS, et al: Fewer-than-four ports versus four ports for lapa-
Bass EB, et al: Cost-effectiveness of laparoscopic cholecystectomy roscopic cholecystectomy, Cochrane Database Syst Rev (2):CD007109,
versus open cholecystectomy, Am J Surg 165(4):466–471, 1993. 2014.
Berci G, et al: Operative Biliary Radiology, Baltimore, 1981, Williams Horton M, Florence MG: Unusual patterns following dropped gall-
& Wilkins, pp 110–116. stones during laparoscopic cholecystectomy, Am J Surg 175:375–
Bingham J, et al: Biliary complications associated with laparoscopic 379, 1998.
cholecystectomy—an analysis of common misconceptions, Ulster Hull D, et al: Management of cholelithiasis in heart and lung transplant
Med J 69(2):106–111, 2000. patients: with review of laparoscopic cholecystectomy, Conn Med
Boline G, et al: Cholecystectomy in the potential heart transplant 58:643–647, 1994.
patient, J Heart Lung Transplant 10(2):269–274, 1991. Jayaraman S, et al: Getting started with robotics in general surgery with
Breitenstein S, et al: Robotic-assisted versus laparoscopic cholecystec- cholecystectomy: the Canadian experience, Can J Surg 52(5):374–
tomy: outcome and cost analyses of a case-matched control study, 378, 2009.
Ann Surg 247(6):987–993, 2008. Jones D, et al: The influence of intraoperative gallbladder perforation
Cates J, et al: Biliary complications of laparoscopic cholecystectomy, on long-term outcome after laparoscopic cholecystectomy, Surg
Am Surg 59:243–247, 1993. Endosc 9:977–980, 1995.
Chapman WC, et al: Bile duct injuries 12 years after the introduction Kamiński JP, et al: Robotic versus laparoscopic cholecystectomy inpa-
of laparoscopic cholecystectomy, J Gastrointest Surg 7(3):412–416, tient analysis: does the end justify the means?, J Gastrointest Surg
2003. 18(12):2116–2122, 2014.
Chen TH, et al: Variations of the cystic artery in Chinese adults, Surg Kane R, et al: The outcomes of elective laparoscopic and open chole-
Laparosc Endosc Percutaneous Tech 10:154–157, 2000. cystectomy, J Am Coll Surg 180:136–145, 1995.
Chmielecki DK, et al: Laparoscopic cholecystectomy is the preferred Kozarek RA, et al: Endoscopic treatment of biliary injury in the era of
approach in cirrhosis: a nationwide, population-based study, HPB laparoscopic cholecystectomy, Gastrointest Endosc 40(1):101–106,
(Oxford) 14(12):848–853, 2012. 1994.
Croce E, et al: Laparocholecystectomy: 6,865 cases from Italian insti- Lai ECH, et al: Prospective randomized comparative study of single
tutions, Surg Endosc 8:1088–1089, 1994. incision laparoscopic cholecystectomy versus conventional four-port
Csikesz NG, et al: Surgeon volume metrics in laparoscopic cholecys- laparoscopic cholecystectomy, Am J Surg 202(3):254–258, 2011.
tectomy, Dig Dis Sci 55:2398–2405, 2009. Landman MP, et al: The long-term effect of bile duct injuries on
Cuschieri A, et al: The European experience with laparoscopic chole- health-related quality of life: a meta-analysis, HPB (Oxford)
cystectomy, Am J Surg 161:383–388, 1991. 15(4):252–259, 2013.
Dan AG, et al: Two-trocar cholecystectomy by strategic laparoscopy Legorreta A, et al: Increased cholecystectomy rate after introduction of
for improved cosmesis (SLIC), JSLS 17(4):578–584, 2013. laparoscopic cholecystectomy, JAMA 270:1429–1432, 1993.
de Goede B, et al: Meta-analysis of laparoscopic versus open cholecys- Livingston EH, Rege RV: A nationwide study of conversion from lapa-
tectomy for patients with liver cirrhosis and symptomatic cholelithia- roscopic to open cholecystectomy, Am J Surg 188:205–211, 2004.
sis, Br J Surg 100(2):209–216, 2013. Lo CM, et al: Prospective randomized study of early vs. delayed lapa-
Ding YM, et al: New classification of the anatomic variations of cystic roscopic cholecystectomy for acute cholecystitis, Ann Surg 227:461–
artery during laparoscopic cholecystectomy, World J Gastroenterol 466, 1998.
13(42):5629–5634, 2007. Majeed AW, et al: Randomised, prospective, single-blind comparison
Dorrence HR, et al: Acquired abnormalities of the biliary tract from of laparoscopic versus small-incision cholecystectomy, Lancet 347:
chronic gallstone disease, J Am Coll Surg 189(3):269–273, 1999. 989–994, 1996.
Dunn D, et al: Laparoscopic cholecystectomy in England and Wales: Marks JM, et al: Single-incision laparoscopic cholecystectomy is associ-
results of an audit by the Royal College of Surgeons of England, Ann ated with improved cosmesis scoring at the cost of significantly
R Coll Surg Engl 76(4):269–275, 1994. higher hernia rates: 1-year results of a prospective randomized, mul-
Eisenstat S: Abdominal wall abscess due to spilled gallstones, Surg ticenter, single-blinded trial of traditional multiport laparoscopic
Laparosc Endosc 3:485–486, 1993. cholecystectomy vs single-incision laparoscopic cholecystectomy,
Elshaer M, et al: Subtotal cholecystectomy for “difficult gallbladders”: J Am Coll Surg 216(6):1037–1047, 2013.
systematic review and meta-analysis, JAMA Surg 150(2):159–168, Mayumi T, et al: Results of the Tokyo Consensus Meeting Tokyo
2015. Guidelines, J Hepatobiliary Pancreat Surg 14(1):114–121, 2007.
Escarce J, et al: Falling cholecystectomy thresholds since the intro McGinn FP, et al: Randomized trial of laparoscopic cholecystectomy
duction of laparoscopic cholecystectomy, JAMA 273:1581–1585, and mini-cholecystectomy, Br J Surg 82:1374–1377, 1995.
1995. McMahon AJ, et al: Laparoscopic versus minilaparotomy cholecystec-
Fendrick A, et al: Asymptomatic gallstones revisited: is there a role for tomy: a randomised trial, Lancet 343(8890):135–138, 1994.
laparoscopic cholecystectomy?, Arch Fam Med 2:959–968, 1993. Memon MA, et al: The outcome of unretrieved gallstones in the peri-
Fenster L, et al: What symptoms does cholecystectomy cure? Insights toneal cavity during laparoscopic cholecystectomy. A prospective
from an outcome measurement project and review of literature, Am analysis, Surg Endosc 13(9):848–857, 1999.
J Surg 169:533–538, 1995. Mihaljevic AL, et al: Multicenter double-blinded randomized con-
Fletcher DR, et al: Complications of cholecystectomy: risks of the trolled trial of standard abdominal wound edge protection with sur-
laparoscopic approach and protective effects of operative cholan gical dressings versus coverage with a sterile circular polyethylene
giography: a population based study, Ann Surg 229(4):449–457, drape for prevention of surgical site infections: a CHIR-Net trial,
1999. Ann Surg 260(5):730–737, 2014.
Flum DR, et al: Intraoperative cholangiography and risk of common Moore MJ, Bennett CL: The learning curve for laparoscopic cholecys-
bile duct injury during cholecystectomy, JAMA 289:1639–1644, tectomy. The Southern Surgeons Club, Am J Surg 170(1):55–59,
2003. 1995.
Nenner R, et al: Increased cholecystectomy rates among Medicare Steck T, et al: Prevalence and management of cholelithiasis in heart
patients after the introduction of laparoscopic cholecystectomy, transplant patients, J Heart Lung Transplant 10:1024–1032, 1991.
J Community Health 19(6):409–415, 1994. Steele RJ, et al: Introduction of laparoscopic cholecystectomy in a large
Newman HF, Northrup JD: Extrahepatic biliary tract anatomy, West J teaching hospital: independent audit of the first 3 years, Br J Surg
Surg Obstet Gynecol 71:59–68, 1963. 82(7):968–971, 1995.
Perera MT, et al: Risk factors for litigation following major transec- Steiner CA, et al: Surgical rates and operative mortality for open and
tional bile duct injury sustained at laparoscopic cholecystectomy, laparoscopic cholecystectomy in Maryland, N Engl J Med 330(6):
World J Surg 34(11):2635–2641, 2010. 403–408, 1994.
Perera MT, et al: Specialist early and immediate repair of post- Strasberg SM, Brunt LM: Rationale and use of the critical view of
laparoscopic cholecystectomy bile duct injuries is associated with an safety in laparoscopic cholecystectomy, J Am Coll Surg 211(1):132–
improved long-term outcome, Ann Surg 253(3):553–560, 2011. 138, 2010.
Pesce A, et al: Bile duct injury during laparoscopic cholecystectomy Strasberg SM, Helton WS: An analytical review of vasculobiliary injury
without intraoperative cholangiography: a retrospective study on in laparoscopic and open cholecystectomy, HPB (Oxford) 13:1–14,
1,100 selected patients, Dig Surg 29(4):310–314, 2012. 2011.
Poggio JL, et al: A comparison of laparoscopic and open cholecystec- Strasberg SM, et al: An analysis of the problem of biliary injury during
tomy in patients with compensated cirrhosis and symptomatic gall- laparosopic cholecystectomy, J Am Coll Surg 180:101–125, 1995.
stone disease, Surgery 127(4):405–411, 2000. Suh SW, et al: Accidental gallbladder perforation during laparoscopic
Puente SG, et al: Radiological anatomy of the biliary tract: variations cholecystectomy: does it have an effect on the clinical outcomes?,
and congenital abnormalities, Word J Surg 7(2):271–276, 1983. J Laparoendosc Adv Surg Tech A 22(1):40–45, 2012.
Ransohoff D, Gracie W: Treatment of gallstones, Ann Intern Med Sun S, et al: Three-port versus four-port laparoscopic cholecystectomy:
119:606–619, 1993. meta-analysis of randomized clinical trials, World J Surg 33(9):1904–
Ransohoff D, et al: Prophylactic cholecystectomy or expectant manage- 1908, 2009.
ment for silent gallstones: a decision analysis to assess survival, Ann Svoboda S, et al: Robotic single-site cholecystectomy in the obese:
Intern Med 99:199–204, 1983. outcomes from a single institution, Surg Obes Relat Dis 11(4):882–
Rao A, et al: Safety of outpatient laparoscopic cholecystectomy in the 885, 2015.
elderly: analysis of 15,248 patients using the NSQIP database, J Am Tagge E, et al: Impact of laparoscopic cholecystectomy on the manage-
Coll Surg 217(6):1038–1043, 2013. ment of cholelithiasis in children with sickle cell disease, J Pediatr
Ros A, et al: Laparoscopic cholecystectomy versus mini-laparotomy Surg 29:209–212, 1994.
cholecystectomy: a prospective, randomized, single-blind study, Ann Tsimoyiannis EC, et al: Different pain scores in single transumbilical
Surg 234(6):741–749, 2001. incision laparoscopic cholecystectomy versus classic laparoscopic
Rosemurgy A, et al: Does the cost of robotic cholecystectomy translate cholecystectomy: a randomized controlled trial, Surg Endosc 24(8):
to a financial burden? Surg Endosc 29(8):2115–2120, 2015. 1842–1848, 2010.
Rosenmüller MH, et al: Expertise-based randomized clinical trial of Tuveri M, Tuveri A: Laparoscopic cholecystectomy: complications and
laparoscopic versus small-incision open cholecystectomy, Br J Surg conversions with the 3-trocar technique: a 10-year review, Surg
100(7):886–894, 2013. Laparosc Endosc Percutan Tech 17(5):380–384, 2007.
Soper N: Laparoscopic cholecystectomy, Curr Prob Surg 28:585–655, Ure B, et al: Long-term results after laparoscopic cholecystectomy, Br
1991. J Surg 82:267–270, 1995.
Soper NJ, et al: Comparison of early postoperative results for laparo- Vidovszky TJ, et al: Single-site robotic cholecystectomy in a broadly
scopic versus standard open cholecystectomy, Surg Gynecol Obstet inclusive patient population: a prospective study, Ann Surg 260(1):
174(2):114–118, 1992. 134–141, 2014.
Soper N, et al: Diagnosis and management of biliary complications of Way LW, et al: Causes and prevention of laparoscopic bile duct injuries:
laparoscopic cholecystectomy, Am J Surg 165:663–669, 1993. an analysis of 252 cases from a human factors and cognitive perspec-
Spinoglio G, et al: Single-site robotic cholecystectomy (SSRC) versus tive, Ann Surg 237(4):460–469, 2003.
single-incision laparoscopic cholecystectomy (SILC): comparison of
learning curves. First European experience, Surg Endosc 26(6):1648–
1655, 2012.
CHAPTER 36A
Stones in the bile duct: clinical features and open
surgical approaches and techniques
Kevin N. Shah and Bryan Marshall Clary
OVERVIEW originate within the CBD and are predominantly brown-

pigment (calcium bilirubin) stones. Primary stones occur in
The first successful common bile duct (CBD) exploration by patients with congenital absence of the gallbladder and in those
Thornton in 1889 and the introduction of catheter-based whose CBD had been cleared at the time of prior cholecystec-
biliary decompression by Courvoisier and Kehr marked the tomy. CBD stones that occur in the immediate postcholecys-
initial efforts in treating choledocholithiasis. Open cholecystec- tectomy period should be assumed to be secondary stones that
tomy and bile duct exploration were performed commonly as are the result of an incompletely cleared CBD. Although sec-
the standard treatment for patients with choledocholithiasis for ondary stones are the most commonly observed CBD stones,
many years with good success and low rates of morbidity and particularly in Europe and North America, primary stones are
mortality. During this era of open operative interventions, the encountered more commonly in Asia. This is associated with
percentage of retained stones was only 1% to 3%, and long- the high incidence of intrahepatic bile duct stones seen primar-
term follow-up revealed that revisional surgery was necessary ily in Southeast Asian countries, Taiwan, Hong Kong, and Sin-
in only about 10% of the patients. (Gonzalez et al, 1997; Ham- gapore (Kim et al, 1995) (see Chapter 39). The relative
marström et al, 1995; Neoptolemos et al, 1987; Targarona prevalence of intrahepatic bile duct stones in all gallstone cases
et al, 1996). in Taiwan is extremely high (>50%), and coexisting intrahe-
In the last several decades, however, there has been a shift patic and extrahepatic bile duct stones are found in approxi-
away from routine open cholecystectomy and CBD exploration mately 70% of these.
with improvements in noninvasive imaging and increasing Inflammatory, infective, and congenital typical presentations
sophistication of percutaneous and endoscopic interventions of choledocholithiasis include biliary colic, jaundice, cholangitis
(see Chapters 19, 29, 30, and 36C) . Beyond the widespread (see Chapter 43), and pancreatitis (see Chapter 55). Of these,
availability of endoscopic retrograde cholangiopancreatography pain from biliary colic tends to be the most common symptom-
(ERCP) the increased use of laparoscopy and minimally inva- atic manifestation of CBD stones. In many cases, the intermit-
sive techniques has made open CBD exploration an infre- tent obstruction and passage of CBD stones will result in
quently used tool. Although the majority of cholecystectomies fluctuating elevation of bilirubin and liver function enzymes. If
are now performed laparoscopically (see Chapter 35), laparo- untreated for a long period of time, these recurrent episodes
scopic CBD exploration (LCBDE; see Chapter 36B) has not may lead to secondary biliary cirrhosis. In contrast to the inter-
been similarly embraced. This is in part because of the wide mittent obstruction that results in biliary colic, persistent CBD
availability of ERCP but is also in part because the technical obstruction can result in cholangitis, which presents with Char-
demands of LCBDE do not lend itself to routine use by most cot’s classic triad (fever/rigors, jaundice, and right upper quad-
general surgeons. Surveys of general surgeons practicing in a rant pain) or Reynold’s pentad (Charcot’s triad plus hypotension
rural area of the United States demonstrated that the preferred and altered mental status) (see Chapter 43). Pancreatitis is the
approach to choledocholithiasis was ERCP (75%), followed by second most frequent symptomatic presentation of CBD stones
laparoscopic (21%) and open (4%) exploration. Time con- (see Chapter 55), and depending on the timing of cholangiog-
straints and lack of equipment were the main factors preventing raphy, CBD stones can be identified in up to 50% of these
the application of the laparoscopic technique for treating cho- patients. Patients who have symptomatic bile duct stones are
ledocholithiasis (Bingener & Schwesinger, 2006). at risk of experiencing further symptoms or complications if left
Despite these considerations, there remain indications for untreated. More than one half of patients who had retained bile
operative choledochotomy and, more specifically, for open duct stones experienced recurrent symptoms during a follow-up
exploration. This chapter presents a review of the clinical fea- period of 6 months to 13 years (Johnson & Hosking, 1987),
tures of choledocholithiasis with an emphasis on the technical and 25% developed serious complications (Caddy & Tham,
aspects of open CBD exploration. 2006). The potential for serious sequelae related to CBD stones
makes the identification and ultimate treatment of patients with
ORIGIN OF CHOLEDOCHOLITHIASIS CBD stones of great importance.
The potential long-term sequelae of untreated clinically sig-
CBD stones are broadly classified by their location of origin nificant stones has led some to advocate for routine intraopera-
(see Chapter 32). Secondary stones, those that originate in the tive cholangiography at the time of cholecystectomy so that
gallbladder and migrate into the bile duct, are the most clearance of the duct at the time of surgery can be ensured.
common. Chemically, these stones tend to be cholesterol The clinical significance and natural history of asymptomatic
or black-pigment stones. Primary CBD stones, in contrast, CBD stones, however, is unpredictable, as many small stones
585
will pass spontaneously without incident. During the era of clinical evidence of ascending cholangitis, and bilirubin greater
open cholecystectomy, the practice of routine cholangiography than 4 mg/dL were identified as “very strong” predictors;
was common, and studies from this period demonstrated an dilated CBD on TUS (>6 mm) and bilirubin between 1.8 and
incidence of choledocholithiasis approaching 10% to 15% in 4 mg/dL were “strong” predictors; abnormal LFTs other than
patients without any clinically evident common duct involve- bilirubin, age older than 55 years, and clinical gallstone pan-
ment (Coelho et al, 1984; DenBesten & Berci, 1986; Doyle creatitis were “moderate” factors. Using these factors, patients
et al, 1982; Ganey et al, 1986; Girard, 2000; Hampson et al, are stratified into high, intermediate, and low likelihood for
1981; Lygidakis, 1983; McSherry & Glenn, 1980). This has choledocholithiasis groups. Presence of any “very strong” or
caused some to advocate for the utility of routine intraoperative both “strong” predictors leads to classification as high likeli-
cholangiography, even in patients without clinical signs of cho- hood for choledocholithiasis. Patients with no predictors are
ledocholithiasis. Proponents of a more selective approach to low likelihood, and all others are intermediate.
cholangiography, however, note that the percentage of clinically In patients identified to be at elevated risk for having CBD
significant stones is lower than the 10% to 15% of patients who stones, a thoughtful and efficient approach to subsequent
will have cholangiographic findings of CBD stones when imaging and intervention is required. TUS is the most com-
routine cholangiography is used. monly used initial diagnostic tool for suspected biliary stones
The use of selective intraoperative cholangiography in lapa- because it is noninvasive, widely available, and inexpensive (see
roscopic cholecystectomy series demonstrates similar findings Chapter 15). If CBD stones are detected, specificity is high
(Fogli et al, 2009). Collins and colleagues (2004) identified (95% to 100%); however, TUS has low sensitivity (25% to
filling defects consistent with stones in 4.6% of patients on 60%) for detection of CBD stones and is highly operator
intraoperative cholangiogram. In these patients, access was dependent (Amouyal et al, 1994; Sugiyama & Atomi, 1997).
maintained for the performance of postoperative cholangio- As mentioned previously, indirect evidence such as the pres-
grams. At 48 hours, 26% of patients had a normal cholangio- ence of gallstones or biliary ductal dilation with a CBD diam-
gram, and an additional 26% had evidence for passage of the eter greater than 6 mm in the appropriate clinical setting can
stones by 6 weeks. Twenty-two patients (2.2%) had persistent be predictive of bile duct stones.
CBD stones at 6 weeks after laparoscopic cholecystectomy and Like TUS, standard computed tomography (CT) scanning
underwent ERCP for retrieval (Collins et al, 2004). has a low sensitivity for the detection of bile duct stones and is
most useful in documenting biliary dilation or excluding mass
PREOPERATIVE DIAGNOSIS lesions as a cause of biliary obstruction (see Chapter 18).
Newer techniques of CT cholangiography use contrast agents
In the absence of clinical signs such as cholangitis or pancre- excreted in the biliary tree, which combined with high-resolution
atitis, preoperative identification of choledocholithiais typically helical scans and three-dimensional reconstructions can give
relies on serum liver function tests (LFTs) and imaging studies. accurate and detailed information about the biliary tree (Cabada
The utility of LFTs in predicting the presence of CBD stones Giadàs et al, 2002; Maniatis et al, 2003). The sensitivity of this
has been demonstrated by a number of groups (Peng et al, technique can be as high as 97%, and the specificity is 75% to
2005; Sgourakis et al, 2005). Serum bilirubin and alkaline 96% (Anderson et al, 2008; Cabada Giadàs et al, 2002; Gibson
phosphatase are typically the most commonly used laboratory et al, 2005; Kim et al, 2007; Kondo et al, 2005; Maniatis et al,
values; however, a raised γ-glutamyltransferase level has been 2003; Polkowski et al, 1999; Soto et al, 2000; Zandrino et al,
suggested to be the most sensitive and specific laboratory indi- 2005). Although these data suggest accuracy comparable to
cator of CBD stones. A value greater than 90 U/L has been magnetic resonance cholangiopancreatography (MRCP) (see
proposed to indicate a high risk of choledocholithiasis, with Chapter 19), helical CT cholangiography is limited by several
sensitivity and specificity of 86% and 74.5%, respectively (Peng issues: (1) possible allergic reactions to the contrast agents (as
et al, 2005). high as 15% in one series using intravenous iotroxate) (Gibson
When used in conjunction, clinical examination, laboratory et al, 2005); (2) suboptimal ductal contrast opacification in the
studies, and ultrasonography (typically the first-line imaging presence of significant jaundice (Polkowski et al, 1999; Soto
modality) (see Chapter 15) are sensitive in 96% to 98% and et al, 1999); and (3) limited visualization of intrahepatic duct
specific in 40% to 75% for identification of patients with cho- branches (Chopra et al, 2000; Gibson et al, 2005).
ledocholithiasis (Alponat et al, 1997; Koo & Traverso, 1996; MRCP has been the gold standard for noninvasive biliary
Trondsen et al, 1998). Liu and colleagues (2001) stratified imaging since its introduction in 1991 (Wallner et al, 1991) and
precholecystectomy patients into four groups of descending risk has been recommended by some as the preoperative modality
of choledocholithiasis based on guidelines incorporating clinical of choice for the detection of CBD stones (Hallal et al, 2005;
evaluation, serum chemistry analysis, and transabdominal Shanmugam et al, 2005; Taylor et al, 2002; Topal et al, 2003)
ultrasound (TUS). The occurrence of choledocholithiasis in (see Chapter 17). MRCP provides precise anatomic detail of
these groups (group 1, extremely high risk; group 2, high risk; the biliary tract and has a sensitivity of 81% to 100% and a
group 3, moderate risk; group 4, low risk) was 92.6%, 32.4%, specificity of 92% to 100% in detecting choledocholithiasis
3.8%, and 0.9%, respectively. Triaging patients in this manner (Hallal et al, 2005; Verma et al, 2006). The accuracy of MRCP
resulted in preoperative identification of choledocholithiasis by in diagnosing CBD stones is comparable with that of ERCP
ERCP in 92.3% of the patients who were subsequently referred (see Chapters 19 and 20) and intraoperative cholangiogram
for endoscopic clearance. Similarly, the American Society of (IOC) (see Chapter 23) (Hallal et al, 2005; Vargghese et al,
Gastrointestinal Endoscopy (ASGE) identified several predic- 2000) but may be somewhat less sensitive than endoscopic
tors of choledocholithiasis (ASGE Standards of Practice Com- ultrasound (EUS) for small stones (Verma et al, 2006). As a
mittee, et al, 2010) in their guidelines for the use of endoscopy diagnostic test, MRCP has largely replaced ERCP, once con-
in the evaluation of CBD stones. CBD stones seen on TUS, sidered the gold standard of preoperative bile duct imaging,
A. Gallstones and Gallbladder Chapter 36A Stones in the bile duct: clinical features and open surgical approaches and techniques 587
because the nonselective use of ERCP in all patients with sus- management (laparoscopic cholecystectomy and intraoperative
pected choledocholithiasis detects CBD stones in less than 50% cholangiogram; CBD exploration in those with CBD stones)
(Behrns et al, 2008; Petrov et al, 2008). Additionally, indis- over preoperative ERCP, followed by laparoscopic cholecystec-
criminate use of ERCP exposes over half of patients to unneces- tomy. Several prospective randomized controlled trials com-
sary procedure-related morbidity and mortality (Demartines pared these two treatment strategies (Cuschieri et al, 1999;
et al, 2000); thus ERCP is better used as therapeutic interven- Rogers et al, 2010; Sgourakis et al, 2005) and found the two
tion rather than a diagnostic one. Although MRCP is currently groups had equivalent success rates of duct clearance and
the most accurate noninvasive imaging modality for choledo- patient morbidity, but a significantly shorter hospital stay was
cholithiasis, it may miss stones smaller than 5 mm in diameter reported with the single-stage laparoscopic treatment. Cusch-
and can underestimate the number of stones detected (Varg- ieri and colleagues (1999) concluded that precholecystectomy
ghese et al, 2000). Furthermore, it is expensive when compared endoscopic clearance should be reserved for higher-risk patients
with TUS or CT, not readily available at smaller facilities, and (those with multiple comorbidities, cholangitis, severe pancre-
may not be technically feasible in obese patients or those with atitis), whereas fit patients (American Society of Anesthesiolo-
significant claustrophobia. gists categories I and II) should be treated with a single-stage
EUS has also emerged as an alternative to ERCP and MRCP laparoscopic approach.
for preoperative assessment of bile duct stones and was initially When laparoscopic CBD exploration is not part of the sur-
described around 1990 (Amouyal et al, 1989; Edmundowicz geon’s armamentarium, the decision to perform a precholecys-
et al, 1992) (see Chapter 16). The overall diagnostic perfor- tectomy ERCP should be weighed carefully because it is not
mance of EUS has been evaluated in numerous prospective without complications. Although most post-ERCP complica-
studies since then, and in two meta-analyses (Garrow et al, tions are mild to moderate in severity (Andriulli et al, 2007),
2007; Tse et al, 2008) the pooled sensitivity and specificity of the risk of severe complications, such as pancreatitis, bleeding,
EUS were 89% to 94% and 94% to 95%, respectively. Ran- infection, and perforation, need to be weighed against the likeli-
domized controlled trials have also demonstrated that EUS- hood that ERCP will find clinically relevant CBD stones.
guided ERCP can avoid unnecessary ERCP in up to two thirds ERCP has an overall complication rate of 10% and a mortality
of patients when compared with using ERCP alone for diagno- rate less than 0.5% (Cotton, 1993; Davis et al, 1997; Delorio
sis of choledocholithiasis (Petrov & Savides, 2009). Further- et al, 1995; Ponsky, 1992). Additionally, the increased financial
more, selective use of ERCP based on EUS findings resulted cost of preoperative ERCP should be considered when evaluat-
in reduced risk of overall complications and post-ERCP pan- ing its role in treatment of suspected stones.
creatitis. When taken in sum, these data suggest that the ERCP Even using criteria such as elevated LFTs and imaging find-
is best reserved as a therapeutic measure for patients with a ings, accurately predicting which patients will have clinically
high probability of CBD stones, rather than as an initial diag- relevant choledocholithiasis is difficult. Several studies have
nostic test. shown that negative ERCP was performed in 40% to 70% of
patients because most of these biochemical and radiographic
abnormalities were the result of transient biliary obstruction
TIMING AND SEQUENCE OF INTERVENTIONS secondary to stones that ultimately passed into the duodenum
(Cotton, 1993; Cuschieri et al, 1996; Delorio et al, 1995; Stain
Suspected Choledocholithiasis Prior to et al, 1991; Ponsky, 1992). As mentioned earlier in this chapter,
Cholecystectomy the ASGE published a system to grade likelihood of choledo-
In patients with suspected choledocholithiasis, selecting and cholithiasis into high, intermediate, and low likelihood groups
sequencing the most appropriate of laparoscopic, open, and based on risk factors, including LFTs, bile duct size on imaging,
endoscopic therapeutic modalities can be challenging. Prior to and presence of gallstone pancreatitis or cholangitis (ASGE
the popularization of laparoscopy, precholecystectomy endo- Standards of Practice Committee, et al, 2010). Treatment
scopic clearance of the CBD was uncommon. Several studies strategies for patients stratified to the low and high likelihood
did not reveal any morbidity or mortality advantage with pre- groups are generally agreed upon. Patients who are in the low
operative endoscopic sphincterotomy (Heinerman et al, 1989; risk for choledocholithiasis group should go directly to laparo-
Neoptolemos et al, 1988; Stain et al, 1991), and one actually scopic cholecystectomy (with or without IOC) because CBD
showed an increased rate of morbidity in patients who under- clearance is unlikely to be necessary.
went preoperative ERCP, followed by open cholecystectomy, For patients in the high-risk group, particularly those
compared with the group that was treated with single-stage needing biliary decompression for treatment of acute cholangi-
open cholecystectomy and CBD exploration (Stain et al, 1991). tis (Cuschieri et al, 1999; Leung, 2003) and in patients with
A systematic review (Martin et al, 2006) demonstrated that severe gallstone pancreatitis and evidence of persistent choledo-
open surgery resulted in less mortality, lower incidence of cholithiasis, precholecystectomy ERCP is warranted. Patients
primary treatment failure, and fewer procedures required per with multiple medical comorbidities, limited life expectancy, or
patient. Given these data, single-stage cholecystectomy and bile other issues that would make them poor surgical candidates,
duct exploration (as indicated) was the preferred approach in ERCP with endoscopic sphincterotomy (ES) and biliary
the era of open cholecystectomy. decompression can sometimes be used as the definitive man-
The advent of laparoscopy saw a rise in the popularity of agement without cholecystectomy (see Chapter 29). This strat-
preoperative ERCP CBD clearance for patients with suspected egy is not ideal for patients who are acceptable surgical
choledocholithiasis, in part because laparoscopic CBD is more candidates, however, because there is a risk of recurrent biliary
technically challenging (see Chapters 36B and 36C). For sur- symptoms if cholecystectomy is not performed. In a prospective
geons who are comfortable with laparoscopic CBD exploration, randomized trial published in 1995 by Hammarström and col-
there are some data that support a single-stage laparoscopic leagues, an expectant policy after ES was compared with open
cholecystectomy combined with CBD exploration. It was require further intervention (Roslyn, 1993). Retained calculi
reported that 20% of the patients after ES alone needed cho- occur rarely after open cholecystectomy without CBD explora-
lecystectomy during follow-up (Hammarström et al, 1995). A tion (Bergdahl & Holmlund, 1976), whereas incidence in those
prospective randomized trial of high-risk patients performed by who undergo open cholecystectomy with concomitant CBD
Targarona and colleagues (1996) comparing ES and subse- exploration is slightly higher but still reported to be less than
quent open cholecystectomy to ES alone resulted in similar 5% (Dayton et al, 1984; Kappes et al, 1982; Roslyn, 1993).
findings. They noted patients who underwent elective open Retained CBD stones occur with higher frequency after positive
cholecystectomy had significantly fewer recurrent biliary symp- CBD exploration than after a negative one. The rate of recur-
toms (6% vs. 21%) and needed fewer readmissions (4% vs. rence increases to approximately 20% after a second operation
23%) than patients who did not undergo surgery after ES. on the biliary tract for choledocholithiasis (Saharia et al, 1977;
In contrast to the general consensus of how to treat high- Way, 1973), and this rate increases after subsequent reopera-
and low-likelihood patients, the question of how to treat tion (Allen et al, 1981).
intermediate-risk patients has been the subject of some debate.
A recent randomized clinical trial (Iranmanesh et al, 2014) Treatment
examined an upfront cholecystectomy and IOC strategy against Endoscopic and percutaneous methods remain the preferred
preoperative ERCP and subsequent cholecystectomy for modalities when managing recurrent or retained CBD stones
intermediate-risk patients. Fifty patients were randomized to (see Chapters 30 and 36C). The open surgical approach is
each group, and differences in length of stay, number of sub- reserved for patients who have failed nonoperative treatments.
sequent CBD interventions, morbidity, mortality, and quality Decision making is further influenced by clinical presentation,
of life were analyzed. No significant difference was found in condition of the patient, institutional expertise, and presence
morbidity or quality of life; however, patients who underwent or absence of a T-tube.
cholecystectomy as the initial procedure had a significantly
shorter length of stay (median, 5 days vs. 8 days; P < .001) and RETAINED STONES IN THE PRESENCE OF A T-TUBE. The rise in popu-
fewer common duct investigations. Based on these findings, the larity of laparoscopic biliary surgery has also meant decreasing
authors suggested that initial cholecystectomy with intraopera- use of the T-tube. Even when laparoscopic CBD exploration is
tive cholangiogram may be the preferred approach. In any performed, primary closure has been shown to be safe (Dong
patient being considered for precholecystectomy ERCP, strong et al, 2014; El-Geidie, 2010; Gurusamy et al, 2013), and
consideration should be given to the use of EUS (with selective routine use of postexploration T-tubes is not common (see
ERCP) or MRCP because these studies are less invasive and Chapters 31 and 42). If a T-tube is present, it provides nonin-
have minimal risk compared with that of an ERCP. vasive options for accessing the biliary tree postoperatively. In
the presence of a T-tube, retained CBD stones in the immedi-
Common Bile Duct Exploration at Time of ate postoperative period can be managed with observation,
Open Cholecystectomy mechanical extraction, or ES.
Laparoscopy has become routine in much of the world, and as If choledocholithiasis is noted within the first 4 to 6 weeks
laparoscopic experience has grown, surgeons have become postoperatively, no treatment is typically necessary, assuming
increasingly more comfortable using laparoscopic CBD explo- there is no evidence of obstruction or cholangitis. In the initial
ration when choledocholithiasis is noted intraoperatively. In weeks after CBD exploration, 10% to 25% of retained stones
areas of the developing world, however, where access to endo- found on postoperative cholangiography will pass spontane-
scopic, radiologic, and laparoscopic expertise is limited, open ously into the duodenum, and no further treatment is required.
cholecystectomy and bile duct exploration remains a mainstay If calculi persist after 4 to 6 weeks, treatment options include
of treatment. Even in settings where laparoscopy and endos- radiologic approach through the T-tube tract (see Chapter 30)
copy are readily available, however, there will still be some or ERCP (see Chapter 29). Because of its high success rate and
patients in whom an open approach to CBD exploration may low morbidity and mortality, nonoperative mechanical extrac-
be required (see Chapters 36B and 36C). Principal among tion through the T-tube tract is an attractive treatment choice.
these include (1) patients with large or impacted CBD stones A success rate of 95% has been reported with a morbidity rate
and those who have failed previous endoscopic interventions, of only 4% (Mazzariello, 1978). Burhenne (1980) reported no
in whom biliary enteric drainage is indicated; (2) those with deaths in 661 patients. When complications do occur, they can
anatomic considerations that preclude endoscopic treatment, be treated medically in most instances, and only 0.2% of cases
such as prior gastric resection or duodenal diverticulae; and (3) have required surgery (Mazzariello, 1978).
patients who require an open approach for cholecystectomy, ES also has been shown to be effective in the management
including those with Mirizzi syndrome, biliary-enteric fistula, of retained stones in the early postoperative period after explo-
severe cholecystitis, or a high index of suspicion for cancer. ration of the CBD with a T-tube still in place (Hammarström
et al, 1996; O’Doherty et al, 1986). Although ES has the con-
siderable advantage that it can be carried out as soon as retained
Postcholecystectomy Choledocholithiasis
stones are discovered, treatment may be unnecessary in some
(See Chapter 38) patients because stones may pass spontaneously. Some authors
Incidence (Lambert et al, 1988) have suggested that mechanical stone
The majority of initial operations for gallstone disease, with or extraction through the T-tube tract is superior to ES because
without demonstrated choledocholithiasis, are curative, but of its high success rate and lower morbidity profile, although
some patients will develop sequelae of choledocholithiasis post modern endoscopic equipment may mitigate some of the post-
cholecystectomy. Approximately 1% to 2% of all patients who ERCP hemorrhagic complications seen in earlier series
undergo cholecystectomy have stones left in the CBD that (Christoforidis et al, 2004). Regardless, the safety and efficacy
of percutaneous intervention through the T-tube makes it an choledocholithotomy. Tompkins and Pitt (1982) and Cameron
ideal choice for initial postoperative interventions, and ES is (1989) emphasized, however, that concomitant biliary drainage
best used in the early postoperative period before a T-tube tract should not be regarded as mandatory procedure in all patients
is well formed, when the patient is clinically unstable, the with retained or recurrent stones. In general, biliary-enteric
T-tube is inappropriate in size and position, or mechanical drainage at reoperation is appropriate in the following scenar-
extraction through the T-tube has failed. If these techniques ios: (1) stricture or stenosis of the distal bile duct or sphincter
fail, operative management can be undertaken with the expec- of Oddi, (2) marked dilation of the duct of 2 cm or more, (3)
tation of a high success rate and acceptable morbidity and multiple or primary bile duct stones, (4) inability to remove all
mortality (Cameron, 1989; Girard & Legros, 1981). stones from the duct, and (5) a third operation.
Transduodenal sphincteroplasty, choledochoduodenos-
RETAINED OR RECURRENT STONES IN THE ABSENCE OF A T-TUBE. ES tomy, and choledochojejunostomy are effective methods of
is the procedure of choice and should be attempted first in biliary enteric drainage (Braasch et al, 1980; Johnson & Harding
patients without a T-tube in place (Cameron, 1989; Sivak, Rains, 1978; Jones, 1978) (see Chapter 31). With the wide
1989) (see Chapters 29 and 36C). Most reports of ES indicate availability of ERCP, operative sphincteroplasty is rarely
a success rate in achieving overall clearance of stones from the required because ES is sufficient in most cases. In the presence
CBD of more than 85% (Cotton, 1984; Dasari et al, 2013; of a long distal CBD stricture, ES is not an appropriate choice
Lambert et al, 1991). Although early complication rates for ES because it does not address the primary obstructive issue. For
range from 5% to 15%, emergency surgery is uncommonly ducts smaller than 1 to 1.5 cm in diameter, sphincteroplasty is
required and most complications can be managed conserva- the preferred operative approach as this avoids possible anas-
tively (Cotton, 1984; Dasari et al, 2013; Escourrou et al, tomotic stricture formation, but it does carry a greater risk of
1984). Hemorrhage, pancreatitis, cholangitis, and perforation postoperative pancreatitis. Occasionally, recurrent or primary
are the most frequent complications, and mortality usually is stones will be seen in patients with dilated ducts and a widely
reported at 0.5% to 2% (Lambert et al, 1991; Sivak, 1989). patent sphincter after sphincteroplasty or sphincterotomy.
Long-term complication rates, mainly from stenosis or new In such cases, choledochoduodenostomy or Roux-en-Y cho-
stones or both, are low (<10%), and most complications can ledochojejunostomy is necessary. Side-to-side or end-to-side
be managed endoscopically (Cotton, 1984; Escourrou et al, choledochoduodenostomy and end-to-side Roux-en-Y cho-
1984; Hammarström et al, 1996; Sivak, 1989). ledochojejunostomy are excellent drainage options for CBDs
If ES is unsuccessful, percutaneous transhepatic rendezvous larger than 1.5 cm and offer better decompression of an
techniques can sometimes aid in duct clearance, particularly if extremely large duct. In the context of previous biliary pancre-
there is difficulty cannulating the ampulla. Often, however, atitis, patients who present for reoperation with multiple stones
failures of endoscopic management will be the result of large and an incompletely cleared proximal biliary system may be
impacted stones or anatomic issues that are not ameliorated by better served with an end-to-side choledochoduodenostomy as
a percutaneous approach. In such settings, operative manage- opposed to a side-to-side technique because it minimizes the
ment is the most reasonable alternative (Cameron, 1989). chance of stones dropping distally and causing recurrent pan-
Reoperation for retained stones can be performed safely, with creatitis. Sump syndrome is an uncommonly observed compli-
operative mortality less than 2% (Girard & Legros, 1981). cation of choledochoduodenostomy and should be managed
Miller and colleagues (1988) reported 237 patients with CBD initially by endoscopic modalities. If ERCP fails to improve
stones treated by CBD exploration or ES. Success was higher symptoms, the choledochoduodenostomy can be converted to
and mortality was lower for the operatively managed group. a Roux-en-Y choledochojejunostomy.
The complication rate was similar, but the complications
tended to be more serious and more apt to require surgery in Clinical Experience With Reoperation
the ES group. A systematic review performed by Dasari and Girard (2000) reviewed all patients who underwent reoperation
colleagues (2013) found that duct clearance in patients under- for retained or recurrent choledocholithiasis at the Maisonneuve-
going open bile duct exploration was superior to ES, but it Rosemont Hospital between 1969 and 1990. Eighty-five patients
should be noted that most series that compare open with preoperatively confirmed choledocholithiasis underwent a
cholecystectomy/CBD exploration to ES are from the era of total of 88 operations. Eighty-five of these operations were
open surgery, which also corresponds to the early days of ERCP second procedures, and three patients required a third opera-
and ES. Therefore caution should be used in extrapolating tion. Three types of bile duct reoperation were performed:
these data to the modern endoscopic experience. Nevertheless, choledocholithotomy with T-tube drainage (64 patients), cho-
these findings reinforce that surgery can be a valuable, effective, ledocholithotomy with side-to-side choledochoduodenostomy
and safe tool in the treatment of recurrent/retained CBD stones, (15 patients), and choledocholithotomy with transduodenal
even if confined to the subset of patients who fail ES. sphincteroplasty (6 patients). Choledocholithotomy with
When reoperation is required for retained CBD stones, the T-tube drainage in 1 patient and choledocholithotomy with
optimal procedure is complete removal of all stones via cho- side-to-side choledochoduodenostomy in 2 patients were per-
ledocholithotomy, choledochoscopy, placement of a T-tube (in formed at a third operation.
many cases), and completion cholangiography. This procedure The average hospital stay was 9.3 days. There were no
is adequate for most patients, and the overall failure rate is deaths in the series despite the fact that 43 of 85 patients were
has been reported as low as 3% (Girard & Legros, 1981). older than age 60 years and 44 patients had associated risk
Others, however, have reported significantly higher failure rates factors. Six minor complications were observed, none of
(Allen et al, 1981; Saharia et al, 1977), which has prompted which necessitated urgent surgery. Two patients (3%) of the
some authors (Allen et al, 1981; Lygidakis, 1982) to recom- 64 who had choledocholithotomy with T-tube drainage devel-
mend biliary-enteric drainage in all patients with previous oped recurrent bile duct stones 4 and 5 years after a second
TABLE 36A.1 Mortality Rates of Biliary Reoperation for SURGICAL TECHNIQUES FOR EXPLORATION OF
Retained or Recurrent Bile Duct Stones THE COMMON BILE DUCT
Reference No. Operations No. Deaths (%) The principal techniques for open exploration of the CBD will
be detailed in this section. Broadly, the goals of CBD explora-
Saharia et al, 1977 30 0
tion include complete clearance of calculi from the biliary
Jones, 1978 22 0
system and establishment of free flow of bile into the gut. The
McSherry & Glenn, 1980 341 7 (2%) preferred approach to CBD exploration is typically through a
Allen et al, 1981 47 1 (2%) supraduodenal choledochotomy, with the transduodenal/
De Almeida et al, 1982 22 1 (5%) transampullary route reserved for patients with impacted stones
DenBesten & Berci, 1986 86 2 (2.3%) that cannot be removed readily from above. Stones impacted
Girard, 2000 88 0 at the ampulla can be broken down and removed through a
TOTAL 920 15 (1.6%) supraduodenal approach; however, a transduodenal sphinctero-
plasty is generally less traumatic.
Clearance of the biliary tree should be confirmed by per-
forming postexploratory choledochoscopy and cholangiogra-
phy. The value of choledochoscopy has been confirmed by
operation, and side-to-side choledochoduodenostomy was many authors (Dayton et al, 1984; Kappes et al, 1982; Nora
performed. et al, 1977). Postexploratory cholangiography should also be
Taking Girard’s (2000) experience with that of other performed prior to closure of the abdomen, not only because
reported series, there have been 15 deaths among 920 patients it can locate missed stones, but also because it may reveal
submitted for reoperation for recurrent bile duct stones (Table unsuspected disruption of the biliary ductal system. If cholan-
36A.1). In one of these series (McSherry & Glenn, 1980), giography technique is meticulous, issues with false positives
choledocholithotomy was done for retained or recurrent stones from air bubbles and poor opacification of the entire system
in 341 patients; 7 patients died after the procedure, resulting can be largely eliminated to provide consistent and reliable
in a mortality of 2%. If three of these deaths that occurred after cholangiograms.
urgent operations for cholangitis or pancreatitis are excluded, The selective use of biliary-enteric drainage procedures is
however, and only the elective procedures are considered, another method to decrease the incidence of subsequently
only 4 patients died (1%) after repeat choledocholithotomy. symptomatic retained stones. Although we do not recommend
These results show that the overall mortality rate for retained routine biliary-enteric decompression at initial operation, it
or recurrent stones is less than 2%, with most deaths occurring should be considered carefully in patients with multiple duct
in elderly patients. This mortality rate is comparable to that stones, particularly in those with large stones; dilated ducts; and
of ES. in elderly patients. If these conditions pertain in an elderly or
To summarize, ES has become the first-line therapy for poor-risk patient, choledochoduodenostomy may obviate reex-
retained or recurrent bile duct stones, but surgery can be per- ploration. Other indications include (1) the presence of irre-
formed safely with low mortality and morbidity when required. trievable intrahepatic stones, (2) proven ampullary stenosis, or
Surgery remains a critical component of the armamentarium (3) an impacted ampullary stone.
that can be used to treat recurrent bile duct stones. As with
most things in modern medicine, a multidisciplinary approach
Supraduodenal Choledochotomy and Exploration of the
to recurrent CBD stones is important to properly select and
sequence the numerous options now available. Gastroenterolo- Common Bile Duct
gists, radiologists, and surgeons should work together closely Exposure
to assess the most appropriate intervention for an individual The liver is retracted superiorly with a broad-bladed, slightly
patient. In making the choice between open surgery, laparo- curved retractor, such as a Hartmann (“sweetheart”) retractor.
scopic surgery, percutaneous therapies, and ES, the surgeon This retractor should be deep enough to displace the liver but
must consider not only the published data but institutional not so curved as to traumatize it. A pack should be put over
expertise and experience. In a patient with a retained stone and the hepatic flexure of the colon down to the hepatorenal pouch
a T-tube in place, percutaneous extraction through the T-tube and the medial part of the duodenum. This pack is retracted
tract or ES should be attempted first. In the absence of a by a similar, broad-bladed retractor to prevent the colon or
T-tube, ES should be attempted first. If unsuccessful or con- duodenum from obscuring vision (Fig. 36A.1). The lesser
traindicated, operative management is a reasonable alternative. omentum and stomach are retracted to the left after placement
Surgical intervention has a high success rate and acceptable of another pack. The gallbladder is usually removed before
rates of mortality and morbidity. Before operation, the surgeon exploration of the CBD because it may obstruct vision, although
must make an accurate diagnosis of retained stones by using a in some cases it can be a useful aid to retraction. Palpation of
combination of MRCP, ERCP, and US. These findings should the CBD and handling of its lower part during exploration and
be confirmed via intraoperative cholangiography and complete subsequent choledochoscopy cannot be done properly without
clearance of the biliary tree documented with completion chol- having performed a classic Kocher maneuver (Fig. 36A.2).
angiography and choledochoscopy. Most patients do not
require biliary-enteric drainage, but certain patients, particu- Choledochotomy
larly those with multiple or incompletely cleared calculi, large A distal vertical supraduodenal choledochotomy is generally
ducts (>2 cm), and distal CBD strictures, will benefit from preferred for several reasons. First, because a choledochoduo-
drainage procedure. denostomy may be required (see Chapter 31), the opening
FIGURE 36A.1. Exposure of the common bile duct by packs and

