Diseases of the Abdomen and Pelvis 2010-2013

Diagnostic Imaging and Interventional Techniques

J. Hodler • G.K. von Schulthess • Ch.L. Zollikofer (Eds)

DISEASES OF THE ABDOMEN

AND PELVIS 2010-2013
DIAGNOSTIC IMAGING
AND INTERVENTIONAL TECHNIQUES

42nd International Diagnostic Course

in Davos (IDKD)
Davos, March 21-26, 2010

including the
Nuclear Medicine Satellite Course “Diamond”
Davos, March 20-21, 2010

Pediatric Satellite Course “Kangaroo”

Davos, March 20-21, 2010

IDKD in Greece

presented by the Foundation for the

Advancement of Education in Medical Radiology, Zurich
Editors

J. HODLER G. K. VON SCHULTHESS

Radiology, Nuclear Medicine,
University Hospital, University Hospital,
Zurich, Switzerland Zurich, Switzerland

CH. L. ZOLLIKOFER
Kilchberg/Zurich, Switzerland

ISBN 978-88-470-1636-1 e-ISBN 978-88-470-1637-8

DOI 10.1007/978-88-470-1637-8

Springer Dordrecht Heidelberg London Milan New York

Library of Congress Control Number: 2010922809

© Springer Verlag Italia 2010

This work is subject to copyright. All rights are reserved, whether the whole or part of the material
is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation,
broadcasting, reproduction on microfilms or in other ways, and storage in data banks. Duplication of
this publication or parts thereof is only permitted under the provisions of the Italian Copyright Law
in its current version, and permission for use must always be obtained from Springer. Violations are
liable for prosecution under the Italian Copyright Law.

The use of general descriptive names, registered names, trademarks, etc., in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.

Product liability: The publisher cannot guarantee the accuracy of any information about dosage and
application contained in this book. In every individual case the user must check such information by
consulting the relevant literature.

Cover design: Simona Colombo, Milan, Italy

Typesetting: C & G di Cerri e Galassi, Cremona, Italy
Printing and binding: Grafiche Porpora, Segrate (MI), Italy
Printed in Italy

Springer-Verlag Italia S.r.l., Via Decembrio 28, 20137 Milan

Springer is a part of Springer Science+Business Media (www.springer.com)
 IDKD 2010-2013

Preface

The International Diagnostic Course in Davos (IDKD) offers a unique learning

experience for imaging specialists in training as well as for experienced radio-
logists and clinicians wishing to be updated on the current state of the art and the
latest developments in the fields of imaging and image-guided interventions.
This annual course is focused on organ systems and diseases rather than on
modalities. This year’s program deals with diseases of the abdomen and pelvis.
During the course, the topics are discussed in group workshops and in plenary
sessions with lectures by world-renowned experts and teachers. While the work-
shops present state-of-the-art summaries, the lectures are oriented towards future
developments. Accordingly, this Syllabus represents a condensed version of the
contents presented under the 20 topics dealing with imaging and interventional
therapies in abdominal and pelvic diseases. The topics encompass all the relevant
imaging modalities including conventional X-rays, computed tomography, nu-
clear medicine, ultrasound and magnetic resonance angiography, as well as
image-guided interventional techniques.
The Syllabus is designed to be an “aide-mémoire” for the course participants
so that they can fully concentrate on the lecture and participate in the discussions
without the need of taking notes.
Additional information can be found on the IDKD website: www.idkd.org

J. Hodler
G.K. von Schulthess
Ch.L. Zollikofer
 IDKD 2010-2013

Table of Contents

Workshops
Emergency Radiology of the Abdomen: The Acute Abdomen . . . . . . . . . . . . . . . . . . . . . . . 3
Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

Trauma of the Abdomen and Pelvis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Philip J. Kenney, Stuart E. Mirvis

Diseases of the Esophagus and Stomach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Marc S. Levine, Ahmed Ba-Ssalamah

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis 28

Marc J. Gollub

Diseases of the Small Bowel, Including the Duodenum – MRI . . . . . . . . . . . . . . . . . . . . . . 32

Karin A. Herrmann

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases . . . . . . . . . . 37

Jaap Stoker, Richard M. Gore

CT Colonography: Updated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Daniel C. Johnson, Michael Macari

Imaging of Diffuse and Inflammatory Liver Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Pablo R. Ros, Rendon C. Nelson

Focal Liver Lesions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Wolfgang Schima, Richard Baron

Imaging Diseases of the Gallbladder and Bile Ducts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Angela D. Levy, Celso Matos

Diseases of the Pancreas, I: Pancreatitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Thomas Helmberger

Diseases of the Pancreas, II: Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Ruedi F. Thoeni

Adrenal Imaging and Intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

William W. Mayo-Smith, Isaac R. Francis

Renal Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Richard H. Cohan, Ronald J. Zagoria

Urinary Tract Obstruction and Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

Parvati Ramchandani, Julia R. Fielding
VIII Table of Contents

Benign Diseases of the Female Genital Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

Caroline Reinhold, Rahel A. Kubik-Huch

Malignant Diseases of the Female Genital Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Evis Sala, Susan Ascher

Magnetic Resonance Imaging of Prostate Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa,
Caroline Hoeks, Jurgen J. Futterer

Imaging of the Male Pelvis: The Scrotum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Brent J. Wagner

Spread of Metastatic Disease in the Abdomen and Pelvis . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

James A. Brink, Ali Shirkhoda

Abdominal Vascular Disease: Diagnosis and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Johannes Lammer

Non-vascular Abdominal Disease: Diagnosis and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi

An Approach to Imaging the Acute Abdomen in the Pediatric Population . . . . . . . . . . . 167

Alan Daneman, Simon G. Robben

Imaging Uronephropathies in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

Jeanne S. Chow, Fred E. Avni

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging . . . . . . . . . . . . . . . . . . . . 183

Gerald Antoch

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging . . . . . . . . . . . . . . . . . 190

Thomas F. Hany

Nuclear Medicine Satellite Course “Diamond”

Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals . . . . . 199
Angelika Bischof Delaloye

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal

and Genitourinary Tract Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Ariane Boubaker

PET in Hepatobiliary-Pancreatic Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri

PET in Tumors of the Digestive Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

Thomas F. Hany

Tumors of the Adrenergic System: Imaging and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . 226

Cornelis A. Hoefnagel

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy . . . . . . . . . . . . . . . . . . . . 231

Dik J. Kwekkeboom
Table of Contents IX

Pediatric Satellite Course “Kangaroo”

Imaging Cystic Kidneys in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Fred E. Avni

Understanding Duplication Anomalies of the Kidney . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Jeanne S. Chow

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management . . . . . 247

Alan Daneman

Pediatric Intestinal Ultrasonography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

Simon G. Robben
 IDKD 2010-2013

List of Contributors

Allegri V., 215 Hoefnagel C.A., 226

Antoch G., 183 Hoeks C., 125
Ascher S., 119 Hulsbergen-van der Kaa C., 125
Avni F.E., 174, 239 Johnson D.C., 48
Ba-Ssalamah A., 22 Kenney P.J., 14
Barentsz J.O., 125 Kubik-Huch R.A., 110
Baron R., 63 Kwekkeboom D.J., 231
Bartolozzi C., 162 Lammer J., 154
Battaglia V., 162 Levine M.S., 22
Bischof Delaloye A., 199 Levy A.D., 75
Boubaker A., 205 Macari M., 48
Bozzi E., 162 Maffione A.M., 215
Brink J.A., 146 Marincek B., 3
Bruel J.-M., 3 Matos C., 75
Chow J.S., 174, 243 Mayo-Smith W.W., 96
Cohan R.H., 99 Mirvis S.E., 14
Daneman A., 167, 247 Nelson R.C., 50
Fanti S., 215 Ramchandani P., 104
Fielding J.R., 104 Reinhold C., 110
Francis I.R., 96 Robben S.G., 167, 252
Futterer J.J., 125 Ros P.R., 50
Gollub M.J., 28 Sala E., 119
Gore R.M., 37 Schima W., 63
Hany T.F., 190, 219 Shirkhoda A., 146
Heiken J.P., 3 Stoker J., 37
Heijmink S.W.T.P.J., 125 Thoeni R.F., 89
Helmberger T., 81 Wagner B.J., 142
Herrmann K.A., 32 Zagoria R.J., 99
WORKSHOPS