retractors.
should be positioned in the lowest part of the supraduodenal

CBD such that an anastomosis can be created easily and
without tension (Figs. 36A.3 and 36A.4). Second, a distal
choledochotomy leaves the maximal amount of bile duct proxi-
mally so that it may be used in the future for an additional
procedure (e.g., repair of a stricture). Third, the usual distance
from this point to the papilla measures 7 cm or more, which is
the exact length of the rigid choledochoscope sometimes used
for CBD exploration; this is less of an issue when using a flex-
ible scope, which is typically much longer.
The anatomy of the cystic duct is highly variable, and care
must be exercised to open the correct duct (see Chapter 2). A B
cystic duct lying anterior or closely applied to the CBD can be FIGURE 36A.2. A, The gallbladder has been removed. The dotted line
easily opened in error, particularly if dilated. Bile is aspirated indicates the incision in the retroperitoneum to allow mobilization of the
by gentle suction, and a specimen should be sent for culture duodenum by the Kocher maneuver (B).
(Fig. 36A-5).
Exploration of the Duct gently withdrawn through the papilla, and then the balloon is
All efforts must be made to minimize trauma related to the reinflated immediately. Passage back through the ampulla can
exploration, and rigid instruments should be avoided if possi- be detected by a sudden easing of the pull on the catheter. At
ble, as false passages into the duodenum and pancreas can be this point, the syringe is held in the surgeon’s nondominant
easily created (Gunn, 1983; Orloff, 1978). Grasping forceps of hand, and the degree of balloon inflation is controlled by the
any type can catch the duct wall and result in delayed stricture thumb and the plunger. With gentle traction superiorly from
formation. Use of the Fogarty balloon catheter can avoid these long forceps held in the surgeon’s dominant hand, the catheter
issues and has been found to be suitable for CBD exploration is gradually pulled up to the choledochotomy site (Fig. 36A.8),
(Fogarty et al, 1968; Fox & Gunn, 1984). with care being taken to prevent any stone slipping into the
The Fogarty probe is held in long forceps with the surgeon’s proximal biliary tree. If the traction is anterior rather than
dominant hand and introduced into the CBD (Fig. 36A.6), superior, there is the risk of lacerating the opening into the duct
The catheter is then passed into the duodenum. The balloon (Fig. 36A.9), and the risk is increased when the opening in the
is inflated, and the catheter is withdrawn until it impinges duct is longitudinal. The procedure is repeated until the distal
against the papilla (Fig. 36A.7). The balloon is identified in the duct is considered to be clear.
second part of the duodenum by palpation, and the location is Next, this procedure is repeated for the proximal ducts by
noted as the site of duodenotomy should a sphincteroplasty reinserting the catheter upward into each of the main hepatic
become necessary. Stones can usually be felt against the shaft ducts. The degree of balloon inflation is of great importance
of the catheter within the duct. The balloon is deflated and here, as overinflation will result in damage to the ducts and
FIGURE 36A.3. The common bile duct is opened just above the
duodenum to leave room for a choledochoduodenostomy if
necessary. FIGURE 36A.4. Two fine absorbable stay sutures are used to lift and
render the common bile duct (CBD) tense for an incision about 1 to
2 cm long, depending on the size of the duct and the size of the stones.
If the CBD is not made tense, damage can be done to the posterior
wall, or an irregular incision can be made.
A B
FIGURE 36A.5. A, The cystic duct may lie anterior to the common bile duct (CBD) and may be opened in error. B, The cystic duct may run parallel
to the CBD with a low entrance, mimicking a dilated duct.
A B
FIGURE 36A.6. A Fogarty catheter is fed into the duct with forceps
by using the dominant hand. The operator’s nondominant hand is used
to grasp the mobilized duodenum, which allows palpation of the passage
of the catheter and of any stones within the intrapancreatic portion of
the common bile duct.
underinflation risks missing stones. Correct inflation can be

achieved by inflating the balloon until the tension of the syringe
plunger can be felt in the fingers. This tension is maintained as
the catheter is withdrawn into the gradually widening duct. It
is important to remove the stone when it appears at the cho-
ledochotomy opening and to avoid letting it fall into another
part of the duct.
The next step in CBD exploration is to irrigate the duct
generously with saline. Small stones, sludge, and debris can be
flushed into the duodenum or back into the choledochotomy
opening by irrigating the ductal system. Finally, the Fogarty
C D
catheter is passed again into the duodenum, and the balloon is
inflated and retracted against the papilla. While the catheter is FIGURE 36A.7. A, The Fogarty catheter is attached to a syringe, and
the balloon is inflated in the duodenum. B, The Fogarty catheter is
held in the surgeon’s dominant hand, the index and middle
retracted with the balloon against the papilla. C and D, The balloon is
fingers of the other hand are placed posterior to the duodenum deflated and gently withdrawn until it slips through the papilla; the
with the thumb anterior; this allows palpation of the duct balloon is then reinflated.
against the wall of the catheter for any residual stones.
Postexploratory Investigations orifices of the smaller biliary radicles. Although flexible scopes
Following exploration, the surgeon must make every effort to carry a higher cost and are more difficult to maintain compared
ensure that the duct system is normal using choledochoscopy with rigid scopes, they allow for less traumatic choledochos-
and cholangiography (see Chapter 23). copy, are more versatile because of their increased length, and
allow introduction of therapeutic instruments through the
Choledochoscopy working channel. Some surgeons experienced in choledochos-
Choledochoscopy is the established method to ensure that the copy recommend exploration of the CBD and removal of
duct system is normal. Modern instruments are small enough stones under direct vision by using the choledochoscope (Berci,
to allow visualization of the major right and left hepatic ducts 2000). This can be facilitated by use of a stone basket through
and intermediate hepatic ducts and to allow visualization of the the working port of a flexible choledochoscope. The use of
A B
FIGURE 36A.8. A, The balloon is withdrawn gently, revealing the stone. B, Long forceps can be used to obstruct the common hepatic duct to
prevent the stone from slipping upward.
adequate clearance of the biliary system. With proper technique

and use of fluoroscopy, T-tube cholangiography is an excellent
tool for detecting residual stones after CBD exploration (Fig.
36A.10). The cystic duct stump, a possible location of residual
stones, can be delineated, and incorrect placement of the
T-tube can be detected to prevent complications after the oper-
ation. In case of residual stones, the T-tube has to be removed,
and the duct needs to be explored again; this necessitates a
second suture of the CBD, which is a disadvantage of T-tube
cholangiography.
T-TUBE DRAINAGE. The standard practice is to use a T-tube to

allow spasm or edema of the sphincter to settle after the trauma
of the exploration. Failure to drain the duct might theoretically
result in a buildup of pressure in the extrahepatic ductal system
and cause leakage at the disruption of the closure of the duct,
along with biliary peritonitis. As noted earlier in this chapter,
several series have not found any decreased risk for bile leak
with placement of a T-tube and have in fact noted increased
operative times and length of stay (Gurusamy et al, 2013).
Routine use of a T-tube has, therefore, been questioned. The
main role of T-tube placement, therefore, is not to prevent a
bile leak but rather to allow an avenue for subsequent treatment
of retained stones in high-risk patients. In the event of a retained
stone, the T-tube can be useful later for interventional radio-
logic techniques through the tract created by the tube (Fig.
FIGURE 36A.9. Angled traction on the Fogarty catheter can result in 36A.11) or even direct choledochoscopy. The size of the T-tube
tearing of the lower end of the choledochotomy. should be adapted to the diameter of the CBD, and 14 Fr is
the smallest size that should be used. Use of a smaller tube will
grasping or biopsy forceps should be avoided if possible because not result in a satisfactory tract for subsequent interventional
these instruments can damage the bile duct. radiology procedures.
T-Tube Cholangiography T-TUBE PLACEMENT. First, the limbs of the T-tube must be short-
After insertion of a T-tube and closure of the choledochotomy, ened (Fig. 36A.12A). T-tubes can become obstructed, particu-
T-tube cholangiography should be performed to ensure larly if they are tight fitting, and they can be difficult to extract.
This situation can be avoided by cutting off a strip of the wall

(Fig. 36A.12B). The practice of dividing the back wall of the
T-tube makes subsequent interventional radiology more diffi-
cult because the guidewire lodges in the posterior defect. This
problem can be avoided by making the length of the T-tube
appropriate or by limiting the division of the T-tube. The modi-
fied T-tube is held in Desjardin forceps, which conveniently
grasps the T-junction of the tube, allowing it to be slipped into
the choledochotomy (Fig. 36A.13). The long limb of the tube
is placed at the lower end of the opening, and repair is begun
just above the upper apex of the incision by using continuous
or interrupted absorbable fine sutures. The final stitch should
close the opening against the T-tube (Fig. 36A.14).
AVOIDING PROBLEMS IN THE CLOSURE OF THE CHOLEDOCHOTOMY.

When closing the choledochotomy over a T-tube, care must be
taken to ensure that the wall of the T-tube is not caught in one
of the sutures or accidentally affixed to the CBD wall. If this
occurs, there is a risk of laceration of the CBD when the T-tube
is eventually removed. The proximal limb of the T-tube should
be shortened so that it does not enter and obstruct one of the
hepatic ducts (Fig. 36A.15). The distal limb should be similarly
shortened so that it does not enter the duodenum because, if
the tube does, it can act as a siphon. Furthermore, a tube
extending through the papillary orifice may incite pancreatitis.
The correct position of the T-tube is with the long limb emerg-
ing under the costal margin laterally (Fig. 36A.16). This posi-
FIGURE 36A.10. A postexploration T-tube cholangiogram identifies a
tion facilitates radiologic techniques for later postoperative
residual stone (arrow) in the intrahepatic bile ducts that was missed
during operative exploration of the bile duct. removal of stones should this be necessary. A suction drain is
A B
FIGURE 36A.11. A, Faceted retained distal common bile duct (CBD) stone on T-tube cholangiography 1 week after operation. The T-tube is inserted
in the common hepatic duct above the confluence of the cystic duct. B, The retained CBD stone is ensnared in the wire basket before extraction.
This stone measured 5 mm in diameter and was extracted intact through the sinus tract of a 14-Fr T-tube.
6 cm
A B
FIGURE 36A.12. A, The T-tube is modified by shortening the limbs to prevent proximal obstruction and distal entry into the duodenum. B, A T-tube
is modified by removing half the diameter to prevent obstruction and enable easy removal.
FIGURE 36A.14. The choledochotomy closure is begun above, with

the T-tube emerging at the lower end of the repair.
FIGURE 36A.13. The T-tube is introduced by Desjardin forceps.

A B C
FIGURE 36A.15. A, The suture must not catch the tube as shown here. B, The T-tube limb should not enter the hepatic duct, as shown here.
C, The distal end of the T-tube should not enter the duodenum, as shown here.
easily through the papilla into the duodenum. The T-tube

should be left to external gravity drainage until this occurs.
Once it appears that bile is flowing into the duodenum, a
T-tube cholangiogram can be taken about 5 to 7 days postop-
eratively. If it appears normal, the tube is removed on day 7 or
8 by gentle traction.
If there are residual stones or unclear findings on T-tube
cholangiography in the first 1 to 2 weeks after surgery, interven-
tion does not necessarily need to be undertaken unless the
patient has signs of cholangitis or rising bilirubin. A repeat
cholangiogram several days later will often show spontaneous
passage of the stone. If a stone is still seen in an otherwise well
patient, the patient can safely be sent home with a sealed drain-
age system and instructions to open the drain to a bag in
the event of any problems. After about 5 weeks, further chol-
angiography is carried out. If the stone is still present, it is
extracted via interventional radiology or endoscopic papillot-
omy (Fig. 36A.17).
FIGURE 36A.16. The T-tube should be brought out lateral to the
wound, and a closed-suction drain should be placed in the hepatorenal Transduodenal Sphincteroplasty
space beneath the liver.
The role of transduodenal sphincteroplasty in the treatment of
choledocholithiasis centers predominantly around management
placed on the right, within the abdomen as high as possible in of impacted stones at the ampulla. It also has a role in treating
the hepatorenal pouch. patients with anatomy that prevents ES (e.g., Billroth II gas-
trectomy), failures of ES, and sometimes in patients with pan-
POSTOPERATIVE MANAGEMENT. Initially, bile is allowed to drain creatitis where a drainage procedure of the duct of Wirsung is
freely into a bile bag to allow any spasm or edema of the sphinc- indicated (Lehman & Sherman, 1998). Similarly, hydatid cyst
ter to settle before testing the suture line of the choledochot- remnants and membranes can be readily extracted from the
omy. The volume drained externally should decrease as the bile CBD. Exploration may extend to the left and right hepatic
flow through the ampulla improves. Persistently elevated ducts, and angled Randall forceps are useful for this purpose.
volume of externally drained bile should raise concerns for Sphincteroplasty consists of suturing the outer edge or both
continued distal obstruction or to the distal T-tube limb lying edges of a surgical sphincterotomy to avoid possible future
within the duodenum. Similarly, there is a problem if there is stenosis of the incision. The stitches achieve hemostasis of the
no external drainage of the bile or if bile drains around the incision margins and help to avoid possible leakage of the duo-
T-tube, and may indicate that the tube is blocked or dislodged denal contents should the excision extend beyond the common
from the duct. Issues with T-tube drainage should be evaluated portion of the sphincter, incurring the risk of retroduodenal
by T-tube cholangiography. If cholangiography does not reveal perforation. Transduodenal sphincteroplasty is contraindicated
any issues, management is aimed at waiting for the bile to flow in the presence of a large CBD (>2 cm) or where there is a long
FIGURE 36A.18. A section of the sphincter of Oddi. Note the distinc-

tion between the papilla, the common portion of the sphincter, and the
sphincters of the common bile duct and duct of Wirsung.
FIGURE 36A.17. Balloon extraction of a single stone (arrow).
suprasphincteric stricture. It also should not be attempted in PYOGENIC CHOLANGITIS (SEE CHAPTER 43). If papillary stenosis or
the presence of a duodenal diverticulum or where there is severe CBD stones or both exist together with cholangitis, transduo-
periampullary inflammation. denal sphincteroplasty can be an excellent procedure for defini-
tive biliary drainage.
Indications
The most common indications for transduodenal sphinctero- CHRONIC PANCREATITIS AND ACUTE GALLSTONE PANCREATITIS (SEE
plasty relate to bile duct stones and cholangitis (see CHAPTERS 55 TO 58). In chronic pancreatitis, some authors report
Chapter 43). good long-term results with transduodenal sphincteroplasty
alone (Hacaim et al, 1994) or in addition to transpapillary
STONES IMPACTED IN THE DISTAL AMPULLARY REGION. An impacted septectomy (Moody et al, 1983) or with other drainage proce-
stone is often readily palpable, and the incision may be made dures of the duct of Wirsung (Kestens et al, 1996). The pres-
safely using the stone as a guide. In such cases, extraction ence of a stone at the lower end of the CBD or pancreatic duct
through a supraduodenal choledochotomy is often impossible may cause biliary pancreatitis, and transduodenal sphinctero-
without undue risk of creating a false passage and without plasty with clearance of the CBD is a treatment option.
significant risk of postoperative pancreatitis.
Technique
MULTIPLE AND RECURRENT COMMON BILE DUCT STONES. In cases of Sphincteroplasty consists of the incision of the common portion
multiple and recurrent CBD stones, sphincteroplasty should of the sphincter of Oddi (Fig. 36A.18) with partial suture of
provide long-term biliary drainage. When 20 or more stones the incision margin. Using this procedure, the sphincters of the
are removed from the CBD, it is probable that one or more CBD and the duct of Wirsung are not involved, and their func-
stones are still present (Stain et al, 1991). In this situation, tions are not impaired. The procedure also can be called a
choledochoduodenostomy or sphincteroplasty yields excellent subtotal lower sphincteroplasty (Fig. 36A.19) (Stefanini et al,
results. 1977). The approach to the sphincter of Oddi is through a
minimal duodenotomy in the second part of the duodenum.
PAPILLARY STENOSIS. Papillary stenosis is encountered less fre-
quently than in the past. When it is found at operation, trans- Preparation, Position of the Patient, and Incision
duodenal sphincteroplasty ensures good biliary drainage and Preoperative preparation is routine. The patient is placed in a
prevents restenosis (Ramirez et al, 1994). ES is technically suc- supine position on a radiotransparent operating table. A trans-
cessful in only 60% to 80% of cases, and the mortality rate verse incision below the right costal margin is preferred (Vogt
exceeds 1% (Seifert et al, 1982). In addition, sphincterotomy & Hermann, 1981). This incision allows optimal light and
for papillary stenosis is five times more likely to lead to reste- excellent access; it is particularly suitable in obese patients, and
nosis than if the same procedure is performed for calculi the incidence of postoperative incisional hernia is probably
(Tzovaras & Rowlands, 1998). lower than with vertical and oblique incisions. The incision of
FIGURE 36A.20. Transverse subcostal incision offers excellent expo-

sure with a low incidence of postoperative incisional hernia.
FIGURE 36A.19. Subtotal lower sphincteroplasty involves only the
papilla, whereas the common bile duct and the duct of Wirsung are
preserved.
the abdominal wall follows a transverse line, from the midaxil-

lary to the median line at the level of the 11th and 12th ribs
(Fig. 36A.20).
Preparation of the Operative Field and Exposure

The abdomen is opened, and a large retractor is positioned at
the upper margin of the wound. The hepatic flexure of the
colon is displaced inferiorly, and the stomach is displaced to
the left by means of two surgical pads. Viscera are maintained
in this position with two large, curved Deaver retractors. For
the performance of the sphincteroplasty, extended mobilization
of the duodenum and pancreas (Kocher maneuver) is manda-
tory (Moody et al, 1983). The assistant surgeon displaces the
second portion medially and forward, and the peritoneum is
incised posteriorly along the curved lateral margin of the duo-
denum. The mesocolon of the right colic flexure is cleared
inferiorly. At this point, the assistant surgeon also should dis-
place the duodenum superiorly (Fig. 36A.21). Access is pro-
vided to the avascular space between the posterior aspect of the
head of the pancreas anteriorly and the perinephric fat and
inferior vena cava posteriorly; elevation of the structures should
reach the left margin of the inferior vena cava. It is important
to expose and mobilize the third portion of the duodenum to
allow easy access to the papilla and for closure of the duode-
notomy without tension (see Fig. 36A.21).
Duodenotomy
Duodenotomy is performed in the lateral duodenal wall by
surgical diathermy. The cut is 10 to 15 mm long immediately FIGURE 36A.21. It is important to expose and mobilize the third
above the inferior knee of the duodenum, with the surgeon portion of the duodenum to easily identify and operate on the papilla
taking account of the fact that the papilla usually is located at and allow facile closure of the duodenotomy.
2
3
15 mm
1
3
FIGURE 36A.23. If the papilla is not easily seen, it can often be pal-
pated as a small elevation in the medial duodenal wall. If the papilla is
FIGURE 36A.22. The duodenotomy is performed above the junction still not easily seen, a small Nélaton catheter is introduced through the
of the second and third portions of the duodenum; the surgeon takes cystic duct stump until it is seen to protrude from the papillary orifice.
into account that the papilla usually is located at the junction of the upper The forefinger, introduced through the duodenotomy, detects the papilla
two thirds and lower third of the second part of the duodenum. as a small, thick elevation.
the junction of the lower third with the upper two thirds of the (Fig. 36A.24) (Partington, 1977). A Nélaton catheter (4 to 5
second portion of the duodenum (Fig. 36A.22). The duodenal Fr) is introduced from the outside or via the cystic duct. Fol-
incision may be longitudinal or transverse; both types are suit- lowing the line of the catheter—and avoiding plastic catheters,
able, provided that the suture of such incisions is always trans- which melt when surgical diathermy is applied—the surgeon
verse. We prefer a longitudinal incision, because if the retractor makes a cut using surgical diathermy. This cut is made superi-
on the duodenum widens the duodenotomy, this occurs longi- orly (at the 11 o’clock position) for 4 to 5 mm (Fig. 36A.25).
tudinally. In the case of a transverse duodenotomy, any inad- We prefer surgical diathermy because the instrument ensures
vertent extension would cause a transverse enlargement of the good hemostasis. When a sample for biopsy is required, it
wound. should be obtained with a scalpel and be taken only from the
outer margin of the incision. Possible bleeding from the cut,
Identification of the Papilla usually modest, can be arrested with a stitch. After sphincter-
After the duodenal incision, the papilla is readily shown on the otomy, two or three stitches are placed between the duodenal
medial duodenal wall in 15% to 20% of patients. It appears as mucosa and the wall of the CBD on the outer margin by using
a roundish elevation with a central orifice. When the papilla is an atraumatic needle and fine sutures. Traction is applied to
not readily visible, it should be detected by displacement and these sutures, and incision of the sphincter is extended for
flattening of the mucosal folds. This should be done with great another 6 to 7 mm with sutures placed every 2 to 3 mm, all
care to avoid tearing of the mucosa, which would hinder good laterally, until the entire common tract of the sphincter of Oddi
exposure. Identification of the papilla under direct vision is has been incised (Fig. 36A.26). The incision is complete when
possible in 80% of patients. If this is not the case, digital palpa- it is 10 to 12 mm long and an appropriate forceps can be easily
tion can be used running the forefinger, introduced through the introduced (Fig. 36A.27). Its entry into the CBD allows an
duodenotomy, across the medial duodenal wall. The papilla is abundant flow of bile owing to distension of its sphincter.
identified as a small elevation. If digital palpation fails, a small Sutures should be placed only on the outer margin of the
(5 to 6 Fr) Nélaton catheter can be introduced via the cystic sphincterotomy to prevent the risk of damage to the duct of
duct stump and advanced downward to emerge at the papilla Wirsung. The opening of the duct of Wirsung usually is identi-
(Fig. 36A.23). This maneuver should never be performed with fied as a small orifice from which clear, colorless pancreatic
rigid catheters because this may result in the formation of false juice flows.
passages.
Sometimes a very small papilla is detected, and its catheter- Instrumental Exploration of the Common Bile Duct
ization is difficult or impossible. In such cases, the orifice is After sphincteroplasty, instrumental exploration of the CBD
probably that of the duct of Santorini. The major papilla should and extraction of stones is performed (Speranza et al, 1982).
be searched for in a lower position. An angled Randall forceps is introduced into the CBD, and the
bile duct is carefully explored. The maneuver should be
Sphincteroplasty repeated several times to extract all stones. The next step is to
After the papilla has been identified, it is exposed by gentle rinse with saline solution, introduced under slight pressure with
extraction with an Allis or similar clamp. This clamp is applied a Nélaton catheter (8 to 9 Fr) and abruptly withdrawn so that
laterally, never medially, to avoid trauma to the duct of Wirsung small fragments flow downstream with the siphoning (Fig.
FIGURE 36A.25. With a Nélaton catheter as a guide, a cut is made

using surgical diathermy on the medial wall of the duodenum extending
superiorly and slightly externally (11 o’clock position). With diathermy,
good hemostasis is achieved.
FIGURE 36A.24. The duodenotomy is kept open by a suitable retrac-
tor placed in the upper margin of the duodenal incision. The papilla is
exposed by gentle traction with an Allis clamp placed laterally, never
medially, to avoid trauma to the duct of Wirsung.
FIGURE 36A.26. After sphincterectomy is performed, several stitches

are placed between the duodenal mucosa and the wall of the common
bile duct (CBD) by using an atraumatic needle and 3-0 suture. Sutures
should be placed only on the outer margin of the sphincterotomy to FIGURE 36A.27. Sphincteroplasty is completed when the incision is
avoid the risk of damaging the duct of Wirsung, which in its distal portion 10 to 21 mm long, and a Randall forceps can be introduced easily into
runs inferiorly and medially along the length of the CBD. the common bile duct to extract stones or other foreign bodies.
36A.28). Other means to extract stones from the CBD are with
a Fogarty catheter and Dormia basket. The problem of residual
stones is best prevented with choledochoscopy; the endoscope
is introduced via the sphincteroplasty.
Duodenal Closure
As already emphasized, initial longitudinal duodenotomy
always should be closed transversely to avoid stenosis of the
duodenum. First, the superior and inferior margins of the inci-
sions are approximated, and the resulting gaps are sutured with
two extramucosal, nonabsorbable purse-string sutures. Three
or four nonabsorbable seromuscular sutures are added (Fig.
36A.29). The suture should not be under tension, and for this
reason, preliminary extended mobilization of the duodenum
and pancreas is mandatory. The operation is now complete,
and the wound is closed without abdominal drains.
Comment
It is not necessary to perform transduodenal sphincteroplasty
combined with supraduodenal choledochotomy, which has a
higher associated mortality rate (Scheridan et al, 1987). The
cystic duct remnant may be used to introduce a Nélaton cath-
eter to assist recognition of the papilla. There is no need to
insert a T-tube, which lengthens the hospital stay and may
predispose the patient to stenosis and infection of the CBD
(Ratych et al, 1991; Scheridan et al, 1987).
Review of Reported Results

In a retrospective analysis of 25,541 transduodenal sphinctero-
plasties performed by 130 surgeons in different countries, early
transduodenal sphincteroplasty–related complications were
bleeding (0.65%), acute pancreatitis (0.60%), dehiscence of 50 mL
the duodenal closure (0.55%), and cholangitis (0.50%), for an FIGURE 36A.28. After the extraction of stones, the common bile duct
overall morbidity rate of 2.3% and a mortality rate of 0.8% is rinsed with saline solution introduced under slight pressure with a
(Negro et al, 1984). A retrospective study (Sellner et al, 1988) Nélaton catheter and with subsequent siphoning so that small fragments
found that the factors affecting mortality in 2.1% of 333 patients can run downstream.
A B C
FIGURE 36A.29. The duodenotomy always should be closed transversely to avoid stenosis of the duodenum. A, The superior and inferior angles
are approximated. B, The resulting lateral gaps are sutured with two extramucosal nonabsorbable purse-string sutures. C, Three to four nonabsorb-
able seromuscular stitches are added as a second layer. The sutures should not be under tension.
(but only in 0.9% for sphincterectomy-related complications) with transampullary septectomy (Kelly & Rowlands, 1996)
were age older than 70 years and a bilirubin level greater has led to good long-term results in patients with chronic and
than 85 mmol/L, diabetes, renal failure, and coagulopathy. acute recurrent pancreatitis, even in cases with pancreas divisum
The mortality rate increased when supraduodenal choledo- and in selected patients with abdominal pain of hepatobiliary
chotomy was combined with transduodenal sphincteroplasty origin.
and when a T-tube was used (Sellner et al, 1988). Transduo-
denal sphincteroplasty alone (Hacaim et al, 1994) or associated References are available at expertconsult.com.
A. Gallstones and Gallbladder Chapter 36A Stones in the bile duct: clinical features and open surgical approaches and techniques 603.e1
REFERENCES Demartines N, et al: Evaluation of magnetic resonance cholangiogra-

phy in the management of bile duct stones, Arch Surg 135:148–152,
Allen B, et al: Management of recurrent and residual common duct 2000.
stones, Am J Surg 142:41–47, 1981. DenBesten L, Berci G: The current status of biliary tract surgery: an
Alponat A, et al: Predictive factors for synchronous common bile duct international study of 1072 consecutive patients, World J Surg 10:
stones in patients with choledocholithiasis, Surg Endosc 11:928–932, 116–122, 1986.
1997. Dong ZT, et al: Primary closure after laparoscopic common bile duct
Amouyal P, et al: Endosonography: promising method for diagnosis of exploration versus T-tube, J Surg Res 189:249–254, 2014.
extrahepatic cholestasis, Lancet 2:1195–1198, 1989. Doyle PJ, et al: The value of routine preoperative cholangiography: a
Amouyal P, et al: Diagnosis of choledocholithiasis by endoscopic ultra- report of 4000 cholecystectomies, Br J Surg 69:617–619, 1982.
sonography, Gastroenterology 106:1062–1067, 1994. Edmundowicz SA, et al: Preliminary experience using endoscopic
Anderson SW, et al: Detection of biliary duct narrowing and choledo- ultrasonography in the diagnosis of choledocholithiasis, Endoscopy
cholithiasis: accuracy of portal venous phase multidetector CT, 24:774–778, 1992.
Radiology 247:418–427, 2008. El-Geidie AA: Is the use of T-tube necessary after laparoscopic cho-
Andriulli A, et al: Incidence rates of post-ERCP complications: a sys- ledochotomy? J Gastrointest Surg 14:844–848, 2010.
tematic survey of prospective studies, Am J Gastroenterol 102:1781– Escourrou J, et al: Early and late complications after endoscopic
1788, 2007. sphincterotomy for biliary lithiasis with and without the gallbladder
ASGE Standards of Practice Committee, et al: The role of endoscopy in situ, Gut 25:598–602, 1984.
in the evaluation of suspected choledocholithiasis, Gastrointest Endosc Fogarty FJ, et al: Evaluation of an improved operating technique in
71:1–9, 2010. common duct surgery, Am J Surg 116:117–182, 1968.
Behrns KE, et al: Early ERCP for for gallstone pancreatitis: for whom Fogli L, et al: Laparoscopic cholecystectomy without intraoperative
and when? J Gastrointest Surg 12:629–633, 2008. cholangiography: audit of long-term results, J Laparoendosc Adv Surg
Berci G: Choledochoscopy. In Blumgart LH, Fong Y, editors: Surgery Tech A 19:191–193, 2009.
of the liver and biliary tract, ed 3, Philadelphia, 2000, Saunders, Fox J, Gunn AA: Common bile duct exploration by balloon catheter,
p 491. J R Coll Surg Edinb 29:81–84, 1984.
Bergdahl L, Holmlund DEW: Retained bile duct stones, Acta Chir Ganey JB, et al: Cholecystectomy: clinical experience with a large
Scand 142:145–149, 1976. series, Am J Surg 151:352–357, 1986.
Bingener J, Schwesinger WH: Management of common bile duct Garrow D, et al: Endoscopic ultrasound: a meta-analysis of test per-
stones in a rural area of the United States, Surg Endosc 20:577–579, formance in suspected biliary obstruction, Clin Gastroenterol Hepatol
2006. 5:616–623, 2007.
Braasch JW, et al: Refractory primary common bile duct stone disease, Gibson RN, et al: Accuracy of computed tomographic intravenous
Am J Surg 139:526–530, 1980. cholangiography (CT-IVC) with iotroxate in the detection of cho-
Burhenne HJ: Percutaneous extraction of retained biliary tract stones: ledocholithiasis, Eur Radiol 15:1634–1642, 2005.
661 patients, AJR Am J Roentgenol 1334:889–898, 1980. Girard RM: Stones in the common bile duct: surgical approaches. In
Cabada Giadàs T, et al: Helical CT cholangiography in the evaluation Blumgart LH, Fong Y, editors: Surgery of the liver and biliary tract,
of the biliary tract: application to the diagnosis of choledocholithia- ed 3, Philadelphia, 2000, Saunders, p 737.
sis, Abdom Imaging 27:61–70, 2002. Girard RM, Legros G: Retained and recurrent bile duct stones: surgical
Caddy GR, Tham TC: Gallstone disease: symptoms, diagnosis and or nonsurgical removal? Ann Surg 199:21–27, 1981.
endoscopic management of common bile duct stones, Best Pract Res Gonzalez JJ, et al: Biliary lithiasis in the elderly patient: morbidity and
Clin Gastroenterol 20:1085–1101, 2006. mortality due to biliary surgery, Hepatogastroenterology 44:1565–
Cameron JL: Retained and recurrent bile duct stones: operative man- 1568, 1997.
agement, Am J Surg 158:218–221, 1989. Gunn AA: Cholecystectomy and exploration of the common duct. In
Chopra S, et al: Helical CT cholangiography with oral cholecysto- Dudley HA, et al, editors: Rob and Smith operative surgery, ed 4,
graphic contrast material, Radiology 214:596–601, 2000. London, 1983, Butterworth, p 616.
Christoforidis E, et al: Endoscopic management of retained bile stones Gurusamy KS, et al: T-tube drainage versus primary closure after open
with an indwelling T-tube, Surg Endosc 18:1582–1586, 2004. common bile duct exploration, Cochrane Database Syst Rev
Coelho JCU, et al: Incidence of common bile duct stones in patients (6):CD005640, 2013.
with acute and chronic cholecystitis, Surg Gynecol Obstet 158:76–80, Gurusamy KS, et al: T-tube drainage versus primary closure after lapa-
1984. roscopic common bile duct exploration, Cochrane Database Syst Rev
Collins C, et al: A prospective study of common bile duct calculi in (6):CD005641, 2013.
patients undergoing laparoscopic cholecystectomy: natural history of Hacaim AG, et al: Long-term results of surgical management of
choledocholithiasis revisited, Ann Surg 239:28–33, 2004. chronic pancreatitis, Am J Surg 60:306–308, 1994.
Cotton PB: Endoscopic management of bile duct stones (apples and Hallal AH, et al: Magnetic resonance cholangiopancreatography accu-
oranges), Gut 25:587–597, 1984. rately detects common bile duct stones in resolving gallstone pan-
Cotton PB: Endoscopic retrograde cholangiopancreatography and creatitis, J Am Coll Surg 200:869–875, 2005.
laparoscopic cholecystectomy, Am J Surg 165:474–478, 1993. Hammarström LE, et al: Long-term follow up of a prospective ran-
Cuschieri A, et al: EAES ductal stone study: preliminary findings of domized study of endoscopic versus surgical treatment of bile duct
multi-center prospective randomized trial comparing two-stage vs calculi in patients with gallbladder in situ, Br J Surg 82:1516–1521,
single-stage management, Surg Endosc 10:1130–1135, 1996. 1995.
Cuschieri A, et al: EAES multicenter prospective randomized trial Hammarström LE, et al: Long-term follow up after endoscopic treat-
comparing two-stage vs single-stage management of patients with ment of bile duct calculi in cholecystectomized patients, World J Surg
gallstone disease and ductal calculi, Surg Endosc 13:952–957, 1999. 20:272–276, 1996.
Dasari BV, et al: Surgical versus endoscopic treatment of bile duct Hampson LG, et al: Common bile duct exploration: indications and
stones, Cochrane Database Syst Rev (3):CD003327, 2013. results, Can J Surg 24:455–457, 1981.
Davis WZ, et al: ERCP and sphincterotomy in the context of laparo- Heinerman PM, et al: Selective ERCP and preoperative stone removal
scopic cholecystectomy: academic and community practice patterns in bile duct surgery, Ann Surg 209:267–272, 1989.
and results, Am J Gastroenterol 92:597–601, 1997. Iranmanesh P, et al: Initial cholecystectomy vs sequential common duct
Dayton MT, et al: Incidence of complications with operative choledo- endoscopic assessment and subsequent cholecystectomy for sus-
choscopy, Am J Surg 147:139–145, 1984. pected gallstone migration: a randomized clinical trial, JAMA
de Almeida AM, et al: Definitive decompression of the biliary tree- 312:137–144, 2014.
preferred approach to management of choledocholithiasis and/or Johnson AG, Harding Rains AJ: Prevention and treatment of recurrent
associated pathology, Am J Gastroenterol 77(12):941–946, 1982. bile duct stones by choledochoduodenostomy, World J Surg 2:487–
Delorio JR, et al: Acute biliary pancreatitis: the roles of laparoscopic 496, 1978.
cholecystectomy and endoscopic retrograde cholangiopancreatogra- Johnson AG, Hosking SW: Appraisal of the management of bile duct
phy, Surg Endosc 9:392–396, 1995. stones, Br J Surg 74:555–560, 1987.
Jones SA: The prevention and treatment of recurrent bile duct stones Peng WK, et al: Role of liver function tests in predicting common bile
by transduodenal sphincteroplasty, World J Surg 2:473–485, 1978. duct stones in acute calculous cholecystitis, Br J Surg 92:1241–1247,
Kappes SK, et al: Intraoperative endoscopy: mandatory for all common 2005.
duct operations? Arch Surg 117:603–607, 1982. Petrov MS, Savides TJ: Systemic review of endoscopic ultrasonography
Kelly SB, Rowlands BJ: Transduodenal sphincteroplasty and transam- versus endoscopic retrograde cholangiopancreatography for sus-
pullary septectomy for papillary stenosis, HPB Surg 9:199–207, pected choledocholithiasis, Br J Surg 96:967–974, 2009.
1996. Petrov MS, et al: Early endoscopic retrograde cholangiopancreato
Kestens PJ, et al: Traitement chirurgical de la pancréatite chronique graphy versus conservative management in acute biliary pancreatitis:
avec atteinte céphalique predominante par double derivation wir- a meta-analysis of randomized trials, Ann Surg 247:250–257,
sungienne et repermeabilization canaliere céphalique, Ann Chir 50: 2008.
853–864, 1996. Polkowski M, et al: Helical computed tomographic cholangiography
Kim MH, et al: Primary intrahepatic stones, Am J Gastroenterol versus endosonography for suspected bile duct stones: a prospec-
90:540–548, 1995. tive blinded study in non-jaundiced patients, Gut 45:744–749,
Kim HJ, et al: Multidetector computed tomography cholangiography 1999.
with multiplantar reformation for the assessment of patients with Ponsky JL: Endoscopic management of common bile duct stones, World
biliary obstruction, J Gastroenterol Hepatol 22:400–405, 2007. J Surg 16:1060–1065, 1992.
Kondo S, et al: Detection of common bile duct stones: comparison Ramirez P, et al: Choledochoduodenostomy and sphincterotomy in the
between endoscopic ultrasonography, magnetic resonance cholangi- treatment of choledocholithiasis, Br J Surg 81:121–123, 1994.
ography, and helical-computed-tomographic cholangiography, Eur J Ratych RE, et al: Transduodenal exploration of the common bile duct
Radiol 54:271–275, 2005. in patients with nondilated ducts, Surg Gynecol Obstet 173:49–53,
Koo KP, Traverso LW: Do predictive indicators predict the presence 1991.
of common bile duct stones during laparoscopic cholecystectomy? Rogers SJ, et al: Prospective randomized trial of LC+LCBDE vs
Am J Surg 171:495–499, 1996. ERCP/S+LC for common bile duct stone disease, Arch Surg 145:28–
Lambert ME, et al: Endoscopic removal of retained stones after biliary 33, 2010.
surgery, Br J Surg 75:896–898, 1988. Roslyn JL: Calculous biliary disease. In Greenfield LJ, editor: Surgery:
Lambert ME, et al: Endoscopic sphincterotomy: the whole truth, Br J scientific principles and practice, Philadelphia, 1993, Lippincott, pp
Surg 78:473–476, 1991. 936–953.
Lehman GA, Sherman S: Diagnosis and therapy of pancreas divisum, Saharia PC, et al: Primary common duct stones, Ann Surg 1985:598–
Gastrointest Endosc Clin N Am 8:55–77, 1998. 602, 1977.
Leung JW: Does the addition of endoscopic sphincterotomy to stent Scheridan WG, et al: Morbidity and mortality of common bile duct
insertion improve drainage of the bile duct in acute suppurative exploration, Br J Surg 71:1095–1099, 1987.
cholangitis? Gastrointest Endosc 58:570–572, 2003. Seifert E, et al: Lanseitresulate nach endoskipischer Sphinkterotomie,
Liu TH, et al: Patient evaluation and management with selective use Dtsch Med Wochenschr 107:610–614, 1982.
of magnetic resonance cholangiography and endoscopic retrograde Sellner FJ, et al: Factors affecting mortality in transduodenal sphinc-
cholangio-pancreatography before laparoscopic cholecystectomy, teroplasty, Surg Gynecol Obstet 167:23–27, 1988.
Ann Surg 1:33–40, 2001. Sgourakis G, et al: Predictors of common bile duct lithiasis in laparo-
Lygidakis NJ: A prospective randomized study of recurrent choledo- scopic era, World J Gastroenterol 11:3267–3272, 2005.
cholithiasis, Surg Gynecol Obstet 155:679–684, 1982. Shanmugam V, et al: Is magnetic resonance cholangiopancreatography
Lygidakis NJ: Incidence and significance of primary stones of the the new gold standard in biliary imaging? Br J Radiol 78:888–893,
common bile duct in choledocholithiasis, Surg Gynecol Obstet 2005.
157:434–436, 1983. Sivak MV: Endoscopic management of bile duct stones, Am J Surg
Maniatis P, et al: Virtual CT cholangiography in patients with choledo- 158:228–240, 1989.
cholithiasis, Abdom Imaging 28:536–544, 2003. Soto JA, et al: Choledocholithiasis: diagnosis with oralcontrast-
Martin DJ, et al: Surgical versus endoscopic treatment of bile duct enhanced CT cholangiography, AJR Am J Roentgenol 172:943–948,
stones, Cochrane Database Syst Rev (2):CD003327, 2006. 1999.
Mazzariello RM: A fourteen-year experience with nonoperative instru- Soto JA, et al: Diagnosing bile duct stones: comparison of unenhanced
ment extraction of retained bile duct stones, World J Surg 2:447–455, helical CT, oral contrast-enhanced CT cholangiography, and MR
1978. cholangiography, AJR Am J Roentgenol 175:1127–1134, 2000.
McSherry CK, Glenn F: The incidence and causes of death following Speranza V, et al: Transduodenal papillostomy as a routine procedure
surgery for nonmalignant biliary tract disease, Ann Surg 191:271– in managing choledocholithiasis, Arch Surg 117:875–878, 1982.
275, 1980. Stain SC, et al: Choledocholithiasis: endoscopic sphincterotomy or
Miller BM, et al: Surgical versus endoscopic management of common common bile duct exploration? Ann Surg 213:627–633, 1991.
bile duct stones, Ann Surg 207:135–141, 1988. Stefanini P, et al: Long-term results of papillostomy. In Delmont J,
Moody FG, et al: Transduodenal sphincteroplasty and transampullary editor: The sphincter of oddi, Basel, 1977, Karger, pp 206–212.
septectomy for postcholecystectomy pain, Ann Surg 197:627–636, Sugiyama M, Atomi Y: Endoscopic ultrasonography for diagnosing
1983. choledocholithiasis: a prospective comparative study with ultraso-
Negro P, et al: Le risque opératoire de la sphinctérotomie oddienne: nography and computed tomography, Gastrointest Endosc 45:143–
resultats d’une enquête internationale (25, 541 cas), J Chir 121:133– 146, 1997.
139, 1984. Targarona EM, et al: Randomised trial of endoscopic sphincterotomy
Neoptolemos JP, et al: Prospective randomized study of preoperative with gallbladder left in situ versus open surgery for common bile
endoscopic sphincterotomy versus surgery alone for common bile duct calculi in high-risk patients, Lancet 347:926–929, 1996.
duct stones, Br Med J 294:470–474, 1987. Taylor AC, et al: Prospective assessment of magnetic resonance chol-
Neoptolemos JP, et al: Controlled trial of urgent endoscopic retrograde angiopancreatography for noninvasive imaging of the biliary tree,
cholangiopancreatography and endoscopic sphincterotomy versus Gastrointest Endosc 55:17–22, 2002.
conservative treatment for acute pancreatitis due to gallstones, Tompkins PK, Pitt HA: Surgical management of benign lesions of the
Lancet 2:979–983, 1988. bile ducts, Curr Probl Surg 19:327–398, 1982.
Nora PF, et al: Operative choledochoscopy: results of a prospective Topal B, et al: The value of magnetic resonance cholangiopancreatog-
study in several institutions, Am J Surg 133:105–110, 1977. raphy in predicting common bile duct stones in patients with gall-
O’Doherty DP, et al: Endoscopic sphincterotomy for retained common stone disease, Br J Surg 90:42–47, 2003.
bile duct stones in patients with t-tube in situ in the early postopera- Trondsen E, et al: Prediction of common bile duct stones prior to
tive period, Br J Surg 73:454–456, 1986. cholecystectomy. A prospective validation of a discriminant analysis
Orloff M: Importance of surgical technique in prevention of retained function, Arch Surg 133:162–166, 1998.
and recurrent bile duct stones, World J Surg 2:403–410, 1978. Tse F, et al: EUS: a meta-analysis of test performance in suspected
Partington PF: Twenty-three years of experience with sphincterotomy choledocholithiasis, Gastrointest Endosc 67:235–244, 2008.
and sphincteroplasty for stenosis of the sphincter of Oddi, Surg Tzovaras G, Rowlands BJ: Diagnosis and treatment of sphincter of
Gynecol Obstet 145:161–168, 1977. Oddi dysfunction, Br J Surg 85:588–595, 1998.
A. Gallstones and Gallbladder Chapter 36A Stones in the bile duct: clinical features and open surgical approaches and techniques 603.e3
Vargghese JC, et al: Diagnostic accuracy of magnetic resonance Wallner BK, et al: Dilated biliary tract: evaluation with MR cholangi-
cholangio-pancreatography and ultrasound compared with direct ography with a T2-weighted contrast-enhanced fast sequence, Radi-
cholangiography in the detection of choledocholithiasis, Clin Radiol ology 181:805–808, 1991.
55:25–35, 2000. Way LW: Retained common duct stones, Surg Clin North Am 53:1139–
Verma D, et al: EUS vs MRCP for detection of choledocholithiasis, 1147, 1973.
Gastrointest Endosc 64:248–254, 2006. Zandrino F, et al: MR versus multislice CT cholangiography in evalu-
Vogt DP, Hermann RE: Choledochoduodenostomy, choledochojeju- ating patients with obstruction of the biliary tract, Abdom Imaging
nostomy or sphincteroplasty for biliary and pancreatic disease, Ann 30:77–85, 2005.
Surg 193:161–168, 1981.
CHAPTER 36B
Stones in the bile duct: minimally invasive
surgical approaches
Zeljka Jutric, Chet W. Hammill, and Paul D. Hansen
OVERVIEW used and the sequence in which they are used depend on the
specific clinical scenario. In addition, the capability and experi-
Choledocholithiasis is identified in 10% to 18% of patients ence of the available personnel will affect the treatment algo-
undergoing cholecystectomy, either preoperatively, intraopera- rithm. The most common clinical scenarios encountered by
tively, or postoperatively (Dasari et al, 2013) (see Chapters 32 surgeons include known or suspected stones prior to cholecys-
and 37). The majority of cases of choledocholithiasis are sec- tectomy, the diagnosis of stones intraoperatively, and stones
ondary, the result of stones that originally formed in the gall- identified subsequent to cholecystectomy.
bladder. Primary choledocholithiasis, 10% to 15% of cases
(with a higher incidence in Asian countries), results from the CLINICAL SCENARIOS
formation of stones within the common bile duct (CBD) (Ko
& Lee, 2002). Choledocholithiasis can lead to a variety of clini- Preoperative Choledocholithiasis
cal manifestations, including biliary colic, obstructive jaundice, Ascending cholangitis (see Chapter 43), gallstone pancreatitis
ascending cholangitis (see Chapter 43), and pancreatitis (see (see Chapters 54 and 55), symptomatic cholelithiasis with evi-
Chapters 54 and 55). These clinical scenarios can vary signifi- dence of choledocholithiasis on imaging or serum liver function
cantly in presentation and morbidity, ranging from minimal to tests, are all potential indicators that CBD stones are present.
critical illness and death. If left untreated, chronic choledocho- For CBD stones suspected at the time of initial clinical presen-
lithiasis can also cause inflammatory strictures, recurrent infec- tation, the decision making is often based on surgeon prefer-
tions, or cirrhosis. ence and institutional capabilities. When CBD stones are highly
The management of choledocholithiasis has changed radi- likely or established with noninvasive imaging, the two primary
cally over the last several decades. Historically, stones in the therapeutic strategies are ERCP followed by laparoscopic cho-
CBD were removed at the time of open surgical exploration. lecystectomy or laparoscopic cholecystectomy with intraopera-
Even after the introduction of endoscopic retrograde cholangio- tive cholangiogram. In the latter case, laparoscopic CBD
pancreatography (ERCP) in the 1970s, laparotomy remained exploration or postoperative ERCP are performed if CBD
the mainstay for CBD exploration (Kroger et al, 1975). At that stones are identified on the cholangiogram.
time, the tools and techniques used for surgical clearance of the In general, ERCP, which is available in essentially all urban
CBD were superior to those available via ERCP, and as long settings and in a growing number of smaller communities, is
as a laparotomy was used to perform the cholecystectomy, more widely available than surgeons skilled at laparoscopic
minimally invasive treatment of choledocholithiasis was unnec- CBD exploration (see Chapters 20 and 29). ERCP is also
essary. It was not until the introduction of the laparoscopic highly effective, with successful clearance of the CBD in more
cholecystectomy in the late 1980s that the role for an associated than 95% of cases (Kum & Goh, 1996). In most centers, CBD
less invasive method of treating choledocholithiasis became a stones up to 1.5 cm can be extracted, and centers that use
priority. lithotripsy can extract stones as large as 3 cm in diameter
During the initial adoption of laparoscopic cholecystectomy, (Lenze et al, 2014). Therefore, in settings where laparoscopic
most general surgeons did not have the skill set, experience, or CBD exploration is not available, the debate centers on preop-
equipment to facilitate a laparoscopic CBD exploration. As the erative versus postoperative ERCP. The central argument
skills, experience, and tools have developed, the advantages and against an ERCP-first strategy is that 80% of known or sus-
disadvantages of laparoscopic CBD exploration versus ERCP pected CBD stones will have passed prior to the intervention,
have been a frequent source of debate. The objective of this making the risks and cost of the intervention unnecessary
chapter is to discuss minimally invasive surgical techniques for (Urbach et al, 2001). The risks of ERCP include acute pancre-
managing choledocholithiasis, including indications and tech- atitis (3% to 7%), bleeding (0.3% to 1.4%), ascending infection
nical aspects of laparoscopic transcystic, transcholedochal, and (1.4%), and perforation (0.6%). The mortality associated with
transduodenal CBD exploration, as well as laparoscopic biliary- ERCP is 0.2% to 0.9% (Orenstein et al, 2014). Proponents of
enteric bypass procedures and laparoscopic-assisted ERCP. performing preoperative ERCP argue that, if ERCP is unsuc-
cessful, a CBD exploration can be performed at the time of
INDICATIONS cholecystectomy without requiring a third procedure.
In most patients, our preferred strategy is laparoscopic cho-
The standard of care for the management of most CBD stones lecystectomy with selective intraoperative cholangiogram (see
is minimally invasive: using either laparoscopic, endoscopic, or Chapter 23). Laparoscopic CBD exploration or ERCP is per-
percutaneous techniques. The minimally invasive techniques formed when CBD stones are identified. However, in patients
604
A. Gallstones and Gallbladder Chapter 36B Stones in the bile duct: minimally invasive surgical approaches 605
with cholangitis, ongoing pancreatitis, or persistent hyperbili- unsuccessful or cannot be performed due to altered anatomy.
rubinemia (>4 U/dL), we recommend preoperative ERCP ERCP fails to cannulate the CBD in 1% to 2% of cases. In this
(Chan et al, 2008). This allows for both clearance of the bile situation, laparoscopic CBD exploration is necessary. Similarly,
duct and assessment of additional or alternative pathologies, if stone extraction fails due to the size of the stones, laparo-
such as impacted stones, strictures, or malignancy. In the case scopic exploration is a reasonable next step. In the case of
of ascending cholangitis, broad-spectrum antibiotics and urgent impacted or large stones that cannot be removed endoscopically
biliary decompression with ERCP is typically the best option, or surgically, biliary-enteric bypass should be considered (see
as it is the least invasive and both diagnostic and therapeutic Chapter 36A).
(Hui et al, 2001). If ERCP is unavailable or not possible (e.g., Laparoscopic CBD exploration or laparoscopic-assisted
history of Roux-en-Y gastric bypass), percutaneous transhe- ERCP may be the only options in patients with a history of
patic biliary decompression should be considered (see Chapter gastrointestinal surgery. With the introduction and rapid expan-
30). Surgical options, laparoscopic or open, are indicated when sion of Roux-en-Y gastric bypass procedures, it is becoming
less invasive methods are not immediately available and the increasingly common to have patients present in whom an
patient’s condition warrants immediate biliary decompression ERCP is either challenging or not possible.
(see Chapter 36A). In both cholangitis and pancreatitis, once
the patient has sufficiently recovered, cholecystectomy should
be completed during the same hospitalization as early recur- TECHNIQUES
rence of symptoms is common, and can lead to significant
morbidity (Chang et al, 2000; Li et al, 2010).
Laparoscopic Transcystic Common Bile
Endoscopic ultrasound (see Chapter 16) and magnetic reso- Duct Exploration
nance cholangiopancreatography (MRCP) (see Chapter 19) The primary method for performing laparoscopic CBD explo-
are additional diagnostic tools that may be helpful in these situ- ration is through the cystic duct. Transcystic exploration is
ations. Endoscopic ultrasound is highly sensitive for CBD typically performed using a standard laparoscopic cholecystec-
stones, without the risk of pancreatitis, allowing ERCP to be tomy trocar configuration. An additional trocar may be placed
used selectively (Artifon et al, 2009). MRCP is a relatively in the midclavicular line to facilitate access to the cystic duct
accurate and noninvasive test that can identify the presence or (Fig. 36B.1A). The gallbladder is left in situ to provide liver
absence of choledocholithiasis, as well as provide anatomic retraction and countertraction on the cystic duct, allowing
delineation prior to surgery. Some centers advocate MRCP to
avoid the risk of unnecessary ERCP; however, this may be
associated with an increase in cost.
Intraoperative Choledocholithiasis (See Chapters 31
and 36A)
When choledocholithiasis is diagnosed on intraoperative chol- Cholangio
angiogram, and the surgeon’s experience and diameter of the scope port
cystic duct are favorable, our preference is to proceed with 5
5
transcystic CBD exploration. If clearance cannot be achieved 5
via the transcystic route, the options are to proceed with tran- 10 mm
5
scholedochal CBD exploration versus a postoperative ERCP.
Intraoperative ERCP, although uncommonly used, is also avail-
able at some institutions (Moreels, 2014; Schreiner et al, 2012).
Our preference is to proceed with postoperative ERCP if the
stones are less than 2 cm and do not appear to be impacted.
ERCP avoids complications associated with a transcholedochal
A
exploration, including biliary stricture and bile leak (Verbesey
& Birkett, 2008). When the probability of endoscopic clearance
is questionable or low, we proceed with a transcholedochal
exploration.
A number of retrospective studies have shown that laparo-
scopic cholecystectomy with CBD exploration as a single oper-
ative procedure is more cost-effective and results in shorter
hospital length of stay than laparoscopic cholecystectomy and
ERCP (Cuschieri et al, 1999; Nathanson et al, 2005; Rogers 5 5
et al, 2010). A recent prospective randomized controlled trial If needed
5
found that both options were highly effective and equivalent in
10 10
overall cost, although the hospital length of stay was lower for
laparoscopic CBD exploration (Rogers et al, 2010).
Postoperative Choledocholithiasis B
Choledocholithiasis identified after a cholecystectomy or when FIGURE 36B.1. A, Trochar placement for transcytic and transchole-
there is not a plan to perform a cholecystectomy is typically dochal common bile duct (CBD) exploration. B, Trochar placement for
managed with ERCP; however, occasionally endoscopy is transduodenal CBD exploration and biliary-enteric bypass.
easier passage of wires, catheters, and the cholangioscope. The removed, and the duct is again vigorously irrigated. Finally, a
first and easiest step in facilitating passage of stones through cholangiogram is performed to assess for residual stones.
the ampulla is pharmacologic relaxation of the sphincter of If ampullary dilation is unsuccessful, choledochoscopy is the
Oddi by using glucagon (Petelin, 1993). After giving 1 mg of next step. Some surgeons prefer to proceed straight to choledo-
intravenous glucagon and waiting 3 minutes, the duct is vigor- choscopy without attempting ampullary dilation because it
ously irrigated with saline or contrast in an attempt to flush the allows direct visualization of the bile duct and stones. A chol-
stones through the ampulla (Ponce et al, 1989). This technique angioscope or ureteroscope is passed transcystically into the
is most successful for sludge and stones less than 4 mm (Kroh common duct (Fig. 36B.2B). If the cystic duct is too small, the
& Chand, 2008). same balloon used for ampullary dilation can be used to gently
If flushing is unsuccessful, two additional techniques are dilate the cystic duct until it is large enough to allow the chol-
used: dilation of the ampulla and choledochoscopy. Ampullary angioscope to pass. Once the stones are identified, they can be
dilation is performed by passing a wire through the cystic duct gently pushed through the ampulla with the tip of the cholan-
and into the duodenum under fluoroscopic guidance. A 4- to gioscope, or they can be snared. To snare the stones, we recom-
6-mm diameter, 4-cm long ureteral balloon is advanced over mend a helical stone retrieval basket, which is passed through
the wire and positioned across the ampulla. After fluoroscopic the working channel of the cholangioscope. Under direct visu-
confirmation of the balloon’s position, it is inflated, dilating the alization, the basket is passed beyond the stone. Once the
ampulla (Fig. 36B.2A) Following dilation, the balloon is basket is distal to the stone, it is opened and withdrawn until
the stone is captured; the snare is then closed, and the stone is
removed through the cystic duct. Care is taken to avoid causing
injury with the snare, as it is a stiff instrument, and ductal
perforation or penetration into the pancreas is possible. Some
authors recommend blind or fluoroscopically guided wire
basket stone capture and extraction; in our experience, this has
proven less successful and has a higher risk of injury. There are
two options to manage large stones that cannot be extracted
through the cystic duct. Large stones that are soft, can often
be crushed by tightening the snare around them, and the debris
can then be snared individually or pushed through the ampulla
into the duodenum. For large stones that are too hard to crush,
a laser or mechanical lithotripsy catheter can be passed through
the cholangioscope to fracture the stone under direct visualiza-
tion. Finally, in bile ducts with significant inflammation or
strictures, a biopsy should be considered. Malignancy may be
an inciting event in the development of ductal debris or ductal
obstruction. Complications of transcystic CBD exploration, as
observed by Paganini and colleagues (2007), include bile leak
(1%), acute pancreatitis (0.5%), and rupture of the cystic duct
(6.8%).
The success rate of a laparoscopic transcystic approach to
choledocholithiasis by experienced surgical teams is 80% to
90% (Kroh & Chand, 2008). There are a number of reasons a
transcystic CBD exploration can fail. In some cases, the cystic
duct is too small, tortuous, or even obliterated, and passage of
a catheter into the CBD is difficult or impossible. Occasionally,
the cystic duct inserts into the very distal common duct or at
an acute angle, making access to the proximal common duct
challenging. Large stones, impacted stones, and significant
inflammation can all increase the difficulty of transcystic bile
duct clearance. In such cases, the surgeon must decide whether
to proceed with transcholedochal exploration or defer to post-
operative ERCP. If the CBD is less than 1 cm in diameter, and
in the absence of large impacted stones, our preference is to
proceed with postoperative ERCP due to the high rate of
success and less potential morbidity.
Laparoscopic Transcholedochal Common Bile
Duct Exploration
Laparoscopic choledochotomy is indicated when the transcys-
tic approach fails or in the presence of multiple, large, or
FIGURE 36B.2. A, Transcystic balloon dilation of the sphincter of impacted stones. Choledochotomy should be avoided if the
Oddi. B, Transcytic examination of the bile duct with the CBD is less than 1 cm in diameter (Verbesey & Birkett, 2008).
choledochoscope. The CBD is accessed by exposing the second portion of the
duodenum and mobilizing the duodenum down to the level of Our preference is to close the ductotomy with simple inter-
the pancreas. Excessive mobilization of the CBD should be rupted sutures, spaced evenly to avoid duct ischemia. We gener-
avoided to preserve blood supply to the proximal duct. Stay ally use a 5-0 monofilament, slowly absorbable suture. If a
sutures may be placed on either side of the choledochotomy T-tube is used, the ductotomy is closed around the base of the
(preserving the 3 and 9 o’clock blood supply) to allow for trac- tube by using a few sutures adjacent to the side of the T-tube,
tion on the CBD. A longitudinal choledochotomy is then made and the tube is exteriorized through a lateral trocar site. A final
on the distal CBD; the choledochotomy should be long enough cholangiogram through the T-tube is performed at the conclu-
for removal of the largest stone and passage of the choledocho- sion of the case. Tubes are left in place for 3 weeks to promote
scope. Typically, a 1-cm incision is sufficient to meet both of formation of an inflammatory tract around the tube. The
these goals. T-tube can then be gently removed, and the tract will collapse
A choledochoscope is passed directly into the bile duct, and and seal spontaneously. If stones are discovered postopera-
a thorough exploration of the proximal and distal duct is per- tively, percutaneous stone extraction via the T-tube is success-
formed (Fig. 36B.3). Stones may be flushed out of the duc- ful in 95% of cases (Burhenne, 1980). Surgical drains are not
totomy, removed directly with atraumatic graspers, or extracted placed routinely unless there is an increased risk of bile leak or
using a snare passed through the cholangioscope as described when a T-tube is used. Once the patient is tolerating a diet and
in the transcystic approach. There is a mechanical advantage there is no evidence of obstruction, we clamp the T tube. The
to removing impacted stones directly through a choledochot- surgical drain is removed if there is no evidence of a bile leak
omy, although stones impacted in the head of the pancreas may after the T-tube has been clamped.
require a transduodenal approach. The overall success rate of laparoscopic choledochotomy is
After clearance of the duct and completion of the explora- 83% to 96%, with a morbidity rate of 5% to 10%, and mortality
tion, the ductotomy is closed primarily or occasionally over a rate of 1% (Verbesey & Birkett, 2008). As mentioned previ-
T-tube. Historically, T-tubes have been used to decompress the ously, this procedure is not without complications. Bile leak
biliary tree and were thought to minimize bile leaks. They also (reported as high as 14%) and postoperative CBD strictures
allow postoperative percutaneous access to the CBD. T-tubes, are the most feared complications (Nathanson et al, 2005).
however, have potential morbidity, including inadvertent dis- Choledochotomy has similar rates of pancreatitis (7.3% vs.
placement, erosion, cholangitis, nutritional deficiencies from 8.8%), retained stones (2.4% vs. 4.4%), reoperation (7.3% vs.
bile loss (Gurusamy et al, 2013). In addition, T-tubes can cause 6.6%), and overall morbidity (17% vs. 13%) as ERCP (Nathan-
pain and be problematic to manage. son et al, 2005). Major advantages of the transcholedochal
If clearance of the duct has been achieved with confidence approach include easier access to both the upper and lower
and there are no concerns for distal obstruction, our preference ductal systems, and extraction of any size stones. Conversion
is to primarily close the choledochotomy without a T-tube. to an open procedure must be considered in difficult cases,
Several randomized trials have concluded that primary closure although challenging laparoscopic cases are frequently also
of the duct does not result in a higher rate of bile leaks (Zhang challenging open cases; referral to a specialty center should
et al, 2009). These studies have uniformly shown a decreased always be considered before converting to a laparotomy. In
length of hospital stay, shorter operative times, lower hospital patients with a high risk of recurrent stone disease or formation
expenses, and earlier return to normal activity in patients who of a biliary stricture, due to inflammation or a small duct, a
did not have a T-tube (Mangla et al, 2012). In addition, due definitive bypass should be considered.
to the high success rates of duct clearance by choledochotomy,
the need for T-tubes to provide percutaneous access for retained
Laparoscopic Transduodenal Sphincterotomy and
stones is not commonly necessary (Zhang et al, 2009). Common Bile Duct Exploration
For impacted stones refractory to clearance by endoscopy or
CBD exploration, laparoscopic transduodenal exploration with
sphincterotomy is an alternative modality for duct clearance
(see Chapter 36A). This procedure can typically be performed
with a 4 or 5 trocar technique, and we have found it useful to
place the trocars lower than the typical approach to cholecys-
tectomy. An additional camera trocar in the right lower quad-
rant will also provide a better angle for visualization of the
ampullary reconstruction (see Fig. 36B.1B).
To gain access to the duodenum, the right colon is mobi-
lized, and a Kocher maneuver is completed. With the duode-
num elevated, a surgical sponge posterior to the head of the
pancreas may be useful. This will elevate the duodenum and
absorb enteric fluids leaking into the field once the duode-
notomy is created. A longitudinal incision is made in the
antimesenteric wall of the second portion of the duodenum
with electrocautery or ultrasonic shears, to expose the major
duodenal papilla. If the papilla is not easily located, and the
cystic and CBDs are patent, a cholangiogram catheter or
wire can be inserted via the cystic duct and passed through
FIGURE 36B.3. Transcholedochal examination of the bile duct with the ampulla to aid in identification. Once the papilla is identi-
the choledochoscope. Stay sutures are used for retraction. fied, the CBD is intubated with a wire or silicone tube.
A B
FIGURE 36B.4. A, Transduodenal sphincterectomy at the 11 o’clock
position. B, Suturing of the bile duct mucosa to the duodenal mucosa
after sphincterectomy.
Electrocautery or ultrasonic shears are then used to create a

sphincterotomy at the 11 o’clock position on the papilla
(Makary & Elariny, 2006). This facilitates a transduodenal
CBD exploration by using cholangiocatheters, balloons, or a
cholangioscope. Once clearance of the CBD is achieved, the
mucosa of the bile duct is sutured to the duodenum with
interrupted 5-0 monofilament slowly absorbable sutures (Fig.
36B.4) (Makary & Elariny, 2006). The duodenotomy is then
closed to complete the case. The risk of complications from
transduodenal sphincterotomy is reported to be similar to
ERCP (Carboni et al, 2001). We use surgical drains selec-
tively, and if used, they are removed within a few days of the
operation if there is no evidence of bile leak.
Laparoscopic Biliary-Enteric Bypass FIGURE 36B.5. A, Orientation of the choledochotomy and duodenot-
In patients at moderate to high risk for recurrent CBD stones, omy for anastomosis. B, Creation of the choledochoduodenostomy.
or in patients with distal strictures, biliary-enteric bypass may
provide the most durable result. The two primary options for
bypass are choledochoduodenostomy or Roux-en-Y hepaticoduodenal mobility after kocherization. In these cases, we
jejunostomy. Currently, there are inadequate comparative proceed with laparoscopic Roux-en-Y hepaticojejunostomy.
data to recommend one technique of biliary-enteric anasto- To perform a laparoscopic choledochoduodenostomy we
mosis over another (see Chapters 31 and 36A). There are a use a trocar placement similar to that described for the trans-
number of retrospective studies comparing the two anastomo- duodenal sphincterotomy (see Fig. 36B.1B). A Kocher maneu-
ses that show equivalent outcomes and morbidity (Luu et al, ver is performed to mobilize the duodenum enough to perform
2013; Narayanan et al, 2013; Santore et al, 2011). Choledo- a tension-free anastomosis to the CBD. The degree of kocher-
choduodenostomy has been criticized in the past due to ization ultimately depends on individual anatomy and mobility
concern of reflux of enteric contents into the CBD leading to of the duodenum. Next, approximately 2 cm of the anterior
chronic inflammation and recurring bouts of cholangitis, surface of the distal CBD is exposed. The dissection plane
referred to as the sump syndrome (Khajanchee et al, 2012; should remain anterior to the duct to avoid injury of the blood
Tang et al, 2003). Results of analysis by several authors supply to the proximal duct. A thorough laparoscopic ultra-
suggest that symptoms associated with sump syndrome may sound examination should be performed to identify the CBD
actually be a mechanical problem associated with a narrow and verify the location of any stones. A 1.5-cm supraduodenal,
anastomosis (de Almeida et al, 1996; Degenshein, 1974; longitudinal choledochotomy is made in the anterior wall of the
Madden et al, 1970). duct by using electrocautery or ultrasonic shears. A choledo-
Our current preference is laparoscopic choledochoduode- choscope is then inserted through the choledochotomy, and a
nostomy as it is technically easier, leads to a more physiologic thorough examination is performed. Biopsies should be taken
reconstruction, and importantly, allows direct access to the if there are any concerns for a neoplastic process. A 1-cm lon-
biliary system should further evaluation or manipulation of the gitudinal duodenotomy is made in the adjacent postbulbar
duct be necessary (Moraca et al, 2002; O’Rourke et al, 2004). duodenum. The duodenotomy is created shorter than the duc-
However, choledochoduodenostomy may not be possible in the totomy due to the inevitable stretching of the duodenotomy
setting of significant duodenal inflammation or inadequate (Fig. 36B.5A).
We perform a single-layer, interrupted choledochoduode-