Emergency Radiology of the Abdomen: The Acute Abdomen
Jean-Michel Bruel1, Borut Marincek2, Jay P. Heiken3
1 Medical Imaging Department, Hôpital Saint-Eloi, CHRU de Montpellier, Montpellier, France
2 Instituteof Diagnostic Radiology, University Hospital, Zurich, Switzerland
3 Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO, USA
Introduction
The term “acute abdomen” defines a clinical syndrome
characterized by a history of hitherto undiagnosed ab-
dominal pain lasting less than one week. A large range of
disorders, from benign, self-limited diseases to condi-
tions that require immediate surgery, can cause acute ab-
dominal pain. Eight conditions account for over 90% of
patients who are referred to the hospital and who are seen
on surgical wards with acute abdominal pain: acute ap-
pendicitis, acute cholecystitis, small bowel obstruction
(SBO), urinary colic, perforated peptic ulcer, acute pan-
creatitis, acute diverticular disease, and non-specific,
non-surgical abdominal pain (dyspepsia, constipation).
Imaging Techniques
Clinical assessment of the acute abdomen is often diffi-
cult because the findings of the physical examination
and laboratory investigations are often non-specific.
Traditionally, plain abdominal radiographs have served
as the initial imaging approach; however, because of
their diagnostic limitations, plain radiographs now play
only a limited role in this clinical setting. Currently, the
major indications for plain radiography are to determine
the presence of bowel obstruction, perforated viscus,
urinary tract calculi, or a foreign body. The convention-
al radiographic examination consists of supine and ei-
ther upright or left lateral decubitus images. Computed
tomography (CT) now serves as the imaging test of
choice for most adult patients with acute abdominal
pain. It has been shown to be superior to plain radio-
graphy for diagnosing nearly all causes of acute abdom-
inal pain. Several studies have demonstrated that the use
of CT to evaluate patients with acute abdominal pain in-
creases the accuracy of clinical diagnosis by >20% and
results in management changes in up to 60% of patients.
The major obstacles to replacing plain abdominal radio-
graphy with CT are its higher cost, more limited avail-
ability, and higher radiation dose. The use of CT in pa-
tients with an acute abdomen requires careful attention
to the CT protocol.
IDKD 2010-2013
1. The multidetector CT (MDCT) image data volume
should be obtained from the dome of the diaphragm
to the inferior aspect of the pubic symphysis (with
thin collimation and a short acquisition time). CT da-
ta are reconstructed with thin, overlapped slices for
multiplanar (coronal and sagittal planes) reformation
and 3- to 5-mm thick contiguous axial slices. The
image series is sent to a dedicated workstation and/or
PACS.
2. Vascular enhancement by iodinated contrast medium
(CM) is mandatory in most cases, unless contraindi-
cated. Special attention should be paid to older pa-
tients and those with metabolic disorders (dehydration
in SBO) in assessing the renal impact of CM admin-
istration. In general, the most helpful scanning phase
is the late portal phase (70 s), but other scanning phases
are useful in selected circumstances: arterial phase
in bowel ischemia, bleeding, or visceral infarction;
delayed phase (3 min) for assessing the lack of en-
hancement in a patient with suspected acute mesen-
teric ischemia. A pre-contrast CT scan allows demon-
stration of calcifications, lithiasis, and acute or sub-
acute hemorrhage; multiphasic scanning should be
used only for specific indications in order to limit ra-
diation dose.
3. Changes in the CT protocol should be decided accord-
ing to the clinical conditions and/or the preliminary re-
sults of the CT examination. In selected cases, colo-
rectal opacification and/or image acquisition with the
patient in the prone position may be helpful to clarify
equivocal findings. In most cases, the systematic use
of ingested oral contrast is not recommended.
4. The method of image evaluation is critical to optimize
interpretation. Additional window level (M) and width
(W) settings are useful to identify tiny bubbles of ex-
traluminal gas (CT lung windows) or hyperattenuation
from recent hemorrhage (narrow CT windows). The
systematic use of multiplanar reformation (MPR), par-
ticularly in the coronal plane, is recommended, and 3D
imaging may be helpful.
Ultrasonography (US) is the initial imaging technique of
choice for patients with suspected acute cholecystitis or
acute gynecological abnormalities. It also is the primary
4 Jean-Michel Bruel, Borut Marincek, Jay P. Heiken
method for evaluating pregnant women and pediatric pa-
tients. Although less sensitive and specific than CT, US
is an excellent imaging test for diagnosing acute appen-
dicitis, when employed by experienced individuals. It can
also be used to evaluate the presence or absence of the
layered structure of the digestive tract wall or to assess
the structure of a lesion identified at CT.
Until recently, magnetic resonance imaging (MRI) has
played a very limited role in patients with acute abdomi-
nal pain; however, it is now established in the imaging of
pregnant women with abdominal pain who have had a
negative or equivocal US examination. Recent studies as-
sessing the use of MRI to evaluate all patients with acute
lower abdominal pain have shown promising results. MRI
may also have a role in patients with biliary diseases
and/or pancreatitis.
The differential diagnosis in a patient with an acute ab-
domen is influenced greatly by the nature and location of
the pain. Therefore, the imaging strategies for acute pain
localized to an abdominal quadrant should be discussed
separately from those for acute pain that is diffuse or lo-
calized to the flank or epigastric region.
Acute Pain in an Abdominal Quadrant
In many cases, acute abdominal pain can be localized to
one either the right upper, left upper, right lower, or left
lower abdominal quadrant.
Right Upper Quadrant
Acute cholecystitis is by far the most common disease to
involve the right upper quadrant. Other important dis-
eases that can have a clinical presentation similar to that
of acute cholecystitis are pyogenic or amebic liver ab-
scess, spontaneous rupture of a hepatic neoplasm (usual-
ly hepatocellular adenoma or carcinoma), hepatitis, and
myocardial infarction.
The preferred imaging method for evaluating patients
with acute right upper abdominal pain is US. It is a reli-
able technique for establishing the diagnosis of acute cal-
culous cholecystitis. The imaging criteria include the de-
tection of gallstones, the sonographic Murphy sign, gall-
bladder wall thickening ≥3 mm, and pericholecystic flu-
id. The association of three of these signs is highly sug-
gestive of acute cholecystitis. Isolated gallbladder wall
thickening may be secondary to other conditions, such as
gallbladder adenomyomatosis, gallbladder carcinoma,
HIV cholangitis, sclerosing cholangitis, acute hepatitis,
cirrhosis, ascites, portal hypertension, hypoproteinemia,
pancreatitis, and cardiac failure. In acute calculous chole-
cystitis, typically a calculus obstructs the cystic duct. The
trapped concentrated bile irritates the gallbladder wall,
causing increased secretion, which in turn leads to dis-
tention and edema of the wall. The rising intraluminal
pressure compresses the vessels, resulting in thrombosis,
ischemia, and subsequent necrosis and perforation of the
wall. Gallbladder perforation and complicating peri-
cholecystic abscess typically occur adjacent to the gall-
bladder fundus because of the sparse blood supply. CT
may be useful for confirmation of the sonographic diag-
nosis, but usually is not necessary. Emphysematous
cholecystitis is a rare complication of acute cholecystitis
that generally is associated with diabetes mellitus. US or
CT demonstrates gas in the wall and/or lumen of the gall-
bladder, which implies underlying gangrenous changes
(Fig. 1). Acalculous acute cholecystitis accounts for only
approximately 5% of cases of acute cholecystitis but is
especially common in patients in the intensive care unit.
Prolonged bile stasis results in increased viscosity of the
bile that ultimately leads to functional cystic duct ob-
struction.
Both US and CT are accurate techniques for diagnos-
ing liver abscesses. US usually demonstrates a round or
oval hypoechoic mass with low-level internal echoes. Al-
though the lesion may mimic a solid hepatic mass, the
presence of through transmission is a clue to its cystic na-
ture. Pyogenic liver abscesses most commonly are the re-
sult of seeding from appendicitis or diverticulitis or direct
extension from cholecystitis or cholangitis. Amebic ab-
scesses result from primary colonic involvement, with
seeding through the portal vein. In most cases, the US ap-
pearances of pyogenic and amebic abscess are indistin-
guishable. The CT appearances of pyogenic and amebic
abscesses also overlap substantially. Amebic abscesses
are cystic masses of low attenuation. An enhancing wall
and a peripheral zone of edema surrounding the abscess
are common but not universally present. Extrahepatic ex-
tension of the amebic abscess with involvement of the
chest wall, pleura, or adjacent viscera is a frequent find-
ing. Whereas amebic abscesses usually are solitary and
unilocular, pyogenic abscesses may be multiple or multi-
Fig. 1. Diabetic patient with emphysematous cholecystitis and gan-
grene of the gallbladder. CT shows air-fluid level in the gallblad-
der lumen and air in the gallbladder wall (arrows)
Emergency Radiology of the Abdomen: The Acute Abdomen
loculated and may demonstrate an irregular contour.
Some pyogenic abscesses have a mixed cystic and solid
appearance on US, CT, or MRI; rarely, they appear com-
pletely solid. A small percentage of hepatic abscesses,
particularly those secondary to Klebsiella infection, are
associated with portal vein thrombosis.
Spontaneous rupture of a hepatocellular carcinoma
with subsequent hemoperitoneum is a frequent complica-
tion in countries with a high incidence of this tumor, but
is less commonly seen in Western countries. Subcapsular
location and tumor necrosis have been implicated in the
pathogenesis. US, and especially CT, are the most useful
techniques for diagnosing a ruptured hepatocellular car-
cinoma, which appears as a peripheral or subcapsular
mass. Transcatheter embolization of either the tumor or
the bleeding hepatic artery is the treatment of choice.
Spontaneous hemorrhage within a hepatocellular adeno-
ma occurs most commonly in women taking oral contra-
ceptives. Capsular rupture with subsequent hemo-
peritoneum is an uncommon complication. On CT, high-
density intraperitoneal fluid confirms the diagnosis of
hemoperitoneum. Extravasation of CM, when present, is
indicative of active bleeding.
Left Upper Quadrant
Although infrequent, acute left upper quadrant pain is
most often seen in splenic infarction, splenic abscess,
gastritis, and gastric or duodenal ulcer. US is most fre-
quently used for screening, while CT enables accurate
further evaluation. The diagnosis of gastric pathology is
established by endoscopy, with imaging playing a minor
role.
Common causes of splenic infarction include bacteri-
al endocarditis, portal hypertension, and marked
splenomegaly. Pancreatitis or tumors that extend into the
splenic hilum can also result in infarction. Splenic in-
farction may be focal or global. Typical focal splenic in-
farcts appear as peripheral wedge-shaped defects, hypo-
echoic or isoechoic at US and hypoattenuating at CT.
Most splenic abscesses are secondary to hematogenous
dissemination of infection, e.g., bacterial endocarditis or
tuberculosis. Intravenous drug abusers and immunocom-
promised individuals are predominantly affected. US and
CT are sensitive, but the specificity of either one is low.
On US, most abscesses appear as hypo- or anechoic,
poorly defined lesions; on CT, they typically appear as
rounded lesions of low attenuation and with rim en-
hancement. Spontaneous splenic rupture can occur in pa-
tients with hematological malignancy or secondary to
rapid splenic enlargement from viral infections such as
mononucleosis.
Right Lower Quadrant
Acute appendicitis is not only the most frequent cause of
acute right lower quadrant pain, it is also the most com-
monly encountered cause of an acute abdomen. Other
5
diseases that can present with acute right lower quadrant
pain include acute terminal ileitis (Crohn’s disease), ty-
phlitis, right-sided colonic diverticulitis and, in women,
pelvic inflammatory disease, complications of ovarian
cyst (hemorrhage, torsion, and leakage), endometriosis,
or ectopic pregnancy. Less common causes of right low-
er quadrant pain include segmental infarction of the
greater omentum, mesenteric adenitis, epiploic ap-
pendagitis, perforated cancer, and ileal or Meckel’s di-
verticulitis.
The diagnosis of acute appendicitis is uncertain in up
to one-third of patients. Thus, pre-operative imaging
plays an important role in confirming or excluding the
diagnosis. With the increasing use of medical imaging
to evaluate patients with suspected acute appendicitis,
the rate of both false-positive (unnecessary appendecto-
my) and false-negative (leading to complications from
perforated appendicitis) diagnoses has decreased. The
standard surgical teaching is that patients with typical
clinical findings should undergo immediate appendec-
tomy without pre-operative imaging. Nevertheless, at
most medical centers pre-operative imaging is obtained
even when the clinical presentation is typical. The most
specific CT finding of acute appendicitis is a thick-
walled appendix that contains an appendicolith (Fig. 2).
The inflamed appendix often is dilated and fluid-filled.
Additional helpful findings are stranding of the peri-
appendiceal fat and thickening of the cecal apex. Find-
ings that indicate appendiceal perforation include peri-
appendiceal abscess, extraluminal gas, a right lower
quadrant inflammatory mass, a defect in the appen-
diceal wall, and SBO.
In the evaluation of suspected acute appendicitis in
children, pregnant women, and women of reproductive
age, US is an important imaging option. Demonstration
of a swollen, noncompressible appendix >7 mm in di-
ameter with a target configuration is the primary sono-
graphic criterion (Fig. 3). Additional helpful US find-
ings are “MacBurney’s sign” (maximum tenderness
found with graded compression of the inflamed appen-
dix) and demonstration of an appendicolith. These US
signs may also be demonstrated by transvaginal high-
resolution US. The advantages of US include the lack
of ionizing radiation, relatively low cost, and wide-
spread availability. However, US requires considerable
skill and is difficult to perform in obese patients, pa-
tients with severe pain, and patients likely to have a
complicating periappendiceal abscess. When the sono-
graphic findings are unclear, CT can provide a rapid
and definitive diagnosis. Due to its exceptional accura-
cy, CT has emerged in many centers as the primary
imaging test for patients with suspected acute appen-
dicitis.
In a small percentage of patients, diverticulitis mani-
fests itself as a right-sided condition. Right-sided colonic
diverticula are often congenital, solitary, and true diver-
ticula, unlike sigmoid diverticula. In right-sided divertic-
ulitis the normal appendix should be visible.
6 Jean-Michel Bruel, Borut Marincek, Jay P. Heiken
a c
a b c
Left Lower Quadrant
Diverticulitis is the most common cause of acute left low-
er quadrant abdominal pain. The condition occurs in
10–20% of patients with diverticulosis and most common-
ly involves the sigmoid colon. CT is very sensitive and ap-
proaches 100% specificity and accuracy in the diagnosis or
exclusion of diverticulitis; it has therefore largely replaced
barium enema examinations. CT is also very useful in es-
tablishing the presence of pericolic complications. The CT
diagnosis of acute diverticulitis is based on the identifica-
tion of segmental colonic wall thickening and pericolic in-
Fig. 2 a-c. Acute appendicitis. Axial (a, b)
and sagittal (c) views of multidetector CT
demonstrate a dilated appendix in retroce-
cal position, a calcified appendicolith at
the base of the appendix (arrowheads), and
inflammatory changes of the mesenteric
fat (arrow)
Fig. 3 a-c. Ultrasonography of acute appen-
dicitis in a 12-year-old girl. Oblique (a, b)
and transverse (c) views show swollen ap-
pendix (diameter 10 mm, arrowheads)
with a target configuration
flammatory changes, such as fat stranding, inflammatory
mass, gas bubbles, or free fluid (Figs. 4, 5). Complications
of acute diverticulitis include abscess, fistula (most com-
monly colovesical), SBO, peritonitis, septic thrombo-
phlebitis, colonic obstruction, and ureteral obstruction.
In patients with left lower quadrant pain, alternative diag-
noses that should be considered are colitis (infectious/
inflammatory or ischemic), colonic carcinoma, epiploic ap-
pendagitis, neutropenic colitis, functional colonic disor-
ders, and extragastrointestinal disorders (pyelonephritis or
gynecological diseases). Epiploic appendagitis is a clinical
condition mimicking acute colonic diverticulitis, with focal
Emergency Radiology of the Abdomen: The Acute Abdomen
Fig. 4. Sigmoid diverticulitis with pericolic abscesses. CT shows
fine linear strands within pericolic fat, diverticula filled with air,
barium, or fecal material, circumferential bowel thickening, and
frank abscesses (arrows)
a These are important causes of acute lower abdominal
b PID most commonly results from an ascending infection,
Fig. 5 a, b. Perforated sigmoid diverticulitis. a CT demonstrates tiny
bubbles of extraluminal gas within the left subphrenic space (ar-
rows, 1) and hepatic pedicle (arrow, 2). The determination of a
pneumoperitoneum from the perforated sigmoid diverticulitis (b)
requires dedicated window settings. b Coronal reformat shows the
patterns indicating the severity of the local extent of this case of
sigmoid diverticulitis: thickened bowel wall with diverticula asso-
ciated with dramatic changes of the pericolic fat and thickened root
of the sigmoid mesocolon. Note the tiny bubble of extraluminal gas
(arrow) within the hepatic pedicle
7
exquisite lower abdominal pain. It is diagnosed with CT (or
US) by the demonstration of an ovoid lesion within the
pericolonic fat, surrounded by inflammatory changes and
abutting the colonic wall. As this disease resolves sponta-
neously within a few days, its correct diagnosis on CT im-
ages is important to avoid unnecessary surgery.
Distinguishing sigmoid diverticulitis from carcinoma
is a major differential diagnostic consideration but is of-
ten difficult on CT images despite CT findings sugges-
tive of colon carcinoma, such as marked thickening of the
colonic wall, focal thickening (length <5 cm), and the
presence of pericolonic lymph nodes. The distinction be-
tween perforated colon carcinoma and diverticulitis may
be especially difficult because many patients with colon
carcinoma also have diverticulosis. Therefore, it is crucial
to perform colonoscopy 3-4 weeks after the onset of
acute diverticulitis to rule out a colonic carcinoma.
Gynecological Disorders
pain (right lower quadrant, left lower quadrant, or central
pelvic) in female patients, particularly young women.
Acute uterine and ovarian disorders, pelvic inflammatory
disease, endometriosis, and ectopic pregnancy are diag-
nostic considerations in women of child-bearing age who
present with an acute abdomen.
Pelvic inflammatory disease (PID) is the most frequent
gynecological cause of an acute abdomen. Its manifesta-
tions may include salpingitis, tubo-ovarian abscess, and
peritonitis. The most common causative organisms are
Neisseria gonorrhoeae and Chlamydia trachomatis. Since
it usually involves both adnexa. US is the imaging test of
choice for patients with suspected PID; however, CT is
helpful when the clinical diagnosis is equivocal. CT find-
ings may include fluid within the cul-de-sac, endometrial
thickening, fluid within the uterine cavity, and thickening
and dilation of the fallopian tubes (pyosalpinx) (Fig. 6).
Fig. 6. Pelvic inflammatory disease. Transaxial CT image shows
thickened, dilated fluid-filled fallopian tubes indicative of bilateral
pyosalpinx
8 Jean-Michel Bruel, Borut Marincek, Jay P. Heiken
With progression to tubo-ovarian abscesses, unilateral or
bilateral cystic adnexal masses may be seen, usually in as-
sociation with pyosalpinx. In advanced cases of PID, the
intrapelvic spread of purulent material may result in peri-
tonitis. Patients with Fitz-Hugh-Curtis syndrome present
with right upper quadrant pain due to perihepatic inflam-
mation from intraperitoneal exudates stretching between
the liver capsule and the peritoneum, which can mimic
carcinomatosis on CT.
In adult patients, ovarian torsion (adnexal torsion) usu-
ally occurs in association with an ovarian mass, which
acts as a fulcrum to potentiate torsion. Teratoma is the
most common cause of ovarian torsion. The typical pre-
sentation is non-specific, consisting of acute lower ab-
dominal pain associated with nausea, vomiting, and
leukocytosis. If ovarian torsion is suspected, Doppler US
is the initial imaging test of choice; however, because the
clinical presentation usually is non-specific, CT is often
the first imaging test requested. A CT finding helpful in
making the diagnosis of ovarian torsion is an enlarged
ovary that is displaced from its normal location. Sec-
ondary signs include a thickened fallopian tube, a twist-
ed vascular pedicle, hemoperitoneum, and deviation of
the uterus toward the affected side.
Complications of ovarian cysts such as hemorrhage
and rupture also can cause acute lower abdominal pain.
Hemorrhagic cysts contain fluid that is high in attenua-
tion, sometimes with a fluid-fluid level. In a small per-
centage of patients, the cyst may rupture, resulting in
hemoperitoneum. Correlation with β-human chorionic
gonadotropin (hCG) levels is important as a ruptured
ectopic pregnancy may present with similar clinical and
imaging features.
Endometriosis is characterized by the presence of
functioning endometrial tissue outside of its normal in-
trauterine location. It presents as acute abdominal pain in
only a small percentage of women and is usually caused
by rupture or hemorrhage of an endometrioma or by tor-
sion of an ovary that contains endometrial implants. On
CT, endometriomas have a variable appearance, ranging
from cystic to solid adnexal masses.
Ectopic pregnancy remains the leading cause of death
during the first trimester of pregnancy, with a mortality
of 9-14%. The main risk factors for ectopic pregnancy in-
clude a history of ectopic pregnancy, tubal surgery, and
PID. The initial evaluation of patients suspected of hav-
ing an ectopic pregnancy requires quantitative measure-
ment of serum β-hCG level and transvaginal US. The lat-
ter should be used to search for the presence of an ad-
nexal mass (with or without highly specific signs such as
adnexal gestational sac with a live embryo, suggested by
the demonstration of cardiac activity and/or a “tubal ring
sign”), hematosalpinx, pelvic free fluid (highly sugges-
tive if heterogeneous), and hemoperitoneum, enlarged
uterus (with a pseudo-gestational sac), and symmetrical-
ly enlarged ovaries. The diagnosis of ectopic pregnancy
often is difficult, since transvaginal US may be normal in
up to 25% of these patients and adnexal abnormalities
may be found in numerous alternative diagnoses. The di-
agnosis of ectopic pregnancy is based on an association
of the β-hCG level with the transvaginal US findings:
1. a normal β-hCG level rules out an ectopic pregnancy;
2. an intrauterine gestational sac with a live embryo (car-
diac activity) rules out an ectopic pregnancy (but in
patients with assisted reproduction by ovulation induc-
tion the rate of “heterotopic pregnancy”, defined as the
simultaneous occurrence of an intrauterine and an ex-
trauterine pregnancy, has been reported to be 1–3%);
3. an elevated β-hCG level (>1000 IU/L) without an in-
trauterine gestational sac, associated with an abnormal
adnexal pattern and/or heterogeneous pelvic fluid, in-
dicates an ectopic pregnancy.
Acute Abdomen with Diffuse Pain
Any disorder that irritates a large portion of the gastroin-
testinal (GI) tract and/or the peritoneum can cause diffuse
abdominal pain. The most common disorder is gastro-
enterocolitis. Other important disorders are bowel obstruc-
tion, ischemic bowel disease, and GI tract perforation.
Bowel Obstruction
Bowel obstruction is a frequent cause of abdominal pain
and accounts for approximately 20% of surgical admis-
sions for acute abdominal conditions. The small bowel
is involved in 60-80% of cases. Frequent causes of SBO
are postoperative adhesions, hernias, and neoplasms.
Mechanical large bowel obstruction is most commonly
due to colorectal carcinoma, but volvulus and diverticuli-
tis are also important causes. Colonic volvulus most
commonly involves the sigmoid region, followed by the
cecum.
The diagnosis of bowel obstruction is established on
clinical grounds and usually confirmed with plain ab-
dominal radiographs. Due to the diagnostic limitations of
plain radiography, CT is increasingly used to establish the
diagnosis, identify the site, level, and cause of obstruc-
tion, and determine the presence or absence of associat-
ed bowel ischemia. CT can be useful for differentiating
between simple and closed-loop obstruction. Closed-loop
obstruction is a form of mechanical bowel obstruction in
which two points along the course of the bowel are ob-
structed at a single site. It is usually secondary to an ad-
hesive band or a hernia. Since a closed loop tends to in-
volve the mesentery and is prone to produce a volvulus,
it represents the most common cause of strangulation.
However, only colonic volvulus is associated with classic
features on plain abdominal radiography.
CT is particularly reliable in higher grades of bowel
obstruction. It has proved useful in characterizing bowel
obstruction from various causes, including adhesions,
hernia, neoplasm, extrinsic compression, inflammatory
bowel disease, radiation enteropathy, intussusception,
gallstone ileus, or volvulus. The essential CT finding of
Emergency Radiology of the Abdomen: The Acute Abdomen
bowel obstruction is the delineation of a transition zone
between the dilated and decompressed bowel. Careful in-
spection of the transition zone and luminal contents usu-
ally reveals the underlying cause of obstruction. Howev-
er, the presumed point of transition from dilated to non-
dilated bowel can be difficult to determine in the transax-
ial plane. MDCT facilitates this task by providing the ra-
diologist with a volumetric data set that can be viewed in
the transaxial, sagittal, or coronal plane or any combina-
tion of the three. These MPR views centered on the an-
ticipated transition point help to determine the site, level,
and cause of obstruction.
Mechanical obstruction of the small bowel (SBO) has
to be differentiated from paralytic ileus, large bowel ob-
struction (LBO), and non-obstructive massive distention
of the colon.
Paralytic ileus is a common problem after abdominal
surgery. It may be diffuse or localized and has numerous
causes, e.g., secondary to ischemic conditions, inflam-
matory or infectious disease, abnormal electrolyte,
metabolite, drug or hormonal levels, or innervation de-
fects.
In LBO, CT demonstrates distension of the large bow-
el to the point of obstruction, with collapse of the distal
large bowel. The distal small bowel loops may also be
distended if the ileocecal valve is incompetent. Luminal
obstruction by a colonic carcinoma and colonic volvulus
(Fig. 7) are the main causes of LBO. Perforation of the
cecum, due to gross distention resulting in ischemia of
the cecal wall, is the main complication of severe LBO.
A massively dilated colon may be seen in toxic mega-
colon and Ogilvie’s syndrome. In toxic megacolon sec-
ondary to severe colitis, CT demonstrates a thickened
wall with “thumbprinting” caused by wall edema and in-
flammation. Ogilvie’s syndrome is an acute pseudo-
obstruction with dilation of the colon in the absence of a
colonic transition zone. It may occur in severely ill pa-
tients after surgery and/or with neurological disorders,
serious infections, cardiorespiratory insufficiency, and
metabolic disturbances. Drugs that disturb colonic motil-
ity (e.g., anticholinergics or opioid analgesics) contribute
to the development of this condition.
a b
Fig. 7 a, b. Cecal volvulus. Trans-
axial CT (a) and coronal vol-
ume-rendered (b) CT demon-
strate a markedly dilated cecum
(C) in the left side of the pelvis.
The arrow points to the area of
twist of the ascending colon.
Note the dilated small bowel
loops due to the proximal
colonic obstruction
9
Ischemic Bowel Disease
Predominant causes of bowel ischemia are arterial or ve-
nous occlusion and hypoperfusion. Occlusive disease in-
volves the mesenteric arteries, most commonly the supe-
rior mesenteric artery, in the large majority of cases.
Bowel ischemia secondary to venous thrombosis is much
less common. The only direct sign of vascular impair-
ment of the bowel is diminished bowel wall enhancement,
which is due to inadequate arterial inflow to the bowel.
Increased bowel wall enhancement may be seen in some
cases secondary to reactive hyperemia or compromised
venous outflow. Other CT findings are direct visualiza-
tion of a thrombus in the superior mesenteric artery or
vein. Bowel distention and bowel wall edema are non-
specific findings and may accompany inflammatory or
infectious causes. Bowel distention reflects the interrup-
tion of peristaltic activity in ischemic segments. Non-
occlusive acute mesenteric ischemia usually is due to
hypoperfusion secondary to severe cardiac disease, but
also occurs in patients with end-stage renal or hepatic
disease. An important radiographic manifestation of non-
occlusive acute mesenteric ischemia due to low arterial
flow is mesenteric arterial vasoconstriction. A less com-
mon form of non-occlusive acute mesenteric ischemia is
severe vasculitis, which often affects younger individuals.
Mesenteric vasculitis usually results in bowel edema and
mucosal hyperenhancement. The small and the large
bowel often are both involved. The duodenum is involved
in approximately one-quarter of patients.
The most common mechanical cause of bowel is-
chemia is obstruction. A closed-loop SBO is more likely
than other types of obstruction to result in vascular com-
promise (strangulation obstruction). Strangulation ob-
struction has a reported prevalence of 5-40% and is a
predominantly venous disease. The most frequent abnor-
mality seen on CT is bowel wall thickening. The thick-
ened bowel wall sometimes is associated with a target
sign, consisting of alternating layers of high and low at-
tenuation within the thickened bowel wall, which results
from submucosal edema and hemorrhage. The bowel seg-
ment proximal to an obstruction can become ischemic
10
due to severe distention. CT findings that suggest sub-
sequent infarction are non-enhancement of the bowel
wall, gas in the bowel wall, mesenteric or portal veins,
edema/hemorrhage in the mesentery adjacent to thick-
ened and/or dilated bowel loops, and ascites (Fig. 8).
Perforation of the Gastrointestinal Tract
Gastrointestinal perforation usually causes localized pain
initially, and culminates in diffuse pain if peritonitis de-
velops. Gastroduodenal perforation associated with pep-
tic ulcer disease or a necrotic neoplasm has become less
frequent in recent decades due to earlier diagnosis and
improved therapy. At the same time, the incidence of gas-
troduodenal perforation resulting from endoscopic in-
strumentation has increased. Perforation of the small
bowel is relatively uncommon but may be secondary to a
foreign body, small bowel diverticulitis, or trauma. Spon-
taneous rupture of the colon is more frequent and can oc-
cur when the colon becomes markedly dilated proximal
to an obstructing lesion (tumor, volvulus) or when the
bowel wall is friable (ischemic or ulcerative colitis,
necrotic neoplasm). Fiberoptic colonoscopy with or with-
out biopsy is another cause of colonic perforation.
Pneumoperitoneum can be recognized by the presence
of subdiaphragmatic gas on an upright chest radiograph
or an upright or left lateral decubitus abdominal radio-
graph. A large pneumoperitoneum generally is indicative
of colonic perforation, whereas moderate quantities of
free gas are seen with gastric perforation. Small bowel
perforation usually results in either a limited amount of
peritoneal gas or none, because the small bowel usually
does not contain gas. Detection of subtle pneumo-
a c
Jean-Michel Bruel, Borut Marincek, Jay P. Heiken
peritoneum is often difficult. As CT is far more sensitive
than conventional radiography in demonstrating a small
pneumoperitoneum, it has become the imaging test of
choice when the results of conventional radiography are
equivocal. Viewing the CT images at “lung window”
settings improves the demonstration of small amounts of
extraluminal gas.
Retroperitoneal perforations (duodenal loop beyond
the bulbar segment, or involving the appendix; posterior
aspect of the ascending and descending colon, or the rec-
tum below the peritoneal reflection) tend to be contained
locally and remain clinically silent for several hours or
days. Retroperitoneal gas has a mottled appearance and
may extend along the psoas muscles. In contrast to in-
traperitoneal gas, retroperitoneal gas does not move
freely when the patient’s position is changed from supine
to upright for plain abdominal radiographs.
Acute Abdomen with Flank or Epigastric Pain
Acute flank or upper abdominal pain radiating to the
back is commonly a manifestation of retroperitoneal
pathology, especially urinary colic, acute pancreatitis, or
leaking abdominal aortic aneurysm.
Urinary Colic
For decades, intravenous urography was the primary
imaging technique used to evaluate patients with sus-
pected urinary colic. Plain abdominal radiography and
US may be useful for patients with a contraindication to
radiation or iodinated intravenous CM. However, because
Fig. 8 a-c. Strangulating small bowel ob-
struction due to an adhesive band that de-
veloped after cholecystectomy and appen-
dectomy. Axial (a, b) and sagittal (c) views
of multidetector CT show bowel wall thick-
ening, enhancing bowel wall with submu-
cosal edema, and/or hemorrhage giving a
target sign (long arrows), indicating is-
chemia. Non-enhancement of the bowel
wall of the jejunum (short arrow) corre-
sponds to segmental infarction. A Ascites
Emergency Radiology of the Abdomen: The Acute Abdomen
of the low sensitivity of abdominal radiographs and US
in the detection of urinary tract calculi, the role of unen-
hanced CT has become well established over the past 15
years. On CT, virtually all ureteral stones are radiopaque,
regardless of their chemical composition. Uric acid
stones have attenuation values of 300-500 Hounsfield
units (HU), and calcium-based stones >1000 HU. In ad-
dition to the direct demonstration of a ureteral stone, sec-
ondary signs of ureterolithiasis, including hydroureter,
hydronephrosis, perinephric stranding, and renal enlarge-
ment, may be visible (Fig. 9). Perinephric stranding and
edema result from reabsorbed urine infiltrating the per-
inephric space along the bridging septa of Kunin. The
more extensive the perinephric edema shown on unen-
hanced CT, the higher the degree of urinary tract ob-
struction. Focal periureteral stranding resulting from a lo-
cal inflammatory reaction or irritation and induced by the
passage of a stone helps to localize subtle calculi. Occa-
sionally, a repeat CT examination using intravenous CM
may be required, particularly if infectious complications
are suspected. For the diagnosis of such complications
(pyelonephritis), CT is helpful as it reveals a “striated
nephrogram” after CM administration, as well as global
enlargement of the kidney, renal and/or perirenal ab-
scesses, or emphysematous pyelonephritis.
b c
Fig. 9 a-c. Right-sided urinary colic. Axial
(a), coronal (b), and oblique (c) views of
multidetector CT show dilatation of the
proximal urinary tract due to a ureteral
stone (arrow). Note slight secondary peri-
uretral and perinephric stranding
11
When no stone is detected, an alternative diagnosis
must be established. Non-calculus urinary tract abnor-
malities causing symptoms of colic include acute
pyelonephritis, renal cell carcinoma, acute renal vein
thrombosis, spontaneous dissection of the renal artery,
and renal infarction. Extraurinary diseases, such as retro-
cecal appendicitis, diverticulitis, SBO, pancreatitis, gyne-
cological disorders, and retroperitoneal hemorrhage, may
also simulate acute urinary colic.
Acute Pancreatitis
An important disease causing upper abdominal pain is
acute pancreatitis. US may be helpful for the demonstra-
tion of choledocolithiasis as a cause of acute pancreatitis
and for the follow-up of known fluid collections.
Since the CT findings correlate well with the clinical
severity of acute pancreatitis, CT has become the imag-
ing test of choice to stage the extent of disease (CT sever-
ity index of Balthazar) and to detect complications. The
initial CT should be performed 48-72 h after disease on-
set (a CT examination performed too early in the course
of the disease may not demonstrate any abnormality).
Pancreatic enlargement due to interstitial parenchymal
edema may progress to pancreatic exudate collecting in
12 Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

Fig. 10 a-c. Severe acute pancreatitis. Axial

(a, b) and coronal (c) views of MDCT
demonstrate an enlarged pancreatic area.
a Only small areas of enhanced pancreatic
parenchyma (p) remain within b extensive
pancreatic necrosis (n). Note the tiny
hyperattenuating calculi within the
gallbladder and the lower bile duct (arrows).
c The large amount of pancreatic necrosis
contrasts sharply with the minimal extra-
pancreatic changes at this phase

the abdominal ligaments and potential spaces surround-

ing the pancreas. The pancreatic parenchyma may under-
go necrosis or hemorrhage (Fig. 10). Severe pancreatitis
is often complicated by infection of the necrotic site
and/or thrombosis of the splenic and portal vein. Identi-
fying infectious complications in patients with pancreati-
tis may be difficult; the demonstration of gas bubbles
within peripancreatic collections is neither sensitive nor
specific. CT-guided fluid aspiration for bacteriological
examination is often the best technique to establish the
presence of infection.
Acute pancreatic and peripancreatic fluid collections
may evolve into pseudocysts, which exhibit defined
walls. A pseudocyst can erode peripancreatic vessels,
thus causing bleeding or the formation of a pseudo-
aneurysm (Fig. 11). Fig. 11. Recurrent pancreatitis. CT shows a pseudocyst of the pan-
creatic tail (PC) and a pseudoaneurysm of the gastroduodenal
Leaking Abdominal Aortic Aneurysm artery (PA). The splenic vein is indicated by the arrowhead

One of the most life-threatening diagnoses in patients

with acute flank pain is a leaking abdominal aortic or il-
iac artery aneurysm. When a patient with suspected rup-
ture of an abdominal aortic aneurysm is hemodynamical-
ly unstable, US is the initial imaging technique. The ex-
amination can be performed rapidly with portable equip-
ment in the emergency room. However, para-aortic hem-
orrhage is poorly diagnosed by US. Instead, in hemo-
dynamically stable patients, non-contrast-enhanced CT is
the initial imaging test of choice as it is almost always
able to demonstrate a para-aortic hematoma, if present,
and may show additional findings helpful in establishing
the diagnosis, such as a high-attenuating crescent sign. If
endoluminal stent graft repair of the aorta is planned,
contrast-enhanced CT should be performed.
Conclusions
The imaging evaluation of patients with an acute abdomen
has changed dramatically in the past decade. Plain ab-
dominal radiographs largely have been replaced with US
and CT. MDCT permits a rapid examination with high
Emergency Radiology of the Abdomen: The Acute Abdomen
diagnostic accuracy. Close cooperation with the referring
physician prior to imaging remains essential for rapid and
accurate diagnosis, as the character and location of the pa-
tient’s abdominal pain strongly influences the differential
diagnosis and the choice of initial imaging test.
Suggested Reading
Ahn SH, Mayo-Smith WW, Murphy BL et al (2002) Acute non-
traumatic abdominal pain in adult patients: abdominal radiog-
raphy compared with CT evaluation. Radiology 225:159-64
Balthazar EJ, Robinson DL, Megibow AJ, Ranson JH (1990) Acute
pancreatitis: value of CT in establishing prognosis. Radiology
174:331-336
Freeman AH (2001) CT and bowel disease. Br J Radiol 74:4-14
Gore RM, Miller FH, Pereles FS et al (2000) Helical CT in the
evaluation of the acute abdomen. AJR 174:901-913
Mindelzun RE, Jeffrey RB (1997) Unenhanced helical CT for eval-
uating acute abdominal pain: a little more cost, a lot more in-
formation. Radiology 205:43-47
13
Novelline RA, Rhea JT, Rao PM, Stuk JL (1999) Helical CT in
emergency radiology. Radiology 213:321-339
Paulson EK, Jaffe TA, Thomas J et al (2004) MDCT of patients with
acute abdominal pain: a new perspective using coronal refor-
mations from submillimeter isotropic voxels. AJR 183:899-906
Singh AK, Gervais DA, Hahn PF et al (2005) Acute epiploic ap-
pendagitis and its mimics. Radiographics 25:1521-34
Smith RC, Varanelli M (2000) Diagnosis and management of acute
ureterolithiasis. AJR 175:3-6
Stoker J, van Randen A, Lameris W, Boermeester MA
(2009) Imaging Patients with acute abdominal pain. Radiolo-
gy 253:31-46
Taourel P, Kessler N, Lesnik A et al (2003) Helical CT of large
bowel obstruction. Abdom Imaging 28:267-275
Urban BA, Fishman EK (2000) Tailored helical CT evaluation of
acute abdomen. Radiographics 20:725-749
Werner A, Diehl SJ, Farag-Soliman M, Düber C. (2003) Multi-slice
spiral CT in routine diagnosis of suspected acute left-sided
colonic diverticulitis: a prospective study of 120 patients. Eur
Radiol 13:2596-2603
Wiesner W, Khurana B, Ji H, Ros PR (2003) CT of acute bowel is-
chemia. Radiology 226:635-650