nostomy with 4-0 absorbable sutures. The posterior sutures are
tied internally, whereas the anterior layer is tied externally.
There are several techniques that facilitate creation of the anas-
tomosis. Tying the sutures as they are placed can limit access
during placement of the subsequent sutures, however leaving
too many sutures untied can lead to tangling of the sutures. To
avoid these problems, we typically place 2 to 3 sutures at a time,
tying the deeper sutures after the shallower sutures are placed.
Alternating dyed and undyed sutures, as well as clipping the
suture tails also helps in identification of the sutures and pre-
vents tangling (Fig. 36B.5B). Final inspection of the anastomo-
sis should confirm that there is no leakage of bile from the
anastomosis, and no tension on the duodenum or duct. Similar
to laparoscopic sphincterotomy, drains are used selectively and
removed early. We also use upper gastrointestinal contrast
studies selectively before starting a diet in patients who are at
high risk of an anastomotic leak.
Trocar placement for laparoscopic Roux-en-Y hepaticojeju-
nostomy is similar to that used for sphincterotomy and cho-
ledochoduodenostomy, although a slightly lower placement
may facilitate access to and division of the jejunum. Access to
the common hepatic duct may require mobilization of the right
colon and kocherization of the duodenum. Laparoscopic ultra-
sound examination is again used to facilitate identification of
the portal structures. We typically prefer an end-to-side hepati-
cojejunostomy (Fig. 36B.6A), although this requires circum
ferential dissection and division of the duct distally. In cases
where circumferential dissection is not possible or contrain
dicated (e.g., dense fibrotic scarring in the porta hepatis), a
side-to-side hepaticojejunostomy should be considered (Fig.
36B.6B).
To create the Roux limb, the jejunum is divided by using an
endostapler, approximately 20 cm distal to the ligament of
Trietz. The mesentery is then divided to allow the Roux limb
to reach the bile duct without tension. The limb is preferentially
passed antecolic; however, if this results in too much tension,
FIGURE 36B.6. A, End-to-side hepaticojejunostomy. B, Side-to-side
it can be passed retrocolic through the transverse mesocolon hepaticojejunostomy.
just to the right of the middle colic vessels. The jejunojejunos-
tomy is created in the infracolic position 30 to 40 cm distal to
the future bilioenteric anastomosis.
The hepaticojejunostomy is created in a single layer by using through the abdominal wall and through the stomach adjacent
4-0 or 5-0 absorbable sutures. Depending on the size of the to the future gastrotomy site with a Keith needle or transfascial
duct, this anastomosis may be created by using interrupted suture passing device. These sutures are used to retract and
sutures (for ducts <1 cm), in a fashion similar to that described stabilize the stomach. A gastrotomy is then made in the location
for the choledochoduodenostomy, or running sutures (for ducts that will allow easy passage of the endoscope through the
>1 cm). Surgical drains are used selectively. pylorus. A 15-mm radially dilating trocar with a balloon tip or
a single-port device is inserted through the abdominal wall and
Laparoscopic-Assisted Endoscopic Retrograde directly into the gastrotomy. The endoscope can then be placed
Cholangiopancreatography through the trocar or single-port device into the stomach to
In patients who have undergone a previous Roux-en-Y gastric perform the ERCP. After the conclusion of the ERCP, the
bypass procedure, access to the ampulla of Vater by using gastrotomy is closed with 3-0 absorbable suture. Surgical drains
double- and single-balloon endoscopes can result in successful are rarely indicated after this procedure.
performance of ERCP in nearly 80% of patients (Moreels,
2013). However, the majority of centers do not currently have CONCLUSION
access to these technologies (Lopes et al, 2009). In these cases,
laparoscopic access to the distal gastric remnant is necessary to The first laparoscopic biliary surgery was performed almost
perform an intraoperative ERCP and/or EUS. One study has 30 years ago. During the past three decades, substantial pro-
shown this procedure to be more cost effective than double- gress has been made in surgical expertise, experience, and
balloon enteroscopy (Schreiner et al, 2012). technology. Our understanding of how and when to apply these
At laparoscopy, the gastric remnant is identified, and the minimally invasive tools has greatly reduced the suffering of
front wall of the stomach is exposed. We place two sutures patients with choledocholithiasis. We now have the ability to
diagnose biliary stone disease, differentiate it from neoplastic and the capabilities of their team; they also must understand
processes, and treat it with high levels of success and decreasing the tools and talent available within their institutions. Finally,
morbidity. this knowledge must be applied to the particular clinical sce-
We have reviewed a wide range of approaches to the man- nario of each individual patient.
agement of patients with CBD stones. In managing this group
of patients, surgeons must assess their own surgical expertise References are available at expertconsult.com.
C. Biliary Infection and Infestation Chapter 36B Stones in the bile duct: minimally invasive surgical approaches 610.e1
REFERENCES Madden JL, et al: Choledochoduodenostomy: an unjustly maligned

surgical procedure?, Am J Surg 119:45–54, 1970.
Artifon EL, et al: Prospective randomized trial of EUS versus ERCP- Makary MA, Elariny HA: Laparoscopic transduodenal sphinctero-
guided common bile duct stone removal: an interim report (with plasty, J Laparoendosc Adv Surg Tech 16:629–632, 2006.
video), Gastrointest Endosc 69:238–243, 2009. Mangla V, et al: A randomized trial comparing the use of endobiliary
Burhenne HJ: Garland lecture. Percutaneous extraction of retained stent and T-tube for biliary decompression after laparoscopic
biliary tract stones: 661 patients, AJR Am J Roentgenol 134:889–898, common bile duct exploration, Surg Laparosc Endosc Percutan Tech
1980. 22:345–348, 2012.
Carboni M, et al: Transduodenal sphincterotomy in laparoscopic era, Moraca RJ, et al: Long-term biliary function after reconstruction of
World J Surg 25:1357–1359, 2001. major bile duct injuries with hepaticoduodenostomy or hepaticoje-
Chan T, et al: Total bilirubin is a useful predictor of persisting common junostomy, Arch Surg 137:889–893, discussion 893–894, 2002.
bile duct stone in gallstone pancreatitis, Am Surg 74:977–980, 2008. Moreels TG: ERCP in the patient with surgically altered anatomy, Curr
Chang L, et al: Preoperative versus postoperative endoscopic retro- Gastroenterol Rep 15:343, 2013.
grade cholangiopancreatography in mild to moderate gallstone Moreels TG: Endoscopic retrograde cholangiopancreatography in
pancreatitis: a prospective randomized trial, Ann Surg 231:82–87, patients with altered anatomy: how to deal with the challenges? World
2000. J Gastrointest Endosc 6:345–351, 2014.
Cuschieri A, et al: E.A.E.S. multicenter prospective randomized trial Narayanan SK, et al: Hepaticoduodenostomy versus hepaticojejunos-
comparing two-stage vs single-stage management of patients with tomy after resection of choledochal cyst: a systematic review and
gallstone disease and ductal calculi, Surg Endosc 13:952–957, 1999. meta-analysis, J Pediatr Surg 48:2336–2342, 2013.
Dasari BV, et al: Surgical versus endoscopic treatment of bile duct Nathanson LK, et al: Postoperative ERCP versus laparoscopic cho-
stones, Cochrane Database Syst Rev (12):CD003327, 2013. ledochotomy for clearance of selected bile duct calculi, a randomized
de Almeida AC, et al: Choledochoduodenostomy in the management trial, Ann Surg 242:188–192, 2005.
of common duct stones or associated pathology—an obsolete Orenstein SB, et al: Technical aspects of bile duct evaluation and
method?, HPB Surg 10:27–33, 1996. exploration, Surg Clin North Am 94:281–296, 2014.
Degenshein GA: Choledochoduodenostomy: an 18 year study of 175 O’Rourke RW, et al: Laparoscopic biliary reconstruction, Am J Surg
consecutive cases, Surgery 76:319–324, 1974. 187:621–624, 2004.
Gurusamy KS, et al: T-tube drainage versus primary closure after lapa- Paganini AM, et al: Thirteen years’ experience with laparoscopic tran-
roscopic common bile duct exploration, Cochrane Database Syst Rev scystic common bile duct exploration for stones. Effectiveness and
(6):CD005641, 2013. long-term results, Surg Endosc 21:34–40, 2007.
Hui CK, et al: Acute cholangitis–predictive factors for emergency Petelin JB: Laparoscopic approach to common duct pathology, Am J
ERCP, Aliment Pharmacol Ther 15:1633–1637, 2001. Surg 165:487–491, 1993.
Khajanchee YS, et al: Outcomes following laparoscopic choledocho- Ponce J, et al: Effect of intravenous glucagon and glucagon-(1-21)-
duodenostomy in the management of benign biliary obstruction, J peptide on motor activity of sphincter of Oddi in humans, Dig Dis
Gastrointest Surg 16:801–805, 2012. Sci 34:61–64, 1989.
Kroger W, et al: [First experiences with endoscopic retrograde cholan- Rogers SJ, et al: Prospective randomized trial of LC+LCBDE vs
giopancreaticography (ERCP) (author’s transl)], Radiol Diagn (Berl) ERCP/S+LC for common bile duct stone disease, Arch Surg 145:28–
16:315–324, 1975. 33, 2010.
Kroh M, Chand B: Choledocholithiasis, endoscopic retrograde chol- Santore MT, et al: Hepaticoduodenostomy vs hepaticojejunostomy for
angiopancreatography, and laparoscopic common bile duct explora- reconstruction after resection of choledochal cyst, J Pediatr Surg
tion, Surg Clin North Am 88:1019–1031, vii, 2008. 46:209–213, 2011.
Kum CK, Goh PM: Preoperative ERCP in the management of Schreiner MA, et al: Laparoscopy-assisted versus balloon enteroscopy-
common bile duct stones before laparoscopic cholecystectomy, Eur assisted ERCP in bariatric post-Roux-en-Y gastric bypass patients,
J Surg 162:205–210, 1996. Gastrointest Endosc 75:748–756, 2012.
Lenze F, et al: Clearance of refractory bile duct stones with extracor- Tang CN, et al: Laparoscopic choledochoduodenostomy: an effective
poreal shockwave lithotripsy: higher failure rate in obese patients, drainage procedure for recurrent pyogenic cholangitis, Surg Endosc
Scand J Gastroenterol 49:209–214, 2014. 17:1590–1594, 2003.
Li VK, et al: Optimal timing of elective laparoscopic cholecystectomy Urbach DR, et al: Cost-effective management of common bile duct
after acute cholangitis and subsequent clearance of choledocholithia- stones: a decision analysis of the use of endoscopic retrograde chol-
sis, Am J Surg 200:483–488, 2010. angiopancreatography (ERCP), intraoperative cholangiography, and
Lopes TL, et al: Laparoscopy-assisted ERCP: experience of a high- laparoscopic bile duct exploration, Surg Endosc 15:4–13, 2001.
volume bariatric surgery center (with video), Gastrointest Endosc Verbesey JE, Birkett DH: Common bile duct exploration for choledo-
70:1254–1259, 2009. cholithiasis, Surg Clin North Am 88:1315–1328, ix, 2008.
Luu C, et al: Choledochoduodenostomy as the biliary-enteric bypass Zhang WJ, et al: Laparoscopic exploration of common bile duct with
of choice for benign and malignant distal common bile duct stric- primary closure versus T-tube drainage: a randomized clinical trial,
tures, Am Surg 79:1054–1057, 2013. J Surg Res 157:e1–e5, 2009.
CHAPTER 36C
Stones in the bile duct: endoscopic and
percutaneous approaches
Demetrios J. Tzimas and Satish Nagula
HISTORICAL OVERVIEW 3. High suspicion of CBD stones: cholelithiasis, dilated CBD,

and abnormal liver biochemical tests
In the 1970s and 1980s, endoscopic retrograde cholangiopan- 4. Worsening gallstone pancreatitis
creatography (ERCP) transformed the diagnostic approach to 5. Recurrent CBD stones or gallstone pancreatitis, nonsurgical
suspected biliary disease and jaundice (see Chapters 20 and candidate for cholecystectomy
29). Similarly, in the years since it was first performed in
humans (Classen & Demling, 1974; Kawai et al, 1974), endo- ENDOSCOPIC TECHNIQUES
scopic sphincterotomy (ES) has had a dramatic impact on the
management of biliary disease, specifically in the treatment of An endoscopy service to treat CBD stones must have access to
common bile duct (CBD) stones. Approximately 150,000 an appropriate endoscopy facility and high-quality fluoroscopy.
endoscopic biliary sphincterotomies are performed annually in The endoscopy team must be fully cognizant of all basic
the United States, and the widespread availability of this pro- ERCP maneuvers, less frequently used techniques, and poten-
cedure has made endoscopic stone extraction the primary tial complications and their management. It is essential to
modality for the management of choledocholithiasis. Interest in explain the nature of the procedure to the patient and to outline
ERCP and endoscopic sphincterotomy as definitive therapy for the purpose, benefits, advantages, alternatives, and potential
CBD stones grew in the 1990s after the introduction of lapa- hazards.
roscopic cholecystectomy. Patient-related factors, clinical judg- Upon successful deep biliary cannulation with the sphinc-
ment, availability of expertise, and current evidence from tertome, a cholangiogram is initially performed, which defines
clinical trials must be combined to decide on an endoscopic, the ductal anatomy and the extent of the stone burden. ES is
percutaneous, or surgical approach. Although ERCP as a diag- usually the first therapeutic step in stone extraction. Balloon
nostic modality has been replaced by noninvasive imaging dilation of the biliary sphincter is an alternative to ES, but this
modalities such as magnetic resonance cholangiopancreatogra- has fallen out of favor due to increased risks of severe post-
phy (MRCP) (see Chapter 19), it remains the major nonopera- ERCP pancreatitis (Baron & Harewood, 2004; Bergman et al,
tive tool for the management of biliary diseases such as 1997; Disario et al, 2004). Standard pull-type sphincterotomes
choledocholithiasis and obstructive jaundice. allow a vertical incision to be made from the papillary orifice
in a cephalad direction along the intramural course of the CBD
INDICATIONS FOR ENDOSCOPIC THERAPY for a variable length (average, 10 to 15 mm), depending on
local anatomy, the degree of CBD dilation, and the size of stone
Patients with choledocholithiasis may present with asymptom- to be removed (Fig. 36C.1). The incision is produced by the
atic stones on noninvasive imaging or direct cholangiography controlled application of monopolar electrocautery delivered by
or with a variety of clinical symptoms, such as cholestasis, pain, a generator specifically designed for endoscopic use. It is fun-
cholangitis, and pancreatitis (see Chapter 32). In the early days damental to ES technique that complete control of wire tension
of ES–at a time when few endoscopy centers could offer the and electrocautery be maintained at all times. “Smart” genera-
technique and criticisms by surgical experts were common–it tors incorporate a pulsed generator (Erbe, Tubingen, Germany;
was considered justifiable only in elderly postcholecystectomy ConMed Endoscopic Technologies, Billerica, MA) with
patients with recurrent or retained bile duct stones who were feedback-controlled power output, thus avoiding a “zipper
at high risk of serious complications from open surgical CBD effect” and reducing pancreatitis and bleeding.
exploration or reexploration (Blumgart & Wood, 1978). The Occasionally, a precut sphincterotomy, also referred to as an
impressive successes of ES in this group, combined with access papillotomy, is needed to initiate ES when the standard
expanded availability, a low rate of complications, and strong instrument cannot be inserted deeply. This incision is often
patient preference, has led to ERCP becoming the primary needed when cannulation has been prevented by an impacted
modality for the management of CBD stones. stone. The needle-knife is more useful in this situation because
The endoscopist now must consider several clearly defined the intramural CBD is usually grossly distended and easily
conditions for which endoscopic management may be indicated incised, starting from the papilla and extending cephalad.
in patients with definite or suspected bile duct stones (Early Needle-knife fistulotomy is a variant of this technique; the inci-
et al 2012; Maple et al, 2011): sion is begun above the papilla to form a choledochoduodenal
1. Acute cholangitis fistulotomy. This technique is similar in efficacy to precut
2. Visualized CBD stone on abdominal ultrasound, endoscopic sphincterotomy, but more often it requires mechanical litho-
ultrasound (EUS), computed tomography (CT), MRCP, or tripsy (ML) and may have a lower rate of pancreatitis (Mavro-
intraoperative cholangiogram giannis et al, 1999). Patients with Bilroth II partial gastrectomy
611
A B
FIGURE 36C.2. Endoscopic retrograde cholangiopancreatography in
a patient with Billroth II partial gastrectomy showing insertion of a cath-
eter (curved arrow) (A) and placement of a guidewire and short biliary
endoprosthesis (solid arrow) (B) immediately before needle-knife sphinc-
terotomy and demonstration of common bile duct stones (open arrows).
FIGURE 36C.1. Endoscopic retrograde cholangiopancreatography
showing cholangiography with dilated bile duct, a single duct stone just
below the endoscope, a guidewire, and a sphincterotome in position
during sphincterotomy (arrows).
(Fig. 36C.2) and Roux-en-Y bypass operations present special

problems to the endoscopist, and numerous methods have been
described to obtain successful cannulation (Lin et al, 1999;
Wright et al, 2002) and removal of CBD stones (Bergman et al,
2001).
It is standard practice to attempt stone extraction from the
CBD immediately after ES. The two accessory instruments
used most commonly for this are the Dormia-type basket (Fig.
36C.3) and the Fogarty-type balloon (Figs. 36C.4 and 36C.5),
which are greater than 90% successful in clearing the CBD.
Occlusion cholangiography is performed after stone extraction
to confirm complete ductal clearance.
Difficult Stones
The most difficult circumstances encountered during endo-
scopic stone removal result from technical difficulties with
achieving deep biliary cannulation or in performing ES. These
difficulties are sometimes because of an inaccessible papilla
related to aberrant anatomy or unfavorable duodenal or papil-
lary structures, such as a periampullary diverticulum, or prior
surgery, such as Bilroth II or Roux-en-Y reconstruction. Tech-
niques have been described for the unique challenge of selective
bile duct cannulation in a patient with a Bilroth II partial gas-
trectomy (Lin et al, 1999). The performance of ES is also a
challenge because the visualized anatomy is inverted. In espe-
cially difficult cases, needle-knife sphincterotomy with a stent, FIGURE 36C.3. Endoscopic retrograde cholangiopancreatography
nasobiliary drain, or guidewire used as a guide for cutting showing large stone in the proximal bile duct (arrows) and basket extrac-
may be an option, or specially designed reverse-direction tion of a distal bile duct stone (chevrons) after endoscopic
accessories. The literature on techniques of cannulation and sphincterotomy.
A. Gallstones and Gallbladder Chapter 36C Stones in the bile duct: endoscopic and percutaneous approaches 613
sphincterotomy in Roux-en-Y reconstructions is limited (Wright to the stones, and the presence of coexisting pathology (e.g.,
et al, 2002), but some success has been reported with balloon stricture or tumor). Stones that are likely to be more difficult
enterosocopy or overtube-assisted ERCP (Kikuyama et al, to extract and may require adjuvant techniques to remove them
2009; Koornstra et al, 2008). are those that appear larger than the endoscope on radiographic
When ES has been successfully performed, extraction may imaging (usually >15 mm); stones that are numerous or hard
be hindered by a variety of stone factors—including size, in consistency; stones that are square, piston shaped, or faceted
number, consistency, shape, and location of stones—and ductal that tightly fit the bile duct or that are packed against each
factors such as contour and diameter at the level of and distal other; intrahepatic stones; or stones located proximal to a stric-
ture or narrowed distal bile duct or in a sigmoid-shaped duct.
Methods that have been developed to dilate the papillary
orifice, reduce stone size, and facilitate endoscopic removal
include endoscopic papillary large balloon dilation (EPLBD),
ML, intracorporeal lithotripsy with laser or electrohydraulic
probes, extracorporeal shock-wave lithotripsy (ESWL), and
chemical contact dissolution therapy. Treatment options
must be discussed jointly by the endoscopist, surgeon, and
interventional radiologist when difficulties are encountered
(Fig. 36C.6).
Endoscopic Papillary Large Balloon Dilation

EPLBD in conjunction with ES has greatly improved the rates
of ductal clearance of difficult stones without the need for
advanced ERCP lithotripsy techniques. The first report of
EPLBD (12 to 20 mm) with ES was described in 2003, when
a retrospective review demonstrated that 38 out of 40 patients
with large stones previously unable to be removed with stan-
dard techniques had successful ductal clearance with accept-
able complication rates (Ersoz et al, 2003). A prospective study
that randomized 90 patients to either ES with EPLBD or ES
with ML had similar rates of ductal clearance (97.7 % vs.
A B 91.1%, P = .36) but a significantly higher rate of complications
FIGURE 36C.4. Endoscopic retrograde cholangiopancreatography in the ML group (20% vs. 4.4%), which included 6 patients
showing a nondilated bile duct containing a single distal stone (white with cholangitis and 1 with perforation (Stefanidis et al, 2011).
arrow) (A) and extraction balloon (black arrow, right) (B) placed above the The rates of pancreatitis and postsphincterotomy bleeding were
stone immediately before its removal after endoscopic sphincterotomy. the same in both groups (2%).
A B
FIGURE 36C.5. Endoscopic retrograde cholangiopancreatography series showing technique of balloon dilation of the papilla for extraction of small
stones. A, Initial cholangiography with demonstration of three small stones (long arrows, left) and placement of a guidewire and insertion of an 8-mm
dilating balloon located between two radiopaque markers (small arrows, right). B, Inflation of the dilating balloon (left) followed by insertion of an
extraction basket for stone removal (right).
prevent perforation and bleeding. (3) Inflate the balloon gradu-

ally to recognize an occult stricture. (4) Discontinue balloon
inflation if resistance is met in the presence of a persistent
balloon waist. (5) Do not inflate the balloon beyond the
maximal size of the upstream dilated CBD. (6) Do not hesitate
to convert into alternative stone removal methods, such as ML
or electrohydraulic lithotripsy (EHL), if there is difficulty in
removing stones.
If performed correctly (with an incomplete ES and then
EPLBD to the size of the CBD), ES with EPLBD is a safe and
effective method for removal of multiple or large stones in the
CBD, which leads to shorter procedure times with decreased
use of advanced lithotripsy techniques.
Mechanical Lithotripsy
Removal of large CBD stones is a challenge for the most skilled
endoscopists. ML (see Chapter 27) (Leung & Tu, 2004; Leung
et al, 2001) remains an excellent option for stones that cannot
be removed by conventional techniques because it can be used
safely and effectively during the initial endoscopic procedure.
Mechanical lithotripters are modifications of standard Dormia
baskets and possess great tensile strength (Fig. 36C.7). The
reinforced basket is opened in the CBD, and the stone is
entrapped within the braided wires. This procedure can be
performed through the endoscope instrumentation channel, or
FIGURE 36C.6. Endoscopic retrograde cholangiopancreatography it can be done after the endoscope has been removed from the
showing a dilated bile duct containing multiple faceted stones in a post patient and a metal sheath has been extended over the inner
cholecystectomy patient positioned such that standard extraction tech- Teflon catheter. The end of the metal sheath is attached to a
niques might be difficult.
winding mechanism, which retracts the basket when cranked
and impales the stone against the rigid distal end of the metal
sheath leading to stone fracturing. The stone fragments can be
Further research has shown that endoscopic sphincterotomy removed with the same basket or a standard retrieval basket or
with EPLBD can obviate the need for ML. A recent review of balloon. In experienced centers, this technique allows removal
seven studies included 902 patients and compared ES with of more than 90% of difficult bile stones that are refractory to
EPBLD versus ES with standard techniques (Madhoun et al, standard extraction techniques, but multiple procedures may
2014). The authors found no differences in ductal clearance be required to achieve complete ductal clearance (Akcakaya
between the EPLBD and standard-techniques groups (98% vs. et al, 2009; Chang et al, 2005; Shaw et al, 1993; Van Dam &
95%, P = .6), but patients in the EPLBD group needed less Sivak, 1993).
ML. The use of EPLBD had a risk reduction of 0.58 (0.32 to
0.74) in terms of adverse events. Other Lithotripsy Modalities
Although the studies mentioned in the previous sections For the 5% of patients with biliary stones resistant to ES and
demonstrate the safety of EPLBD, there has been concern for ML, other methods are available, including intracorporeal tech-
the rare but serious complications, such as bleeding, perfora- niques (laser or electrohydraulic probes) and ESWL (Adamek
tion, and pancreatitis related to stretching the ampullary orifice et al, 1996). The choice between these methods or surgery
to such a large size. A recent review of 33 publications that depends largely on availability and local expertise.
included 2924 total procedures compared the complication
rates of EPLBD with a large ES, with limited ES, and without Electrohydraulic Lithotripsy
ES (Kim et al, 2013). The rate of adverse events was less than Since its development during the 1950s in the former Soviet
10% in each group and not significantly different between the Union as a method to fragment rocks during mining, EHL has
three groups. The rates of severe complications such as pan- been adapted for medical use in the treatment of nephrolithiasis
creatitis (all <4%) and perforation (all <0.5%) were acceptable and, more recently, biliary tract calculi. The electrohydraulic
and also not significantly different among the three groups. The probe consists of two coaxially isolated electrodes at the tip of
rate of bleeding, however, was highest in the large ES group, a flexible catheter, which is capable of delivering electric sparks
at 4.1%, which was significantly higher than the limited-ES and in short, rapid pulses leading to sudden expansion of the sur-
no-ES groups (1.3% vs. 1.9%, respectively). rounding liquid environment and generating pressure waves
A multicenter retrospective study that included 946 patients that result in stone fragmentation (Picus, 1990). Direct chol-
who underwent EPBLD for CBD stones greater than 10 mm angioscopy with either mother-daughter cholangioscopes
in size sought to determine the predictive factors of adverse (Hixson et al, 1992) or single-operator cholangioscopy (SOC)
events in EPLBD (Park et al, 2013). Based on their findings systems, such as the SpyGlass (Boston Scientific, Natick, MA)
authors gave these recommendations for safe EPLBD: (1) Indi- direct visualization system, are used to target the stones to
cation should be patients with a dilated CBD without distal facilitate stone contact with the electrode and to avoid ductal
CBD strictures. (2) Avoid full-ES immediately before LBD to trauma or perforation (Aljebreen et al, 2014; DiSario et al,
A B
FIGURE 36C.7. Endoscopic retrograde cholangiopancreatography sequence showing use of transendoscopic mechanical lithotripsy. A, Positioning
of the mechanical lithotripsy basket with its metallic sheath in the proximal bile duct (left) and its slow withdrawal toward the distally placed stone to
entrap it (right). B, Process of lithotripsy after stone entrapment within the basket (left) as the basket wires cut through the stone (right) to produce
stone fragmentation.
2007). Continuous saline irrigation is used with the bipolar with normal saline is needed to provide a medium for the
electrode placed at the surface of the stone to provide a media transfer of energy and to help clear stone fragments (Lee et al,
for shock-wave energy transmission, to flush away debris, and 2012). Despite the fragmentation of stones, standard tech-
to maintain adequate visualization (DiSario et al, 2007). niques such as balloon sweep or mechanical lithotripsy may still
Reports document complete stone clearance after multiple ses- be required to completely clear the duct of all debris. A study
sions in 86% of patients (Adamek et al, 1996; Arya et al, 2004; published in 2007 described how the holmium laser can frag-
Hixson et al, 1992; Siegel et al, 1990), and in a prospective ment stones regardless of their composition, whether they are
nonrandomized trial, EHL was comparable to ESWL in stone cholesterol, pigment, or calcium stones (Yates et al, 2007). The
clearance (Adamek et al, 1996). holmium laser has a high absorption coefficient in water and
The largest study of EHL performed to date is a retrospec- therefore has a better safety margin and has more than 100
tive review of 94 patients with difficult biliary stones referred times the energy absorption than the neodymium laser (Maydeo
to a tertiary hospital (Arya et al, 2004). The authors used a et al, 2011). Despite the safety profile, to prevent bile duct
mother-daughter cholangioscope and achieved fragmentation injury, it is necessary to use direct visual control while perform-
in 96% of the patients, with an eventual stone clearance rate of ing holmium laser lithotripsy, which also allows real-time
90%. Seventy-six percent of these patients required only one assessment of any biliary injuries (Hochberger et al, 2003;
EHL session, and most patients did not need additional ses- Sauer et al, 2013).
sions to help clear residual stones or debris. Complications Cholangioscopy was classically performed using mother-
included cholangitis (14%), pancreatitis (1%), and hemobilia daughter scopes with two endoscopists, but several recent
(1%) that was successfully treated with epinephrine. Another studies demonstrate the safety and efficacy of the SpyGlass
retrospective review of 26 patients examined EHL for difficult SOC for laser lithotripsy with the holmium : YAG laser. A pro-
biliary stones by using a single-operator duodenoscope-assisted spective study in 2011 examined 60 patients with choledocho-
cholangioscope (Farrell et al, 2005). All of the patients achieved lithiasis who either failed therapy with conventional methods
complete ductal clearance without any complications, and 58% or were referred for management of potentially difficult stone
of the patients only needed one session of EHL. A recent ret- removal (Maydeo et al, 2011). Complete ductal clearance
rospective review studied 13 patients who underwent EHL with using the Spyglass SOC system with the holmium : YAG laser
the SpyGlass SOC and demonstrated a ductal clearance rate of was achieved in 50 patients (83.3%) in one session, and the
100% with a mean number of EHL procedures at 1.6 per remaining patients achieved ductal clearance after one addi-
patient (Aljebreen et al, 2014). There was only one complica- tional session. Complications included fever in 3 patients
tion of cholangitis, treated conservatively with antibiotics. (although these patients were already admitted with cholangi-
tis), postprocedure pain requiring hospital admission in 4
Laser Lithotripsy patients, and a biliary stricture in 1 patient who developed a
Reports of the use of a holmium : yttrium-aluminum-garnet stricture proximal to the stone, which was successfully treated
(holmium : YAG) laser lithotripsy for choledocholithiasis was with dilation using a 10-Fr biliary stent for 3 months. The
first published in 1998 (Burdick et al, 1998; Das et al, 1998). authors were not sure if the stricture was caused by the laser
During holmium laser therapy, continuous ductal irrigation therapy or the stone itself. Holmium laser therapy has also been
used to treat patients with cystic duct stones (CDS) and Mirizzi
syndrome (MS) using the Spy Glass SOC. A retrospective
study at a single center described 34 patients: 31 with MS, 3
with CDS (Bhandari et al, 2016). All of the Mirizzi syndrome
patients had successful removal in one session, and 2 of the
CDS patients needed a second session for stone removal.
Extracorporeal Shock-Wave Lithotripsy

ESWL with a variety of lithotripsy machines is now an accepted
alternative to endoscopic fragmentation of difficult bile duct
stones. In contrast to intracorporeal techniques, direct contact
with the stone is unnecessary. Most centers localize stones with
fluoroscopic focusing during contrast perfusion of the bile duct
through an endoscopically placed nasobiliary catheter or per-
cutaneous drain (Gordon et al, 1991; White et al, 1998).
Ponchon and colleagues (1990) reported ESWL success with
an ultrasound localization system, although it was less effective
when multiple stones were present. Several large series (Bland
et al, 1989; Gilchrist et al, 1997; Sackman et al, 2001; Sauer-
bruch & Stern, 1989) indicated success rates for ESWL stone
fragmentation of 53% to 91% and duct clearance in 58% to
90%. Minor complications are common and include biliary
pain, hemobilia, transient liver function test elevations, and
cutaneous petechiae. Overall, with the use of endoscopic tech- FIGURE 36C.8. Endoscopic retrograde cholangiopancreatography
niques such as ML, EHL, laser lithotripsy, and ESWL, one showing placement of a 10-Fr diameter endoprosthesis for the tempo-
report showed successful stone removal in 98% of 217 patients, rary management of cholangitis caused by several bile duct stones.
with only 5 patients requiring surgery (Schumacher et al,
1998). However, given the high efficacy of current EHL
technology/Spyglass SOC, ESWL is rarely used for bile duct does not support the concept of permanent plastic stent therapy
stones. except for patients with severe comorbidity and a short life
expectancy.
Endoprosthesis Placement The fully covered self-expanding metal stent (FC-SEMS)
In the few situations in which stone extraction is incomplete or was intended to improve patency of metal stents by preventing
impossible, a nasobiliary tube or, more commonly, an endo- tissue ingrowth in malignant bile duct obstruction. However,
prosthesis (Fig. 36C.8) must be inserted to provide biliary the silicone covering on the stent has allowed for delayed stent
decompression and prevent stone impaction in the distal CBD. removal and thus has subsequently been successfully used in
This is a temporizing therapy to allow the patient’s clinical an off-label fashion for benign biliary diseases, such as benign
condition to improve, until complete stone clearance is achieved biliary strictures and complex bile duct stones (Deviere et al,
via additional endoscopic maneuvers or surgery (see Chapters 2014; Tarantino et al, 2012). It has been postulated that the
29 and 36). friction between the stones and the stent reduces the stone size,
Nasobiliary tubes are rarely tolerated beyond a few days. and radial dilating force of the stent across the papilla further
Furthermore, problems with tube placement, such as acciden- assists in the clearance of choledocholithiasis (Garcia-Cano
tal dislodgment, have led to the alternative therapy of tempo- et al, 2013; Katsinelos et al, 2003). A retrospective review
rary biliary endoprosthesis placement (Kiil et al, 1989; studied 36 patients with complex biliary stones who had incom-
Rustgi & Schapiro, 1991). In a poor-risk surgical patient, ES plete ductal clearance despite the use of advanced extraction
and long-term placement of a plastic biliary endoprosthesis techniques (Cerefice et al, 2011). All patients had successful
has been proposed as a nonsurgical alternative (Cotton et al, biliary drainage after the initial SEMS was placed. Thirty-three
1987; Foutch et al, 1989; Nordback, 1989; Soomers et al, patients (94%) had complete ductal clearance after a mean of
1990). Of 84 patients intentionally treated with permanent 2.2 ERCP sessions. There were no immediate or delayed com-
plastic stents for endoscopically irretrievable stones and fol- plications related to FC-SEMS placement or removal, and only
lowed for a mean of 3 years, 49 (58%) developed biliary com- 4 patients had spontaneous migration of stents that were
plications, and 9 died as a result of complications. Most of the deemed clinically insignificant. Another retrospective review
patients had a long, symptom-free interval, however, before evaluated 29 patients with long-term FC-SEMS placement for
complications developed, supporting stenting only as a short- the management of difficult CBD stones that could not be
term treatment (Bergman et al, 1995; Maxton et al, 1995). In removed by standard or advanced techniques (Garcia-Cano
a randomized study with a shorter mean follow-up (1.5 years) et al, 2013). All patients had successful procedures in terms of
that compared placement of a plastic biliary stent with ES biliary drainage, and the stents were left in place for a median
as definitive therapy with stone clearance by means of the of 200 days. After ERCP, 13 of the patients refused repeat
basket, balloon, or mechanical lithotripter, the patients with procedures, and therefore only 16 patients returned for success-
stents had a significantly greater rate of cholangitis (36%) than ful stent removal. Complete ductal clearance was achieved in
the patients managed by a conventional endoscopic duct clear- 15 of 16 (93.7%) patients after the second ERCP. The 13
ance approach (14%). The high risk of long-term complications patients who refused repeat ERCP were followed for more than
6 months without any complications related to stent placement. 1985; Nakajima et al, 1979; Reiter et al, 1978; Safrany, 1978;
In patients in whom complete ductal clearance is not feasible Schumacher et al, 1998; Seifert et al, 1982; Sherman et al,
on the index ERCP, FC-SEMS placement can greatly improve 1991; Siegel, 1981; Vaira et al, 1989) with individual and col-
the chances of success on the subsequent ERCP while also lected series of 430 to 9041 patients range from 75% to 96%
minimizing the need for advanced lithotripsy techniques. for duct clearance, with a median value of 91%.
Dissolution Therapy Complications of Endoscopic Therapy