Trauma of the Abdomen and Pelvis
Philip J. Kenney1, Stuart E. Mirvis2
1 University of Arkansas for Medical Sciences, Little Rock, AR, USA
2 University of Maryland School of Medicine, Baltimore, MD, USA
Introduction
Trauma is a major health problem in all age groups but it
is especially true in the young, due to high-velocity trans-
portation, altercations, including with weapons and re-
sulting in penetrating injuries, as well as falls and sports-
related injuries. In addition, both the elderly and pregnant
women are vulnerable to trauma. Improvements in the
management of trauma include more rapid rescue, better
organization of trauma centers, and advances in treat-
ment. Current trends include increased non-operative
management of trauma-related injuries, accurate imaging-
based diagnosis, and greater emphasis on the efficient but
cost-effective use of imaging. One aspect of the trend to
non-operative care is the desire to avoid non-therapeutic
surgery; this is possible if imaging can identify those
patients who require surgery. Another is the realization
that non-operative care can result in better long-term
outcome, such as splenic salvage.
Computed Tomography vs. Ultrasound
Controversy exists about the appropriate use of comput-
ed tomography (CT) vs. ultrasound (US), although each
modality has its advantages and disadvantages [1, 2]. In
general, CT has the best statistical accuracy for detect-
ing, characterizing, and excluding injuries. In modern
high-volume trauma centers, the CT apparatus must be
located in the trauma suite such that even unstable pa-
tients can be examined quickly without compromise.
This allows for the efficient use of CT in rapid and ac-
curate diagnostics and obviates the need for outmoded
studies, such as diagnostic peritoneal lavage. CT is also
more reliable at excluding injury, allowing the patient to
be discharged home and avoiding the expense of obser-
vation in hospital.
However CT may be overused; indeed, in one study
only three of 100 patients had alterations of clinical man-
agement due to follow-up CT [3]. US can detect signifi-
cant injury which can then be appropriately treated; con-
versely, low-risk patients with normal sonograms may be
observed and possibly avoid CT [2]. However, patients
IDKD 2010-2013
with abnormal US findings often require further evalua-
tion with CT. In a large study, US had 86% sensitivity and
98% specificity but with 43 false-negative and 23 inde-
terminate studies, including six splenic, one liver, one
renal, one pancreatic, and one bowel injury [2].
In traumatized pregnant women, US should be the first
examination as it can evaluate the pregnancy, document-
ing fetal death or viability. US is nearly as accurate in de-
tecting the abnormal presence of fluid in pregnant patient
as in non-pregnant patients [4]. If US shows fluid or oth-
er injury, CT is justified for further evaluation (Fig. 1).
The best outcome for the fetus is assured by best care of
the mother. The radiation risk is reasonable if there is life-
threatening injury, such that prompt diagnosis and treat-
ment are paramount [5].
Urinary Tract
The nearly universal use of CT has altered the assessment
of urinary tract trauma. While significant hematuria has
been shown to be the best indicator of urinary tract injury,
presently the decision to perform CT has little to do with
the presence or absence of hematuria. CT is a primary in-
vestigation, after standard radiographs, in those with sig-
nificant mechanisms of injury or any signs or symptoms
of significant injury. Intravenous urography has been re-
placed by CT (Fig. 2).
The ability of US to evaluate renal injury is limited [6]
whereas CT has excellent negative predictive value for re-
nal injury. CT also accurately indicates the presence and
type of renal injury [7]. Renal contusion appears as an ill-
defined region of diminished enhancement. Segmental re-
nal infarction is identified as a wedge-shaped, well-defined
area of non-enhancement. Renal artery occlusion can be
accurately diagnosed by its complete lack of either contrast
enhancement or excretion by the kidney, usually with little
to no associated hematoma. Angiography is thus not need-
ed, and conservative therapy is most often used today. Most
renal injuries are lacerations, with simple lacerations lim-
ited to the cortex and deep lacerations extending into the
collecting system, which may show extravasation. Delay
scans of 2-10 min aid in demonstrating or excluding
Trauma of the Abdomen and Pelvis 15

a b

Fig. 1 a-c. A pregnant woman suffered a high-speed motor vehicle

collision. a US demonstrates intrauterine fetus (heart motion doc-
umented) and free pelvic fluid. b US shows perisplenic fluid with
hypoechoic defect. c CT confirms splenic laceration; note the high-
er density of the perisplenic hematoma compared to the rim of flu-
id about liver (sentinel clot sign) c

a b

Fig. 2 a, b. Hematuria
and left upper quadrant
pain after a football-
related injury. a Intra-
venous urogram shows
no abnormality. b Sub-
sequent CT for persis-
tent pain showed free
fluid in the pelvis and
extensive splenic lacera-
tion with extensive
“blush”. Surgery con-
firmed a grade 4
splenic injury

extravasation (Fig. 3), although in most cases small
amounts of extravasation will resolve with conservative
therapy. Subcapsular hematoma is delimited by the renal
cortex and may deform the renal surface; perinephric
hematoma extends from the renal surface to fill Gerota’s
space but does not deform the renal contour, although it
may displace the kidney. CT is excellent at demonstrating
the extent of hematoma and in evaluating enlargement on
follow-up scans [7]. Renal fracture indicates a single com-
plete fracture plane, often extending through the collecting
system; multiple planes of disruption are seen in a shat-
tered kidney. CT can also diagnose avulsion of the uretero-
pelvic junction (UPJ) or ureteral injury, demonstrating lack
of opacification of the ureter, retroperitoneal water attenu-
ation collections adjacent to the pelvis or ureter, and pos-
sibly extravasation of contrast on delay scans (Fig. 4) [7].
16
a a
b
Fig. 3 a, b. Hematuria after fall from a power line. a Initial CT shows
left renal laceration with perinephric hematoma. b Delay image
shows no leak from the collecting system
Ureteral injuries, including UPJ avulsion, are uncom-
mon. They can occur either with penetrating trauma or
high-velocity blunt trauma and have no specific signs or
symptoms, but can be detected with CT. Routine CT
scans show subtle suggestive signs, such as perinephric
and peripelvic stranding or fluid, that indicate the need
for delay scans, if not routinely done, to demonstrate ex-
travasation. In a study based on over 4,000 trauma pa-
tients, CT enabled the correct identification of seven of
eight UPJ avulsions [8].
The AAST Organ Injury severity scale for the kidney
includes lesions with different appearances in each cate-
gory (1: contusion, small subcapsular hematoma; 2: <1 cm
laceration without extravasation; 3: >1 cm laceration
without extravasation; 4: deep laceration with extravasa-
tion or main renal artery or vein injury; 5 shattered kid-
ney or UPJ avulsion) and has been shown to correlate
with need for surgery and outcome [9].
Urethral injuries are predominantly seen in males. An-
terior urethral ruptures most commonly occur due to
straddle injury. Posterior urethral ruptures most often are
Philip J. Kenney, Stuart E. Mirvis
b
Fig. 4 a, b. Routine trauma CT image shows fluid and stranding
about the right ureter. a Note that contrast filling of the ureter has
not yet occurred. Delay image shows extravasation medial to the
kidney typical, of UPJ avulsion. b Note the intact parenchyma
due to compressive force and resultant pubic bone frac-
tures, although both anterior and posterior urethral injury
can result from penetrating injury. Retrograde urethro-
graphy is the only accurate diagnostic imaging procedure.
If a urethral injury is strongly suspected, a urethrogram
should be performed before passage of a catheter (Fig. 5).
However, in patients with moderate risk, a urethral
catheter may be gently passed so that the patient may go
on to CT. A pericatheter urethrogram may then be per-
formed after any other injuries have been stabilized. Five
types of urethral injuries are recognized. In type 1, the
posterior urethra is stretched but intact; in type 2, there is
a tear of the membranous urethra above the urogenital di-
aphragm; in type 3, the posterior urethral tear is above
and below the urogenital diaphragm; type 4 is defined as
a bladder-neck injury and type 5 as an anterior urethral
injury [10].
Bladder injuries consist of contusions and ruptures;
classically, they have been detected with standard radi-
ographic cystography (Fig. 6). They may be extraperi-
toneal, most commonly, intraperitoneal, less commonly,
or combined in about 5%. While CT with only intra-
venous contrast may fail to identify extravasation from
Trauma of the Abdomen and Pelvis
Fig. 5. Blunt trauma resulted in pubic rami fractures. Retrograde
urethrogram reveals type 3 posterior urethral rupture
Fig. 6. Gross hematuria after gunshot wound to the pelvis. Standard
cystogram shows extraperitoneal rupture (X marks entry site, O exit)
a ruptured bladder, several studies have shown very
high accuracy for CT-cystography (Fig. 7), which is
now the standard in our institutions. In patients with
suspected bladder rupture (primarily those with gross
hematuria, over 25 RBC/hpf, with pelvic fractures or
unexplained pelvic fluid), a standard CT with the
17
Fig. 7 a, b. Gross hematuria following motor-vehicle collision re-
sulting in extensive pelvic fractures. a Standard CT shows pelvic
fluid but no extravasation. b CT cystogram documents extraperi-
toneal bladder rupture (note the clot in the bladder)
bladder catheter clamped is performed. If there is no
extravasation, the bladder is drained and then re-filled
with 300-500 mL of dilute contrast and the pelvis
rescanned. Bladder ruptures are virtually always associ-
ated with fluid or hematoma in the pelvis, but such
blood or fluid may be due to splenic or other injuries or
to pelvic fracture. Extravasation confined to the lower
pelvis and not outlining bowel loops (and which may
extend up the retroperitoneum) indicates extraperi-
toneal rupture, which most often is managed conserva-
tively. Extravasation high near the dome and outlining
bowel loops or extending to the gutters or higher indi-
cates intraperitoneal rupture, which is more often man-
aged surgically [11]. In a study of 495 patients with po-
tential pelvic injuries, CT-cystography detected 98% of
the bladder injuries while standard cystography detect-
ed 95%. Of the patients with bladder injury (65%
extraperitoneal, 35% intraperitoneal, 5 combined), 89%
had gross hematuria and pelvic fracture, 9% had gross
hematuria with no pelvic fracture, and one patient had
microscopic hematuria and pelvic fracture [12]. CT-
cystography carried out with multidetector CT allows
for reformatting, which can more clearly demonstrate
the point of leakage and allows more accurate charac-
terization of the type of injury [13].
18
Bowel and Mesenteric Injuries
Bowel and mesenteric injuries are found in about 5% of
patients undergoing surgery for trauma and are seen in
0.7% of all traumatized patients [1, 14]. The mechanism
of injury is direct compressive force, including from seat-
belts, although deceleration may play a role. Morbidity
and mortality can occur, with peritonitis and abscess re-
sulting if the injury is missed. Clinical signs and symp-
toms are non-specific. Although diagnosis by CT is not
as straightforward as is the case for other abdominal or-
gan injuries, CT is the most accurate diagnostic modali-
ty, with >90% sensitivity and specificity reported [14,
15]. The use of orally administered contrast is now some-
what controversial; while extravasation of oral contrast
can be a very specific sign of bowel injury, it is rarely
seen. Contrast administration delays performance of the
scan, and most bowel injuries will be evident based on
other signs. There is no one CT sign that is both sensitive
and specific for bowel or mesenteric injury. Focal bowel
wall thickening, mesenteric stranding, interloop fluid, and
hematoma are common but less specific, particularly for
surgically important injuries (Fig. 8). Active bleeding,
vessel beading, abrupt termination of mesenteric vessels,
and bowel wall defect are more specific but less sensitive
signs [16]. Active bleeding is seen as a focal extralumi-
nal collection with attenuation similar to that of the aor-
ta at the same level and different from the adjacent or-
gans. Free air is considered a good sign of perforated
bowel, but in fact it has limited value. It is infrequently
seen in those with bowel injury and may represent air
tracking into the peritoneum from thoracic injuries. In a
study reported in 2008, free air had a sensitivity of 24%
albeit a specificity of 95%. There were three false-posi-
tives with intraperitoneal air instead resulting from supra-
diaphragmatic or bladder injuries [16]. If a single finding
Fig. 8. Motor-vehicle collision. Focal hematoma and thickening of
cecum; at surgery, cecal laceration found
Philip J. Kenney, Stuart E. Mirvis
is noted, the likelihood of injury is low; a combination of
findings, particularly free fluid without obvious source in
combination with focal bowel wall thickening and/or
mesenteric stranding, is very suggestive of bowel injury
and such patients should be explored or followed very
carefully [14-16]. In our practice, we have found that per-
forming a repeat abdominal-pelvic CT 4-6 h after the ad-
mission scan can be helpful in patients with suspicious
but non-diagnostic findings for full-thickness bowel in-
jury, by demonstrating injury progression such as the de-
velopment of free air, increasing intraperitoneal fluid, or
stability of findings. Of course, management decisions
are made in conjunction with any evolution of the clini-
cal findings.
Splenic Injuries
The spleen is the most frequently injured abdominal
organ in blunt trauma. There may be signs of blood loss
or left upper quadrant pain, but the diagnosis largely
rests on imaging or surgical exploration. A trend to non-
operative management is supported by evidence that
long-term health is better in those who have had splenic
function preserved. This necessitates accurate non-
invasive diagnosis and is aided by signs predictive of the
success or failure of conservative management.
Splenic injuries can cause free fluid, perisplenic or
elsewhere, which can readily be detected by sonography.
Splenic injury may alter echo-texture: lacerations may be
anechoic if there is rapid bleeding, but more commonly
are more echogenic than normal spleen [2]. With such
findings on sonography, the decision whether to further
evaluate with CT or to proceed to surgery can be made
on clinical grounds. Splenic injuries may be missed by
sonography, particularly if they are not associated with
free fluid. In one large study, there were 43 false-negative
sonograms, including six splenic ruptures that required
surgery [2].
CT is quite sensitive in the detection of splenic injuries
[17]. Subcapsular hematoma is seen as a crescentic, low-
attenuation, peripheral rim; intraparenchymal hematoma
as a rounded area within the spleen with low attenuation
and no enhancement. Lacerations are common, appearing
as linear or branching low-attenuation lesions that often
extend to the surface; if so, they are often associated with
perisplenic or free fluid. Hemoperitoneum tends to be of
higher attenuation close to the source of bleeding; thus,
when the spleen is the source, the collection adjacent to
the spleen may be higher in attenuation than elsewhere, a
finding referred to as the sentinel clot sign. Lacerations
may involve the vasculature. There can be devasculariza-
tion of the spleen by hilar injury, or active extravasation
into the peritoneal cavity, or a confined area of extrava-
sation (pseudoaneurysm) (Fig. 9). Both types of extrava-
sation indicate that non-operative management may not
succeed, although angiographic embolization may control
the bleeding and allow splenic salvage [18].
Trauma of the Abdomen and Pelvis 19

a b

Fig. 9 a, b. Blunt trauma. a Routine trauma CT

image shows area of hyperdensity (arrow-
head). b Delay image shows that the hyper-
intensity is no longer visible; the lesion has
become isodense with blood pooling, indi-
cating confined pseudoaneurysm rather than
free active bleeding

A number of schemes have been devised to grade
splenic injury on CT in an attempt to predict outcome,
with variable correlation with the need for surgery [1].
One of the commonest is the AAST scoring system. In a
large study, failure of non-operative management corre-
lated with splenic injury grade: the failure rate was <10%
with grades 1 or 2, while one-third of the grade 4 injuries
and three-fourths of the grade 5 injuries required surgery
[19]. Nevertheless, in occasional cases of low-grade
injury, the patient suffered delayed rupture, while some
high-grade injuries have been successfully managed con-
servatively. Attempts have been made to develop CT-
based criteria that may be more predictive: one referred
to three key features: devascularization, laceration of
>50% of the parenchyma; contrast blush >1 cm; or large
hemoperitoneum; however, further study showed that this
approach also had limited predictive value with poor sen-
sitivity although fair specificity [20]. The additional find-
ing of traumatic pseudoaneurysm or active extravasation
(which does not confer a specific stage in the AAST scor-
ing system) increased the likelihood of failure of non-
operative management, regardless of grade [21]. Delayed
images can help distinguish between active bleeding,
which persists as a hyperdense area, and confined vascu-
lar injury (pseudoaneurysm), which washes out [22].
Patients with active bleeding are more likely to require
surgery or other forms of intervention.
of the liver and the difficulty in clearly imaging all por-
tions of the organ ultrasonographically.
Injuries to the liver include contusion, seen on CT as
an ill-defined area of low attenuation; subcapsular
hematoma, a crescentic collection limited by the capsule;
and intraparenchymal hematoma, a collection of blood
within a liver laceration. Laceration is commonest, seen
as linear or branching low-attenuation regions, some-
times with jagged margins, that can extend to the hepat-
ic surface or to vessels. Superficial lacerations are <3 cm
in size. Periportal low attenuation is usually edema, a dis-
tended inferior vena cava and renal veins, and subserosal
edema of the gallbladder wall, but on occasion may rep-
resent blood tracking along the portal veins (Fig. 10). It
is rare that periportal low attenuation is the only sign of
liver injury, and patients with only this finding should be
managed conservatively [24].
Liver injuries may require surgery but most can
be managed non-operatively. The liver, with its dual
blood supply, is relatively resistant to infarction and
has considerable functional reserve. Grading systems