Contact chemical dissolution of stones has been attempted by Despite the disparate indications and selection of patients
perfusing the CBD with solvents administered via an indwell- among centers, the overall incidence of ERCP-related compli-
ing nasobiliary tube, percutaneous transhepatic catheter, cho- cations seems to be remarkably consistent, between 5% to 10%
lecystostomy tube, or an existing T-tube. The initial results (Andruilli et al, 2007; Cotton, 1984; Cotton & Vallon, 1981;
with these agents were disappointing because of incomplete Freeman et al, 1996; Geenen et al, 1981; Leese et al, 1985;
stone dissolution and complications. A semisynthetic vegetable Masci et al, 2001; Seifert et al, 1982; Siegel, 1981; Vander-
oil, monooctanoin, composed of 70% glycerol-1-monooctano- voort et al, 1996). A recent review of all major prospective
ate and 30% glycerol-1,2-dioctanoate, was used experimen- ERCP trials (16,855 patients) reveal the following specific com-
tally for the dissolution of CBD stones beginning in 1977. plication rates: acute pancreatitis, 3.5% (range, 1.0% to 8.7%);
Results collected from 222 clinicians treating 343 patients with cholecystitis or cholangitis, 1.4% (range, 0% to 5%); acute
CBD stones between 1977 and 1983 reported a success rate hemorrhage from sphincterotomy site, 1.3% (range, 0.3% to
for complete stone dissolution of only 25.6% and an additional 6.2%); perforation, 0.6% (range, 0% to 6.2%), with a small
partial success rate of 28% (Palmer & Hoffmann, 1986). numbers of other rare complications, such as impacted basket
Serious adverse events leading to discontinuation of treatment and gallstone ileus. Complication rates must be interpreted
occurred in 5% of patients, including hemorrhage from duo- with caution because definitions of hemorrhage, acute pancre-
denal ulceration, acute pancreatitis, jaundice, pulmonary atitis, cholangitis, and perforation often differ, although many
edema, acidosis, anaphylaxis, septicemia, and leukopenia, but studies use consensus definitions (Cotton et al, 1991).
no deaths were reported. The use of organic solvents, such as Post-ERCP pancreatitis is defined as a rise in serum amylase
the aliphatic ether methyl tert-butyl ether (Allen et al, 1985), to at least three times the upper limit of normal with accompany-
also has been disappointing, with complete stone dissolution ing typical pain of pancreatitis, leading to either a hospital admis-
achieved in only 30% to 45% and an unacceptable com sion or prolongation of current hospitalization. It is important
plication rate related to systemic absorption from spillover of to recognize that isolated asymptomatic hyperamylasemia after
solvent into the duodenum and intrahepatic bile ducts ERCP is a common and expected finding. Post-ERCP pancre-
(Brandon et al, 1988; Diaz et al, 1992; Kaye et al, 1990; atitis is managed like any typical bout of acute pancreatitis (see
Murray et al, 1988; Neoptolemos et al, 1990). Expectations of Chapters 55 and 56), and although most attacks are mild and
developing a solvent-chelating agent (ethylenediaminetetraace- self-limited, clinicians must remain vigilant for severe pancreati-
tic acid) for pigment stones have not been realized. As a result tis. Wire-guided cannulation decreases the rate of post-ERCP
of its low efficacy and morbidity, contact dissolution therapy pancreatitis by approximately 50% compared with conventional
has not assumed an important role in patients with refractory contrast-guided approaches (Tse et al, 2013). Risk factors for
CBD stones, and newer agents with better methods for instil- post-ERCP pancreatitis include suspected sphincter of Oddi
lation are awaited. dysfunction, prior history of post-ERCP pancreatitis, female
patients, young age, difficult cannulation, and contrast injec-
RESULTS OF ENDOSCOPIC THERAPY tions into the pancreatic duct (Freeman et al, 1996, 2001;
Neoptolemos et al, 1989; Vandervoort et al, 2002). Prophylactic
Successful endoscopic treatment of CBD stones requires an pancreatic ductal stent placement and periprocedural adminis-
adequate ES, which is now achieved in greater than 90% of tration of rectal nonsteroidal antiiflammatory drugs, such as
attempts in most reported series, with noticeable improvement diclofenac and indomethacin, each reduce the risk of post-ERCP
as experience increases (Blumgart & Wood, 1978; Cotton, pancreatitis by 50% (Elmunzer et al, 2012). The combined
1984; Cotton & Vallon, 1981; Geenen et al, 1981; Leese et al, protective effect of pancreatic stents and rectal nonsteroidal
1985; Schumacher et al, 1998; Seifert et al, 1982; Siegel, antiinflammatory drugs is the subject of ongoing study. A new,
1981). Most experts now would expect to extract stones in at highly potent protease inhibitor, nafamostat mesylate, has shown
least 90% of successful sphincterotomies. Failure to extract or significant efficacy in early trials; however, larger clinical studies
pass stones may be due to the size or number of stones within are needed (Park et al, 2011).
the duct or unfavorable duct diameter, usually in its retropan- Postsphincterotomy bleeding is often recognized immedi-
creatic segment. Stones 10 mm in diameter do not give rise to ately after the sphincterotomy, but some patients may have
many problems, but in general, with size greater than 15 to delayed bleeding. Although there is a paucity of data, use of
20 mm, the chance of retention increases. Interpretation of antiplatelet agents does not appear to increase the risk of bleed-
success rates necessitates care because centers with greater ing. Controlled sphincterotomy technique with the use of
expertise are more likely to be referred difficult cases that may blended current, while avoiding the “zipper” cut, is a recom-
be failures from attempts elsewhere, biasing results. Patient mended method to prevent bleeding. In patients with delayed
groups also vary considerably from unit to unit and country to bleeding, symptoms are similar to any routine upper gastroin-
country, reflecting different referral patterns, selection of testinal bleed, including hemodynamic changes and melena.
patients, and attitudes toward endoscopic therapy. Results from Mild cholestasis may be evident due occlusion of the biliary
centers around the world (Cotton, 1984; Cotton & Vallon, orifice with blood clots. Mild to moderate bleeding can often
1981; Freeman et al, 1996; Geenen et al, 1981; Leese et al, be controlled with endoscopic techniques, including balloon
learned to prevent or salvage this situation. These maneuvers

include (1) avoiding basket closure during initial attempts to
extract a large stone to prevent impaling the basket wires in the
stone surface, (2) converting the standard basket into a crush-
ing type by replacing the handle with a mechanical lithotripter,
and (3) extending the ES incision by removing the duodeno-
scope over the impacted basket catheter and reintroducing it
alongside the catheter, introducing a second duodenoscope, or
passing a sphincterotome along the same instrument channel
as the impacted basket catheter when using large-channel
(3.7 mm or 4.2 mm) endoscopes.
Death after ERCP is a rare event, with a mortality rate
directly attributable to the procedure ranging between 0% and
0.94%, with an average of 0.3% (Andruilli et al, 2007), with
roughly equal distribution of causes between hemorrhage, pan-
creatitis, cholangitis, and perforation.
Long-Term Morbidity
FIGURE 36C.9. Computed tomographic scan of the stomach, with Long-term follow-up of 1 to 15 years after ERCP with ES in
oral contrast administration showing extensive retroperitoneal, intraperi- postcholecystectomy patients (Cotton, 1984; Escourrou et al,
toneal, and subcutaneous air caused by a perforation from endoscopic 1984; Hawes et al, 1990; Ikeda et al, 1988; Rosch et al, 1981;
sphincterotomy. The patient made an uneventful recovery on conserva-
Seifert et al, 1982; Sivak, 1989) has demonstrated that more
tive treatment.
than 90% of patients are well on symptomatic review, and 7%
to 11% have significant symptoms secondary to recurrent
stones (5%), with or without stenosis of the ES site (1.5% to
tamponade of the sphincterotomy site, injection of dilute epi- 3%) and cholangitis (2%). Most of these long-term complica-
nephrine (1 : 10,000), bipolar cautery, and placement of hemo- tions are amenable to further endoscopic treatment.
static clips (Grimm & Soehendra, 1983; Leung et al, 1995;
Wilcox et al, 2004). Temporary placement of FC-SEMS can LAPAROSCOPIC AND PERCUTANEOUS
provide durable hemostasis through long-term tamponade of APPROACHES TO BILE DUCT STONES
the bleeding site, with efficacy demonstrated in a small case
series (Debenedet et al, 2013). In rare cases of major arterial Laparoscopic Common Bile Duct Exploration
hemorrhage, endoscopic view of the papillary area is obscured Laparoscopic CBD exploration (see Chapter 36B) was first
by blood, precluding any further endoscopic therapies. In these performed in the early 1990s and typically performed at expe-
patients, angiography with superselective embolization of the rienced tertiary referral centers (Khoo et al, 1996; Martin et al,
active bleeding site has been shown to be highly effective 1998; Rhodes et al, 1995). Ductal exploration may be accom-
(Maleux et al, 2014). Accordingly, surgical management for plished through the cystic duct or directly through a choledo-
post-ERCP bleeding has become uncommon in hospitals with chotomy. The transcystic route is the least invasive and generally
interventional radiology services. does not require any ductal manipulation or drainage proce-
Duodenal perforation is relatively rare and is either a small dure, whereas choledochotomy requires either closure of the
retroperitoneal perforation related to the sphincterotomy or a duct over a T-tube or primary closure of the choledochotomy
large duodenal perforation from the shaft of the scope. The with or without a biliary stent placed in an antegrade fashion
perforation may be asymptomatic and noticed only as retroperi- without need for a T-tube (Huang et al, 1996; Rhodes et al,
toneal gas (Fig. 36C.9) or extravasation of radiographic con- 1995). Bile duct clearance rates average 90%, with a median
trast material, but even in a symptomatic patient, conservative rate of conversion to open operation of 4% (Tranter & Thomp-
treatment is often effective, with spontaneous resolution and son, 2002). Complication rate is 2.5%, with a median mortality
avoidance of potentially difficult surgery. Occasionally, this rate of 1% (Strasberg et al, 1995).
complication presents late after ES with a retroperitoneal col- Multiple randomized clinical trials have been performed
lection of bile or pus in the flank or inguinal region (Leese et al, comparing single-stage laparoscopic CBD exploration at the
1985; Neoptolemos et al, 1984) and requires percutaneous or time of cholecystectomy versus a two-stage approach with
surgical drainage. ERCP preceding or following laparoscopic cholecystectomy
Cholangitis is confined almost completely to patients in (Cuschieri et al, 1996; Ding et al, 2014; Rhodes et al, 1998;
whom CBD clearance has not been achieved, and measures Rogers et al, 2010). The results of these trials have been strik-
should be directed at providing adequate bile drainage (e.g., by ingly similar, demonstrating similar ductal clearance rates for
nasobiliary catheter or endoprosthesis) and administering par- both groups (75% to 95%) with comparable rates of complica-
enteral antibiotics. Gallstone ileus is a rare complication, but tions and mortality. In studies that examined hospital param-
its recognition needs to be emphasized because symptoms may eters, the single-stage surgical approach offered lower length of
be obscure in elderly patients, and they occur many days after stay and reduce hospital costs. Length of stay and associated
ERCP and stone release; treatment is along standard surgical hospital costs can be reduced with improved coordination
lines (see Chapter 43). between the surgeon and the endoscopist. Although the single-
The impaction of an extraction basket occurs rarely in expe- stage approach appears at least equivalent to the two-stage
rienced hands because many endoscopic maneuvers have been approach, implementation of this strategy is restricted to centers
with significant expertise in laparoscopic bile duct exploration. 2015). Endoscopic treatment can be performed safely in preg-
Conversely, most facilities have ready access to ERCP services, nant patients (Jamidar et al, 1995; Kahaleh et al, 2004;
and thus a two-stage approach is the most common strategy in Simmons et al, 2004; Tham et al, 2003) using techniques to
the United States. minimize fluoroscopy. A recent retrospective study using the
Nationwide Inpatient Sample demonstrated a higher risk of
Percutaneous Approach post-ERCP pancreatitis in pregnant patients (12%) compared
In the 5% to 10% of patients for whom the endoscopic approach with control patients (5%); multivariable analysis confirmed
is unsuccessful at clearing the CBD of stones, two nonsurgical pregnancy was an independent risk factor for post-ERCP pan-
approaches are available: a rendezvous procedure and a complete creatitis (Inamdar et al, 2016).
percutaneous procedure (see Chapter 30). If the papillary region
can be reached endoscopically, yet deep biliary cannulation is Patients With Gallbladder in Situ
unable to be achieved, then the rendezvous procedure may be In elderly patients or in patients with significant comorbidity,
used. Although typically used for patients with obstructive a deliberate decision often is made to leave the gallbladder in
jaundice from pancreaticobiliary malignancy, this technique situ after ERCP and CBD stone removal. The short- and long-
can be used for choledocholithiasis in patients with surgically term results and complications of ERCP with ES in patients
altered anatomy, such as a Bilroth II, or in patients with chal- with gallbladders do not differ from those in postcholecystec-
lenging papillary anatomy, such as a large periampullary diver- tomy patients (Kaw et al, 2002). The risk of deferring chole-
ticulum. This involves a two-stage procedure with the cystectomy was examined in a study of 120 patients with known
introduction of a percutaneous guidewire through the bile ducts gallstones who underwent successful ERCP with bile duct
and papilla into the duodenum in the first stage, followed by clearance; the study excluded patients who were not surgical
an ERCP (Dowsett et al, 1989; Martin, 1994; Tomizawa et al, candidates, and patients were randomized to laparoscopic cho-
2014; Verstandig et al, 1993). In one report of rendezvous lecystectomy within 6 weeks after ERCP or a wait-and-see
procedures for choledocholithiasis, 0.9% (15/1753) of ERCPs policy with median 2.5-year follow-up. Nearly half the patients
were unsuccessful at biliary cannulation, usually owing to duo- in the wait-and-see group developed biliary-related problems,
denal diverticula or Bilroth II anatomy (Calvo et al, 2001). including biliary pain and cholecystitis, which led to cholecys-
Three patients underwent surgery, and 93% of the remaining tectomy or repeat ERCP or both, whereas none of the patients
patients underwent a successful rendezvous procedure. There in the cholecystectomy group experienced biliary-related prob-
was one complication with a retroperitoneal perforation that lems. The wait-and-see group also had a higher rate of conver-
required surgical management, and during follow-up, only one sion to an open procedure at the time of surgery (Boerma et al,
patient developed recurrent choledocholithiasis, requiring a 2002). A meta-analysis of randomized trials comparing the
repeat rendezvous procedure. In the multicenter prospective wait-and-see approach with elective cholecystectomy confirmed
trial of endoscopic biliary sphincterotomy complications these findings, with a higher risk of biliary pain (relative risk,
(Freeman et al, 1996), the combined endoscopic-percutaneous 14.6), cholangitis (relative risk, 2.5) in patients who deferred
approach was a risk factor for the development of complica- cholecystectomy (McAlister et al, 2007). The risk of these
tions, with a high rate of complications at 22.6% and with 6.5% future biliary complications needs to be balanced with operative
classified as severe. EUS-guided rendezvous procedures are risk in patients with significant underlying comorbidity.
emerging as a viable alternative to percutaneous approaches for
bile duct access in patients with unsuccessful biliary cannula- Suspected Choledocholithiasis
tion (Iwashita et al, 2012) (see Chapter 29). In patients with suspected choledocholithiasis, it is imperative
The complete percutaneous approach is labor intensive and to perform a risk assessment for the presences of CBD stones
typically requires multiple sessions. It has been established as prior to cholecystectomy. The following algorithm accounts for
treatment for hepatolithiasis (Yeh et al, 1995) (see Chapters 39 the relative efficacy, safety, and cost-effectiveness of each pro-
and 44), but it may be used for choledocholithiasis. The procedure and imaging study (Tse et al, 2004). Patients with any
cedure involves initial establishment of a transhepatic fistula, very strong predictor of CBD stones (visualized CBD stone on
followed by stone extraction under fluoroscopy or cholangios- imaging, cholangitis or bilirubin > 4) or two strong predictors
copy 7 to 8 days after the fistula forms. In a series of 31 patients of CBD stones (dilated CBD > 6 mm, bilirubin between 1.8
with failed endoscopic procedures, percutaneous biliary access and 4) should undergo preoperative ERCP. Moderate predic-
was achieved in all patients, and stone clearance was complete tors are any other abnormal liver function tests, age older than
in 87% after a mean of 5.6 sessions (Van der Velden et al, 55 years, and gallstone pancreatitis. Patients without any very
2000). All patients underwent balloon dilation of the papilla, strong, strong, or moderate predictors are deemed low risk for
and most patients required additional means of stone fragmen- CBD stones and should proceed directly to cholecystectomy.
tation, including ML, EHL, and ESWL. Complications includ- All remaining patients are at intermediate risk for CBD stones
ing pancreatitis and bacteremia occurred in 9.7% and did not and should either undergo preoperative MRCP or EUS, fol-
require surgical intervention. The 4 patients who did not lowed by ERCP if CBD stones are visualized. Alternatively,
respond to percutaneous management underwent surgery. depending on local expertise, these intermediate-risk patients
can proceed with laparoscopic cholecystectomy with intraop-
SPECIFIC CLINICAL SCENARIOS erative cholangiogram; patients can then proceed with laparo-
scopic bile duct exploration or undergo postoperative ERCP.
Pregnancy In patients who undergo preoperative ERCP, it is important to
Symptomatic choledocholithiasis during pregnancy poses a minimize the time between the ERCP and cholecystectomy, to
diagnostic and therapeutic challenge. Biliary tract stone disease reduce the risk of recurrent CBD stones prior to surgery. Sub-
occurs in approximately 1.5% of pregnancies (Ellington et al, sequent studies have found this algorithm to be moderately
accurate (60% to 70%) for predicting the presence of CBD suggesting the possibility that early stone extraction by surgical
stones; the remaining patients did not have CBD stones on or endoscopic techniques would halt progression of the acute
ERCP despite having criteria suggestive of choledocholithiasis event and prevent the development of future attacks in the
(Adams et al, 2015; Sethi et al, 2016). short term.
Early surgical therapy in cases of acute biliary pancreatitis
Acute Cholangitis (ABP) has been challenged owing to the high operative morbid-
Acute cholangitis (see Chapter 43) resulting from CBD stones ity and mortality. Results and conclusions from numerous
traditionally was managed by supportive measures and paren- series comparing early and late surgical therapy in gallstone
teral antibiotics, followed by early surgery if improvement was pancreatitis are difficult to interpret. The mortality rates range
slow or absent. In early reports, the mortality from emergency from 2% to 67%, studies are retrospective with frequent com-
surgery ranges from 12% to 16%, with higher rates for elderly parisons to historical controls, and stratification for severity of
patients (Boey & Way, 1980; Cotton, 1984; Thompson et al, illness has not been used (Acosta et al, 1978; Kim et al, 1988;
1982). The only randomized trial of emergency endoscopic Osborne et al, 1981; Ranson, 1979). In one study, Kelly and
versus surgical management of severe calculous cholangitis (Lai Wagner (1988) prospectively randomized 165 patients with
et al, 1992) showed a threefold difference in mortality rate gallstone pancreatitis to early or delayed surgery. In the group
(10% vs. 32%; P < .03) in favor of ERCP. In patients who are with severe pancreatitis, mortality was 48% after urgent opera-
hemodynamically stable, it is reasonable to proceed with ES tive intervention compared with 11% mortality rate in patients
during ERCP with removal of all calculi. Care must be taken in whom surgery for gallstones was delayed for more than 48
to minimize aggressive contrast injection (particularly during hours. In contrast, patients with mild pancreatitis had mortality
balloon-occluded cholangiography) to reduce the potential risk rates of 3.3% and 0%, respectively. Another study of moderate
of bacteremia from the procedure. In patients with hemody- to severe gallstone pancreatitis with peripancreatic fluid collec-
namic compromise, procedure duration can be minimized tions confirmed these results, with complications in 44% of
using a two-stage approach, placing a plastic biliary stent patients in the early surgery group compared with 6% in the
without sphincterotomy to achieve biliary decompression, delayed surgery group (Nealon et al, 2004).
which leads to resolution of cholangitis (Hui et al, 2003). After The results of these aforementioned studies favor avoidance
the patient’s clinical status has improved, a second ERCP can of early operative intervention in the acute phase of biliary
be performed with ES, allowing complete ductal clearance of pancreatitis, as the previous surgical dictum was that inflam-
CBD stones. mation and edema from pancreatitis can distort biliary anatomy,
which would complicate surgeries and predispose patients to
Gallstone Pancreatitis bile duct injury (da Costa et al, 2015). However, recent robust
Acute pancreatitis (see Chapters 55 and 56) resulting from clinical data have challenged this thought process, leading to a
gallstones in the ampulla of Vater was first reported by Opie in paradigm shift favoring cholecystectomy during the initial hos-
1901. From his observations in this study, an “obstructive pitalization for ABP, once the acute inflammatory process has
theory” was derived to explain the mechanism responsible for improved.
gallstone pancreatitis. Current evidence suggests that transient A U.K. study retrospectively reviewed admissions to the
stone impaction in the common channel of the pancreatic duct hospital while awaiting an elective outpatient cholecystectomy
and CBD causes increased pancreatic ductal pressure with after an episode of biliary pancreatitis (Cameron et al, 2004).
associated inappropriate activation of pancreatic enzymes Of the 58 patients awaiting laparoscopic cholecystectomy, 21%
(Hirano et al, 1993). In support of the obstructive theory, gall- had unplanned readmissions if waiting for more than 28 days;
stones can be recovered from the feces of 85% to 95% of no patients who had cholecystectomy within 28 days had recur-
patients (Acosta & Ledesma, 1974; Kelly, 1980a, 1980b), and rent admissions. The presence of a sphincterotomy did not
the incidence of CBD stones is 80% in patients undergoing affect readmission rates despite the fact that no patients with
urgent operative or endoscopic intervention compared with a ES returned with ABP, as they instead returned with cholecys-
5% to 30% incidence when the procedure is delayed (Acosta titis or biliary colic. These results were echoed by another
et al, 1978; Kelly, 1980a, 1980b; Ranson, 1979; Stone et al, European study that looked retrospectively at 80 patients with
1981). ABP who had ERCP with ES (Mador et al, 2014). Authors
The Ranson, Imrie, Glasgow, and Acute Physiology, Age, found a 60% rate of recurrent biliary complications (pancreati-
and Chronic Health Evaluation (APACHE II) assessments tis, symptomatic choledocholithiasis, colic) in patients who
provide well-established criteria for assessing the severity of delayed cholecystectomy versus 2% in the group who under-
pancreatitis and predicting local adverse events—such as necro- went early cholecystectomy (P < .0001). A recent multiinstitu-
sis, hemorrhage, infection, and pseudocyst formation—and tional randomized control trial of 266 patients with mild
systemic complications of acute respiratory distress syndrome, gallstone pancreatitis has been performed. In this study, 129
disseminated intravascular coagulation, distant fat necrosis, patients were randomly assigned to same-admission cholecys-
and renal failure (Banks, 1991). Stratifying patients by severity tectomy within 3 days, whereas 137 were randomized to inter-
based on these criteria has been helpful in directing appropriate val cholecystectomy within 25 to 30 days (da Costa et al,
management. Most patients experience mild pancreatitis result- 2015). Readmission for gallstone related complications (pan-
ing from transient impaction of a stone in the ampulla, followed creatitis, cholecystitis, cholangitis, jaundice, colic) were signifi-
by spontaneous migration into the duodenum. These patients cantly more common in the interval group than the
do well with conservative therapy alone and are unlikely to same-admission group (17% vs. 5%, P = .002). These results
benefit from urgent intervention. In contrast, it has been pro- remained significant when comparing patients with endoscopic
posed that more severe cases of pancreatitis result from persis- sphincterotomy in a subgroup analysis. As with the previous
tent stone impaction or choledocholithiasis with infected bile, retrospective studies, there was no difference in length of stay,
difficulty of surgery, conversions from laparoscopic to open Urgent ERCP resulted in a reduction of biliary sepsis from 12%
surgery, or health care use between the two groups. in the conservatively treated patients to 0% in patients undergo-
Because of this paradigm shift in the surgical management ing ERCP. Overall, no significant differences were found in the
of biliary pancreatitis, along with the change in guidelines rec- incidences of either local (10.3% vs. 12.2%) or systemic (10.3%
ommending cholecystectomy on the index admission, a retro- vs. 14.3%) complications. In the subset of patients with severe
spective review recently studied how the implementation of an pancreatitis and CBD stones, urgent ERCP resulted in a
acute-care surgery (ACS) service for biliary disease affected decrease in the combined incidence of local and systemic com-
outcomes in biliary pancreatitis (Murphy et al, 2015; Tenner plications to 21% and mortality rate to 5.3% compared with
et al, 2013). The rate of index cholecystectomy increased from the conservatively managed group, in which the rates were
2.4% to 67% (P < .001) after the implementation of the ACS 68.8% and 25%, respectively. Patients with biliary obstruction
service, which correlated with a decrease in readmission rate and cholangitis were also included in this study. A third ran-
for biliary related disease from 16.8% to 7.3% (P = .04). domized trial (Nowak et al, 1995) studied 280 patients with
Although an ACS service is not a new concept, the study dem- ABP who all underwent urgent ERCP within 24 hours of onset
onstrates the importance of inpatient consultation and prompt of their disease; nearly 25% were found to have impacted stones
evaluation by a surgical service that adheres to the current at the papilla, which were treated by immediate sphincterot-
management guidelines. omy. The remaining 205 patients were randomized to conven-
The data regarding cholecystectomy in patients with severe tional treatment or sphincterotomy regardless of the
ABP with complications such as multiorgan failure or necrosis cholangiographic findings. Compared with conventional treat-
are not as robust, and a recent Cochrane Review states that ment, the authors showed a significant advantage for patients
there is currently no evidence to support or refute early lapa- treated endoscopically with respect to morbidity (17% vs. 36%;
roscopic cholecystectomy for patients with severe acute pancre- P < .001) and mortality (2% vs. 13%; P < .001). Further review
atitis (Gurusamy et al, 2013). Usually a delay in surgery in of these findings is not possible, as the study was only published
these patients is secondary to critical illness or while awaiting in abstract form.
other surgical or endoscopic treatments for complications such Although the aforementioned studies support early ERCP
as a cyst-gastrostomy for a pseudocyst or debridement of in ABP, the patients included in these studies had associated
walled-off pancreatic necrosis. In patients who are too ill or conditions that would require an ERCP (cholangitis, jaundice),
have limiting comorbidities (severe coronary artery disease, and therefore the question of a biliary sphincterotomy in all
cirrhosis) to tolerate any type of surgery, there is a potentially patients with ABP cannot be answered by these data. A ran-
protective effect of an endoscopic sphincterotomy in preventing domized multicenter European study avoided these confound-
further biliary complications, although the data for the role of ing variables by examining the role of early ERCP with
ES in lieu of a cholecystectomy for a high-risk patient in the sphincterotomy in patients with ABP without cholangitis or
absence of choledocholithiasis are limited (May et al, 1991). jaundice (Folsch et al, 1997). Two hundred thirty-eight patients
In the acute setting, an endoscopic approach to biliary pan- were randomly assigned to early ERCP (within 72 hours) or
creatitis offers the theoretical advantage of immediate relief of conservative treatment. The overall mortality and complication
ampullary obstruction and ductal clearance without the risks of rates were similar between the two groups regardless of the
surgery. As animal models and human studies have suggested severity of pancreatitis, but the rate of serious respiratory failure
that the duration of biliary obstruction is a critical factor in was higher in the invasive group (P = .03). A recent meta-
determining the severity of pancreatitis, early resolution of analysis of the seven known well-designed, randomized,
obstruction with ERCP would theoretically ameliorate the controlled trials on this topic confirms these same findings
course of the disease (Kapetanos et al, 2010). The timing of (Uy et al, 2009).
ERCP in ABP has been controversial, and until recently, many According to the most recent high-quality published data,
studies did suggest a role for early ERCP (Fogel et al, 2014). the current practice management guidelines state that early
Three oft-quoted studies in previous guidelines and reviews ERCP is not needed in patients with ABP who lack the labora-
gave the antecedent evidence that early ERCP (within 72 hours) tory or clinical evidence of ongoing biliary obstruction or chol-
and sphincterotomy definitively reduced complications in ABP. angitis. In the absence of jaundice or cholangitis, noninvasive
The Leicester investigators (Neoptolemos et al, 1986) were methods such as MRCP and EUS should be used to screen for
the first to perform a controlled study to evaluate the efficacy choledocholithiasis (Fogel et al, 2014; Tenner et al, 2013).
of early ERCP in biliary pancreatitis. A group of 121 patients
with suspected gallstone pancreatitis was randomized to ERCP CONCLUSIONS
with stone extraction within 72 hours of hospitalization or
conventional nonendoscopic management. Patients with mild Endoscopic management of choledocholithiasis is widely
pancreatitis had similar outcomes regardless of treatment strat- accepted as a highly effective therapy for CBD stones. Endo-
egy. In the group of patients with severe pancreatitis, ERCP scopic techniques are well established, and accessories have
significantly reduced the morbidity from 61% to 24% and been developed to enhance success and safety. ERCP with ES
reduced the length of hospitalization from 17 days to 9.5 days, is the standard of care in the management of CBD stones in
but mortality was not different between the two groups. It is most clinical situations, regardless of the presence or absence
noteworthy that biliary obstruction and jaundice were not of the gallbladder. Advancement in endoscopic techniques
exclusion criteria in this study, and 11 of the patients had con- allows the management of complex bile duct stone disease. The
current cholangitis. A second study (Fan et al, 1993) prospec- endoscopic removal of stones in the perioperative period has
tively randomized 195 consecutive patients with pancreatitis of been shown to be effective, minimizing the need for CBD
any cause to urgent ERCP with ES or initial conservative treat- exploration. Patients with acute cholangitis should be consid-
ment with selective ERCP only if there is clinical deterioration. ered for urgent endoscopic management. ERCP is generally not
indicated in the management of gallstone pancreatitis in the suggest that the enthusiastic speculations of the early 1970s—
absence of associated obstructive jaundice or cholangitis. Inte- that in stone disease, the surgeon should treat the gallbladder,
grated endoscopic therapy for biliary disease is well established and the endoscopist should treat the bile duct—has come to
in centers where surgeons and endoscopists work closely pass 4 decades later.
together and provide each other with a suitable forum for criti-
cal evaluation of alternative therapeutic techniques. As argu- References are available at expertconsult.com.
ments and clinical practice come full circle, it is tempting to
A. Gallstones and Gallbladder Chapter 36C Stones in the bile duct: endoscopic and percutaneous approaches 622.e1
Cotton PB, Vallon AG: British experience with duodenoscopic sphinc-

REFERENCES terotomy for removal of bile duct stones, Br J Surg 68:373–375,
1981.
Acosta JM, Ledesma CI: Gallstone migration as a cause of acute pan- Cotton PB, et al: Endoscopic stenting for long-term treatment of large
creatitis, N Engl J Med 290:484–487, 1974. bile duct stones: 2- to 5-year follow up, Gastrointest Endosc 33:411–
Acosta JM, et al: Early surgery for acute gallstone pancreatitis: evalu- 412, 1987.
ation of a systematic approach, Surgery 83:367–370, 1978. Cotton PB, et al: Endoscopic sphincterotomy complications and their
Adamek HE, et al: Management of retained bile duct stones: a prospec- management: an attempt at consensus, Gastrointest Endosc 37:383–
tive open trial comparing extracorporeal and intracorporeal litho- 393, 1991.
tripsy, Gastrointest Endosc 44:40–47, 1996. Cuschieri A, et al: EAES ductal stone study: preliminary findings of
Adams MA, et al: Predicting the likelihood of a persistent bile duct multicenter prospective randomized trial comparing two-stage versus
stone in patients with suspected choledocholithiasis: accuracy of single-stage management, Surg Endosc 10:1130–1135, 1996.
existing guidelines and the impact of laboratory trends, Gastrointest da Costa D, et al: Same-admission versus interval cholecystectomy for
Endosc 82:88–93, 2015. mild gallstone pancreatitis (PONCHO): a multicentre randomised
Aljebreen AM, et al: Efficacy of Spyglass-guided electrohydraulic litho- controlled trial, Lancet 386:1261–1268, 2015.
tripsy in difficult bile duct stones, Saudi J Gastroenterol 20:366–370, Das A, et al: Treatment of biliary calculi using holmium:yttrium alu-
2014. minium garnet laser, Gastrointest Endosc 48:207–209, 1998.
Allen MJ, et al: Rapid dissolution of gallstones by methyl-tert-butyl Debenedet AT, Elta GH: Post-sphincterotomy bleeding: fully covered
ether, N Engl J Med 312:217–220, 1985. metal stents for hemostasis, F1000Res 2:171, 2013.
Akcakaya A, et al: Mechanical lithotripsy and/or stenting in manage- Ding G, et al: Single-stage vs. two-stage management for concomitant
ment of difficult common bile duct stones, Hepatobiliary Pancreat Dis gallstones and common bile duct stones: a prospective randomized
Int 8:524–528, 2009. trial with long-term follow-up, J Gastrointest Surg 18:947–951, 2014.
Andruilli A, et al: Incidence rates of post-ERCP complications: a sys- Deviere J, et al: Successful management of benign biliary strictures with
tematic survey of prospective studies, Am J Gastroenterol 102:1781– fully covered self-expanding metal stents, Gastroenterology 147:385–
1788, 2007. 395, 2014.
Arya N, et al: Electrohydraulic lithotripsy in 111 patients: a safe and Diaz D, et al: Methyl tert-butyl ether in the endoscopic treatment of
effective therapy for difficult bile duct stones, Am J Gastroenterol common bile duct radiolucent stones in elderly patients with naso-
99:2330–2334, 2004. biliary tube, Dig Dis Sci 37:97–100, 1992.
Banks PA: Predictors of severity in acute pancreatitis, Pancreas 6(Suppl Disario JA, et al: Endoscopic balloon dilation compared with sphinc-
1):7–12, 1991. terotomy for extraction of bile duct stones, Gastroenterology 127:
Baron TH, Harewood GC: Endoscopic balloon dilation of the biliary 1291–1299, 2004.
sphincter compared to endoscopic biliary sphincterotomy for removal DiSario J, et al: Biliary and pancreatic lithotripsy devices, Gastrointest
of CBD stones during ERCP: a meta-analysis of randomized, con- Endosc 65:750–756, 2007.
trolled trials, Am J Gastroenterol 99:1455–1460, 2004. Dowsett JF, et al: Endoscopic biliary therapy using the combined per-
Bergman JJ, et al: Biliary endoprostheses in elderly patients with endo- cutaneous and endoscopic technique, Gastroenterology 96:1180–
scopically irretrievable common bile duct stones: report on 117 1186, 1989.
patients, Gastrointest Endosc 42:195–201, 1995. Early DS, et al: Appropriate use of GI endoscopy, Gastrointest Endosc
Bergman JJ, et al: Randomized trial of endoscopic balloon dilation 75:1127–1131, 2012.
versus endoscopic sphincterotomy for removal of bile duct stones, Ellington SR, et al: Recent trends in hepatic diseases during pregnancy
Lancet 349:1124–1129, 1997. in the United States, 2002–2010, Am J Obstet Gynecol 212:524.e1–
Bergman JJ, et al: A randomized trial of endoscopic balloon dilation 524.e7, 2015.
and endoscopic sphincterotomy for removal of bile duct stones in Elmunzer BJ, et al: A randomized trial of rectal indomethacin to
patients with a prior Billroth II gastrectomy, Gastrointest Endosc prevent post-ERCP pancreatitis, N Engl J Med 366:1414–1422,
53:19–26, 2001. 2012.
Bhandari S, et al: Usefulness of single-operator cholangioscope-guided Ersoz G, et al: Biliary sphincterotomy plus dilation with a large balloon
laser lithotripsy in patients with Mirizzi syndrome and cystic duct for bile duct stones that are difficult to extract, Gastrointest Endosc
stones: experience at a tertiary care center, Gastrointest Endosc 2016; 57:156–159, 2003.
Jan 5 [Epub ahead of print]. Escourrou J, et al: Early and late complications after endoscopic
Bland KI, et al: Extracorporeal shock wave lithotripsy of bile duct sphincterotomy for biliary lithiasis, with and without the gallbladder
calculi, Ann Surg 209:743–755, 1989. “in situ.”, Gut 25:598–602, 1984.
Blumgart LH, Wood CB: Endoscopic treatment of biliary tract dis- Fan ST, et al: Early treatment of acute biliary pancreatitis by endo-
eases, Lancet 2:1249, 1978. scopic papillotomy, N Engl J Med 328:228–232, 1993.
Boerma D, et al: Wait-and-see policy or laparoscopic cholecystectomy Farrell JJ, et al: Single-operator duodenoscope-assisted cholangioscopy
after endoscopic sphincterotomy for bile duct stones: a randomized is an effective alternative in the management of choledocholithiasis
trial, Lancet 360:761–765, 2002. not removed by conventional methods, including mechanical litho-
Boey JH, Way LW: Acute cholangitis, Ann Surg 190:264–270, 1980. tripsy, Endoscopy 37:542–547, 2005.
Brandon JC, et al: Common bile duct calculi: updated experience with Fogel E, et al: ERCP for gallstone pancreatitis, N Engl J Med 370:150–
dissolution with ethyl-tertiary-butyl-ether, Radiology 166:665–667, 157, 2014.
1988. Folsch UR, et al: Early ERCP and papillotomy compared with conser-
Burdick J, et al: Holmium laser for treatment of left hepatic duct stone, vative treatment for acute biliary pancreatitis. The German Study
Gastrointest Endosc 48:523–526, 1998. Group on Acute Biliary Pancreatitis, N Engl J Med 336:237–242,
Calvo MM, et al: The rendezvous technique for the treatment of cho- 1997.
ledocholithiasis, Gastrointest Endosc 54:511–513, 2001. Foutch HPG, et al: Endoscopic placement of biliary stents for the
Cameron DR, et al: Delayed cholecystectomy for gallstone treatment of high-risk geriatric patients with common duct stones,
pancreatitis:re-admissions and outcomes, Ann R Coll Surg Engl 86: Am J Gastroenterol 84:527–529, 1989.
358–362, 2004. Freeman ML, et al: Complications of endoscopic biliary sphincterot-
Cerefice M, et al: Complex biliary stones: treatment with removal self- omy, N Engl J Med 335:909–918, 1996.
expandable metal stents: a new approach (with videos), Gastrointest Freeman ML, et al: Risk factors for post-ERCP pancreatitis: a prospec-
Endosc 74:520–526, 2011. tive, multicenter study, Gastrointest Endosc 54:425–434, 2001.
Chang WH, et al: Outcome of simple use of mechanical lithotripsy of Garcia-Cano J, et al: Fully covered self-expanding metal stents in the
difficult common bile duct stones, World J Gastroenterol 11:593–596, management of difficult common bile duct stones, Rev Esp Enferm
2005. Dig 105:7–12, 2013.
Classen M, Demling L: Endoscopische Sphinkterotomie der Papilla Geenen JE, et al: Resumé of a seminar on endoscopic retrograde
Vater, Dtsch Med Wochenschr 99:496–497, 1974. sphincterotomy (ERS), Gastrointest Endosc 27:31–38, 1981.
Cotton PB: Endoscopic management of bile duct stones (apples and Gilchrist AM, et al: Extracorporeal shockwave lithotripsy for common
oranges), Gut 25:587–597, 1984. bile duct stones, Br J Surg 84:29–32, 1997.
Gordon SJ, et al: Successful shockwave lithotripsy of bile duct stones Leung JW, Tu R: Mechanical lithotripsy for large bile duct stones,
using ultrasound guidance, Dig Dis Sci 36:1102–1109, 1991. Gastrointest Endosc 59:688–690, 2004.
Grimm H, Soehendra N: Unterspritzung zur Behandlung der Leung JW, et al: Endoscopic sphincterotomy-induced hemorrhage: a
Papillotomie-Blutung, Dtsch Med Wochenschr 108:1512–1514, 1983. study of risk factors and the role of epinephrine injection, Gastrointest
Gurusamy KS, et al: Early vs delayed laparoscopic cholecystectomy for Endosc 42:550–554, 1995.
acute gallstone pancreatitis (review), Cochrane Database Syst Rev Leung JW, et al: Mechanical lithotripsy in the common bile duct,
(9):CD010326, 2013. Endoscopy 33:800–804, 2001.
Hawes RH, et al: Follow-up 6 to 11 years after duodenoscopic sphinc- Lin LF, et al: ERCP in post-Billroth II gastrectomy patients: emphasis
terotomy for stones in patients with prior cholecystectomy, Gastro- on technique, Am J Gastroenterol 94:144–148, 1999.
enterology 98:1008–1012, 1990. Madhoun M, et al: Endoscopic papillary large balloon dilation reduces
Hirano T, et al: A possible mechanism for gallstone pancreatitis: the need for mechanical lithotripsy in patients with large bile duct
repeated short-term pancreaticobiliary duct obstruction with exo- stones: a systematic review and meta-analysis, Diagn Ther Endosc
crine stimulation in rats, Proc Soc Exp Biol Med 202:246–252, 1993. 2014:309618, 2014.
Hixson LJ, et al: Peroral cholangioscopy with intracorporeal electrohy- Mador B, et al: Early versus delayed cholecystectomy following endo-
draulic lithotripsy for choledocholithiasis, Am J Gastroenterol 87:296– scopic sphincterotomy for mild biliary pancreatitis, Surg Endosc 28:
299, 1992. 3337–3342, 2014.
Hochberger J, et al: Management of difficult common bile duct stones, Maleux G, et al: Embolization of post-biliary sphincterotomy bleeding
Gastrointest Endosc Clin N Am 13:623–634, 2003. refractory to medical and endoscopy therapy: technical results, clini-
Huang SM, et al: An alternative approach of choledocholithotomy via cal efficacy and predictors of outcome, Eur Radiol 24:2779–2786,
laparoscopic choledochotomy, Arch Surg 13:407–411, 1996. 2014.
Hui CK, et al: Does the addition of endoscopic sphincterotomy to stent Maple JT, et al: The role of endoscopy in the management of choledo-
insertion improve drainage of the bile duct in acute suppurative cholithiasis, Gastrointest Endosc 74:731–744, 2011.
cholangitis?, Gastrointest Endosc 58:500–504, 2003. Martin DF: Combined percutaneous and endoscopic procedures for
Ikeda S, et al: Endoscopic sphincterotomy: long-term results in 408 bile duct obstruction, Gut 35:1011–1012, 1994.
patients with complete follow-up, Endoscopy 20:13–17, 1988. Martin LJ, et al: Towards T-tube free laparoscopic bile duct explora-
Inamdar S, et al: Pregnancy is a risk factor for pancreatitis after endo- tion: a methodologic evolution during 300 consecutive procedures,
scopic retrograde cholangiopancreatography in a national cohort Ann Surg 228:29–34, 1998.
study, Clin Gastroenterol Hepatol 14:107–114, 2016. Masci E, et al: Complications of diagnostic and therapeutic ERCP: a
Iwashita T, et al: Endoscopic ultrasound-guided rendezvous for biliary prospective multicenter study, Am J Gastroenterol 96:417–423, 2001.
access after failed cannulation, Endoscopy 44:60–65, 2012. Mavrogiannis C, et al: Needle-knife fistulotomy versus needle-knife
Jamidar PA, et al: Endoscopic retrograde cholangiopancreatography in precut papillotomy for the treatment of common bile duct stones,
pregnancy, Am J Gastroenterol 90:1263–1267, 1995. Gastrointest Endosc 50:334–339, 1999.
Kahaleh M, et al: Safety and efficacy of ERCP in pregnancy, Gastroin- Maxton DG, et al: Retained common bile duct stones after endoscopic
test Endosc 60:287–292, 2004. sphincterotomy: temporary and long-term treatment with biliary
Kapetanos D, et al: Early ERCP in acute biliary pancreatitis: 20 years stenting, Gut 36:446–449, 1995.
of dispute, Ann Gastroenterol 23:27–30, 2010. May GR, Shaffer EH: Should elective endoscopic sphincterotomy
Katsinelos P, et al: The effect of indwelling endoprosthesis on stone replace cholecystectomy for the treatment of high-risk patients with
size or fragmentation after long-term treatment with biliary stenting gallstone pancreatitis?, J Clin Gastroenterol 13:125–128, 1991.
for large stones, Surg Endosc 17:1552–1555, 2003. Maydeo A, et al: Single operator cholanioscopy-guided laser lithostripsy
Kaw M, et al: Management of gallstone pancreatitis: cholecystectomy in patients with difficult biliary and pancreatic ductal stones, Gastro-
or ERCP and endoscopic sphincterotomy, Gastrointest Endosc 56:61– intest Endosc 74:1308–1314, 2011.
65, 2002. McAlister VC, et al: Cholecystectomy deferral in patients with endo-
Kawai K, et al: Endoscopic sphincterotomy of the ampulla of Vater, scopic sphincterotomy, Cochrane Database Syst Rev (4):CD006233,
Gastrointest Endosc 20:148–151, 1974. 2007.
Kaye GL, et al: Methyl-tert-butyl ether dissolution therapy for common Murphy P, et al: Implementation of an acute care surgery service facili-
bile duct stones, J Hepatol 10:337–340, 1990. tates modern clinical practice guidelines for gallstone pancreatitis,
Kelly TR: Gallstone pancreatitis: a prospective randomized trial of the J Am Coll Surg 221:975–981, 2015.
timing of surgery, Surgery 104:600–605, 1980a. Murray WR, et al: Choledocholithiasis: in vivo stone dissolution using
Kelly TR: Gallstone pancreatitis: the timing of surgery, Surgery 88:34– methyl tertiary butyl ether (MTBE), Gut 29:143–145, 1988.
35, 1980b. Nakajima M, et al: Five years experience of endoscopic sphincterotomy
Kelly TR, Wagner DS: Gallstone pancreatitis: a prospective random- in Japan: a collective study from 25 centers, Endoscopy 2:138–141,
ized trial of the timing of surgery, Surgery 104:600–605, 1988. 1979.
Khoo DE, et al: Laparoscopic common bile duct exploration: evolution Nealon WH, et al: Appropriate timing of cholecystectomy in patients
of a new technique, Br J Surg 83:341–346, 1996. who present with moderate to severe gallstone-associated acute pan-
Kiil J, et al: Large bile duct stones treated by endoscopic biliary drain- creatitis with peripancreatic fluid collections, Ann Surg 239:741–
age, Surgery 105:51–56, 1989. 751, 2004.
Kikuyama M, et al: ERCP after Roux-en-Y reconstruction can be Neoptolemos JP, et al: Abdominal wall staining and “biliscrotum” after
carried out using an oblique-viewing endoscope with an overtube, retroperitoneal perforation following endscopic sphincterotomy, Br
Dig Endosc 21:180–184, 2009. J Surg 71:684, 1984.
Kim U, Bosner B: Timing of surgery for acute gallstone pancreatitis, Neoptolemos JP, et al: A prospective study of ERCP and endoscopic
Am J Surg 156:393–396, 1988. sphincterotomy in the diagnosis and treatment of gallstone acute
Kim JH, et al: Endoscopic papillary large balloon dilation for the pancreatitis: a rational and safe approach to management, Arch Surg
removal of bile duct stones, World J Gastroenterol 19:8580–8594, 121:697–702, 1986.
2013. Neoptolemos JP, et al: A multivariate analysis of preoperative risk
Koornstra JJ, et al: ERCP with the balloon-assisted enteroscopy tech- factors in patients with common bile duct stones: implications for
nique: a systematic review, Dig Dis 26:324–329, 2008. treatment, Ann Surg 209:157–161, 1989.
Lai EC, et al: Endoscopic biliary drainage for severe acute cholangitis, Neoptolemos JP, et al: Methyl-tert-butyl-ether for treating bile duct
N Engl J Med 326:1582–1586, 1992. stones: the British experience, Br J Surg 77:32–35, 1990.
Lee TY, et al: Direct cholangioscopy bsed holmium laser lithotripsy of Nordback I: Management of unextractable bile duct stones by endo-
difficult bile duct stones by using an ultrathin upper endoscope scopic stenting, Ann Chir Gynaecol 78:290–292, 1989.
without separate biliary imaging, Photomed Laser Surg 30:31–36, Nowak A, et al: Final results of the prospective, randomized, controlled
2012. study on endoscopic sphincterotomy versus conventional manage-
Leese T, et al: Successes, failures, early complications and their man- ment in acute biliary pancreatitis, Gastroenterology 108:A380, 1995.
agement following endoscopic sphincterotomy: results in 394 con- Opie EL: The etiology of acute hemorrhagic pancreatitis, Bull Johns
secutive patients from a single centre, Br J Surg 72:215–219, 1985. Hopkins Hosp 12:182–188, 1901.
A. Gallstones and Gallbladder Chapter 36C Stones in the bile duct: endoscopic and percutaneous approaches 622.e3
Osborne DH, et al: Biliary surgery in the same admission for gallstone- Simmons DC, et al: Endoscopic retrograde cholangiopancreatography
associated pancreatitis, Br J Surg 68:758–761, 1981. (ERCP) in pregnancy without the use of radiation, Am J Obstet
Palmer KR, Hoffmann AF: Intraductal mono-octanoin for the direct Gynecol 190:1467–1469, 2004.
dissolution of bile duct stones: experience in 343 patients, Gut 27: Sivak MV: Endoscopic management of bile duct stones, Am J Surg
196–202, 1986. 158:228–240, 1989.
Park KT, et al: Is high-dose nafamostat mesilate effective for the pre- Soomers AJ, et al: Endoscopic placement of biliary endoprostheses in
vention of post-ERCP pancreatitis, especially in high risk patients?, patients with endoscopically unextractable common bile duct stones,
Pancreas 40:1215–2119, 2011. Endoscopy 22:24–26, 1990.
Park SJ, et al: Factors predictive of adverse events following endoscopic Stefanidis G, et al: Large balloon dilation vs. mechanical lithotripsy for
papillary large balloon dilation: Results from a multicenter series, the management of large bile duct stones: a prospective randomized
Dig Dis Sci 58:1100–1109, 2013. study, Am J Gastroenterol 106:278–285, 2011.
Picus D: Intracorporeal biliary lithotripsy, Radiol Clin North Am Stone HH, et al: Gallstone pancreatitis: biliary tract pathology in rela-
28:1241–1249, 1990. tion to time of operations, Ann Surg 194:305–312, 1981.
Ponchon T, et al: Extracorporeal lithotripsy of bile duct stones using Strasberg SM, et al: An analysis of the problem of biliary injury during
ultrasonography for stone localization, Gastroenterology 98:726–732, laparoscopic cholecystectomy, J Am Coll Surg 180:101–125, 1995.
1990. Tarantino I, et al: Fully covered self-expandable metallic stents in
Ranson JHC: The timing of biliary surgery in acute pancreatitis, Ann benign biliary strictures: a multicenter study on safety, Endoscopy
Surg 189:654–662, 1979. 44:923–927, 2012.
Reiter JJ, et al: Results of endoscopic papillotomy: a collective experi- Tenner S, et al: American College of Gastroenterology guideline: man-
ence from nine endoscopic centers in West Germany, World J Surg agement of acute pancreatitis, Am J Gastroenterol 108:1400–1415,
2:505–511, 1978. 2013.
Rhodes M, et al: Laparoscopic exploration of the common bile duct: Tham TC, et al: Safety of ERCP during pregnancy, Am J Gastroenterol
lessons learned from 129 consecutive cases, Br J Surg 82:666–668, 98:308–311, 2003.
1995. Thompson JE, et al: Factors in management of acute cholangitis, Ann
Rhodes M, et al: Randomised trial of laparoscopic exploration of Surg 195:137–145, 1982.
common bile duct versus postoperative endoscopic retrograde chol- Tomizawa Y, et al: Combined Interventional interventional radiology
angiography for common bile duct stones, Lancet 351:159–161, followed by endoscopic therapy as a single procedure of patients with
1998. failed initial endoscopic biliary access, Dig Dis Sci 59:451–458, 2014.
Rogers SJ, et al: Prospective randomized trial of LC+LCBDE vs Tranter SE, Thompson MH: Comparison of endoscopic sphincterot-
ERCP/S+LC for common bile duct stone disease, Arch Surg 145:28– omy and laparoscopic exploration of the common bile duct, Br J
33, 2010. Surg 89:1495–1504, 2002.
Rosch W, et al: Long-term follow-up after endoscopic sphincterotomy, Tse F, et al: The elective evaluation of patients with suspected cho-
Endoscopy 13:152–153, 1981. ledocholithiasis undergoing laparoscopic cholecystectomy, Gastroin-
Rustgi AK, Schapiro RH: Biliary stents for common bile duct stones, test Endosc 60:437–438, 2004.
Gastrointest Endosc Clin N Am 1:79–91, 1991. Tse F, et al: Guide wire-assisted cannulation for the prevention of
Sackman M, et al: Extracorporeal shock wave lithotripsy for clearance post-ERCP pancreatitis: a systematic review and meta-analysis,
of bile duct stones resistant to endoscopic extraction, Gastrointest Endoscopy 45:605–618, 2013.
Endosc 53:27–32, 2001. Uy M, et al: Early ERCP in acute gallstone pancreatitis without chol-
Safrany L: Endoscopic treatment of biliary tract diseases, Lancet 2:983– angitis: a meta-analysis, JOP 10:299–305, 2009.
985, 1978. Vaira D, et al: Endoscopic sphincterotomy in 1000 consecutive
Sauer B, et al: Safety and efficacy of laser lithotripsy for complicated patients, Lancet 2:431–434, 1989.
biliary stones using direct choledochoscopy, Dig Dis Sci 58:253–256, Van Dam J, Sivak MV: Mechanical lithotripsy of large common bile
2013. duct stones, Cleve Clin J Med 60:38–42, 1993.
Sauerbruch T, Stern M: Fragmentation of bile duct stones by extra- Van der Velden JJ, et al: Percutaneous treatment of bile duct stones in
corporeal shock waves: a new approach to biliary calculi after failure patients treated unsuccessfully with endoscopic retrograde proce-
of routine endoscopic measures, Gastroenterology 96:146–152, 1989. dures, Gastrointest Endosc 51:418–422, 2000.
Schumacher B, et al: Endoscopic treatment of symptomatic choledo- Vandervoort J, et al: Prospective analysis of risk factors for pancreatitis
cholithiasis, Hepatogastroenterology 45:672–676, 1998. after diagnostic and therapeutic ERCP, Gastrointest Endosc 43:400,
Seifert E, et al: Langzeitresultate nach endoskopischer Sphinkteroto- 1996.
mie: follow-up-Studie aus 25 Zentren in der Bundesrepublik, Dtsch Vandervoort J, et al: Risk factors for complications after performance
Med Wochenschr 107:610–614, 1982. of ERCP, Gastrointest Endosc 56:652–656, 2002.
Sethi S, et al: Prospective assessment of consensus criteria for evalua- Verstandig AG, et al: Combined transhepatic and endoscopic proce-
tion of patients with suspected choledocholithiasis, Dig Endosc dures in the biliary system, Postgrad Med J 69:384–388, 1993.
28:75–82, 2016. White DM, et al: Extracorporeal shock-wave lithotripsy for bile duct
Shaw MJ, et al: Results of a multicenter trial using a mechanical litho- calculi, Am J Surg 175:10–13, 1998.
tripter for the treatment of large bile duct stones, Am J Gastroenterol Wilcox CM, et al: Patterns of bleeding after endoscopic sphincterot-
88:730–733, 1993. omy, the subsequent risk of bleeding, and the role of epinephrine
Sherman S, et al: Complications of endoscopic sphincterotomy: a pro- injection, Am J Gastroenterol 99:244–248, 2004.
spective series with emphasis on the increased risk associated with Wright BE, et al: ERCP in patients with long-limb Roux-en-Y gastro-
sphincter of Oddi dysfunction and nondilated bile ducts, Gastroen- jejunostomy and intact papilla, Gastrointest Endosc 56:225–232,
terology 101:1068–1075, 1991. 2002.
Siegel JH: Endoscopic papillotomy in the treatment of biliary tract Yates J, et al: A comparison of the FREDDY and holmium lasers
disease: 258 procedures and results, Dig Dis Sci 26:1057–1064, during ureteroscopic lithotripsy, Lasers Surg Med 39:637–640, 2007.
1981. Yeh YH, et al: Percutaneous trans-hepatic cholangioscopy and litho-
Siegel J, et al: Endoscopic electrohydraulic lithotripsy, Gastrointest tripsy in the treatment of intrahepatic stones: a study with 5 year
Endosc 36:134–136, 1990. follow-up, Gastrointest Endosc 42:13–18, 1995.
CHAPTER 37
Cholecystolithiasis and stones in the common bile
duct: which approach and when?
Mark P. Callery, Rachel E. Beard, and Lygia Stewart
CHOLECYSTOLITHIASIS contraindicated in all but emergent settings in Child-Pugh class

C patients due to high complication rates (Curro et al, 2005)
Indications for Cholecystectomy (see Chapter 77).
Asymptomatic Gallstones Owing to the association between morbid obesity and cho-
Gallstones are one of the most common pathologies affecting lelithiasis, a high proportion of patients undergoing bariatric
the general population. Ten to 20 percent of the Western popu- surgery have gallbladder pathology. Patients undergoing bariat-
lation has gallstones, and the majority of patients with gall- ric surgery have a higher incidence of cholelithiasis, related both
stones, about 65% to 80%, are asymptomatic (Sakorafas et al, to obesity and rapid weight loss. Studies report a cholelithiasis
2007) (see Chapter 32). Studies of the natural history of silent incidence of 27% to 35% before bariatric operations and a 28%
gallstones have shown that symptoms develop in 1% to 2% of to 71% increase in gallstone formation following bariatric
patients per year. Among patients with asymptomatic gall- surgery (Wudel LJ, 2002). Some surgeons use bile salt medica-
stones, about 10% develop symptoms in 5 years, and about tions during periods of rapid weight change to help prevent
20% develop symptoms by 20 years. Importantly, most patients cholesterol gallstone formation; however, more recent studies
experience symptoms prior to the development of a complica- have shown that this approach is not cost-effective (Benarroch-
tion (Festi et al, 2010; McSherry et al, 1985; National Insti- Gampel et al, 2012). The question of whether or not to perform
tutes of Health [NIH] Consensus Statement Online, 1992; concomitant cholecystectomy at the time of bariatric surgery is
Stewart et al, 1989). Therefore the majority of patients with controversial, but an increasing number of studies suggest that
asymptomatic gallstones can be observed, and surgical inter- prophylactic cholecystectomy in the absence of symptoms is not
vention (laparoscopic cholecystectomy) should be offered only indicated (D’Hondt et al, 2011; Warschkow et al, 2013). In the
when symptoms develop. case of gallstone formation following bariatric surgery without
There are certain groups for which prophylactic cholecys- concomitant cholecystectomy, management should be expect-
tectomy has previously been recommended for asymptomatic ant because the majority of patients remain asymptomatic
gallstones. This is an area of controversy, however, and the (NIH Consensus Statement Online, 1992).
recommendations are changing. These populations include Several factors must be considered for potential solid-organ
solid-organ transplant patients; diabetics; patients with chronic transplant patients with asymptomatic cholelithiasis. In these
liver disease, sickle cell anemia or other chronic hemolytic patients, cholelithiasis is common, immunosuppression may
anemias; patients undergoing bariatric or other gastrointestinal increase infectious morbidity, and morbidity and mortality may
operations; and those with a potentially increased risk of gall- be increased with emergency surgery. This problem was exam-
bladder carcinoma (Table 37.1) (see Chapter 32). ined with a recent decision analysis, using probabilities and
Prophylactic cholecystectomy for asymptomatic cholelithia- outcomes derived from a pooled analysis of published studies
sis was previously recommended for patients with diabetes mel- (Kao et al, 2005). For pancreas and kidney transplant patients
litus. Studies in the late 1960s reported a higher mortality with asymptomatic cholelithiasis, however, expectant manage-
following emergency cholecystectomy in diabetic patients; ment was recommended, an approach that is widely agreed
however, subsequent meta-analysis revealed that diabetes was upon in the literature (Melvin et al, 1998). Kao and colleagues
not an independent variable. Rather, associated risk factors (2005) recommended prophylactic posttransplantation chole-
such as cardiovascular, peripheral vascular, cerebrovascular, or cystectomy for cardiac transplant recipients with asymptomatic
prerenal azotemia were associated with more severe acute cho- cholelithiasis, an approach advocated by other studies as well
lecystitis (Hickman et al, 1988; Stewart et al, 1989). More because of the increased morbidity and mortality that has been
recent series have shown similar complication rates for acute demonstrated with subsequent urgent or emergent cholecystec-
cholecystectomy among diabetic and nondiabetic patients. Dia- tomy compared with the general populace (Kilic et al, 2013).
betic patients with asymptomatic gallstones today are managed This remains an area of debate, however, as other studies have
expectantly. demonstrated that expectant management of asymptomatic
The incidence of gallstones is twice as high in patients with gallstones is safe (Takeyama, 2006).
chronic liver disease. Most of these patients remain asymptom- Asymptomatic gallstones found at an unrelated open gastro-
atic. Operative morbidity and mortality rates for patients with intestinal operation should prompt a cholecystectomy, if expo-
chronic liver disease are also significantly higher. Meta-analyses sure is adequate and if the operation can be done safely. Studies
report no increase in mortality in asymptomatic patients with of expectant management for patients with asymptomatic gall-
an expectant management approach (NIH Consensus State- stones undergoing laparotomy for other conditions have shown
ment Online, 1992; Stewart et al, 1989). Although laparo- a high (up to 70%) incidence of symptoms and/or complica-
scopic cholecystectomy has been shown to be safe in tions from the biliary system, and a significant percentage (up
well-selected Child-Pugh class A and B cirrhotic patients, it is to 40%) of patients require a cholecystectomy within 1 year of
623
from acute vasoocclusive sickle cell crisis (Curro et al, 2007).