Hepatic Injuries
The liver is the second most frequently injured abdomi-
nal organ, accounting for about 20% of abdominal in-
juries [1, 17]. The right lobe is more often affected than
the left, with the posterior right lobe the most commonly
injured segment. Hepatic injuries may be associated with
intraperitoneal hemorrhage, but injury may be confined
to the liver, or hemorrhage may be limited by an intact
capsule. Lacerations involving the bare area may be as-
sociated with retroperitoneal hematoma. US may show
liver lacerations, which appear similar to splenic injuries Fig. 10. Blunt trauma, shock, and aggressive resuscitation followed by
but this modality has limited sensitivity (67%, compared trauma CT. Note the periportal low attenuation tracking throughout
to 93% for CT) [23]. This is in part due to the large size the liver with intact parenchyma and no perihepatic hematoma
20
(e.g., AAST) have less direct correlation with the need
for intervention than is the case for the spleen. Although
not included in the AAST scheme, active extravasation
may predict the need for surgery or angioembolization.
Sub-classification of extravasation can be useful: extra-
vasation into the peritoneal cavity is highly correlated
with the need for intervention; intraparenchymal extra-
vasation with significant hemoperitoneum may also
require intervention; extravasation limited within a
hepatic hematoma without hemoperitoneum usually can
be managed conservatively. Actually, the success rate of
non-operative management is >80%, and clinical signs
of hemodynamic instability dictate the need for inter-
vention more than imaging features. However in one
recent study of 214 patients with hepatic injury, all 14
who showed intraperitoneal contrast extravasation on CT
required surgery [25]. Massive hemoperitoneum in six
compartments also correlated independently with the
need for surgical intervention.
Gallbladder injuries occur in <2% of major blunt ab-
dominal traumas and are usually seen with concurrent
liver injury. Injuries to the gallbladder include wall con-
tusion, intraluminal hemorrhage, laceration, and partial
or complete avulsion. CT findings may consist of focal or
diffuse wall thickening, pericholecystic fluid (blood
and/or bile), ectopic position, intraluminal clot or mucos-
al flaps, and focal mass effect on the adjacent duodenum
(Fig. 11). Full wall thickness tears result in a collapsed
lumen [26].
Fig. 11. Gallbladder rupture. Coronal reformation in blunt trauma
patient shows ectopic location and tear in gallbladder. Note en-
hancement of the gallbladder mucosa. Hematoma fills the gall-
bladder fossa
Philip J. Kenney, Stuart E. Mirvis
Adrenal Injuries
Adrenal injuries are uncommon, seen in about 2% of pa-
tients with blunt abdominal trauma, and are rarely isolat-
ed, but their presence indicates a high-energy mechanism.
In a review of 73 cases, 71% were right adrenal only, 15%
left, and 8% bilateral [27]. The right adrenal is more
prone to injury due to its location between the liver and
spine, resulting in crush injury; also, the short right
adrenal vein can transmit increased pressure. Acute
adrenal hematomas are usually 2-4 cm in diameter, round
or oval lesions with relatively high attenuation (40-60
HU, mean 55), often with stranding. Active bleeding may
be seen and correlates with poor outcome. If there is any
concern that a lesion represents an adrenal adenoma or
pre-existing mass, such as adenoma, repeat exam with
pre- and post-contrast technique can be done, or the pa-
tient simply followed-up at 6-8 weeks as hematomas will
decrease in size and attenuation.
Pancreatic and Duodenal Injuries
Pancreatic and duodenal injuries are also uncommon, to-
gether accounting for about 2% of all abdominal injuries
[28]. The mechanism of injury is anteroposterior com-
pression, with compression against the spine leading to
the actual injury; thus, pancreatic injuries usually are at
the mid-body. Like adrenal injuries, they are also often
associated with other injuries, including hepatic, vascu-
lar, splenic, renal, and gastric. Morbidity and mortality
are relatively high in part due to the multisystem trauma
seen in these patients. Significant complications are com-
mon and include pancreatitis, pseudocysts, abscess, fis-
tulas, and pneumonia. A delay in diagnosis is not unusu-
al as the initial findings can be subtle but it contributes
to high morbidity and mortality. Clinical signs include
upper abdominal pain, with laboratory signs of leukocy-
tosis and elevated amylase, but amylase also may be nor-
mal for 2-48 h in up to 40% of patients with pancreatic
injuries [28].
Duodenal contusion may be present when there is
hematoma or edema limited to the duodenal wall, per-
haps with intramural gas and focal mural thickening.
Duodenal perforation should be diagnosed when there is
an extraluminal retroperitoneal collection of contrast,
gas, or fluid, or loss of continuity of the wall. Stranding
of the retroperitoneal fat can be seen with either condi-
tion. While use of oral contrast may aid in the diagnosis,
it is not absolutely necessary; if needed, additional im-
ages can be obtained with oral contrast if it was not used
initially.
It has been reported that the pancreas appears normal
on CT for the first 12 h after injury in some 40% of pa-
tients, but this conclusion was based on older CT tech-
nology and most trauma scans do not include a pancre-
atic parenchymal phase. However, a recent multicenter
Trauma of the Abdomen and Pelvis
Fig. 12. High-speed deceleration injury with blunt-force trauma.
Note the discrete linear low attenuation defect crossing the entire
diameter of the body of the pancreas, consistent with a pancreatic
transection. Given the extent of the defect, involvement of the pan-
creatic duct is likely
study using 16- or 64-slice scanners showed a sensitiv-
ity of 60 and 47% and a specificity of 95 and 90%, re-
spectively, in 206 subjects [29]. Fracture or laceration of
the pancreatic parenchyma, active hemorrhage in the
pancreas, and hematoma between the splenic vein and
pancreas are specific findings. Lacerations appear as
hypoattenuating linear lesions; if crossing more than
half the diameter, pancreatic duct injury should be sus-
pected (which correlates with increased complications)
(Fig. 12). Hematoma and contusion are diffuse or local-
ized, ill-defined, mixed-attenuation regions within the
pancreas [28].
References
1. Novelline RA, Rhea JT, Bell T (1999) Helical CT of abdomi-
nal trauma. RCNA 37:591-612
2. Dolich MO, McKenney MG, Varela JE et al (2001) Ultra-
sounds for blunt abdominal trauma. J Trauma 50:108-112
3. Shapiro MJ, Krausz C, Durham RM et al (1999) Overuse of
splenic scoring and computed tomographic scans. J Trauma
47:651-658
4. Goodwin H, Holmes JF, Wisner DH (2001) Abdominal ultra-
sound examination in pregnant blunt trauma patients. J Trau-
ma 50:689-693
5. Lowdermilk C, Gavant ML, Qaisi W et al (1999) Screening
helical CT for evaluation of blunt traumatic injury in the preg-
nant patient. Radiographics 19:S243-255. Kenney PJ Invited
commentary S256
6. McGahan JP, Richards JR, Jones CD et al (1999) Use of ul-
trasonography in the patient with acute renal trauma. JUM
18:207-213
7. Kawashima A, Sandler CM, Corl FM et al (2001) Imaging of
renal trauma: a comprehensive review. Radiographics 221:
557-574
21
8. Ortega SJ, Netto FS, Hamilton P et al (2008) CT scanning for
diagnosing blunt ureteral and uretropelvic junction injuries.
BMC Urology 8:3-7
9. Santucci RA, McAninch JW, Safir M et al (2001) Validation
of the American Association for the Surgery of Trauma organ
injury severity scale for the kidney. J Trauma 50:195-200
10. Goldman SM, Sandler CM, Corriere JN et al (1997) Blunt ure-
thral trauma: a unified anatomical mechanical classification. J
Urol 157:85-89
11. Morgan DM, Nallamalla LK, Kenney PJ et al (2000) CT cys-
tography: radiographic and clinical predictors of bladder rup-
ture. AJR 174:89-95
12. Quagliano PV, DeLair SM, Malhotra AK (2006) Diagnosis of
blunt bladder injury: a prospective comparative study of CT
Cystography and conventional retrograde cystography. J Trau-
ma 61:410-422
13. Chan DPN, Abujudeh HH, Cushing GL et al (2006) CT Cys-
tography with multiplanar reformation for suspected bladder
rupture. AJR 187:1296-1302
14. Malhotra AK, Fabian TC, Katsis SB et al (2000) Blunt bowel
and mesenteric injuries: the role of screening CT. J Trauma
48:991-998
15. Butela ST, Federle MP, Chang PJ et al (2001) Performance of
CT in Detection of bowel injury. AJR 176:129-135
16. Atri M, Hanson JM, Grinblat L et al (2008) Surgically impor-
tant bowel and or mesenteric injury in blunt trauma: accuracy
of multidetector CT for evaluation. Radiology 249:524-533
17. West OC (2000) Intraperitoneal abdominal injuries. ARRS
Categorical course syllabus. ARRS, Reston, VA, p 87
18. Davis KA, Fabian TC, Croce MA et al (1998) Improved suc-
cess in nonoperative management of blunt splenic injuries:
embolisation of splenic artery pseudoaneurysms. J Trauma
44:1008-1013
19. Peitsman AB, Heil B, Rivera L et al (2000) Blunt splenic in-
jury in adults: multi-institutional study of the Eastern Associ-
ation for the Surgery of Trauma. J Trauma 49:177-187
20. Cohn SM, Arango JL, Myers JG et al (2009) Computed to-
mography grading systems poorly predict the need for inter-
vention after spleen and liver injuries. American Surgeon
75:133-139
21. Gavant ML, Schurr M, Flick PA et al (1997) Predicting clini-
cal outcome of nonsurgical management of blunt splenic in-
jury: using CT to reveal abnormalities of splenic vasculature.
AJR 168:207
22. Anderson SW, Varghese JC, Lucey BC et al (2007) Blunt
splenic trauma: delayed phase CT for differentiation of active
hemorrhage from contained vascular injury in patients. Radi-
ology 243:88-95
23. Richards JR, McGahan JP, Pali MJ et al (1999) Sonographic
detection of blunt hepatic trauma. J Trauma 47:1092-1097
24. Yoon W, Jeong YY, Kim JK et al (2005) CT in blunt liver trau-
ma. Radiographics 25:87-104
25. Fang J-F, Wong Y-C, Lin B-C et al (2006) CT risk factors for
the need of operative treatment in initially hemodynamically
stable patients after blunt hepatic trauma. J Trauma 61:547-554
26. Wittenburg A, Minotti AJ (2005) CT diagnosis of traumatic
gallbladder injury. AJR 185:1573-1574
27. Sinelnikov AO, Abujudeh HH, Chan D, Novelline RA (2007)
CT manifestations of adrenal trauma: experience with 73 cas-
es. Emerg Radiol 13:313-318
28. Linsenmaier U, Wirth S, Reiser M et al (2008) Diagnosis and
classification of pancreatic and duodenal injuries in emer-
gency radiology. Radiographics 28:1591-1601
29. Phelan HA, Velmahos GC, Jurkovich J et al (2009) An evalu-
ation of multidetector CT in detecting pancreatic injury: re-
sults of a multicenter AAST study. J Trauma 66:641-647
 IDKD 2010-2013

Diseases of the Esophagus and Stomach

Marc S. Levine1, Ahmed Ba-Ssalamah2
1 Department of Radiology, Penn Radiology at the Hospital of the University of Pennsylvania, Philadelphia, PA, USA
2 Department of Radiology, Medical University of Vienna, General Hospital of Vienna, Wien, Austria

Introduction
The esophagus and stomach are susceptible to a wide
spectrum of diseases, including benign and malignant tu-
mors, inflammatory diseases, and other conditions. For
the diagnosis of this large variety of disorders, multi-
modality imaging is required. Barium studies, particular-
ly double-contrast studies, continue to have a major role
in the diagnostic work-up of inflammatory diseases and
in post-operative follow-up, whereas cross-sectional
imaging studies, particularly multidetector computed
tomography CT (MDCT), is used in the pre-operative Fig. 1. Candida esophagitis.
staging of oncological disorders. In this chapter, we Upright, left posterior-oblique
review the most frequent diseases and describe the use of spot image from double-
different imaging modalities for their diagnostic work-up. contrast esophagogram shows
multiple, discrete, plaque-like
lesions in the mid-esophagus.
Note the linear configuration
Gastroesophageal Reflux Disease of the lesions and their sepa-
ration by segments of normal,
intervening mucosa. These
Mild reflux esophagitis may be manifested on double- findings are characteristic of
contrast studies by small, shallow ulcers or granularity of fungal esophagitis
the mucosa in the distal esophagus [1]. In advanced
disease, the esophagus can have an irregular contour,
with serrated margins and decreased distensibility from
ulceration, edema, and spasm. Subsequent scarring can epidemic has led to a more fulminant form of candidiasis,
lead to the development of smooth, tapered, or ring-like characterized by a “shaggy” esophagus that has a grossly
peptic strictures in the distal esophagus, almost always irregular contour due to multiple plaques and pseudo-
above a hiatal hernia [2]. membranes [5]. In contrast, herpes esophagitis is manifest-
Barrett’s esophagus is a well-recognized complication ed by small, shallow ulcers (Fig. 2) [6] whereas esophagi-
of reflux esophagitis, and it is associated with an in- tis due to cytomegalovirus (CMV) or human immunodefi-
creased risk of esophageal adenocarcinoma. The classic ciency virus (HIV) can be associated with the development
radiological features of Barrett’s esophagus include a dis- of giant, flat ulcers [7]. Since CMV ulcers are treated with
tinctive reticular pattern of the mucosa or a high antiviral agents, and HIV ulcers with steroids, endoscopy is
esophageal stricture or ulcer occurring at a discrete dis- required to differentiate CMV- from HIV-mediated
tance from the gastroesophageal junction [3]. esophagitis before treatment in these patients is instituted.

Infectious Esophagitis Drug-Induced Esophagitis

By far the most common cause of opportunistic esophageal Drug-induced esophagitis is caused by various oral med-
infection is Candida albicans. Candida esophagitis is man- ications, including tetracycline, doxycycline, potassium
ifested on double-contrast studies by discrete, plaque-like chloride, quinidine, alendronate sodium, aspirin, and oth-
lesions separated by normal mucosa (Fig. 1) [4]. The AIDS er non-steroidal anti-inflammatory drugs (NSAIDs).
Diseases of the Esophagus and Stomach
Fig. 2. Herpes esophagitis. Up-
right, left posterior-oblique
spot image from double-
contrast esophagogram shows
multiple small ulcers sur-
rounded by radiolucent
mounds of edema in the mid-
esophagus. These findings are
characteristic of viral (espe-
cially herpes) esophagitis
These patients often have a history of ingesting the pills
with little or no water immediately before retiring. As a
result, the pills tend to lodge in the mid-esophagus where
it is compressed by the aortic arch or left main bronchus.
This can result in contact esophagitis, manifested by
small, discrete ulcers in the mid-esophagus [8]. Affected
individuals may present with odynophagia, but there is
rapid clinical improvement after withdrawal of the of-
fending agent.
Idiopathic Eosinophilic Esophagitis
Idiopathic eosinophilic esophagitis (IEE) usually occurs
in adolescent or young men with a long history of com-
pensated dysphagia and occasional food impactions.
These patients often have an atopic history, asthma, or
peripheral eosinophilia. Barium studies may reveal a
“ringed esophagus”, with multiple distinctive ring-like
indentations, or a “small-caliber esophagus”, with diffuse
loss of distensibility of the esophagus in the absence of a
discrete stricture [9]. The diagnosis can be confirmed by
endoscopic biopsies showing >20 eosinophils per high-
powered field. These patients usually have a dramatic
positive response to oral steroids or inhaled steroid prepa-
rations.
Esophageal Carcinoma
Early esophageal cancers classically appear on double-
contrast studies as small, protruded tumors [10]. They
may be plaque-like or small polypoid lesions. Other,
23
Fig. 3. Adenocarcinoma of the esophagus. Prone, right anterior-
oblique spot image from single-contrast esophagogram shows a
polypoid mass (arrows) in the distal esophagus in this patient with
adenocarcinoma arising in Barrett’s esophagus
superficially spreading carcinomas may be manifested by
poorly defined nodules, producing confluent nodularity
of the mucosa [10]. In contrast, the lesions of advanced
esophageal carcinomas are polypoid (Fig. 3), ulcerated,
or infiltrating, with irregular luminal narrowing and
shelf-like borders. Rarely, these tumors have a varicoid
appearance due to submucosal spread of tumor; in such
cases, they can be mistaken for esophageal varices.
Erosive Gastritis
Erosive gastritis is usually manifested on double-con-
trast studies by varioliform erosions with punctate or
slit-like collections of barium surrounded by radiolu-
cent mounds of edema. Varioliform erosions tend to be
located in the gastric antrum and are often aligned on
the crests of the folds. Aspirin and other NSAIDs are by
far the most common cause of erosive gastritis. Occa-
sionally, NSAID-induced erosive gastritis may also
manifest as distinctive linear or serpiginous erosions
clustered together on or near the greater curvature of the
gastric body [11]. It has been postulated that these ero-
sions result from localized mucosal injury, as the dis-
solving NSAID tablets collect by gravity in the most de-
pendent portion of the stomach.
Helicobacter Pylori Gastritis
Gastritis caused by the bacterium Helicobacter pylori can
be diagnosed on barium studies by the presence of thick-
ened folds in the antrum, body, or, less commonly, the
24
Fig. 4. Helicobacter pylori gastritis. Left posterior-oblique spot im-
age from double-contrast upper gastrointestinal examination
shows thickened, irregular folds in the gastric body due to chronic
H. pylori gastritis
fundus of the stomach (Fig. 4) [12]. Other patients with
H. pylori infection may have a polypoid form of gastritis,
with grossly thickened, lobulated folds such that the
lesions resemble those of Menetrier’s disease, lymphoma,
or a submucosally infiltrating carcinoma [12]; thus, en-
doscopy and biopsy are required for a definitive diagnosis.
Gastric Ulcers
Benign gastric ulcers classically appear en face as
round or ovoid collections of barium, often surrounded
by a smooth mound of edema or thin, straight folds ra-
diating to the edge of the ulcer crater [13]. When
viewed in profile, benign ulcers project beyond the
contour of the adjacent gastric wall and are often asso-
ciated with an ulcer mound or collar. In contrast, ma-
lignant gastric ulcers appear en face as irregular ulcer
craters within a discrete mass, sometimes associated
with nodularity or clubbing of adjoining folds due to
tumor infiltration of the folds [13]. When viewed in
profile, malignant ulcers project inside the lumen with-
in a mass that forms acute angles with the gastric wall
rather than the obtuse, gently sloping angles expected
for a benign mound of edema.
Most benign ulcers are located on the lesser curvature
or posterior wall of the gastric antrum or body [13]. Oc-
casionally, benign gastric ulcers may occur on the
greater curvature of the distal stomach, in which case the
vast majority are caused by aspirin or other NSAIDs
[13]. As these NSAID-induced greater-curvature ulcers
enlarge, they can penetrate inferiorly into the transverse
colon, producing a gastrocolic fistula [14]. Ulcer healing
may be seen as a decrease in the size of the ulcer or a
Marc S. Levine, Ahmed Ba-Ssalamah
change in its shape. Usually, ulcer healing produces a
visible scar manifested by a central pit or depression, ra-
diating folds, and/or retraction of the adjacent gastric
wall [13].
Gastric Carcinoma
Advanced gastric carcinomas may appear on barium
studies as polypoid, ulcerated, or infiltrating lesions. Oth-
er primary scirrhous carcinomas can have a “linitis plas-
tica” appearance, with luminal narrowing, irregularly
thickened folds, and nodularity of the mucosa [15]. Scir-
rhous carcinomas classically involve the gastric antrum,
but 40% of these lesions are confined to the gastric body
or fundus (Fig. 5) [15]. Early gastric cancers may be man-
ifested by small polypoid or ulcerated lesions. However,
in the western world, the vast majority of patients with
gastric carcinoma already have advanced lesions at pre-
sentation. As a result, early gastric cancer is unlikely to
be detected as long as barium studies are performed pre-
dominantly on symptomatic patients [16].
Gastric Lymphoma
Chronic H. pylori gastritis can lead to the development of
mucosa-associated lymphoid tissue (MALT) in the stom-
ach. This lymphoid tissue is the precursor of low-grade,
B-cell gastric MALT lymphomas, which, if untreated,
may undergo blastic transformation to more high-grade
lymphomas. Gastric MALT lymphomas may sometimes
be recognized on double-contrast studies by variably
sized, rounded, confluent nodules in the stomach [17]. In
contrast, advanced gastric lymphomas may be manifest-
ed by thickened folds, multiple submucosal masses, ul-
cerated bull’s-eye lesions, or giant, cavitated lesions.
Fig. 5. Scirrhous adenocarcinoma of the stomach. Front spot image
from double-contrast upper gastrointestinal examination shows ir-
regular narrowing of the gastric body and fundus due to infiltration
of the wall by tumor. Note transition (arrows) to uninvolved gas-
tric antrum distally. About 40% of scirrhous carcinomas are con-
fined to the body or fundus of the stomach with antral sparing
Diseases of the Esophagus and Stomach
Gastrointestinal Stromal Tumors
Gastrointestinal stromal tumors (GISTs) are the most com-
mon mesenchymal tumors of the gastrointestinal tract.
Approximately 70% of all GISTs are found in the stomach
and only 2-5% originate from the esophagus. GISTs have a
wide clinical spectrum, ranging from benign, incidentally
detected nodules to large malignant tumors, and must be
distinguished from other mesenchymal tumors [18]. The
most frequent symptom related to gastric or esophageal tu-
mor is dysphagia or heartburn; thus, most patients undergo
endoscopy of the esophagus, stomach and duodenum, with
simultaneous biopsy if tumor is seen [19]. If the histopatho-
logical results reveal a tumor, accurate staging is required.
Endoscopic ultrasound (EUS) can depict the normal
gastric and esophageal wall with its five-layered internal
structures, thus allowing detailed evaluation of the depth
of tumor penetration even in early-stage disease. EUS is
useful in the diagnostic work-up of early cancer, as it can
distinguish between T1a tumors, in which only mucosec-
tomy is needed, and T1b tumors, in which a complete re-
section is indicated [20]. However, because ultrasound
penetration is not deep enough when transducers with
higher frequencies are used to visualize fine structures,
evaluation of deep tumor infiltration may be difficult and
assessment of metastases may be limited by the finite
depth of penetration [21]. EUS also is examiner-depen-
dent, time-consuming, and unable to pass stenotic tumors.
These limitations can be overcome using multidetector
CT technology (MDCT), with its ability to cover a large
volume in a very short scan time with a single breath-
hold. Thin collimation and isotropic voxels allow imaging
of the entire esophagus and stomach with high-quality
multiplanar reformation and 3D reconstruction (Fig. 6)
Fig. 6. Hydro-MDCT of the esophagus in coronal reformation to fol-
low the course of the esophagus, demonstrating normal wall thick-
ening of the esophagus (≤3 mm) and homogeneous enhancement
25
[22, 23]. Furthermore, MDCT with water filling (hydro-
MDCT, HMDCT) provides information about esophageal
and gastric wall infiltration, extramural extent of disease,
lymph node involvement, and distant metastases (Fig. 7)
[23, 24]. Inadequately distended hollow viscera on CT
may hide large lesions and may even mimic pseudole-
sions. Thus, optimal distention of the esophagus and
stomach is a necessary prerequisite for achieving good di-
agnostic imaging. When water is used as the oral contrast
agent, subtle pathology is easier to visualize. Gas or CO2
resulting from the administration of effervescent granules
can be used for hollow-organ distention alone or in com-
bination with water. The use of negative rather than posi-
tive contrast media is preferred, especially if CT angio-
graphy images are needed. Subtle pathology is easier to
visualize, especially when an adequate intravenous con-
trast material bolus is administered [23, 24]. Three-
dimensional reconstructions of CT data sets with multi-
planar reconstruction, curved planar reformations, or
other protocols are mandatory to exploit the full potential
of MDCT. Three-dimensional virtual gastroscopy prov-
ides an endoluminal image similar to conventional fiber-
optic gastroscopy. Therefore, HMDCT is a valuable tool
for the complete staging of gastric and esophageal tumors
and serves as an adjunct to endoscopy.
However, EUS and MDCT are anatomically based di-
agnostic techniques with certain drawbacks. These in-
clude limited sensitivity with false-negative findings due
Fig. 7. Hydro-MDCT of the esophagus in coronal reformation
shows a huge mass with inhomogeneous enhancement in the me-
diastinum arising from the esophageal wall, with infiltration of the
left main bronchus (arrow) and the left diaphragm (arrowhead),
with regard to the T4 tumor. Note the pleural effusion
26 Marc S. Levine, Ahmed Ba-Ssalamah

to non-enlarged, tumor-involved lymph nodes, and limit- a

ed specificity with false-positive findings due to enlarged
lymph nodes not involved by tumor. Furthermore, after
neo-adjuvant chemotherapy and re-evaluation, it is not
possible with these techniques to distinguish between fi-
brotic changes or remaining vital tumor due to their mor-
phological similarities [25].Thus, there is an urgent need
for additional functional examination techniques.
Positron emission tomography (PET) yields physio-
logical information that provides a means to diagnose
cancer based on altered tissue metabolism. PET takes ad-
vantage of the principle that biochemical changes often
precede or are more specific than the structural changes b
associated with any given disease process [26]. There-
fore, PET offers the potential to show early esophageal
cancer or small lymph node metastases before any struc-
tural abnormality is detectable or to exclude the presence
of tumor in an anatomically altered structure. For exam-
ple, FDG-PET can detect metastatic lymph nodes that are
not enlarged on CT, and can help differentiate pathologi-
cal from non-specifically enlarged lymph nodes, which
usually show no uptake of FDG. Tumor uptake of FDG,
measured as the maximal standardized uptake value
(SUVmax) in FDG-PET, even provides a quantitative
estimate of tumor aggressiveness [27]. Fig. 8 a, b. Hydro-MDCT PET scan of an early cervical esophageal
Recent studies demonstrated that FDG-PET can be carcinoma and enlarged right-sided lymph node. a On the axial
HMDCT scan of the upper thorax, the small esophageal cancer is
used not only for pre-treatment staging, but also for as- not clearly seen and the enlarged lymph appears suspicious. b Cor-
sessment of treatment response, detection of recurrence, responding axial fused HMDCT-PET image shows increased ac-
and prediction of survival in patients with adenocarcino- tivity in the region of the small primary tumor (arrow) and an en-
ma of the esophagus or stomach [27]. However, for mu- larged metastatic lymph node (arrowhead)
cinous and signet-ring cell adenocarcinoma of the stom-
ach, it is less useful. This may be due to low or absent
FDG activity in these tumors, which is the result of a high
content of metabolically inert mucus, leading to a re- for patients with esophageal and gastric tumors. HMDCT
duced FDG concentration. Another reason could be the may help distinguish surgical candidates with limited dis-
lack of expression of the glucose transporter Glut-1 on ease from patients in need of pre-operative chemoradia-
the cell membrane of most signet-ring cell and mucinous tion for down-sizing a tumor or from patients who need
adenocarcinomas [28]. palliative therapy for advanced, unresectable tumor.
The spatial resolution of FDG-PET is lower than that When HMDCT shows definite advanced disease with ex-
of CT scans. When metastatic nodes exist around the pri- tensive tumor spread, pre-surgical chemotherapy or ra-
mary tumor, it can be difficult to distinguish uptake in diochemotherapy is used to improve the prognosis. After
these nodes from the intense activity of the primary tu- completion of neo-adjuvant treatment, the tumor can then
mor. However, the advent of PET/CT imaging, enabling be restaged to determine whether surgical resection is in-
co-registration of both anatomical and functional infor- dicated. Thus, pre-operative staging of esophageal and
mation, has overcome this disadvantage and improved the gastric tumors is by far the most important indication for
localization of increased FDG uptake (Fig. 8) [29]. FDG- HMDCT. This technique also plays an important role in
PET/CT shows the extent of disease more accurately than the evaluation of postoperative complications such as fis-
other imaging methods, and this frequently leads to a rad- tulas when the findings on barium studies are equivocal.
ical change in patient management. Combined PET/CT
imaging is therefore a valuable diagnostic tool for the pri-
mary diagnosis of GISTs or assessment of the response References
to therapy [30, 31]. PET/CT scans can also be used in 1. Graziani L, Bearzi I, Romagnoli A et al (1985) Significance of
staging patients with primary gastric lymphoma, as well diffuse granularity and nodularity of the esophageal mucosa at
as for monitoring these tumors after therapy [30]. How- double-contrast radiography. Gastrointest Radiol 10:1-6
ever, the availability of FDG-PET, and FDG-PET/CT in 2. Gupta S, Levine MS, Rubesin SE et al (2003) Usefulness of
barium studies for differentiating benign and malignant stric-
particular, is still limited and their use expensive. tures of the esophagus. AJR 180:737-744
Thus, at present, HMDCT plays a major role as a 3. Levine MS, Kressel HY, Caroline DF et al (1983) Barrett esoph-
triage tool to aid in choosing the appropriate treatment agus: reticular pattern of the mucosa. Radiology 147:663-667
Diseases of the Esophagus and Stomach
4. Levine MS, Macones AJ, Laufer I (1985) Candida esophagitis:
accuracy of radiographic diagnosis. Radiology 154:581-587
5. Levine MS, Woldenberg R, Herlinger H, Laufer I (1987) Oppor-
tunistic esophagitis in AIDS: radiographic diagnosis. Radiology
165:815-820
6. Levine MS, Loevner LA, Saul SH et al (1988) Herpes esophagi-
tis: sensitivity of double-contrast esophagography. AJR 151:57-62
7. Sor S, Levine MS, Kowalski TE et al (1995) Giant ulcers of the
esophagus in patients with human immunodeficiency virus:
clinical, radiographic, and pathologic findings. Radiology
194:447-451
8. Bova JG, Dutton NE, Goldstein HM, Hoberman LJ (1987)
Medication-induced esophagitis diagnosed by double-contrast
esophagography. AJR 148:731-732
9. Zimmerman SL, Levine MS, Rubesin SE et al (2005) Idio-
pathic eosinophilic esophagitis in adults: the ringed esopha-
gus. Radiology 236:159-165
10. Levine MS, Dillon EC, Saul SH, Laufer I (1986) Early
esophageal cancer. AJR 146:507-512
11. Levine MS, Verstandig A, Laufer I (1986) Serpiginous gastric
erosions caused by aspirin and other nonsteroidal antiinflam-
matory drugs. AJR 146:31-34
12. Sohn J, Levine MS, Furth EE et al (1995) Helicobacter pylori
gastritis: radiographic findings. Radiology 195:763-767
13. Levine MS, Creteur V, Kressel HY et al (1987) Benign gastric
ulcers: diagnosis and follow-up with double-contrast radiogra-
phy. Radiology 164:9-13
14. Levine MS, Kelly MR, Laufer I et al (1993) Gastrocolic fistu-
las: the increasing role of aspirin. Radiology 187:359-361
15. Levine MS, Kong V, Rubesin SE et al (1990) Scirrhous carci-
noma of the stomach: radiographic and endoscopic diagnosis.
Radiology 175:151-154
16. White RM, Levine MS, Enterline HT, Laufer I (1985) Early
gastric cancer: recent experience. Radiology 155:25-27
17. Yoo CC, Levine MS, Furth EE et al (1998) Gastric mucosa-as-
sociated lymphoid tissue lymphoma: radiographic findings in
six patients. Radiology 208:239-243
18. Miettinen M, Sarlomo-Rikala M, Sobin LH, Lasota J (2000)
Esophageal stromal tumors: a clinicopathologic, immunohis-
tochemical, and molecular genetic study of 17 cases and com-
parison with esophageal leiomyomas and leiomyosarcomas.
Am J Surg Pathol 24:211-222
27
19. Galmiche JP, Clouse RE, Balint A et al (2006) Functional
esophageal disorders. Gastroenterology 130:1459-1465
20. Jung M (2005) [Mucosectomy as sufficient therapy for early
squamous cell]. Chirurg 76:1018-1024
21. Kutup A, Link BC, Schurr PG et al (2007) Quality control of
endoscopic ultrasound in preoperative staging of esophageal
cancer. Endoscopy 39:715-719
22. Prokop M. New challenges in MDCT (2005) Eur Radiol 15
Suppl 5:E35-45
23. Ba-Ssalamah A, Prokop M, Uffmann M et al (2003) Dedicat-
ed multidetector CT of the stomach: spectrum of diseases. Ra-
diographics 23:625-644
24. Ba-Ssalamah A, Zacherl J, Noebauer-Huhmann IM et al
(2009) Dedicated multi-detector CT of the esophagus: spec-
trum of diseases. Abdom Imaging 34:3-18
25. Westerterp M, van Westreenen HL, Reitsma JB et al (2005)
Esophageal cancer: CT, endoscopic US and FDG PET for as-
sessment of response to neoadjuvant therapy-systematic re-
view. Radiology 236:841-851
26. Luketich JD, Schauer PR, Meltzer CC et al (1997) Role of
positron emission tomography in staging esophageal cancer.
Ann Thorac Surg 64:765-769
27. Cerfolio RJ, Bryant AS (2006) Maximum standardized uptake
values on positron emission tomography of esophageal cancer
predicts stage, tumor biology, and survival. Ann Thorac Surg
82:391-394; discussion 394-395
28. Chen J, Cheong JH, Yun MJ et al (2005) Improvement in pre-
operative staging of gastric adenocarcinoma with positron
emission tomography. Cancer 103:2383-2390
29. Hsu WH, Hsu PK, Wang SJ et al (2009) Positron emission to-
mography-computed tomography in predicting locoregional
invasion in esophageal squamous cell carcinoma. Ann Thorac
Surg 87:1564-1568
30. Suga K, Yasuhiko K, Hiyama A et al (2009) F-18 FDG
PET/CT findings in a patient with bilateral orbital and gastric
mucosa-associated lymphoid tissue lymphomas. Clin Nucl
Med 34:589-593
31. Antoch G, Kanja J, Bauer S et al (2004) Comparison of PET,
CT, and dual-modality PET/CT imaging for monitoring of
imatinib (STI571) therapy in patients with gastrointestinal
stromal tumors. J Nucl Med 45:357-365