TABLE 37.1 Management of Asymptomatic Gallstones
There is also a high incidence of choledocholithiasis in this
Patient Population Management population, and studies have demonstrated that endoscopic
Healthy adults Expectant
retrograde cholangiopancreatography (ERCP) can be safely
used in children to perform sphincterotomy and stone extrac-
Children (without Expectant
hemoglobinopathy or hemolytic tion (Al-Salem et al, 2012). In contrast, children with asymp-
anemia) tomatic gallstones caused by other etiologies can be safely
Diabetes mellitus Expectant managed expectantly, and these gallstones have been shown to
Chronic liver disease Expectant regress in 17% to 34% of cases (Curro et al, 2007).
Concomitant cholecystectomy at Only if symptomatic
Finally, gallstones have a proven association with gallbladder
time of bariatric procedure carcinoma (Tewari, 2006) (see Chapter 49). In a review of 200
Previous bariatric surgery Expectant consecutive calculous cholecystitis specimens, Albores-
Abdominal aortic aneurysm repair Expectant
Saavedra and colleagues (1980) reported that 83% exhibited
epithelial hyperplasia, 13.5% atypical hyperplasia, and 3.5%
Transplant Expectant
Kidney or pancreas Likely cholecystectomy carcinoma in situ. In areas endemic for gallbladder cancer, the
Cardiac posttransplant, but risk of carcinoma increases with larger gallstones: The relative
remains controversial risk rises from 2.4 for stones 2 to 2.9 cm in diameter to 10 for
Undergoing gastrointestinal Incidental cholecystectomy gallstones larger than 3 cm. Native Americans and patients
operation with gallbladder calcification also have a higher incidence of
Hemoglobinopathy/chronic Elective cholecystectomy gallbladder cancer. Elective cholecystectomy has been recom-
hemolytic anemia (sickle cell mended in patients with gallstones greater than 3 cm in diam-
disease, spherocytosis,
elliptocytosis, β-thalassemia) eter, but no proof is available to support that such an approach
is warranted from an oncologic standpoint (Gupta et al, 2004;
High-risk group for gallbladder Consideration for prophylactic
carcinoma (>3 cm gallstones, cholecystectomy, although Mohandas et al, 2006; Tewari 2006).
calcified gallbladder, Native- data from randomized
American race) controlled trials are lacking Symptomatic Gallstones
Approximately 30% of patients with gallstones will develop
symptoms, and once this occurs, cholecystectomy is usually
indicated for both symptomatic improvement and to prevent
the initial operation. Further, no increase in morbidity is associ- further complications. The spectrum of severity characterizing
ated with concomitant cholecystectomy (Klaus et al, 2002; symptomatic gallstones ranges from episodic pain to life-
Stewart et al, 1989). threatening infection and shock (see Chapters 32 and 33).
The management of patients with asymptomatic gallstones
undergoing abdominal aortic aneurysm (AAA) repair has BILIARY COLIC. Biliary colic is the most typical clinical presenta-
recently evolved, especially with the advent of endovascular tion of symptomatic gallstones (see Chapter 13). It usually
aortic procedures. In the past, when AAA repair and cholecys- occurs a few hours after a meal, especially one of high fat or
tectomy were open operations, concomitant cholecystectomy spice content, as a slowly progressive and constant pain that
was recommended to prevent the higher morbidity associated occurs in the epigastrium and right upper quadrant (RUQ) of
with the development of acute cholecystitis in the postoperative the abdomen and often radiates posteriorly to the scapula and
period. Studies reported no increase in graft infection or mor- right shoulder. This visceral pain likely reflects the gallbladder
bidity when cholecystectomy was performed following closure contracting against a cystic duct blocked by an impacted gall-
of the retroperitoneum; however, more recent data show similar stone. If pain persists and escalates, it can herald a worse
mortality rates with or without concomitant cholecystectomy. complication of gallstones, such as cholecystitis, cholangitis, or
Current management is typically expectant, and laparoscopic pancreatitis. Pain often remits after several hours, which can
cholecystectomy can be performed after AAA repair without create a false sense of security in some patients. More than 60%
increased morbidity if symptoms develop, although simultane- of patients will suffer recurrent pain within 2 years of their
ous laparoscopic cholecystectomy and endovascular AAA initial attack, and several studies have indicated that gallstone-
repair has been reported (Cadot et al, 2002; Pitoulias et al, associated complications occur more frequently in patients who
2012). experience biliary colic. Biliary colic is therefore the most
Children with asymptomatic gallstones comprise two main common indication for cholecystectomy.
etiologic groups: those with hemolytic anemia (sickle cell
disease, β-thalassemia, hemoglobinopathies) and those whose CHOLECYSTITIS. Acute cholecystitis occurs in about 20% of
cholelithiasis stems from some other cause (total parenteral patients with symptomatic gallstones (see Chapter 33). The
nutrition, short bowel syndrome, cardiac surgery, leukemia, pathogenesis is prolonged calculous obstruction of the cystic
lymphoma). Expectant management for children with hemo- duct with resulting inflammation. The inflamed gallbladder
lytic anemia is associated with a significant increase in morbid- becomes dilated and edematous, manifested by wall thickening,
ity and postoperative hospital stay, and elective cholecystectomy and an exudate of pericholecystic fluid can develop. If the
is therefore recommended (Curro et al, 2007). For patients gallstones are sterile, the inflammation is initially sterile, which
with sickle cell disease and asymptomatic gallstones, elective can occur in patients with cholesterol gallstones. In other cases,
cholecystectomy is advised because expectant management however, gallstone formation occurs as a result of bacterial
yields more than a twofold increase in morbidity. Further, the colonization of the biliary tree, rendering pigmented gallstones
diagnosis of acute cholecystitis can be difficult to differentiate containing bacterial microcolonies (Stewart et al, 2002). In
A. Gallstones and Gallbladder Chapter 37 Cholecystolithiasis and stones in the common bile duct: which approach and when? 625
these cases, obstruction of the cystic duct results in a contained showed a shorter hospital stay for early cholecystectomy patients
infection of the gallbladder. Research on the pathogenesis of and no significant difference in complication rates or conver-
gallstone-associated infections has shown that patients with sion rates between early laparoscopic cholecystectomy (within
bacteria-laden gallstones have more severe biliary infections. In 7 days) versus delayed laparoscopic cholecystectomy (6 to 12
addition, acute cholecystitis can coexist with choledocholithia- weeks) (Gurusamy & Samraj, 2006). Conversion rates, however,
sis, cholangitis, or gallstone pancreatitis. In the general popula- were 45% among patients randomized to the delayed group,
tion, 5% of patients presenting with cholecystitis have coexisting which required a cholecystectomy between 1 and 6 weeks. For
common bile duct (CBD) stones. In the elderly, however, this patients with significant medical problems, cholecystectomy
figure rises to 10% to 20% (Siegel & Kasmin, 1997). may need to be delayed to maximize medical therapy. Most of
The initial treatment for patients with acute cholecystitis is these patients with acute cholecystitis can be safely managed
intravenous hydration, antibiotics, and bowel rest. Many with antibiotics and bowel rest, with resolution of their acute
patients should be offered early cholecystectomy, but others will illness; they can then undergo an elective cholecystectomy once
benefit from delayed intervention, either following conservative their medical problems have been addressed.
therapy or percutaneous gallbladder drainage. Several factors
govern the approach to patients with acute cholecystitis. One GRADE II ACUTE CHOLECYSTITIS. Patients presenting with grade II
consideration is patient comorbidity; emergency cholecystec- acute cholecystitis are a more diverse group. Many will be well
tomy in patients with significant comorbidities can be associ- managed with early cholecystectomy; this is particularly true
ated with high morbidity (20% to 30%) and mortality (6% to for cases with delayed presentation as their only grade II finding.
30%) rates. Guidelines for the management of acute cholecys- In these cases, laparoscopic cholecystectomy should be per-
titis and acute cholangitis were described at an international formed, if possible, within 7 days of the acute illness. In cases
consensus meeting held in Tokyo in 2006, and updated guide- with severe local inflammation, early gallbladder drainage (per-
lines were then published in 2013 (Takada et al, 2007, 2013). cutaneous or surgical) is recommended as the initial treatment
The Tokyo Guidelines define three levels of severity for acute of choice, followed by elective cholecystectomy once the acute
cholecystitis and serve as a useful tool in the management of inflammation resolves (see Chapters 30 and 34). The key is to
acute cholecystitis (Table 37.2) (Yokoe et al, 2013). identify which patients have such an inflammatory process.
Several studies have correlated such findings as age older than
GRADE I ACUTE CHOLECYSTITIS. Patients presenting with mild 50 years, male sex, presence of diabetes, elevated bilirubin level
grade I acute cholecystitis should be offered early cholecystec- (>1.5 mg/dL), and leukocytosis (while blood cell count >
tomy, performed laparoscopically if possible. Several studies 15,000 mm3) with gangrenous cholecystitis (Nguyen et al,
have documented high success rates for laparoscopic cholecys- 2004). These findings are also frequently associated with a
tectomy when the procedure is performed within 72 hours of severe inflammatory process. Other factors suggestive of a sig-
onset of acute cholecystitis (Hadad et al, 2007; Yamashita et al, nificant inflammatory process include symptoms of gastric
2013). Further, a Cochrane Review of five randomized trials outlet obstruction. Patients with such symptoms should have
cross-sectional imaging, with either computed tomography
(CT) or magnetic resonance imaging, to determine whether a
TABLE 37.2 Tokyo Guidelines (TG13) Severity Grading for Acute severe inflammatory process is present (Fig. 37.1), followed by
Cholecystitis percutaneous cholecystostomy if such is found (Takada et al,
Grade Criteria 2007). In the updated 2013 Tokyo Guidelines, percutaneous
transhepatic gallbladder drainage remains the standard
I: Mild Acute cholecystitis that does not meet the criteria
for a more severe grade. Mild gallbladder
inflammation, no organ dysfunction
II: Moderate The presence of one or more of the following:
1. Elevated white blood cell count (>18,000/mm3)
2. Palpable tender mass in the right upper
abdominal quadrant
3. Duration of complaints > 72 hr
4. Marked local inflammation (biliary peritonitis,
pericholecystic abscess, hepatic abscess,
gangrenous cholecystitis, emphysematous
cholecystitis)
III: Severe Associated with dysfunction of any one of the
following:
1. Cardiovascular system: hypotension requiring
dopamine >5 µg/kg/min or any dose of
norepinephrine
2. Nervous system: decreased level of
consciousness
3. Respiratory system: PaO2/FiO2 ratio < 300
4. Renal system: oliguria, serum creatinine >
2.0 mg/dL
5. Hepatic system: PT-INR > 1.5 FIGURE 37.1. Computed tomographic scan demonstrating a severe
6. Hematologic system: platelet count < inflammatory process in the setting of acute cholecystitis. This patient
100,000/mm3 was treated with percutaneous cholecystostomy, followed by elective
FiO2, Forced inspiratory oxygen concentration; PaO2, partial pressure of oxygen in arterial blood; laparoscopic cholecystectomy once the inflammatory process had
PT-INR, prothrombin time/international normalized ratio. resolved.
drainage method for grade II cholecystitis that does not respond 2006). Biliary injuries are more likely to occur during difficult
to conservative therapy, although techniques such as percutane- laparoscopic operations, no different than with open operations,
ous transhepatic gallbladder aspiration and endoscopic naso- but at a higher incidence (see Chapters 38 and 42). When lapa-
biliary gallbladder drainage were also cited as alternatives roscopic cholecystectomy is performed for acute cholecystitis,
(Tsuyuguchi T, Itoi T, et al, 2013). biliary injuries occur three times more often than during elec-
tive laparoscopic cases and twice as often compared with open
GRADE III ACUTE CHOLECYSTITIS. Patients presenting with grade cholecystectomy for acute cholecystitis. Surgeons should there-
III acute cholecystitis have associated organ dysfunction. fore not hesitate to convert to an open operation if they experi-
Although this occurs rarely, approximately 6% of the time, it ence difficulties with the laparoscopic dissection or are unable
is important because these patients require intensive organ to clearly identify the critical view of safety (Strasberg et al,
support and medical treatment (Yasutoshi et al, 2013). Because 1995). Surgeons should also be aware of certain patient risk
the source of their inflammatory (septic) response and organ factors, including male gender, advanced cholecystitis, the pres-
dysfunction is the severe cholecystitis, percutaneous cholecys- ence of jaundice, and previous abdominal surgery, which are
tostomy (see Chapters 30 and 34) is necessary to treat the associated with an increased risk of conversion to open proce-
severe infection as well as the associated organ dysfunction. dure (Yamashita et al, 2013).
Numerous studies have documented the success of percutane- The decision to perform open cholecystectomy may be dif-
ous cholecystostomy in achieving control of the underlying ficult for some. Over the past 20 years, open cholecystectomy
infection within 24 to 48 hours (Howard et al, 2009). In rare has been far less frequently performed. Trainees during this
cases, urgent cholecystectomy may be required, such as cases period have less experience with open cases (Schulman et al,
with biliary peritonitis as a result of perforation of the gallblad- 2007). The experience and training needed to learn the lapa-
der; but in general, cholecystectomy in the acute phase of grade roscopic operation likely reduces the level of comfort with the
III acute cholecystitis should be avoided (Takada et al, 2007; open technique. Finally, there is the pressure related to patient
Tsuyuguchi et al, 2013). expectation for rapid recovery as well as, perhaps, the hospital
expectation for decreased length of stay and cost, as conversion
UNCOMMON PRESENTATIONS OF ACUTE CHOLECYSTITIS. Acalculous is associated with lengthier stays and increased expense (Lengyel
cholecystitis arises in the absence of cholecystolithiasis, and et al, 2012). Certain scenarios may thus arise that might subtly
associated risk factors include trauma, burns, and gastro- account in part for static biliary injury rates (Khan et al, 2007).
intestinal surgery (Crichlow et al, 2012) (see Chapter 33). Because of inexperience, the surgeon may ignore or resists the
Emphysematous cholecystitis is caused by infection with gas- sensible default option to convert to the open technique, per-
forming anerobes such as Clostridium perfringens. Diabetic sists with the laparoscopic approach, and causes injury. In other
patients are at risk, and the disease can progress quickly to instances, the surgeon overextends laparoscopic experience
profound sepsis. Emergent cholecystectomy is indicated. Gall- when disease severity warrants conversion. To prevent this,
bladder torsion can also occur when the gallbladder is especially patients need to be made fully aware that open cholecystectomy
mobile owing to a connection to the liver by a thin elongated is always a possibility, and the surgeon should not hesitate to
mesentery. Gallbladder perforation can occur as a result of seek help if needed, rather than rely on marginal laparoscopic
gallbladder wall ischemic and resulting necrosis. A localized or open cholecystectomy experience. Conversion from laparo-
perforation can result in formation of a pericholecystic abscess, scopic to open cholecystectomy is not a defeat but rather is
whereas free perforation can lead to biliary peritonitis. A biliary reflective of caution and good judgment (Jenkins et al, 2007;
fistula can also form between the gallbladder and the duode- Wolf et al, 2009).
num as a sequela of cholecystitis, and this can result in a gall- Ideally, a surgeon anticipates the likelihood of conversion on
stone ileus if a stone passes via this fistula and causes a clinical grounds. The anesthesia and operative teams should be
mechanical obstruction at the ileocecal valve (Kimura et al, so notified and prepared. Open-case instruments need to be
2013). readily available, and trocar placement should be along a pre-
drawn right subcostal incision line. Unless there is need to
Cholecystectomy Technique control bleeding, the surgeon enters the RUQ deliberately and
Choosing Laparoscopic Versus Open Techniques is not stressed by the decision to convert. Everyone should be
For typical uncomplicated symptomatic gallstone disease, lapa- ready for what lies ahead, and it should be clear to all that it
roscopic cholecystectomy is the preferred method of removing will be a difficult operation.
the gallbladder (Keus et al, 2006; Yamashita et al, 2013). Since The difficult open cholecystectomy demands adequate expo-
its origin, cholecystectomy rates have increased worldwide, sure, retraction, and identification of anatomy by dissection in
reflecting general acceptance of the laparoscopic technique. the anterior and posterior aspects of the triangle of Calot, fol-
Because the technical aspects of this operation are covered in lowed by dissection of the gallbladder off the liver bed. The
other chapters (see Chapter 35), this section will focus on surgeon achieves conclusive identification of the cystic struc-
concepts of feasibility and safety that relate to disease severity tures as the only two structures entering the gallbladder, elimi-
and the choice between laparoscopic and open cholecystectomy nating the possibility of misidentification (Callery, 2006). As
(Callery, 2006). with the laparoscopic technique, once the critical view is
Laparoscopic cholecystectomy for severe acute and chronic attained, the cystic structures can be ligated and divided. Failure
inflammation is a technically difficult and advanced operation. to achieve this critical view should prompt cholangiography to
Less experienced surgeons must recognize this and seek help define ductal anatomy. Avoidance of ductal injury in the liver
from a more experienced surgeon. Furthermore, the surgeon bed depends upon a combination of patience and staying in the
must understand that conversion to open cholecystectomy may correct plane of dissection, with meticulous technique and expe-
be necessary and is more likely in these cases (Ishizaki et al, rience. In some cases of acute cholecystitis, the gallbladder
“shells out” relatively easily from its edematous hepatic bed. In Silent Common Bile Duct Stones
other cases, and especially in chronic cholecystitis, the dissection Published reports using routine intraoperative cholangiography
of the gallbladder out of the liver bed can be tedious, frustrat- have found that at least 12% of CBD stones are clinically silent
ing, and bloody. Hemostasis can take time and may require an (Murison et al, 1993), and approximately 6% do not exhibit
argon beam, cautery, packing, and topical hemostatics. Subtotal abnormalities in liver function tests (LFTs) or in the diameter
cholecystectomy is always a valid option, especially in patients of the CBD (Majeed et al, 1999). When prospectively followed,
with cirrhosis or in those with severe inflammation that obscures data suggest that more than one third of asymptomatic stones
the anatomy within the porta hepatis. Surgeons should indicate will pass spontaneously after the first 6 weeks after cholecystec-
in operative notes for open and laparoscopic cholecystectomy tomy (Collins et al, 2004).
precisely how they identified the cystic structures for division.
For conversions, they should specify the circumstances, stressing Diagnostic Considerations
safety and surgical judgment. More often than not, the presence of CBD stones is uncertain.
Clinicians use predictive models based on risk factors that
Percutaneous Cholecystostomy include clinical features, abnormal LFTs, jaundice, and CBD
The indications for percutaneous cholecystostomy include dilation. These are very sensitive (96% to 98%) but not very
grade II acute cholecystitis with a severe inflammatory process, specific (0% to 70%) (Koo & Traverso, 1996). The American
grade III acute cholecystitis with associated organ dysfunction, Society of Gastrointestinal Endoscopy (ASGE) provides a
or acute cholecystitis in patients with severe medical morbidity practical strategy that assigns low (<10%), intermediate (10%
that limits surgical options (Hirota et al, 2007; Howard et al, to 50%), or high (>50%) risk of CBD stones by using relevant
2009; Takada et al, 2007; Tsuyuguchi et al, 2013; Yamashita risk factors (ASGE Standards of Practice Committee et al,
et al, 2007). The technical success rate of percutaneous radio- 2010).
logically guided cholecystostomy is 98% to 100% with few
procedure-related complications (mortality and major compli- Imaging Modalities: Why and When
cations, 0% to 6.5%; minor complications, 0% to 20%) Transabdominal Ultrasound
(Howard et al, 2009; Sugiyama et al, 1998). Potential compli- Transabdominal ultrasound (US) is an appropriate initial
cations include intrahepatic hematoma, pericholecystic abscess, modality in the evaluation of CBD stones (Williams et al, 2008)
and biliary peritonitis and pleural effusion caused by puncture (see Chapter 15). US can identify bile duct dilation owing to
of the liver and subsequent migration of the catheter (Yamashita stone obstruction, and it can visualize the actual stone in some
et al, 2013). cases (sensitivity 0.3, specificity 1.0). If the extrahepatic bile
duct diameter is less than 5 mm, CBD stones are exceedingly
Timing of Subsequent Operation for Cholecystitis rare, whereas a diameter greater than 10 mm with signs of
Once the inflammatory process has resolved, elective chole- jaundice predicts the presence of CBD stones in more than
cystectomy can be performed early (within 1 to 7 days) or 90% of cases (Abboud et al, 1996). Axial CT scans have better
delayed (6 to 8 weeks) with excellent success and conversion sensitivity (84%) for choledocholithiasis than US. Helical CT
rates as low as 3% (Akyurek N, 2005). The optimal timing scans outperform conventional nonhelical CT, with 88% sen-
remains controversial due to a lack of randomized controlled sitivity and 73% to 97% specificity (Tseng et al, 2008). In
trials; however, early cholecystectomy following percutaneous terms of availability, cost, and radiation exposure, US prevails
drainage may be preferable if the procedure is free of compli- as the first-line diagnostic.
cations and the patient’s condition improves (Yamashita et al, Magnetic resonance cholangiopancreatography (MRCP) is
2013). Some have reported using percutaneous cholecystos- the most accurate noninvasive modality available. MRCP is the
tomy as definitive treatment for acute cholecystitis in high- standard investigation for CBD stones for patients with inter-
risk, elderly, and debilitated patients. In patients who do not mediate probability or for those who need to be investigated to
have subsequent cholecystectomy, recurrent biliary symptoms exclude other differential diagnoses. MRCP is especially helpful
occur in 9% to 33% (Griniatsos et al, 2008; Sugiyama et al, when anatomic considerations preclude ERCP (status post–
1998). Billroth II gastrectomy, Roux-en-Y biliary bypass, duodenal
stenoses). The drawback of MRCP is its high cost, which chal-
CHOLEDOCHOLITHIASIS lenges its routine use as a more front-line diagnostic modality
(Epelboym et al, 2013).
The clinical clues of CBD stones were recognized during the
Roman Empire by Soranus of Ephesus, who described jaun- Endoscopic Retrograde Cholangiopancreatography
dice, itching, dark urine, and acholic stools. Not all CBD stones ERCP is still considered the gold standard diagnostic modality
render such a classic clinical scenario, but they still carry risk (see Chapters 20, 29, and 36C), although it is invasive, requires
if left unidentified and untreated. More than 85% of CBD radiation, and has significant complications (Andriulli et al,
stones originate in the gallbladder and secondarily migrate into 2007). Observed complications following ERCP include pan-
the CBD. For patients undergoing cholecystectomy for symp- creatitis, hemorrhage, cholangitis, perforation, and a clinically
tomatic gallstones, the prevalence of choledocholithiasis ranges relevant mortality rate (Katsinelos et al, 2014). Routine ERCP
from 10% to 18% (Dasari et al, 2013). Primary CBD stones prior to all laparoscopic cholecystectomies is impractical and
are far less common and are typically associated with conditions unnecessary. When overused, most cholangiograms are normal,
of biliary infection and stasis, such as benign biliary strictures, and costs and complication rates are prohibitive. Even in
sclerosing cholangitis, and choledochal cysts. Primary CBD patients at high risk—those with jaundice, cholestatic LFTs,
stones are more prevalent in Asians and can sometimes be CBD dilation, and a history of pancreatitis—half will not have
related to parasitic infections (Kaufman et al, 1989). CBD stones at the time of ERCP (Daradkeh et al, 2000);
therefore ERCP is now reserved more for its therapeutic than tolerated in most, a 10% rate of complications remains constant
diagnostic strength. for ERCP (Andriulli et al, 2007), with a serious morbidity rate
of 1.5% and an overall mortality rate of 0.2% to 0.5% (Freeman
Endoscopic Ultrasound et al, 1996). When surgical and endoscopic teams are inexpe-
Endoscopic US (EUS) is very sensitive for choledocholithiasis rienced with CBD stones, the perceived need for preoperative
(Garrow et al, 2007) (see Chapter 16), and a meta-analysis ERCP increases. In this setting, ERCP allows laparoscopic
reveals that EUS can reduce unnecessary diagnostic ERCP cholecystectomy to be performed quickly and with confidence
(Tse et al, 2008). A systematic review reveals that patients who that CBD stones are already managed. If ERCP fails, the surgi-
undergo EUS can avoid ERCP in 67% of cases, with fewer cal plan will need to consider intraoperative management of
complications and less pancreatitis compared with those under- choledocholithiasis. Conversely, in centers where successful
going ERCP initially (Petrov & Savides, 2009). The diagnostic sphincterotomy and stone extraction is almost assured, the rate
efficacy of EUS and MRCP compared with ERCP have revealed of preoperative ERCP will be lower. If the surgeon finds a stone
the tests to be quite comparable (Palmucci et al, 2010; Verma at operation, ERCP becomes a reliable postoperative option.
et al, 2006). Current consensus accepts the use of ERCP prior to laparo-
scopic cholecystectomy for patients with a high probability of
Intraoperative Cholangiography choledocholithiasis. It is recommended that patients with a low
Intraoperative cholangiography (IOC) during cholecystectomy or intermediate index of suspicion for choledocholithiasis
can accurately diagnose CBD stones and both minimize and undergo additional imaging techniques (MRCP, EUS, IOC) to
maximize the need for ERCP. The technique can be performed avoid unnecessary biliary instrumentation (ASGE Standards
safely in both open and laparoscopic approaches. Surgeons can of Practice Committee, 2010; Tse et al, 2004; Williams et al,
respond to such findings, flushing the duct to clear stones or 2008).
debris. Open and laparoscopic IOC can successfully be com-
pleted in about 95% of patients (see Chapters 23, 36A, and PERCUTANEOUS TRANSHEPATIC CHOLANGIOGRAPHY. Compared with
36B), with sensitivity for detecting CBD stones between 80% ERCP, percutaneous transhepatic cholangiography (PTC) is
and 92% and specificity of 93% to 97% (Machi et al, 1999). time consuming, more involved, and likely more stressful
Regardless, an ongoing debate remains whether IOC should be for a patient. It is usually reserved for patients in whom ana-
performed routinely or selectively during cholecystectomy. tomic considerations preclude safe ERCP, such as in the
When used routinely, it has high sensitivity and specificity both case of an impossible ampullary cannulation. Experienced
for suspected CBD stones and for the 3% to 4% of stones that PTC groups have reported successful stone removal rates in
are not clinically suspected but may become symptomatic post- more than 90% of cases, with complication rates around 5%
operatively. Other suggested benefits specifically relate to the (Garcia-Vila et al, 2004), although these are hardly the norms.
prevention of bile duct injuries (Fletcher et al, 1999). The PTC stone removal takes time, involving insertions of catheters
randomized trials that have been performed to address this that are upsized over time, before stones are actually retrieved
question are small, and even a systematic review of these trials with stone baskets. Consequently, there are many reports of
was not sufficiently powered to demonstrate a significant benefit attempts to make it easier. Gil and colleagues (2000) have
(Ford et al, 2012). Because no large prospective randomized reported the safety and utility of balloon dilation of the papilla
trial has answered the question of whether routine IOC is ben- in the clearing of CBD stones using occlusion balloon pushing.
eficial,, most practicing surgeons perform IOC selectively. This is also gaining popularity for the laparoscopic surgeon
during IOC.
Symptomatic Common Bile Duct Stones
The symptoms of CBD stones relate to partial or complete Surgical Approaches: Open and Laparoscopic Techniques
biliary obstruction with and without inflammatory complica- The same issues discussed in choosing open versus laparo-
tions, such as cholangitis, hepatic abscesses, or pancreatitis. In scopic cholecystectomy are accentuated when common bile
chronic scenarios, and depending on the extent and duration duct exploration (CBDE) is considered (see Chapters 31, 35,
of biliary obstruction, choledocholithiasis may also lead to sec- and 36). Today, many trainees will graduate residency having
ondary biliary cirrhosis and portal hypertension. Because of the never performed an open CBDE. Their ability to succeed with
uncertain clinical behavior and potential harmful complications, laparoscopic CBDE is quite variable, as many train with a
it is currently accepted that in the great majority of situations, default dependence on ERCP.
CBD stones should be removed, even if they are asymptomatic
(Williams et al, 2008) (see Chapters 32 and 36). Approach to Recurrent Common Bile Duct Stones
In some instances, even following successful decompression
Definitive Treatment Approaches: Biliary Obstruction and stone removal of the common duct with ERCP, patients
Catheter-Based Approaches will continue to present with recurrent choledocholithiasis.
ERCP (SEE CHAPTERS 20, 29, 36B, AND 36C). Before the laparo- Such patients present a management challenge, with varying
scopic era, ERCP was not commonly used, because open surgi- treatment options. One possible endoscopic approach is
cal bile duct clearance was superior to ERCP in terms of balloon dilation of the papilla with concurrent stone removal,
success and morbidity (Martin et al, 2006). This changed as which has been shown to decrease subsequent CBD stone
laparoscopic cholecystectomy emerged and outpaced the abili- recurrences compared with stone extraction alone (Tsai et al,
ties of most surgeons to perform laparoscopic CBD stone 2015; Yoon et al, 2014). For patients who fail nonoperative
removal. Indeed, ERCP captured and has held its role as the treatments, surgical drainage may be necessary (see Chapters
first-line approach to CBD stones, being successful in more 31 and 36A). Described approaches include choledocholithot-
than 90% of patients (Rieger & Wayand, 1995). Although well omy and T-tube drainage, choledochoduodenostomy, and
choledochojejunostomy, and there is evidence indicating that potential complications of an extra procedure (Clayton et al,
choledochoduodenostomy is the most successful approach for 2006; Martin et al, 2006).
preventing future recurrences (Li et al, 2007; Matsushima Although LCBDE can be safely performed through either
et al, 2012; Uchiyama et al, 2003). transcystic or choledochotomy approaches, most surgeons
prefer the transcystic approach. It is feasible in most cases,
CHOLECYSTECTOMY WITH INTRAOPERATIVE CHOLANGIOGRAPHY. Open saves time, does not violate the CBD, and shows no higher
and laparoscopic IOC can successfully be completed in the morbidity than standard laparoscopic cholecystectomy alone
majority of patients by the majority of surgeons (Machi et al, (Hanif et al, 2010; Tinoco et al, 2008). The most consistent
1999). IOC can be performed through the direct insertion of risk factor for failing transcystic stone clearance is the size of
contrast medium into the gallbladder or more often by intubat- the stone. Once stones exceed 5 mm, the likelihood of tran-
ing the cystic duct. Plain radiographs have largely been replaced scystic extraction falls considerably (Stromberg et al, 2008),
by digital fluoroscopic imaging. IOC is common, and most and laparoscopic choledochotomy becomes necessary. However,
surgeons receive sufficient training. Laparoscopic ultrasound many surgeons do not have the laparoscopic dissection and
cholangiography is also efficacious but not broadly used, and suturing expertise to perform this procedure; they rely instead
its utility is limited by its longer learning curve (Stiegmann on ERCP, or they convert to an open operation. Experienced
et al, 1995). Newer techniques such as hyperspectral cholangi- surgeons can remove larger or multiple CBD stones with
ography and near-infrared fluorescence cholangiography hold reported success rates of up to 90% (Lien et al, 2005).
promise and may become more widely used in the future (Bud- The question of how to close the CBD following exploration
dingh et al, 2011). remains a topic of debate. A 2007 Cochrane Review was not
Since the introduction of laparoscopic cholecystectomy, the able to conclude significant differences in outcomes for primary
debate over routine versus selective IOC has been rekindled closure of the CBD versus the routine use of T-tube drainage
because of the increased incidence of CBD injuries and the after open CBDE (Gurusamy et al, 2009). However, two 2013
inability to palpate the CBD during laparoscopy. IOC accu- Cochrane Reviews reported that T-tube drainage results in
rately defines the biliary anatomy and may protect against intra- longer operating times and hospital stays without any apparent
operative bile duct injuries (Fletcher et al, 1999) and may benefit over primary closure in both open and laparoscopic
reduce their severity (Woods et al, 1995). Opponents claim that CBDE (Gurusamy et al, 2013a, 2013b). Another 2012 meta-
routine IOC may lead to bile duct injury and unnecessary CBD analysis of randomized controlled trials confirmed the superior
explorations because of false positives and that it may add time safety and effectiveness of primary closure over T-tube drainage
and costs unnecessarily. after laparoscopic CBDE (Wu, et al, 2012) (see Chapters 31,
Selective IOC relies upon predicting the probability of cho- 36A, and 36B).
ledocholithiasis. In general, patients with a low probability
(normal LFTs, normal CBD diameter) may undergo cholecys- Gallstone Pancreatitis
tectomy with no further preoperative investigation and selective Acute gallstone pancreatitis is the most frequent form of acute
IOC. Patients with intermediate risk (isolated abnormal LFTs pancreatitis in Western countries (see Chapters 54 and 55). The
or CBD dilation) may undergo further preoperative imaging two most commonly accepted mechanisms for the pathogenesis
(MRCP) and routine IOC at an absolute minimum. Patients of gallstone pancreatitis are reflux of bile into the pancreatic
with high risk (jaundice, cholangitis) warrant confirmatory/ duct and transient ampullary obstruction caused by temporary
therapeutic ERCP (ASGE Standards of Practice Committee, impaction of a stone in the ampulla. The disease is mild in
2010) at most centers. In fact, today many patients are triaged approximately 80% of patients, but 20% experience a more
for this purpose, but in years past, many would undergo open severe clinical course that includes complications such as pan-
CBDE. Ultimately, the choice of modality depends on local creatic necrosis, multisystem organ failure, and even death
availability and expertise in minimally invasive treatments (Acosta et al, 1997; Alexakis et al, 2007; NIH, 2002) (see
coupled with considerations of cost and convenience. Chapters 55 and 56).
In many patients, the biliary obstruction has spontaneously
COMMON BILE DUCT EXPLORATION: TRANSCYSTIC VERSUS CHOLEDO- resolved at the time of presentation, so biliary decompression
CHOTOMY ACCESS. When IOC reveals CBD stones, they can be is not needed. These patients should undergo elective chole-
removed during cholecystectomy. The open choledochotomy cystectomy once the pancreatitis has resolved. At the other end
to allow cholangioscopy, flushing, forceps and balloon clear- of the spectrum are patients with gallstone pancreatitis and
ance, and T-tube placement is rarely performed or taught associated acute cholangitis (see Chapter 43). Clear evidence
today. Instead, laparoscopic common bile duct exploration shows that endoscopic biliary drainage is beneficial in patients
(LCBDE) has evolved as an efficient and more commonly used with acute cholangitis; thus these patients should have early
technique, as described in other chapters(see Chapter 36B). biliary decompression.
Several studies have shown LCBDE to be at least as efficient A secondary question is whether patients with gallstone pan-
as preoperative or postoperative ERCP in terms of stone clear- creatitis, without cholangitis, benefit from biliary decompres-
ance, morbidity, mortality, and short hospital stay, and thus sion. Clinical and experimental studies suggest that impacted
LCBDE is recommended for surgeons with appropriate skills ampullary stones and persistent biliopancreatic obstruction are
and facilities (Martin et al, 2006). A 2013 Cochrane Review of associated with a more severe clinical course. In theory, early
randomized trials actually demonstrated the superiority of both endoscopic removal of obstructing ampullary gallstones should
open and laparoscopic CBDE when compared with ERCP in improve outcomes (Acosta et al, 1997; Alexakis et al, 2007).
clearing the CBD, without any associated increased morbidity Between the late 1970s and mid-1980s, urgent surgery with
(Dasari et al, 2013). For capable surgeons, LCBDE is as safe biliary decompression was studied as a treatment of choice for
and efficient as ERCP, thus avoiding the discomfort, costs, and patients with acute gallstone pancreatitis, but this approach was
associated with an increased mortality rate among patients with The severity of the patient’s illness guides the timing of
severe pancreatitis. As such, surgical treatment during the acute intervention. Patients whose biliary obstruction has spontane-
phase of the gallstone pancreatitis is not recommended. ously resolved at the time of presentation and those who have
The role of nonsurgical intervention, prior to definitive sur- mild predicted pancreatitis should have early elective cholecys-
gical therapy, has been examined in several prospective ran- tectomy once their pancreatitis has resolved. Patients with
domized studies in which patients with cholangitis were severe predicted pancreatitis and those with associated cholan-
excluded. Integral to any interpretation of treatment approach gitis should undergo early biliary decompression (ERCP or
is the severity of the gallstone pancreatitis. These studies PTC). Among cases of escalating pancreatitis, biliary decom-
defined severe pancreatitis using a number of systems that pression should be performed within 24 to 72 hours of admis-
included the Ranson criteria (>3), Glasgow criteria (>3), or the sion (Moretti et al, 2008). For cases with associated cholangitis,
Acute Physiology and Chronic Health Evaluation (APACHE) biliary decompression should occur within 24 hours of presen-
II score (>8). The Ranson and Glasgow criteria have the advan- tation. Elective cholecystectomy can then be performed once
tage of ease of use and considerable areas of overlap (Table the severe illness has resolved.
37.3). A meta-analysis analyzed five prospective randomized
studies that examined the use of early biliary decompression in Cholangitis
cases of gallstone pancreatitis without cholangitis (Moretti Clinical findings associated with acute cholangitis include RUQ
et al, 2008). This study reported a significant reduction in abdominal pain, jaundice, fever, and chills—also known as
pancreatitis-related complications in patients with predicted Charcot’s triad (1877). Charcot’s triad demonstrates high speci-
severe pancreatitis (rate difference of 38.5%; 95% confidence ficity but low sensitivity because not all patients with cholangitis
interval, −53% to −23.9%; P < .0001), but no advantage was manifest all findings: 90% develop fever, but only about 50%
seen in cases with mild pancreatitis, and no difference in mor- to 70% of patients develop all three symptoms (Kiriyama et al,
tality rate was noted. More recently, a 2012 Cochrane Review 2013). Reynolds’ pentad (1959)—Charcot’s triad plus shock and
did not demonstrate any differences between early ERCP and altered mental status—represents a form of severe (grade III)
conservative management in pancreatitis of any severity without cholangitis, which can also manifest with multiorgan dysfunc-
evidence of cholangitis or biliary obstruction. However, in tion. Severe cholangitis is reported in 12% to 30% of patients
patients with concurrent cholangitis or biliary obstruction, early with acute cholangitis (Gouma, 2003; Kiriyama et al, 2013).
ERCP significantly reduced mortality and systemic complica- Cholangitis is a localized infection of the biliary tree, and an
tions (Tse et al, 2012). understanding of the pathophysiology of cholangitis guides
treatment decisions. Research into this disease has shown that
bacteria-laden gallstones are often the source of infection.
TABLE 37.3 Ranson and Glasgow Criteria for Severity of Acute These bacteria exist in a bacterial microcolony (biofilm) within
the pigmented matrix of gallstones (Fig. 37.2) (Stewart et al,
Pancreatitis
2002). The bacteria must detach from the biofilm to cause a
Criterion* Action localized infection, and it must reflux into the cholangiovenous
Ranson circulation to cause a more severe illness, including bacteremia
Admission
and organ dysfunction (Raper et al, 1989; Stewart et al, 2003).
Age > 55 years
Cholangiovenous reflux occurs with biliary pressures greater
WBC count > 16,000/mm 3
Admission
than 20 cm H2O, and even at this pressure, bacterial charac-
Glucose > 200 mg/dL Admission teristics (slime production) influence cholangiovenous reflux.
AST > 250 IU/L Admission In addition, bacterial breakdown by complement releases endo-
LDH > 350 IU/L Admission toxin, and this influences the induction of inflammatory cyto-
Increased BUN > 8 mg/dL 48 hr kines that drive the septic manifestations. It is important to note
PaO2 < 60 mm Hg 48 hr that although choledocholithiasis is a common cause of biliary
Calcium < 8.0 mg/dL 48 hr obstruction, benign and malignant biliary stenosis and biliary
anastomotic strictures are also etiologies (Kiriyama et al, 2013)
Base deficit < 4 mEq/L 48 hr
(see Chapters 42, 49, and 51).
Fluid sequestration≥ 6 L 48 hr
The 2013 updated Tokyo Guidelines describe three grades
Glasgow of acute cholangitis (Table 37.4): grade III is associated with
Age > 55 years Admission organ failure, grade II cases should undergo prompt early biliary
WBC count > 15,000/mm3 Admission drainage, and grade I is all others (Kiriyama et al, 2013). All
Glucose > 200 mg/dL Admission patients with suspected cholangitis should be treated with
AST/ALT > 96 IU/L 48 hr intravenous hydration and antibiotics that cover the most
LDH > 219 IU/L 48 hr common biliary organisms: Escherichia coli, Klebsiella spp.,
Enterococcus, Enterobacter cloacae, Pseudomonas spp., and anaero-
BUN > 45 mg/dL Admission
bic pathogens (see Chapter 12). Patients with severe grade III
PaO2 < 76 mm Hg Admission
cholangitis also require organ support and stabilization of organ
Calcium < 8.0 mg/dL 48 hr dysfunction.
Albumin < 3.4 g/dL 48 hr The critical component of the treatment of cholangitis is
*Mortality rates: 0 to 2 = 2%, 3 to 4 = 15%, 5 to 6 = 40%, >7 = 100%. biliary decompression. Because elevated biliary pressure drives
ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; LDH,
cholangiovenous reflux, decompression of the biliary tree with
lactate dehydrogenase; PaO2, partial pressure of oxygen in arterial blood; WBC, white blood cell ERCP is crucial (see Chapter 29); PTC may be used if ERCP
count. is not available (see Chapter 30). Not only does biliary drainage
A cm B
FIGURE 37.2. A, Black-pigment gallstones. B, Scanning electron micrograph of the black-pigment stones demonstrating bacterial microcolonies.
Note the bacterial bridges and three-dimensional nature of the biofilm. (From Stewart L, et al, 2002: The pathogenesis of pigment gallstones in
Western societies: the central role of bacteria. J Gastrointest Surg 6:891-904.)
The Tokyo Guidelines provide a useful tool for the manage-

TABLE 37.4 Tokyo Severity Assessment Criteria for Acute
ment of acute cholangitis (see Table 37.4). Patients with grade
Cholangitis
I cholangitis who respond to medical therapy can be treated
Grade Criteria with ERCP (within 24 hours), followed by definitive surgical
treatment (laparoscopic cholecystectomy), or the surgeon can
I: Mild Does not meet the criteria of grade III (severe)
or grade II (moderate) acute cholangitis at proceed to laparoscopic cholecystectomy with intraoperative
initial diagnosis LCBDE after medical stabilization (Miura et al, 2007; Takada
II: Moderate Associated with any two of the following et al, 2007). Factors guiding these choices include the patient’s
conditions: clinical findings and the surgeon’s experience with LCBDE.
1. Abnormal WBC count (>12,000/mm3, Patients with grade II cholangitis require urgent biliary decom-
<4,000/mm3)
2. High fever (≥39° C)
pression, whereas patients with severe (grade III) cholangitis
3. Age (≥75 years) require urgent endoscopic or percutaneous transhepatic biliary
4. Hyperbilirubinemia (total bilirubin ≥ 5 mg/dL) drainage following stabilization of organ dysfunction. For
5. Hypoalbuminemia (<STD × 0.7) patients with grade II or III cholangitis, the initial therapy
III: Severe Associated dysfunction in at least one of the should emphasize biliary decompression rather than definitive
following organ systems: removal of all CBD stones. Prolonged procedures with exces-
1. Cardiovascular system: hypotension
requiring dopamine >5 µg/kg/min or any
sive manipulation in an attempt to remove large stones should
dose of norepinephrine be avoided in patients with active infectious manifestations.
2. Nervous system: disturbance of Once the acute illness has resolved, early cholecystectomy can
consciousness be performed within 6 weeks of biliary decompression with no
3. Respiratory system: PaO2/FiO2 ratio < 300 increase in postoperative complications (Li et al, 2010).
4. Renal system: oliguria, serum creatinine >
2.0 mg/dL
5. Liver: PT-INR > 1.5
Need for Cholecystectomy After Endoscopic Retrograde
6. Hematologic system: platelet count < Cholangiopancreatography/Sphincterotomy
100,000/mm3 There has been considerable debate over whether cholecystec-
FiO2, Forced inspiratory oxygen concentration; PaO2, partial pressure of oxygen in arterial blood; tomy is required after ERCP/sphincterotomy for patients who
PT-INR, prothrombin time/international normalized ratio; STD, standard deviation; WBC, white initially present with choledocholithiasis (see Chapters 29 and
blood cell count. 36A). Retrospective studies have reported a low incidence of
cholecystectomy among patients with a gallbladder in situ fol-
prevent bacterial cholangiovenous reflux, it has also been shown lowing ERCP/sphincterotomy managed by watchful waiting
to be associated with a marked decreased in bile and serum (10% to 15% over 5 to 14 years) (Schreurs et al, 2004; Sugi-
endotoxin levels. In the setting of acute cholangitis, biliary yama et al, 1998). Many of these studies involved older patient
drainage via a noninvasive procedure is preferable. Drainage populations and patients with multiple medical illnesses. More
can be achieved using ERCP or PTC cannulation of the biliary recently, however, a Cochrane Review of prospective random-
tree. Open surgical drainage may be necessary in medical ized studies reported that elective cholecystectomy is recom-
centers lacking interventional radiology or gastroenterology mended to decrease mortality, recurrent biliary symptoms,
ERCP capability, or in rare cases where anatomic abnormalities and the need for repeat interventions such as ERCP and chol-
limit noninvasive approaches (Tsuyuguchi et al, 2013) (see angiography (McAlister et al, 2007). Other contemporary
Chapter 36A). studies also support this recommendation, particularly if there
is a history of pancreatitis or if more than 6 months have passed historical advances in technology? Patients now benefit from
since sphincterotomy was performed (Archibald et al, 2007). improved imaging and minimally invasive endoscopic and lapa-
roscopic techniques, especially when combined. Indeed, our
CONCLUSION treatment approaches today only occasionally resemble those
of a recent but bygone era in biliary surgery.
In 1970, who could have predicted that the common ailments
of cholecystolithiasis and choledocholithiasis would soon face References are available at expertconsult.com.
A. Gallstones and Gallbladder Chapter 37 Cholecystolithiasis and stones in the common bile duct: which approach and when? 632.e1
REFERENCES Freeman ML, et al: Complications of endoscopic biliary sphincterot-

omy, N Engl J Med 335(13):909–918, 1996.
Abboud PA, et al: Predictors of common bile duct stones prior to Garcia-Vila JH, et al: Balloon sphincteroplasty and transpapillary elimi-
cholecystectomy: a meta-analysis, Gastrointest Endosc 44(4):450– nation of bile duct stones: 10 years’ experience, AJR Am J Roentgenol
455, 1996. 182(6):1451–1458, 2004.
Acosta JM, et al: Effect of duration of ampullary gallstone obstruction Garrow D, et al: Endoscopic ultrasound: a meta-analysis of test per-
on severity of lesions of acute pancreatitis, J Am Coll Surg 184:499– formance in suspected biliary obstruction, Clin Gastroenterol Hepatol
505, 1997. Erratum in: J Am Coll Surg 1997;185:423–424. 5(5):616–623, 2007.
Akyurek N, et al: Management of acute calculous cholecystitis in high- Gil S, et al: Effectiveness and safety of balloon dilation of the papilla
risk patients: percutaneous cholecystotomy followed by early laparo- and the use of an occlusion balloon for clearance of bile duct calculi,
scopic cholecystectomy, Surg Laparosc Endosc Percutan Tech AJR Am J Roentgenol 174(5):1455–1460, 2000.
15(6):315–320, 2005. Gouma DJ: Management of acute cholangitis, Dig Dis 21(1):25–29,
Albores-Saavedra J, et al: The precursor lesions of invasive gallbladder 2003.
carcinoma: hyperplasia, atypical hyperplasia and carcinoma in situ, Griniatsos J, et al: Percutaneous cholecystostomy without interval cho-
Cancer 45:919–927, 1980. lecystectomy as definitive treatment of acute cholecystitis in elderly
Alexakis N, et al: When is pancreatitis considered to be of biliary origin and critically ill patients, South Med J 101:586–590, 2008.
and what are the implications for management? Pancreatology 7(2– Gupta SK, et al: Silent gallstones: a therapeutic dilemma, Trop Gastro-
3):131–141, 2007. enterol 25(2):65–68, 2004.
Al-Salem AH, Issa H: Laparoscopic cholecystectomy in children with Gurusamy KS, Samraj K: Early versus delayed laparoscopic cholecys-
sickle cell anemia and the role of ERCP, Surg Laparosc Endosc Per- tectomy for acute cholecystitis, Cochrane Database Syst Rev
cutan Tech 22(2):139–142, 2012. (4):CD005-440, 2006.
Andriulli A, et al: Incidence rates of post-ERCP complications: a sys- Gurusamy KS, et al: Cholecystectomy for suspected gallbladder dyski-
tematic survey of prospective studies, Am J Gastroenterol 102(8):1781– nesia, Cochrane Database Syst Rev (1):CD007-086, 2009.
1788, 2007. Gurusamy KS, et al: T-tube drainage versus primary closure after open
American Society of Gastrointestinal Endoscopy Standards of Practice common bile duct exploration, Cochrane Database Syst Rev
Committee, Maple JT, et al: The role of endoscopy in the evaluation (6):CDS005640, 2013a.
of suspected choledocholithiasis, Gastrointest Endosc 71(1):1–9, Gurusamy KS, et al: T-tube drainage versus primary closure after lapa-
2010. roscopic common bile duct exploration, Cochrane Database Syst Rev
Archibald JD, et al: The role of prophylactic cholecystectomy versus (6):CD005641, 2013b.
deferral in the care of patients after endoscopic sphincterotomy, Can Hadad SM, et al: Delay from symptom onset increases the conversion
J Surg 50(1):19–23, 2007. rate in laparoscopic cholecystectomy for acute cholecystitis, World J
Benarroch-Gampel J, et al: Cost-effectiveness analysis of cholecystec- Surg 31:1298–1301, 2007.
tomy during Roux-en-Y gastric bypass for morbid obesity, Surgery Hanif F, et al: Laparoscopic transcystic bile duct exploration: the treat-
152(3):363–375, 2012. ment of first choice for common bile duct stones, Surg Endosc
Buddingh KT, et al: Intraoperative assessment of biliary anatomy for 24(7):1552–1556, 2010.
prevention of bile duct injury: a review of current and future patients Hickman MS, et al: Acute cholecystitis in the diabetic: a case-control
safety interventions, Surg Endosc 25(8):2449–2461, 2011. study of outcome, Arch Surg 123:409–411, 1988.
Cadot H, et al: Abdominal aortic aneurysmorrhaphy and cholelithiasis Hirota M, et al: Diagnostic criteria and severity assessment of acute
in the era of endovascular surgery, Am Surg 68:839–843, discussion cholecystitis: Tokyo guidelines, J Hepatobiliary Pancreat Surg 14:78–
843–844, 2002. 82, 2007.
Callery MP: Avoiding biliary injury during laparoscopic cholecystec- Howard JM, et al: Percutaneous cholecystostomy: a safe option in the
tomy: technical considerations, Surg Endosc 20(11):1654–1658, management of acute biliary sepsis in the elderly, Int J Surg 7:94–99,
2006. 2009.
Clayton ES, et al: Meta-analysis of endoscopy and surgery versus Ishizaki Y, et al: Conversion of elective laparoscopic to open cholecys-
surgery alone for common bile duct stones with the gallbladder in tectomy between 1993 and 2004, Br J Surg 93(8):987–991, 2006.
situ, Br J Surg 93(10):1185–1191, 2006. Jenkins PJ, et al: Open cholecystectomy in the laparoscopic era, Br J
Collins C, et al: A prospective study of common bile duct calculi in Surg 94(11):1382–1385, 2007.
patients undergoing laparoscopic cholecystectomy: natural history of Kao LS, et al: Prophylactic cholecystectomy in transplant patients: a
choledocholithiasis revisited, Ann Surg 239(1):28–33, 2004. decision analysis, J Gastrointest Surg 9:965–972, 2005.
Crichlow L, et al: Acute acalculous cholecystitis after gastrointestinal Katsinelos P, et al: Risk factors for therapeutic ERCP-related complica-
surgery, Am Surg 78(2):220–224, 2012. tions: an analysis of 2,715 cases performed by a single endoscopist,
Curro G, et al: Laparoscopic cholecystectomy in Child-Pugh class C Ann Gastroenterol 27(1):65–72, 2014.
cirrhotic patients, JSLS 9(3):311–315, 2005. Kaufman HS, et al: The role of bacteria in gallbladder and common
Curro G, et al: Asymptomatic cholelithiasis in children with sickle cell duct stone formation, Ann Surg 209(5):584–591, 1989.
disease, Ann Surg 245:126–129, 2007. Keus F, et al: Laparoscopic versus open cholecystectomy for patients
Daradkeh S, et al: Selective use of perioperative ERCP in patients with symptomatic cholecystolithiasis, Cochrane Database Syst Rev
undergoing laparoscopic cholecystectomy, Hepatogastroenterology (4):CD006231, 2006.
47(35):1213–1215, 2000. Khan MH, et al: Frequency of biliary complications after laparoscopic
Dasari BV, et al: Surgical versus endoscopic treatment of bile duct cholecystectomy detected by ERCP: experience at a large tertiary
stones, Cochrane Database Syst Rev (12):CD003327, 2013. referral center, Gastrointest Endosc 65:247–252, 2007.
D’Hondt M, et al: Prophylactic cholecystectomy, a mandatory step in Kilic A, et al: Outcomes of cholecystectomy in US heart transplant
morbidly obese patients undergoing laparoscopic Roux-en-Y gastric recipients, Ann Surg 258(2):312–317, 2013.
bypass, J Gastrointest Surg 15(9):1532–1536, 2011. Kimura Y, et al: TG13 current terminology, etiology, and epidemiology
Epelboym I, et al: MRCP is not a cost-effective strategy in the manage- of acute cholangitis and cholecystitis, J Hepatobiliary Pancreat Sci
ment of silent common duct stones, J Gastrointest Surg 17(5):863– 20(1):8–23, 2013.
871, 2013. Kiriyama S, et al: TG13 guidelines for diagnosis and severity grading
Festi D, et al: Natural history of gallstone disease: expectant manage- of acute cholangitis (with videos), J Hepatobiliary Pancreat Sci
ment or active treatment? Results from a population-based cohort 20(1):24–34, 2013.
study, J Gastroenterol Hepatol 25:719–724, 2010. Klaus A, et al: Incidental cholecystectomy during laparoscopic antire-
Ford JA, et al: Systemic review of intraoperative cholangiography in flux surgery, Am Surg 68:619–623, 2002.
cholecystectomy, Br J Surg 99(2):160–167, 2012. Koo KP, Traverso LW: Do preoperative indicators predict the presence
Fletcher DR, et al: Complications of cholecystectomy: risks of of common bile duct stones during laparoscopic cholecystectomy?
the laparoscopic approach and protective effects of operative Am J Surg 171(5):495–499, 1996.
cholangiography—a population-based study, Ann Surg 229(4):449– Lengyel BI, et al: Laparoscopic cholecystectomy: what is the price of
457, 1999. conversion? Surgery 152(2):173–178, 2012.
Li ZF, Chen XP: Recurrent lithiasis after surgical treatment of elderly Schulman CI, et al: Are we training our residents to perform open
patients with choledocholithiasis, Hepatobiliary Pancreat Dis Int gallbladder and common bile duct operations? J Surg Res 142(2):246–
6(1):67–71, 2007. 249, 2007.
Li VK, et al: Optimal timing of elective laparoscopic cholecystectomy Siegel JH, Kasmin FE: Biliary tract diseases in the elderly: management
after acute cholangitis and subsequent clearance of choledocholithia- and outcomes, Gut 41:433–435, 1997.
sis, Am J Surg 200(4):483–488, 2010. Stewart L, et al: Asymptomatic gallstones: nonsurgical versus surgical
Lien HH, et al: Laparoscopic common bile duct exploration with approach, Prob General Surg 6:73–80, 1989.
T-tube choledochotomy for the management of choledocholithiasis, Stewart L, et al: Pathogenesis of pigment gallstones in Western societ-
J Laparoendosc Adv Surg Tech A 15(3):298–302, 2005. ies: the central role of bacteria, J Gastrointest Surg 6(6):891–903,
Machi J, et al: Laparoscopic ultrasonography versus operative cholan- discussion 903–904, 2002.
giography during laparoscopic cholecystectomy: review of the litera- Stewart L, et al: Cholangitis: bacterial virulence factors that facilitate
ture and a comparison with open intraoperative ultrasonography, cholangiovenous reflux and TNFα production, J Gastrointest Surg
J Am Coll Surg 188(4):360–367, 1999. 7:191–199, 2003.
Majeed AW, et al: Common duct diameter as an independent predictor Stiegmann GV, et al: Laparoscopic ultrasonography as compared with
of choledocholithiasis: is it useful? Clin Radiol 54(3):170–172, static or dynamic cholangiography at laparoscopic cholecystectomy:
1999. a prospective multicenter trial, Surg Endosc 9(12):1269–1273, 1995.
Martin DJ, et al: Surgical versus endoscopic treatment of bile duct Strasberg SM, et al: An analysis of the problem of biliary injury during
stones, Cochrane Database Syst Rev (2):CD003-327, 2006. laparoscopic cholecystectomy, J Am Coll Surg 180:101–125, 1995.
Matsushima K, Soybel DI: Operative management of recurren cho- Stromberg C, et al: Stone clearance and risk factors for failure in lapa-
ledocholithiasis, J Gastrointest Surg 16(12):2313–2317, 2012. roscopic transcystic exploration of the common bile duct, Surg
McAlister VC, et al: Cholecystectomy deferral in patients with endo- Endosc 22(5):1194–1199, 2008.
scopic sphincterotomy, Cochrane Database Syst Rev (4):CD006233, Sugiyama M, et al: Is percutaneous cholecystostomy the optimal treat-
2007. ment for acute cholecystitis in the very elderly? World J Surg 22:459–
McSherry CK, et al: The natural history of diagnosed gallstone disease 463, 1998.
in symptomatic and asymptomatic patients, Ann Surg 202:59–63, Takada T, et al: Background: Tokyo guidelines for the management of
1985. acute cholangitis and cholecystitis, J Hepatobiliary Pancreat Surg
Melvin WS, et al: Prophylactic cholecystectomy is not indicated fol- 14:1–10, 2007.
lowing renal transplantation, Am J Surg 175(4):317–319, 1998. Takada T, et al: Updated Tokyo Guidelines for the management of
Miura F, et al: Flowcharts for the diagnosis and treatment of acute acute cholangitis and cholecystitis, J Hepatobiliary Pancreat Sci
cholangitis and cholecystitis: Tokyo guidelines, J Hepatobiliary Pan- 20(1):1–7, 2013.
creat Surg 14(1):27–34, 2007. Takeyama H: Expectant management is safe for cholelithiasis after
Mohandas KM, et al: Cholecystectomy for asymptomatic gallstones heart transplant, J Heart Lung Transplant 25(5):539–543, 2006.
can reduce gall bladder cancer mortality in northern Indian women, Tewari M: Contribution of silent gallstones in gallbladder cancer, J
Indian J Gastroenterol 25(3):147–151, 2006. Surg Oncol 93:629–632, 2006.
Moretti A, et al: Is early endoscopic retrograde cholangiopancreatog- Tinoco R, et al: Laparoscopic common bile duct exploration, Ann Surg
raphy useful in the management of acute biliary pancreatitis? A 247(4):674–679, 2008.
meta-analysis of randomized controlled trials, Dig Liver Dis 40:379– Tsai TJ, et al: Role of endoscopic papillary balloon dilation in patients
385, 2008. with recurrent bile duct stones after endoscopic sphincterotomy, J
Murison MS, et al: Does selective preoperative cholangiography result Chin Med Assoc 78(1):56–61, 2015.
in missed common bile duct stones? J R Coll Surg Edinb 38(4):220– Tse F, et al: The elective evaluation of patients with suspected cho-
224, 1993. ledocholithiasis undergoing laparoscopic cholecystectomy, Gastroin-
National Institutes of Health: Gallstones and laparoscopic cholecystec- test Endosc 60(3):437–448, 2004.
tomy, NIH Consensus Statement Online 10(3):1–20, 1992. Available Tse F, et al: EUS: a meta-analysis of test performance in suspected
at: https://consensus.nih.gov/1992/1992gallstoneslaparoscopy090 choledocholithiasis, Gastrointest Endosc 67(2):235–244, 2008.
html.htm. Tse F, et al: Early routine endoscopic retrograde cholangiopancreatog-
National Institutes of Health: State-of-the-science statement on endo- raphy strategy versus early conservative management strategy in
scopic retrograde cholangiopancreatography (ERCP) for diagnosis acute gallstone pancreatitis, Cochrane Database Syst Rev (5):CD
and therapy, NIH Consens State Sci Statements 19(1):1–23, 2002. 009779, 2012.
Nguyen L, et al: Use of a predictive equation for diagnosis of acute Tseng CW, et al: Can computed tomography with coronal reconstruc-
gangrenous cholecystitis, Am J Surg 188:463–466, 2004. tion improve the diagnosis of choledocholithiasis? J Gastroenterol
Palmucci S, et al: Magnetic resonance cholangiopancreatography and Hepatol 23(10):1586–1589, 2008.
contrast-enhanced magnetic resonance cholangiopancreatography Tsuyuguchi T, Takada T, et al: TG13 indications and techniques for
versus endoscopic ultrasonography in the diagnosis of extrahepatic gallbladder drainage in acute cholecystitis (with videos), J Hepatobili-
biliary pathology, Radiol Med 115(5):732–746, 2010. ary Pancreat Sci 20(1):81–88, 2013.
Petrov MS, Savides TJ: Systematic review of endoscopic ultrasonogra- Tsuyuguchi T, Itoi T, et al: Techniques of biliary drainage for acute
phy versus endoscopic retrograde cholangiopancreatography for sus- cholangitis: Tokyo Guidelines, J Hepatobiliary Pancreat Surg
pected choledocholithiasis, Br J Surg 96(9):967–974, 2009. 14(1):35–45, 2013.
Pitoulias GA, et al: Abdominal aortic aneurysm with symptomatic cho- Uchiyama K, et al: Long-term prognosis after treatment of patients
lelithiasis: report of a case treated by simultaneous endovascular with choledocholithiasis, Ann Surg 238(1):97–102, 2003.
aneurysm repair and laparoscopic cholecystectomy, Surg Laparosc Verma D, et al: EUS vs MRCP for detection of choledocholithiasis,
Endosc Percutan Tech 22(5):e291–e293, 2012. Gastrointest Endosc 64(2):248–254, 2006.
Raper SE, et al: Anatomic correlates of bacterial cholangiovenous Warschkow R, et al: Concomitant cholecystectomy during laparoscopic
reflux, Surgery 105:352–359, 1989. Roux-en-Y gastric bypass in obese patients is not justified: a meta-
Rieger R, Wayand W: Yield of prospective, noninvasive evaluation of analysis, Obes Surg 23(3):397–407, 2013.
the common bile duct combined with selective ERCP/sphincterotomy Williams EJ, et al: Guidelines on the management of common bile duct
in 1390 consecutive laparoscopic cholecystectomy patients, Gastro- stones (CBDS), Gut 57(7):1004–1021, 2008.
intest Endosc 42(1):6–12, 1995. Wolf AS, et al: Surgical outcomes of open cholecystectomy in the lapa-
Sakorafas GH, et al: Asymptomatic cholelithiasis: is cholecystectomy roscopic era, Am J Surg 197(6):781–784, 2009.
really needed? A critical reappraisal 15 years after the introduction Woods MS, et al: Biliary tract complications of laparoscopic cholecys-
of laparoscopic cholecystectomy, Dig Dis Sci 52(5):1313–1325, tectomy are detected more frequently with routine intraoperative
2007. cholangiography, Surg Endosc 9(10):1076–1080, 1995.
Schreurs WH, et al: Endoscopic management of common bile duct Wu X, et al: Primary closure versus T-tube drainage in laparoscopic
stones leaving the gallbladder in situ: a cohort study with long-term common bile duct exploration: a meta-anlysis of randomized clinical
follow-up, Dig Surg 21(1):60–64, discussion 65, 2004. trials, Langenbecks Arch Surg 397(6):909–916, 2012.
A. Gallstones and Gallbladder Chapter 37 Cholecystolithiasis and stones in the common bile duct: which approach and when? 632.e3
Wudel LJ: Prevention of gallstone formation in morbidly obese patients Yokoe M, et al: TG13 diagnostic criteria and severity grading of acute
undergoing rapid weight loss: results of a randomized controlled cholecystitis (with videos), J Hepatobiliary Pancreat Sci 20(1):35–46,
pilot study, J Surg Res 102(1):50–56, 2002. 2013.
Yamashita Y, et al: Surgical treatment of patients with acute cholecys- Yoon HG, et al: Endoscopic papillary large balloon dilation for the
titis: Tokyo guidelines, J Hepatobiliary Pancreat Surg 14:91–97, 2007. management of recurrent difficult bile duct stones after previous
Yamashita Y, et al: TG13 surgical management of cholecystitis, J Hepa- endoscopic sphincterotomy, Dig Endosc 26(2):259–263, 2014.
tobiliary Pancreat Surg 20(1):89–96, 2013.
Yasutoshi K, et al: TG13 current terminology, etiology, and epidemiol-
ogy of acute cholangitis and cholecystitis, J Hepatobiliary Pancreat Sci
20(1):8–23, 2013.
CHAPTER 38
Postcholecystectomy problems
Major Kenneth Lee IV and Charles M. Vollmer Jr.
OVERVIEW pain in 70% to 90% of these patients (Berger et al, 2003;