Small-Bowel Imaging: Pitfalls in Computed Tomography
Enterography/Enteroclysis
Marc J. Gollub
Memorial Sloan Kettering Cancer Center, New York, NY, USA
Computed Tomography Enterography
Computed tomography enterography (CTEG) is a fo-
cused CT scan examination of the small intestine that
combines the advantages of isotropic, thin-section multi-
planar CT; the large volumes of neutral-density oral con-
trast; and rapid administration of intravenous contrast.
Oral contrast agents such as 0.1% barium, polyethylene
glycol (PEG) and methylcellulose contain additives that
inhibit fluid reabsorption and allow maximal bowel dis-
tention. Intravenous contrast is injected at the “enteric
phase” (about 45 s) to provide maximum wall enhance-
ment against the neutral-density lumen (0-30 HU) [1].
Pharmacological manipulation to interrupt small-bowel
spasm and encourage gastric emptying is commonly
used, including glucagon and metoclopramide, respec-
tively. The indications for CTEG include Crohn’s disease
and other enteritides, obscure gastrointestinal bleeding
(OGIB), detection of intestinal masses, and sprue. A
prospective blinded comparison of CTEG with wireless
capsule endoscopy (WCE) using clinical consensus as the
gold-standard found similar sensitivities (82 vs. 83%) for
active small-bowel Crohn’s disease, but CTEG was far
more specific than WCE (89 vs. 53%). Although there is
less experience in using CTEG for OGIB, one study
found a bleeding source in 45% of 22 patients, in three of
whom the source had been missed by initial WCE [2]. No
comprehensive reports of mass detection have been pub-
lished yet, but in our experience CTEG appears to repre-
sent a first-line test for suspected carcinoid. This chapter
will discuss the pitfalls and limitations of CTEG and CT
enteroclysis (CTEC).
Pitfalls of Computed Tomography Enterography
General
Even with the aid of neutral-density oral contrast to as-
sist in mass conspicuity, mildly enhancing masses or
mural inflammation may be subtle and overlooked with-
out proper adjustment of the window and level. We rec-
ommend, in addition to standard abdominal settings, a
IDKD 2010-2013
liver-type setting (W215, L135) and an “enterographic-
type” setting (W430, L155). Since CTEG does not create
an enteral challenge, in contrast to CTEC, its use in the
setting of low-grade small-bowel obstruction should be
discouraged in favor of CTEC or magnetic resonance en-
teroclysis. However, even during a non-obstructive
episode, CTEG may confer an advantage by delineating
a subtle, underlying cause such as a mass or stricture.
Fastidious technique and timing are basic requirements
in CTEG. Inadequate oral intake, early or late scanning,
poor intravenous enhancement (low rate, reduced dose), or
underlying hypo- or hypermotility disorders can interfere
with good distention and wall conspicuity. Some of these
technical pitfalls can be overcome with low-kVp imaging,
close monitoring of the drinking schedule, and pharmaco-
logical manipulation. Poor jejunal distention is a generally
expected limitation compared with other examinations.
Fortunately, many abnormalities are diagnosed at CTEG
solely by virtue of mural hyperenhancement.
Crohn’s Disease
Proof of disease, grading of severity, and assessment of
penetrating disease are important tasks prior to surgical
planning and even prior to medical treatment, since the
therapeutic anti-tumor necrosis factor alpha (TNFα) an-
tibodies (e.g., infliximab) are expensive and have side ef-
fects of infection and drug-induced immune disease [3].
The classic findings at CTEG in patients with active
Crohn’s disease have been well described [2]. The Amer-
ican College of Radiology has deemed CTEG to be the
most appropriate imaging test for Crohn’s disease [4].
Nonetheless, studies testing the accuracy of CTEG com-
pared with other radiological and endoscopic tests have
identified some clear limitations and pitfalls for CTEG in
the setting of suspected Crohn’s disease, including:
1. non-specificity of certain findings, such as mural hy-
perenhancement without skip areas;
2. in known Crohn’s disease, signs of very early disease,
such as aphthous ulcers, which will not be detected un-
less a mucosal study is done;
3. measurement of disease severity, including quantita-
tion of ulcers or fistulas; and
Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis
4. correlation with clinical findings.
Wold et al. found CTEG to be superior to small-bow-
el follow-through for the detection of abscesses and fis-
tulas, with a sensitivity and specificity of 78/83% vs.
62/90%, respectively [5].
Vogel et al. showed that, although CTEG is accurate for
determining the presence or absence of strictures and fis-
tulas (sensitivity 100, 92% respectively), it is less accurate
in determining their number (sensitivity 67%, both). This
may be clinically significant since uncorrected strictures
or fistulas can result in postoperative symptoms [6].
Solem et al. found that the sensitivity of WCE and CTEG
in the detection of active small-bowel Crohn’s disease was
identical (both 83%), and there was no significant differ-
ence with ileocolonoscopy (74%) or small-bowel follow-
through (65%) [7]. Hara et al., investigating disease activ-
ity, showed that findings at CTEG correlated with symp-
toms in 80% of patients, indicating CTEG’s excellence as
a monitoring test [8]. In a study by Higgins et al., clini-
cians suspected only 84% of CTEG-identified strictures
and CTEG excluded strictures in >50% of patients with a
clinically suspected stricture [9]. Furthermore, a survey of
clinicians showed that, following the use of CTEG, clini-
cal management changes 50% of the time [10].
Obscure Gastrointestinal Bleeding
In the search for sources of gastrointestinal (GI) bleeding,
CTEG has found increasing use since it may be a more
sensitive, non-invasive method – using intravenous con-
trast and CT angiography techniques – made possible by
neutral-density oral contrast. However, its application is
limited to being a diagnosis-only test, with no therapeu-
tic capability. Abstracts from the 2009 Radiological So-
ciety of North America (RSNA) meeting in which the de-
tection of bleeding rates using various techniques was
compared suggested equal sensitivity with tagged RBC
and catheter digital subtraction angiography, such that
bleeds of as little as 0.3 mL/min were detected.
The search for bleeding sources in the small bowel
may be subsequent to an apparently negative upper- and
lower-GI endoscopy. It should be kept in mind that
missed gastric and colonic pathology may still be discov-
ered at imaging if these areas are well-filled, since endo-
scopic examinations are performer-dependent, imperfect
gold-standards.
Conventional barium studies do not identify mucosal
erosions or vascular ectasia and have yields of 6% for
small-bowel follow-through and 10% for small-bowel en-
teroclysis [11]. Angioectasias are the most common cause
of OGIB in patients over the age of 50. These lesions are
typically small, may be multiple, and are sessile or slight-
ly raised. They are typically only visible at endoscopy or
potentially at arterial-phase catheter-based or CT angio-
graphy. A major limitation of CTEG is its relative insensi-
tivity to these small, flat, vascular lesions. The most recent
study used optimized thin-sections, three scanning phases
(arterial, enteric, and delayed), rapid boluses of intravenous
29
contrast, and 1800 mL of neutral-density contrast. With
this approach, the majority of patients with a source of
OGIB were detected compared with WCE, surgery, or fol-
low up. Ten of 22 (43%) studies showed positive findings,
including the detection of angioectasias [12]. Huprich et
al. emphasized that, while three-phase CTEG can be done
with the goal of detecting angioectasias and other arterial-
phase dominant lesions, the radiation incurred is substan-
tial (effective dose 59 mSv per exam) such that the relative
risks and benefits have to be weighed, including patient
age, necessity of repeat examinations, and the known risk
of radiation with multiple CT exams, especially in younger
patients [12, 13]. A follow-up prospective study using
three-reader comparisons, presented by Huprich at RSNA
2009, indicated that two out of three phases are often ade-
quate, but conclusive results have not been published. In a
similar study by Hara, using three-phase CTEG, 33% of le-
sions were detected (specificity 89%) and 52% were de-
tectable in retrospect. Some of the missed lesions were in
the stomach and colon, emphasizing the advantage of
CTEG in depicting abnormalities throughout the GI tract.
Currently, the conditions that appear to be undetectable by
CTEG, as proven by other tests, include ulcers, vascular le-
sions, and non-bleeding lesions [14].
Neoplasms
The accuracy of CTEG in the detection of masses such as
carcinoid, adenocarcinoma, lymphoma, gastrointestinal
stromal tumor, and polyps is not known. Capsule en-
doscopy, an examination being used more frequently to
examine the small bowel, has several documented limita-
tions, such that CTEG plays an important ongoing role for
small-bowel mass detection and is especially important
for masses that may have a predominant extraluminal
component [15]. A recent RSNA 2009 Abstract indicated
that masses in the setting of OGIB were better detected at
CTEG than at WCE. Hyperenhancing lesions (e.g., some
melanoma metastases or carcinoid tumors) will be easily
detected with careful observation and windowing; howev-
er, the detection of isoattenuated masses (some melanoma
metastases, polyps, and even some primary adenocarcino-
mas) and smaller masses may require excellent luminal
distention and perusal for secondary findings, such as in-
creased wall thickening, increased luminal caliber, or
complications (intussusception, obstruction). CTEG may
underestimate the number of lesions due to their smaller
size or the lower degree of vascularity of small tumors.
Computed Tomography Enteroclysis
This fused test combines the advantages of enteral chal-
lenge from a catheter small-bowel examination (entero-
clysis) with the isotropic, multiplanar, cross-sectional im-
ages obtained at helical CT. Indications for CTEC include
small-bowel obstruction, small-bowel masses, OGIB,
Crohn’s disease, and malabsorption.
30
Pitfalls of Computed Tomography Enteroclysis
General
Careful attention to technique, including the use of infu-
sion rates that allow uniform bowel distention; catheter
placement in the proper location, with correct balloon in-
sufflation and tip placement; the appropriate administra-
tion of pharmacological agents; and proper use of multi-
detector CT with reformatted images, will obviate many
sources of misinterpretation. Spasm and inadequate filling
should be easily recognized and addressed by adjustment
of the pump flow-rate and use of spasmolytics (glucagon
or Buscopan). Variable window and level settings are ad-
vised to adjust for high-attenuation contrast or, in the in-
terpretation of a neutral-density exam, to appreciate sub-
tle differences in enhancement of the bowel wall.
Crohn’s Disease
In the diagnosis of suspected or known Crohn’s disease,
a comparison of WCE with neutral-density oral contrast
CTEC (nCTEC) in 56 patients showed that 27 could not
undergo WCE due to strictures (≤10 mm) for fear of cap-
sule retention. In the other 41 patients, the limitations of
nCTEC included its inability to detect very early, mini-
mal inflammatory changes or small mucosal lesions such
as villous denudation, aphthoid ulcerations, or erosions.
These were far better detected by WCE than CTEC
(p = 0.004) [16-18]. Since the detection of non-elevated
or non-depressed lesions in Crohn’s disease and superfi-
cial erosions in NSAID enteropathy will be limited, these
entities would probably be best investigated using others
methods, such as push enteroscopy, single- or double-bal-
loon endoscopy, WCE or “air” (C02) double-contrast flu-
oroscopic enteroclysis without CT [19, 20]. In addition,
in late-stage Crohn’s disease, nCTEC may show fewer
fistulae than positive oral contrast CTEC (pCTEC), as
the higher-density contrast may fill the GI tract more
conspicuously.
Masses
In the workup of small-bowel masses or OGIB, several
pitfalls may be encountered. False-positive masses may be
seen in patients with Kerkring fold thickening or even
transient intussusception [21]. Unfortunately, the findings
may be so convincing as to necessitate surgery to prove
the lack of a mass. In Pilleul’s study of 219 patients with
possible small-bowel neoplasms, the overall accuracy was
84.7%. There were five false-positive masses (2.3%) rang-
ing from 6 to 25 mm in size. In two of these, small-bowel
fold thickening was found at surgery and in the others no
mass could be detected [22]. CTEC may also fail to iden-
tify jejunal polyps <10 mm, angioectasias, and ectopic
pancreas 3-5 mm in size. A false-negative rate of 4.1%
was reported as well, as any lesion that is sessile,
Marc J. Gollub
minimally enhancing, or <5 mm in size may be difficult
to detect. As such, the use of CTEC for patients with poly-
posis syndromes (familial adenomatous polyposis and
Peutz-Jeghers) may be somewhat limited [22, 23]. Voder-
holzer et al. found that, compared with WCE, polyps, ero-
sions, angioectasia, and lymphangiectasia were missed by
CTEC but seen on WCE [17]. Careful perusal of thin-sec-
tions at CTEC might prevent overlooking of these lesions.
Careful perusal of wall enhancement is necessary in pa-
tients with Crohn’s disease, to avoid missing more focal,
concurrent masses (secondary malignancy), as these can
be overlooked as well. In the diagnosis of OGIB, there are
few reported series in which CTEC was used. In fact, in
the workup of OGIB, Fillipone et al. reported that in
patients over age 50 OGIB is more commonly caused by
angioectasia than by small-bowel masses. Here too, WCE
may be the more appropriate modality due to the non-
elevated nature of these often subtle, small lesions [18].
Obstruction
In the context of small-bowel-obstruction, Walsh et al.
concluded that CTEC has greater sensitivity and speci-
ficity (89 and 100%, respectively) than conventional CT
(50 and 94%, respectively) [23]. Pitfalls in technique spe-
cific to patients with obstruction include: failure to per-
form suctioning prior to examination of the patient, fail-
ure to place the balloon in the jejunum (as opposed to the
duodenum, for better suctioning and the prevention of
back-flow into the stomach), and failure to properly ad-
just the infusion rate so as to elicit the transition point in
low-grade obstructions. Overzealous infusion can cause
spasm, which may be misinterpreted as a site of stricture
and obstruction. This can be avoided to some extent with
the administration of glucagon or Buscopan. Problem-
solving regarding possible strictures can be accomplished
with repeat limited CT slices through the area of interest;
however, this can be a costly approach given the radia-
tion-burden. The use of multidetector CT with 40 or more
channels can reduce radiation by 10-66% because of
more efficient detector configurations, automatic expo-
sure controls, improved filters, and the availability of im-
age post-processing algorithms [24].
Summary
The workup of small intestinal abnormalities is changing
rapidly due to technological advances in cross-sectional
imaging and endoscopic techniques. Although improve-
ments are still being made, certain early conclusions re-
garding the detection of small-bowel pathologies can be
offered:
1. CTEG, the most rapidly investigated new technique,
appears to be comparable if not superior to most endo-
scopic methods in determining disease presence, sever-
ity, extent, complications, and activity status, as well as
response to treatment; however, pitfalls exist, including
Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis
non-specificity of early findings, the enumeration of
fistulous tracts, and the detection of aphthous ulcers.
2. Patients with OGIB who are stable, i.e., not requiring
surgery or catheter embolization immediately, if at all,
should be offered CTEG using two or more phases to
search for a source of blood – with the known pitfalls
of the technique’s inability to detect shallow gastric or
small-bowel ulcers, small angioectasias, or flat lesions.
3. In the search for small-bowel masses, with or without
OGIB, CTEG appears to perform quite well, but if a
mass is not hyperenhancing, perfect bowel distention
may be required, an achievement not always possible,
especially in the proximal jejunum.
4. CTEC, performed in limited centers in the USA and
perhaps more widely in Europe, is subject to mostly
technique-related pitfalls but has the advantage of an
enteral challenge, which is not available with other
non-catheter/pump methods and which may define its
predominant indication as the best test for low-grade,
intermittent small-bowel obstruction not detectable by
other means.
References
1. Colombel JF, Solem CA, Sandborn WJ et al (2006) Quantita-
tive measurement and visual assessment of ileal Crohn’s dis-
ease activity by computed tomography enterography: correla-
tion with endoscopic severity and C reactive protein. Gut
55:1561-1567
2. Paulsen SR, Huprich JE, Hara AK (2007) CT enterography:
Noninvasive evaluation of Crohn’s Disease and obscure gas-
trointestinal bleed. Radiol Clin N Am 45:303-315
3. Huprich JE, Fletcher JG (2009) CT enterography: Principles,
technique and utility in Crohn’s disease. Eur J Radiol 69:393-397
4. Dave-Verma H, Moore S, Singh A et al (2008) Computed to-
mographic enterography and enteroclysis: pearls and pitfalls.
Curr Probl Diagn Radiol 37:279-287
5. Wold PB, Fletcher JG, Johnson CD et al (2003) Assessment of
small bowel Crohn disease: Noninvasive peroral CT entero-
graphy compared with other imaging methods and endoscopy-
feasibility study. Radiology 229:275-281
6. Vogel J, Moreira A, Baker M (2007) CT enterography for
Crohn’s disease: Accurate preoperative diagnostic imaging.
Dis Colon Rectum 50:1761-1769
7. Solem CA, Loftus Jr EV, Fletcher JG et al (2008) Small-bow-
el imaging in Crohn’s disease: a prospective, blinded, 4-way
comparison trial. Gastrointest Endosc 68:255-266
8. Hara AK, Alam S, Heigh RI et al (2008) Using CT enterogra-
phy to monitor Crohn’s disease activity: a preliminary study.
AJR 190:1512-1516
31
9. Higgins PDR, Caoli E, Zimmerman M et al (2007) Computed
tomographic enterography adds information to clinical man-
agement in small bowel Crohn’s disease. Inflamm Bowel Dis
13:262-268
10. Bruining DH, Siddicki H, Fletcher JG (2008) Clinical benefit
of CT enterography in suspected or established Crohn’s dis-
ease: impact on patient management and physician level of
confidence. Gastroenterology S1211
11. Singh V, Alexander JA (2008) The evaluation and management
of obscure and occult gastrointestinal bleeding. Abdom Imag-
ing 34:311-319
12. Huprich JE, Fletcher JG, Alexander JA (2008) Obscure gas-
trointestinal bleeding: Evaluation with 64-section multiphase
CT enterography – initial experience. Radiology 246:562-571
13. Brenner DJ, Hall EJ (2007) Computed tomography – an increas-
ing source of radiation exposure. N Engl J Med 357:2277-2284
14. Hara AK, Walker FB, Silva AC, Leighton JA (2009) Prelimi-
nary estimate of triphasic CT enterography performance in he-
modynamically stable patients with suspected gastrointestinal
bleeding. AJR 193:1252-1260
15. Postgate A, Despott E, Burling D et al (2008) Significant
small-bowel lesions detected by alternative diagnostic modali-
ties after negative capsule endoscopy. Gastrointest Endosc
68:1209-1214
16. Voderholzer WA, Beinhoelzl J, Rogalla P et al (2005) Small
bowel involvement in Crohn’s disease: a prospective compari-
son of wireless capsule endoscopy and computed tomography
enteroclysis. Gut 54:369-373
17. Voderholzer WA, Ortner M, Rogalla P et al (2003) Diagnostic
yield of wireless capsule enteroscopy in comparison with com-
puted tomography enteroclysis. Endoscopy 35:1009-1014
18. Fillipone A, Cianci R, Milano A et al (2008) Obscure gas-
trointestinal bleeding and small bowel pathology) comparison
between wireless capsule endoscopy and multidetector-row
CT enteroclysis. Abdom Imaging 33:398-406
19. Romano S, De Lutio E, Rollandi GA et al (2005) Multidetec-
tor computed tomography enteroclysis (MDCT-E) with neutral
enteral and IV contrast enhancement in tumor detection. Eur
Radiol 15:1178-1183
20. Maglinte DDT, Lappas JC, Heitcamp DE et al (2003) Techni-
cal refinements in enteroclysis. Radiol Clin North Am 41:
213-229
21. Boudiaf M, Jaff A, Soyer P et al (2004) Small-bowel diseases:
Prospective evaluation of multi-detector row helical CT ente-
roclysis in 107 consecutive patients. Radiology 233:338-344
22. Pilleul F, Penigaud M, Milot L et al (2006) Possible small-
bowel neoplasms: contrast-enhanced and water-enhanced mul-
tidetector CT enteroclysis. Radiology 241:796-801
23. Walsh D, Bender G, Timmons H (1998) Comparison of com-
puted tomography enteroclysis and traditional computed to-
mography in the setting of suspected partial small bowel ob-
struction. Emerg Radiol 5:29-37
24. Mannudeep K, Rizzo SMR, Novelline RA (2005) Technologic
innovations in computer tomography dose reduction: implica-
tions in emergency settings. Emergency Radiology 11:127-128