Fenster et al, 1995, Gui et al, 1998). Impacted gallstones that
Cholecystectomy for the treatment of gallbladder disease has obstruct bile outflow through the cystic duct result in acute
become the most common abdominal operation in westernized cholecystitis. This typically manifests with fevers, right upper
countries. Carl Langenbuch is credited with performing the quadrant pain, leukocytosis, nausea, and vomiting. Laparo-
first open cholecystectomy (OC) in 1882. Laparoscopic chole- scopic cholecystectomy is also the preferred approach in this
cystectomy (LC) was first performed in 1985 (Mühe, 1986) setting, although conversion to an open procedure occurs in
and has now surpassed the open approach as the “gold stan- 10% to 20% of cases (Gutt et al, 2013; Ingraham et al, 2010).
dard” operation due to significant decreases in operative time, Carcinoma of the gallbladder is an uncommon problem,
patient pain, hospital stay, and costs (Begos & Modlin, 1994; with approximately 9850 new cases diagnosed in the United
Keus et al, 2006). At present, approximately one million lapa- States each year (Siegel et al, 2012) (see Chapter 49). Chole-
roscopic cholecystectomies are performed annually in the cystectomy is one component of optimal treatment. Unfortu-
United States (Tsui et al, 2013). Ninety-six percent of all cho- nately, gallbladder cancer is associated with an extremely poor
lecystectomies are performed in a minimally invasive fashion prognosis with 5 year survival rates of 5% to 10% and an overall
(Tsui et al, 2013) (see Chapter 35). median survival of 3 to 6 months from diagnosis (Hueman
Minimally invasive approaches have both revolutionized the et al, 2009). Because of both the poor prognosis of gallbladder
treatment of patients with gallbladder disease for the better and cancer and the difficulty delineating benign from malignant
changed the profile of cholecystectomy. On the whole, chole- gallbladder polyps, cholecystectomy is often used in the treat-
cystectomy patients now tend to be younger and healthier, ment of gallbladder polyps. Gallbladder polyps are present in
whereas patients undergoing open cholecystectomy tend to be 0.3% to 12% of the population, and size is the factor most
older and less well, and the open operation is often performed associated with malignancy (Koga et al, 1988; Sarkut et al,
in higher-risk, complicated scenarios (Khan et al, 2007; Nilsson 2013; Shoup & Fong, 2002).
et al, 2005; Shea et al, 1998) (see Chapter 35). Laparoscopic Several more contentious indications for cholecystectomy
cholecystectomy itself continues to advance; widespread preva- fall under the heading of functional biliary disorders. Patients
lence and familiarity with this procedure has led surgeons to with persistent episodes of prolonged right upper quadrant
experiment with more advanced minimally invasive approaches, abdominal pain without a defined etiology (i.e., cholelithiasis)
such as single-incision and robotic cholecystectomy (Himpens are sometimes evaluated for and diagnosed with one of these
et al, 1998; Navarra et al, 1997). Indications are also broaden- issues. These nebulous conditions include gallbladder dysfunc-
ing, as evidenced by the dramatic increase in the number of tion and sphincter of Oddi dysfunction (SOD). Unfortunately,
procedures performed for acalculous disease (Johanning & although criteria for the diagnosis of these disorders exist
Gruenberg, 1998). (Behar et al, 2006; Geenen et al, 1989), they are not always
The decreasing threshold for cholecystectomy in the laparo- used. Furthermore, the nomenclature for these diagnoses is
scopic era has resulted in increased numbers of rare and unique variable in its application. Various studies suggest that patients
complications, such as common bile duct (CBD) injuries, with motility disorders of the gallbladder may benefit from
spilled gallstones, and port-site hernias. Furthermore, just as cholecystectomy, with reported rates of symptom resolution of
with open cholecystectomy, patients with abdominal pain of 66% to 100% (Chen et al, 2001; Freeman et al, 1975; Gous-
uncertain etiology who undergo LC often continue to be per- sous et al, 2014; Westlake et al, 1990).
sistently symptomatic afterward. This chapter will address the These indications and others warranting cholecystectomy
causes of postcholecystectomy problems (PCPs) along with have been summarized in the Society of American Gastro-
optimal diagnostic and treatment strategies. intestinal and Endoscopic Surgeons guidelines for the applica-
tions of laparoscopic biliary surgery (Overby et al, 2010).
INDICATIONS FOR CHOLECYSTECTOMY Perhaps not surprisingly, low rates of morbidity and mortality
with minimally invasive approaches have extended the applica-
Symptomatic cholelithiasis remains the most common indication of laparoscopic cholecystectomy to patients who have no
tion for cholecystectomy (see Chapter 32). Gallstones are clear indication for the procedure. A recent retrospective series
present in approximately 15% of the population in the United found that approximately 20% of patients underwent cholecys-
States and Europe (Sakorafas et al, 2007). At least half of tectomy for questionable indications, including vague symp-
patients are asymptomatic at diagnosis. In the majority of toms unlikely to be related to gallstones, personal request, and/
symptomatic patients, gallstones provoke epigastric or right or cholelithiasis identified only on FAST (focused assessment
upper quadrant pain that often occurs after fatty meals (see with sonography for trauma) examination (Pulvirenti et al,
Chapter 13). This pattern of symptoms is most commonly 2013). This appears to be one drawback of the technologic
treated with cholecystectomy, which is successful in eliminating advances that have decreased morbidity.
633
OVERALL MORBIDITY OF CHOLECYSTECTOMY these issues is therefore mandatory. PCPs include a heteroge-
neous group of issues ranging from traditional complications
Several large series have quantified the morbidity and mortality that are easily identified and quantified in large studies to more
of laparoscopic and open cholecystectomy in the modern era. nebulous, ill-defined issues that are difficult to detect and esti-
In 1995, Jatzko and colleagues (1995) published a multivariate mate. In our view, these diagnoses can be divided into three
comparison of postcholecystectomy complications demonstrat- groups: technical problems apparent in the intraoperative or
ing higher morbidity (7.7% vs. 1.9%) and mortality (5% vs. perioperative setting (immediate technical problems), technical
1%) for open cholecystectomy versus laparoscopic cholecystec- problems that often take months to years to manifest (delayed
tomy. Soon thereafter, a meta-analysis comparing 98 studies of technical problems), and functional problems that are often
laparoscopic cholecystectomy with 28 studies of open cholecys- unrelated to the operation and are sometimes present preopera-
tectomy measured the rate of mortality for laparoscopic chole- tively (functional problems). Notably, overlooked extrabiliary
cystectomy to be 0.086% to 0.16% versus 0.66% to 0.74% for disorders may be the most common cause of postcholecystec-
an open procedure (Shea et al, 1996). An all-cause morbidity tomy pain (Jaunoo et al, 2010). Table 38.1 summarizes diag-
rate was not calculated in this study. More recently, an analysis noses fitting each group of PCPs.
of 65,511 cholecystectomies performed at American College of
Surgeons–National Surgical Quality Improvement Program
(ACS-NSQIP) hospitals found the rates of morbidity, serious IMMEDIATE TECHNICAL PROBLEMS
morbidity, and mortality for laparoscopic cholecystectomy to
be 3.1%, 1.4%, and 0.27%, respectively (Ingraham et al,
Bile Leak and Bile Duct Injury
2010). The same rates for open cholecystectomy were 17.8%, The incidence of bile leak after laparoscopic cholecystectomy
11.1%, and 2.8%, respectively. Rates of major in-hospital com- appears to be higher than for open cholecystectomy (See Chap-
plications (6.8%) and mortality (0.5%) have also been identi- ters 27 and 42). It has been estimated at 1.1% to 4.0% (Agarwal
fied from the U.S. Nationwide Inpatient Sample (Murphy et al, et al, 2006; Barkun et al, 1997; Binmoeller et al, 1991; Bjork-
2010). Although these series report complications rates for all man et al, 1995; Davids et al, 1992; Kim & Kim, 2014). The
patients who underwent cholecystectomy, the Swiss Associa- majority of bile leaks following laparoscopic cholecystectomy
tion of Laparoscopic and Thoracoscopic Surgeons (SALTS) are from two sources: the cystic duct stump and aberrant
database defined the rates of intraoperative complications branches of hepatic ducts, including ducts of Luschka (Barkun
(7%), postoperative local complications (4%), and postopera- et al, 1997; Bergman et al, 1996; Kim & Kim 2014; Rustagi &
tive systemic complications (2.3%) following laparoscopic cho- Aslanian 2014; Ryan et al, 1998; Tewani et al, 2013; Way et al,
lecystectomy for only those patients diagnosed with acute or 2003). Abdominal pain, fever, ascites, and jaundice are the
chronic cholecystitis (Giger et al, 2006). The mortality rate in most common presentations (Agarwal et al, 2006; Barkun et al,
this study was 0.32%. 1997; Bjorkman et al, 1995; Ferriman, 2000; Kim et al, 2010;
Such national database studies have also been used to iden- Kim & Kim, 2014; Pawa & Al-Kawas, 2009). Bilious drainage
tify risk factors for complications following cholecystectomy. from operative drains or percutaneous drains placed postopera-
Murphy and colleagues (2010) showed on multivariate analysis tively typically confirms the diagnosis. The majority of these
that risk factors for complications following laparoscopic cho- leaks can be successfully managed with drainage alone. Endo-
lecystectomy included advanced age, male gender, and comor- scopic procedures aimed at eliminating the pressure gradient
bidities. Surgeon experience and hospital volume were not risk across the sphincter of Oddi to create preferential flow of bile
factors for morbidity. In their analysis, the SALTS group identi- into the duodenum are commonly used in this setting, and
fied patient characteristics (gender, age, comorbidity, body although multiple studies have reported high success rates asso-
weight), clinical findings (acute vs. chronic cholecystitis), and ciated with endoscopic sphincterotomy in the management of
surgeon experience as predictors of perioperative complication minor bile leaks, ranging from 80% to 100% (Katsinelos et al,
(Giger et al, 2006). 2008; Mavrogiannis et al, 2006; Ryan et al, 1998), there is no
clear proof that these procedures are superior to drainage alone.
POSTCHOLECYSTECTOMY PROBLEMS: It is clear, however, that surgery is rarely required to treat most
OVERVIEW bile leaks and is now largely reserved for the most feared com-
plication of cholecystectomy—major bile duct injury.
Persistence or recurrence of symptoms following cholecystec- Despite the clear advantages of laparoscopic versus open
tomy occurs as much as 40% of the time (Jaunoo et al, 2010; cholecystectomy, CBD injuries have traditionally been seen
Zhou et al, 2003). A comprehensive understanding of both more frequently in the laparoscopic setting (0.3% to 0.5% vs.
potential PCPs and strategies for diagnosing and managing 0.1% to 0.2%) (Southern, 1991; Adamsen et al, 1997; Deziel
TABLE 38.1 Summary of Postcholecystectomy Problems

Immediate Technical Problems Delayed Technical Problems Functional Problems Extrabiliary Disorder
Bile leak Biliary stricture SO dysfunction Peptic ulcer disease

Bile duct injury Spilled gallstones Postcholecystectomy diarrhea Liver disease
Hemorrhage Remnant GB/CD lithiasis GERD
Infection IBS
Retained CBD stone Postsurgical adhesions
Psychosomaticdisorder
CBD, Common bile duct; CD, cystic duct; GB, gallbladder; GERD, gastroesophageal reflux disease; IBS, inflammatory bowel syndrome; SO, sphincter of Oddi.
A. Gallstones and Gallbladder Chapter 38 Postcholecystectomy problems 635
et al, 1993; Harboe & Bardram, 2011; Nuzzo et al, 2005; junction of the cystic and common hepatic ducts (Duca et al,
Tantia et al, 2008; Vecchio et al, 1998; Waage & Nilsson, 2003).
2006). Advanced minimally invasive procedures such as single- Bile duct injuries in which biliary-enteric continuity persists
incision LC may be associated with even higher rates of bile can be managed endoscopically, but the mainstay of treatment
duct injury (Joseph et al, 2012). This may have to do in part for major bile duct injuries remains surgery (see Chapter 42).
with experience. In a series of 350 laparoscopic cholecystecto- Because bile duct injuries are often unrecognized at the time
mies reviewed by Huang and colleagues (1993), CBD injuries of the index operation, definitive repair is often conducted
more often occurred in the first 10 to 15 operations performed remotely. Multiple reports suggest that surgical repair of these
by the surgeon. Similarly, a multiinstitutional report found that injuries is highly successful in both immediate and delayed
90% of bile duct injuries occurred within an individual sur- cases (Lillemoe et al, 2000; Pekolj et al, 2013; Perera et al,
geon’s first 30 cases (Moore & Bennett, 1995). Universal adop- 2011; Sicklick et al, 2005). Large, single-institution reports
tion of the laparoscopic approach now allows the learning curve have demonstrated success rates as high as 98% with low rates
to be more easily surmounted during surgical training. Indeed, of mortality, major morbidity, reoperation, and anastomotic
the most recent data suggest that the rate of bile duct injury leak. This includes more recent reports focused on repairs
may be decreasing with more laparoscopic experience (Grbas performed immediately (Pekolj et al, 2013; Perera et al, 2011).
et al, 2013) and that laparoscopy is no longer associated with The most frequently used options for repair are hepaticojeju-
an increased risk of bile duct injury (Fullum et al, 2013). nostomy (nearly ubiquitous in the delayed setting) and primary
Regardless, a significant fraction of injuries are purely technical repair versus a T-tube. Multiple studies have shown that bile
and unrelated to experience (Archer et al, 2001). duct injuries should be managed by experienced hepatobiliary
The most common reason for major bile duct injury is surgeons because repairs performed by nonhepatobiliary sur-
failure to identify the anatomy of the triangle of Calot (Stras- geons have significantly increased rates of morbidity and failure
berg et al, 1995). Often, the CBD is mistaken for the cystic (Melton & Lillemoe, 2002; Perera et al, 2011; Stewart & Way,
duct. Less commonly, an aberrant duct is misidentified as the 1995). Surgical management of bile duct injuries and strictures
cystic duct. Bile duct injury has been minimized by application is discussed in greater detail elsewhere in this text.
of the “critical view of safety” in laparoscopic cholecystectomy
(Strasberg et al, 1995; Strasberg & Brunt, 2010). The critical Hemorrhage
view requires clearance of fat and fibrous tissue from the tri- Several large series have defined the risk of perioperative bleed-
angle of Calot, separation of the gallbladder from the lower ing in patients undergoing cholecystectomy. An ACS-NSQIP
third of the cystic plate, and that two and only two structures study of 65,511 North American cholecystectomies quantified
be seen entering the gallbladder. The advantage of applying the the risk of bleeding for laparoscopic and open cholecystectomy
critical view is that the surgeon cannot proceed to the division as 0.08% and 0.54%, respectively. The overall rate of bleeding
of structures if the anatomy is misidentified. However, when requiring transfusion in the series was 0.12% (Ingraham et al,
and whether the critical view has been met can be a matter of 2010). A Swedish registry study of 48,010 cholecystectomies
contention. For this reason, the use of intraoperative photog- found that bleeding necessitating transfusion, reoperation, con-
raphy to document the critical view from both anterior and version to open, and/or other measures prolonging hospital stay
posterior perspectives has been suggested (Fig. 38.1) (Sanford occurred in 2.1% of patients (Persson et al, 2012).
& Strasberg, 2014). In addition to the classic direct CBD inju- Bleeding typically occurs from one of several sites in chole-
ries caused by misidentification of the cystic and CBDs, clip- cystectomy. In a retrospective analysis of 9542 consecu-
ping of a “tented” CBD may occur in LC due to excessive tive laparoscopic cholecystectomies by Duca and colleagues
traction of the CBD into the clip itself. These injuries can lead (2003), intraoperative hemorrhage occurred in 224 patients,
to stricture and obstruction of the CBD, and may be avoided for an incidence of 2.3%. The etiology of bleeding was most
by placing the first clip with visualized distance from the commonly from the gallbladder bed and treated with a
Anterior view Posterior view
FIGURE 38.1 Doublet photographs of the critical view of safety. This figure demonstrates complete clearance of the hepatocystic triangle, the
presence of only two structures going into the gallbladder, and separation of the gallbladder from the lower one third of the cystic plate. (Image
courtesy Steven M. Strasberg.)
fibrin-collagen patch in their experience. Tangential lesions of controlled trial of drain use in the setting of acute cholecystitis
the cystic artery or, less commonly, total sectioning of the cystic also showed no difference in intraabdominal infection rates
artery occurred in 95 out of 224 cases of bleeding. In the major- (Park et al, 2015).
ity of cases, the intraoperative bleeding was controlled laparo-
scopically with hemostatic clips. Damage to the hepatic artery Retained Common Bile Duct Stones
occurred in one instance and required immediate conversion The incidence of CBD stones in patients undergoing elective
to an open procedure. Finally, bleeding from the greater cholecystectomy is 5% to 15% (Clayton et al, 2006; Ebner
omentum was seen in 18 cases and controlled laparoscopically et al, 2004; Petelin, 2003; Velanovich et al, 2006a, 2006b;
in 16 cases. Another source of massive intraoperative blood loss Vezakis et al, 2000) (see Chapters 32, 36, and 37). Before lapa-
is from inadvertent incursion into a deep plane of hepatic roscopy became the standard approach to cholecystectomy, the
parenchyma where distal tributaries of the middle hepatic vein presence of CBD stones often had minimal effect on patient
may be encountered. In fact, 10% of patients harbor large management and outcome. This is partly because endoscopic
branches of the middle hepatic vein directly adjacent to the retrograde cholangiopancreatography (ERCP) followed by
gallbladder fossa, which may lead to significant hemorrhage in open cholecystectomy failed to carry significant benefit over
instances of even mild parenchymal dissection (Ball et al, open cholecystectomy with bile duct exploration (Neoptolemos
2006). et al, 1987; Stain et al, 1991; Stiegmann et al, 1992). The shift
It should be recognized that management of profuse bleed- toward laparoscopic cholecystectomy has dramatically altered
ing during cholecystectomy can be fraught with significant the management of patients suspected of having CBD stones.
ramifications. An autopsy study has demonstrated that approxi- Stratified approaches are now generally used in which high-risk
mately 7% of cadavers having undergone cholecystectomy had patients are sent for ERCP, moderate-risk patients for magnetic
evidence of injury to the right hepatic artery or its branches resonance cholangiopancreatography (MRCP), and low-risk
(Halasz, 1991). Although this alone appears to be well toler- patients proceed directly to surgery (Shapey et al, 2012).
ated, combined injuries to the right hepatic artery and bile duct Despite these advances in the identification and clearance of
harbor far more significant consequences (Stewart et al, 2004; CBD stones preoperatively, the incidence of retained CBD
Strasberg & Helton, 2011). Nonetheless, death remains uncom- stones following cholecystectomy still ranges from 0.2% to
mon in this setting. By contrast, extreme vasculobiliary injuries 2.3% (Andrews, 2013; Hamad et al, 2011; Sanjay et al, 2011;
involving injury to a portal vein, hepatic artery, and bile duct Shapey et al, 2012). In these instances, it is often difficult to
are most severe and often result in death (Strasberg & Gouma, decipher whether the retained stones are a consequence of
2012). When seen, these injuries have often occurred despite intraoperative gallbladder manipulation or incorrectly inter-
conversion to an open procedure. The surgeon must be cogni- preted studies.
zant of the fact that the anatomy of the hilum may be severely The majority patients with of retained CBD stones are seen
distorted in the face of severe inflammation, as the fundus- within 6 weeks to 1 year, most commonly with abdominal pain
down technique has been used frequently in these cases. and/or jaundice (Sanjay et al, 2011). Endoscopic sphincterot-
omy is the standard approach for CBD stone extraction, with
Infection a success rate of approximately 87% to 100% (Bergman et al,
Postoperative infection after cholecystectomy is a rare event. In 1997; Granke et al, 1988; Gupta et al, 2008; Kawai et al, 1974;
a 7-year study of 54,504 patients conducted by the Centers for Mo et al, 2002; Ponchon et al, 1989) (see Chapters 29 and
Disease Control and Prevention, the rate of surgical-site infec- 36C). Notably, studies comparing the use of preoperative,
tions (SSI) was significantly lower in the laparoscopic approach intraoperative, and postoperative ERCP in relevant clinical sce-
compared with open cholecystectomy (0.62% vs. 1.82%, narios have not demonstrated superiority of any one approach
respectively) (Richards et al, 2003). The majority of infections (Chang et al, 2000; Wright et al, 2002). Thus no single algo-
in the open approach were seen in superficial spaces, whereas rithm exists for the management of suspected CBD stones.
infections in the laparoscopic setting were less frequent but
were more commonly in organ spaces. As expected, the SSI DELAYED TECHNICAL PROBLEMS
rates were higher in patients taken to the operating room emer-
gently or with an American Society of Anesthesiologists score Spilled Gallstones
of 3 or higher. Few reports of complications from gallstones spilled during
Neither perioperative antibiotic prophylaxis nor routine open cholecystectomy exist (Jacob, et al, 1979; Rothlin et al,
drainage improves the rate of infectious complications follow- 1997). This is because gallbladder spillage during open chole-
ing cholecystectomy. That antibiotic prophylaxis carries no cystectomy is more easily controlled, and dropped gallstones
benefit in lower-risk patients undergoing elective cholecystec- are more likely to be identified and retrieved. Perforation of the
tomy has been shown in multiple studies (Chang et al, 2006; gallbladder during laparoscopic cholecystectomy is more
Harling et al, 2000; Koc et al, 2003; Tocchi et al, 2000; Uludag common and less controlled. Estimates of gallbladder perfora-
et al, 2009). More recently, a gallstone surgery registry was tion and stone spillage range from 6% to 40% in laparoscopic
used to demonstrate that prophylactic antibiotics also carry no cholecystectomy (Brockmann et al, 2002; Helme et al, 2009;
benefit in acute cholecystectomy (Jaafar et al, 2014). Postop- Schafer et al, 1998; Soper & Dunnegan, 1991). It can occur
erative infections occurred in 6.1% of patients who received no for many reasons, including excessive retraction during dissec-
antibiotic prophylaxis compared with 8.4% of patients treated tion, direct puncture with an instrument, and removal of a
with antibiotics. distended gallbladder through a trocar site. This, combined
A Cochrane Review summarizing 12 trials concluded that with increased difficulty in identifying and retrieving spilled
the use of drains in uncomplicated cholecystectomy is unwar- stones, has dramatically increased complications from gallblad-
ranted (Gurusamy et al, 2007), and a recent randomized der perforation in the laparoscopic era.
A. Gallstones and Gallbladder Chapter 38 Postcholecystectomy problems 637
A meta-analysis of eight studies that, with more than 500 et al, 1998; Walsh et al, 2002). In this instance, it is sometimes
laparoscopic cholecystectomies, report on lost gallstones esti- difficult to delineate whether a stone has persisted in a long
mates the incidence of lost stones during LC to be approxi- cystic duct remnant or whether there has been relapse of lithia-
mately 2%; these stones caused complications 8.5% of the time sis in a gallbladder remnant.
(Zehetner et al, 2007). The most frequent complication of The prevalence of remnant cystic duct lithiasis is difficult to
spilled stones is abscess—either intraabdominal or of the estimate. Many reports are surgical series and thus report only
abdominal wall (Dobradin et al, 2013; Horton & Florence, on patients in whom the finding has been managed surgically.
1998; Wilton et al, 1993; Zehetner et al, 2007). Less common However, anecdotal case reports suggest that incomplete gall-
manifestations include fistulas, liver abscesses, and bacteremia. bladder removal is more common in the laparoscopic than open
Stones have even been reported to erode into the chest cavity, era. This may partly reflect the prevalence of subtotal cholecys-
causing empyema and broncholithiasis with expectoration. tectomy in the era of minimally invasive surgery. Laparoscopic
Although the spillage of gallstones is obviously an intraop- subtotal cholecystectomy technique was first reported in the
erative event, complications of spilled gallstones can be termed early 1990s (Bickel & Shtamler, 1993) and has been shown to
a late technical issue in that they often materialize weeks to be relatively safe (Ji et al, 2006; Sharp et al, 2009). However,
months after the procedure. Reoperation for morbidity from a systematic review of the technique used in this procedure
stone spillage has been reported as late as 15 years after chole- shows that a number of surgeons purposefully leave behind a
cystectomy (Arishi et al, 2008; Gooneratne, 2010). Because gallbladder pouch that often is sutured or stapled closed
complications from spilled stones are the exception rather than (Henneman et al, 2013). This tactic is contrary to that described
the rule, we do not recommend converting to an open proce- for open subtotal cholecystectomy in which the posterior wall
dure in the event of gallbladder perforation. However, efforts of the gallbladder is left attached to the liver and the cystic duct
should be made to retrieve stones with suctioning, forceps, and is secured at its origin with a purse-string technique (Bornman
irrigation. It is also a good policy to document the event and & Terblanche, 1985). Although the alternative laparoscopic
to inform the patient that stone spillage occurred in the event approaches to subtotal cholecystectomy are more convenient,
of latent complications and/or confounding imaging findings. they do increase the risk of relapsing lithiasis. In one series, the
Notably, spilled gallstones have even been noted to mimic incidence of cystic duct remnant lithiasis was 4.19% in the
hepatic and peritoneal metastases (Arai et al, 2012; Dasari setting of laparoscopic subtotal cholecystectomy compared
et al, 2009; Rammohan et al, 2012). with 0.02% in conventional laparoscopic cholecystectomy
(Palanivelu et al, 2009).
Gallbladder Remnant and Remnant Cystic Resection of gallbladder and cystic duct remnants due to
Duct Lithiasis lithiasis has been achieved through both laparoscopic and open
It has been suggested that as many as 30% of postcholecystec-

William R. Jarnagin (Auth.) - Blumgart's Surgery of The Liver, Biliary Tract and Pancreas, 2-Volume Set - Elsevier (2016) .Compressed

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

William R. Jarnagin (Auth.) - Blumgart's Surgery of The Liver, Biliary Tract and Pancreas, 2-Volume Set - Elsevier (2016) .Compressed

Uploaded by

Copyright:

Available Formats

Prometheus, chained to the rocky Mount Caucasus, has his liver eaten by the eagle of Zeus.

(Engraving, 1566, possibly after a work by Titian.

Website của chúng tôi: https://doctorplus.club

Library of Congress Cataloging-in-Publication Data

Names: Jarnagin, William R. (Surgeon), editor.

Executive Content Strategist: Michael Houston

Last digit is the print number: 9 8 7 6 5 4 3 2 1

EDITOR-IN-CHIEF Ronald P. DeMatteo, MD, FACS

Ghassan K. Abou-Alfa, MD Jesper B. Andersen, MD

Pietro Addeo, MD Vittoria Arslan-Carlon, MD

N. Volkan Adsay, MD Horacio J. Asbun, MD, FACS

Anil Kumar Agarwal, MCh, FRCS, FACS Béatrice Aussilhou, MD

Peter J. Allen, MD Daniel Azoulay, MD, PhD

Talia B. Baker, MD Marc G. H. Besselink, MD, PhD

Justin M. Burns, MD See Ching Chan, MD, PhD

Giovanni Butturini, MD, PhD William C. Chapman, MD, FACS

C. Ross Carter, MD, FRCS Sean Patrick Cleary, MD, FRCSC

Louise C. Connell, MD Michael D. Darcy, MD

Christopher John DiMaio, MD Imane El Dika, MD

Truman M. Earl, MD Paul T. Fanta, MD

Tomoki Ebata, MD Olivier Farges, MD, PhD

Cristina R. Ferrone, MD David A. Geller, MD

Yuman Fong, MD Brian K. P. Goh, MBBS, MMed(Surgery), MSc,

Gregory J. Gores, MD James J. Harding, MD

Chet W. Hammill, MD, MCR, FACS Alan W. Hemming, MD

Asher Hirshberg, MD, FACS Joseph Ralph Kallini, MD

Adeel S. Khan, MD, FACS Norihiro Kokudo, MD, PhD

Alexis Laurent, MD, PhD Chung-Wei Lin, MD

Giuseppe Malleo, MD, PhD Jessica E. Maxwell, MD, MBA

Ryutaro Mori, MD, PhD Rachel O’Connor

Gérard Pascal, MD Fabio Procopio, MD

James J. Pomposelli, MD Piera Marie Cote Robson, MSN

Flavio G. Rocha, MD Charbel Sandroussi, MMSc, FRACS

Riad Salem, MD, MBA Richard D. Schulick, MD, PhD

Ross W. Shepherd, MD, FRACP, FRCP Stephen B. Solomon, MD

James F. Trotter, MD Alejandra Maria Villamil, MD

Jean-Nicolas Vauthey, MD, FACS Sharon Marie Weber, MD

Jens Werner, MD Chang Jin Yoon, MD, PhD

jaundice could be the result of hepatic disease, it also could

“A case of stones found in the coat of the liver—a woman of

The second quotation also suggests death from gallstones:

FIGURE 0.2 Illustration of the vasculature of the liver. (From Glisson F,

EIGHTEENTH CENTURY TO MODERN TIMES

IMAGING IN HEPATOBILIARY AND

for tumor ablation, arterial embolization, alcohol injection, and

SURGERY OF THE BILIARY TRACT

video laparoscopic cholecystectomy. He did not publish his

hepaticojejunostomy at the French Congress of Surgery in

(1972a, 1972b). Although almost totally abandoned in recent

remarkable fashion. Launois and associates in France (1979)

REFERENCES and biliary tract, ed 3, Philadelphia, 2000, Saunders, pp 1639–

SOX 2 Esophagus and stomach

Primitive foregut tube

Lungs and thyroid: NKX2.1 Pericardial sac

Esophagus and stomach: SOX2

Large intestine: CDXA

Small intestine: CDXC

STM The role of endothelial cells in liver organogenesis has also

BILE DUCT MATURE

based on observations of a number of human diseases termed

regenerative process, suggesting that the biliary and pancreatic

development, further emphasizing the disparate molecular con- TGF-β Signaling

REFERENCES Crompton MR, et al: Identification of a novel vertebrate homeobox

REFERENCES Crompton MR, et al: Identification of a novel vertebrate homeobox

Conwell DL, et al: Analysis of duodenal drainage fluid after cholecys-

REFERENCES Boleslawski E, et al: Hepatic venous pressure gradient in the assessment