Diseases of the Small Bowel, Including the Duodenum – MRI
Karin A. Herrmann
Institute of Clinical Radiology, University Hospitals Munich, Munich, Germany
Introduction
For decades, barium fluoroscopy studies have been the
standard of reference to investigate small bowel diseases.
Since the small bowel was not accessible to endoscopic
techniques, these studies represented the only non-invasive
diagnostic approach to the intestine. Both bowel follow-
through and small bowel enteroclysis yielded fairly good
results, with sensitivities and specificities of, respectively,
98.3% and 99.3% for Crohn’s disease (CD) [1] and 61-95%
for neoplastic disease [2], notably in the assessment of the
intestinal mucosa due to the high spatial resolution ob-
tained with these techniques. However, their limitations are
that they provide almost exclusively intraluminal informa-
tion and are associated with considerably high radiation
exposure, up to 10-18 mSv. The technical advances in
cross-sectional imaging achieved with computed tomogra-
phy (CT) and magnetic resonance imaging (MRI) over the
past 10 years have tremendously improved image quality
in the abdomen, thus encouraging small bowel imaging.
Thin-section multi-detector row CT with 3D multiplanar
reformations and accelerated image acquisition, in addi-
tion to breath-hold techniques in MRI, has fostered imag-
ing of the bowel with no or few limitations. Consequently,
CT and MRI are nowadays considered as state-of-the-art
imaging modalities. Both not only provide insight into the
bowel lumen but also depict mural and extramural patholo-
gy, which is essential for a comprehensive diagnostic as-
sessment and the staging of small bowel diseases. More-
over, CT and MRI convey far more information than ob-
tained with conventional barium techniques [3]. MRI, in
addition, has the advantage of obviating radiation exposure,
and is therefore preferable in children, young individuals, in
pregnancy, and if multiple follow-up studies are required
such as in patients with inflammatory bowel disease.
In order to appropriately assess the intestine, bowel dis-
tention is essential. This can be achieved with contrast
agents that typically have osmotic effects and enhance the
contrast between the bowel lumen and the bowel wall. De-
pending on the mode of application, the corresponding ex-
amination is called MR- or CT-enterography if the contrast
medium is administered orally prior to the examination,
and MR- or CT-enteroclysis if the contrast agent is injected
IDKD 2010-2013
using a nasojejunal tube, aided by an automatic infusion
pump. In CT-enteroclysis, scanning is typically performed
just after completion of the filling process; this results in a
static, high-resolution data set that allows 3D post pro-
cessing. MR-enteroclysis has the advantage that it provides
not only static but also functional information, since repet-
itive scanning can be carried out during the filling process.
The diagnosis of small bowel diseases with CT and
MRI is fundamentally based on morphological criteria.
Distinct imaging findings allow disease detection, diag-
nosis, and staging. This chapter is designed to familiarize
the reader with the typical imaging features indicating in-
flammatory, neoplastic, and ischemic small bowel dis-
eases, with special focus on their MRI appearances.
The Duodenum
Since the upper gastrointestinal tract is readily accessible
to endoscopic techniques, imaging techniques are less im-
portant to establish a diagnosis. Yet, cross sectional imag-
ing is helpful to assess the intra- and extraluminal extent
of disease. Since the duodenum is likely to be involved in
entities affecting surrounding organs, such as the pan-
creas, bile duct, ampulla, and papilla, these should always
be considered in the differential diagnosis of duodenal
pathology. Embryological anomalies include ectopic and
annular pancreas, webs, and choledochoceles. Annular
pancreas occurs in 1:20,000 births and is typically diag-
nosed in the early days after birth due to intestinal ob-
struction. During pregnancy, it can be discovered during
evaluation of the polyhydramnion. Less frequently, it is
found in young adults or incidentally later in life; in both
cases it then has to be distinguished from neoplasm. Oth-
er benign non-neoplastic conditions in the duodenum in-
clude ulcerative and inflammatory disease (i.e., CD) and
pseudoinflammatory polyps. These pathologies are typi-
cally worked up endoscopically.
Neoplasms of the Duodenum
Lipoma, leiomyoma, neurofibroma, adenoma, Brunner
gland adenoma, and polyps are benign neoplasms occur-
Diseases of the Small Bowel, Including the Duodenum – MRI
ring in the duodenum. Patients are typically asymptomatic
until a late stage of disease. While MRI is helpful in char-
acterizing fat-containing lesions, it is less useful in the fi-
nal histopathological diagnosis of most of these neo-
plasms. Malignant neoplasms of the duodenum include
adenocarcinoma, lymphoma, neuroendocrine tumors,
gastrointestinal stromal tumors (GIST), and metastases.
Duodenal adenocarcinoma accounts for only 0.4% of
gastrointestinal tumors. By the time it is clinically appar-
ent, there is often advanced disease. Clinical symptoms are
indicative of high bowel obstruction, vomiting, chronic
bleeding and, if close to the ampulla, obstructive jaundice.
The duodenum is the most common site for adenocarcino-
ma, which may appear as a defined nodular mural mass or
a diffusely infiltrating mass with spiculated borders. An-
other pattern is that of a diffuse or annular wall thickening
causing constrictive narrowing of the lumen. On T1-
weighted imaging, the tumor may be homogeneously
isointense or hypointense compared to the wall and slight-
ly hyperintense or isointense in T2-weighted imaging.
Contrast enhancement is moderate, mostly homogeneous.
Lymphoma: in the duodenum, as in the small and large
bowel, the most frequent form of lymphoma is non-
Hodgkin’s lymphoma (NHL). In 50% of nodal NHLs,
there are concomitant intestinal manifestations. Typical
imaging features of duodenal and intestinal lymphoma are
marked asymmetrical circumferential wall thickening with
mild pre-lesional luminal dilatation. Luminal stenosis is
not always predominant. Lymphoma may also appear as
multiple polypoid intraluminal protrusions or an exophyt-
ic extraluminal mass. Ulceration and fistula formation are
not uncommon. An extraluminal desmoplastic reaction is
unlikely. Associated lymphadenopathy is a helpful diag-
nostic hint but is not always present. Compared to normal
bowel wall, intestinal lymphoma shows moderately in-
creased and homogeneous signal intensity on T2-weighted
images. On T1-weighted imaging, it is isointense or hypo-
intense and exhibits mild and slightly inhomogeneous
contrast enhancement after gadolinium administration [4].
Unlike neuroendocrine tumors (NET) of the small
bowel, the role of CT and MRI in detecting duodenal
NET is minor compared to endoscopic techniques, in-
cluding endoscopic ultrasound. Along with the stomach,
duodenum is a common location for NET. Typically, the
lesions are small; they occasionally can be detected on
CT and MRI as single or multiple hypervascular nodules
in arterial-phase imaging. Luminal distention, e.g., with
hydro-CT of the stomach, is recommended and may help
to achieve a detection rate of up to 89%.
GISTs are most commonly located in the stomach and
small bowel and are described in the following section.
The Small Intestine
Inflammatory Bowel Disease
Currently, the most frequent clinical application of small
bowel MRI is in patients with inflammatory bowel
33
disease, especially CD. CD is a chronic inflammatory au-
toimmune disorder that frequently involves the small bow-
el but may affect the entire gastrointestinal tract. It is most
commonly located in the terminal ileum (40-80%) and the
colon. Involvement of the proximal ileum and the jejunum
is less frequent (22-40%). In imaging, CD manifestations
have been classified by Maglinte et al. in three categories:
acute inflammatory disease, fibrostenotic disease, and fis-
tulizing disease [5]. Due to its chronically recurrent char-
acter, different stages of CD may coexist. Discontinuity of
multiple disease manifestations (skip lesions) is a charac-
teristic feature. The morphological spectrum ranges from
early superficial mucosal disease with disruption, flatten-
ing, thickening, and distortion of the fold pattern, to mild
or pronounced longitudinal or transverse fissures and ul-
cerations resulting in a cobblestone appearance, to trans-
mural disease characterized by wall thickening, stenosis,
and mesenteric hypervascularity. Eventually, extramural
extension of the inflammation into the mesentery may oc-
cur, accompanied by the development of blind sinus tracts,
fistulas, and micro- or macro abscesses.
Crohn’s Disease: Acute Inflammatory Type
A number of imaging findings in MRI have been reported
to determine inflammatory activity in CD: increased small
bowel wall thickening (typically >4 mm), increased mural
contrast enhancement [6], submucosal edema with high
signal intensity on T2-weighted images [7] (Fig. 1), the
presence of deep mucosal ulcers and fissures [8], mesen-
teric hypervascularity (comb sign) [9], and contrast-
enhancing enlarged lymphadenopathy [8]. The most recent
literature describes increased mural thickening, increased
wall signal intensity on T2-weighted fat-saturated images,
and layered mural enhancement, but not mural enhance-
ment alone, as the strongest indicators for active disease.
Layered mural enhancement, however, is also commonly
associated with coexisting fibrostenosis and scar formation
[10]. Early and mild stages of CD may present at MRI as a
focal or regional disruption of the fold pattern, showing on-
ly subtle or no increased enhancement. Superficial erosions
may not be detected at all. In contrast to overt wall thick-
ening and stenosis, these lesions can easily be missed at
MRI especially when distention is suboptimal. This is why
MR-enterography or MR-enteroclysis (MRE), compared to
capsule endoscopy or invasive double balloon endoscopy,
reaches an overall sensitivity of only 75-80% [11, 12]. In
subacute CD and under effective treatment, wall thickening
may persist initially but decreases over time. The hyper-
intense signal intensity from submucosal edema disappears
and becomes intermediate signal intensity. Likewise, the
formerly increased contrast enhancement decreases.
Crohn’s Disease: Fibrostenosing Type
Fibrostenotic lesions are less conspicuous and more dif-
ficult to identify on MRI since they typically show no
wall thickening or increased enhancement. The seemingly
34 Karin A. Herrmann

a b

Fig. 1 a, b. MR-enteroclysis study with bipha-

sic intraluminal contrast in a 32 year-old pa-
tient with Crohn’s disease showing typical
signs of active inflammatory disease. a Wall
thickening, luminal narrowing and sub-
mucosal edema (arrow) is seen on T2-
weighted single-shot fast spin-echo (SSFSE)
(HASTE) imaging in a loop with prominent
active inflammation. b Slightly further up,
the loop shows increased transmural en-
hancement on contrast-enhanced T1-weight-
ed imaging with fat suppression, mesenteric
hypervascularity (arrow), and inflammatory
extramural stranding of the mesentery

normal wall thickness and contrast enhancement in fibro-

stenosis mimics the normal bowel wall. Low signal in-
tensity on T2-weighted images with and without fat sup-
pression has been described [9]. Persistent focal and
fixed circumscript intestinal stenosis and marked pre-
stenotic dilatation are the most conspicuous and reliable
indicators and indirect signs of fibrous strictures.

Crohn’s Disease: Fistulizing Type

Fistulas and peri- extramural abscesses represent the

most severe complications and can be diagnosed accu-
rately with MRI (accuracy 92%). Fistulas are tubular in-
flammatory pathways between bowel loops, bowel, and
other organs or bowel and the abdominal wall. They are
defined as internal fistulas (IF) when located between
bowel loops (entero-enteric) or internal organs (e.g., en-
tero-vesical). External fistulas (EF) are those that abut
the skin (entero-cutaneous). A vast majority of IF remain
asymptomatic or cause non-specific symptoms, especial-
ly if located in distal bowel segments. Complex IF typi- Fig. 2. MR-enteroclysis study of a 45-year-old patient with negative
intraluminal contrast (Lumirem®, Guerbet, France) shows an ex-
cally involve multiple bowel loops [13] and bowel ob- ample of a typical stellate appearance of converging bowel loops
struction due to concomitant stenosis. Abnormal tubular in an entero-enteric fistula involving the cecum, terminal ileum,
tracts within the mesentery containing air or fluid are sigmoid, and the middle ileum. Note the pre-stenotic dilatation of
highly indicative of IF and are well seen on SSFP the bowel loops proximal to it, indicating a mechanical stenosing
(steady-state free precession; e.g., TrueFISP) imaging. effect of the fistula complex (T2-weighted SSFSE)
The outlines of IF are of intermediate signal intensity on
T2-weighted images, similar to the intestinal wall, and
may show moderate enhancement after intravenous
gadolinium. A characteristic, somewhat stellate arrange- Differential Diagnosis of Crohn’s Disease
ment of the bowel loops is referred to as the “star-sign”
and is reportedly a strong indicator for the presence of Although less frequently a reason for imaging than CD, in-
complex IF (Fig. 2) [14]. The inflamed bowel loops con- fectious enteritis is an important differential diagnosis to
verge to a common center, which often is the origin of ab- consider. Infectious enteritis is typically of bacterial origin,
scess formation. Urinary bladder involvement is likely such as Enterococcus, Salmonella, Yersinia, and tubercu-
when there is marked thickening of the bladder wall in lous bacillae. In immunocompromised patients, viral and
the vicinity of an inflamed bowel segment [13]. A con- fungal infections such as Cytomegalovirus (CMV) have to
tiguous inflammatory tract between the inflamed bowel be considered. In patients who have undergone bone mar-
and the bladder wall may or may not be visualized. row transplant, CMV and graft-versus-host reactions are
Diseases of the Small Bowel, Including the Duodenum – MRI
likely. The appearance of infectious enteritis on CT or MRI
is non-specific, the main feature being submucosal edema.
For tuberculosis, mesenteric lymphadenopathy is a major
finding and a helpful clue to the diagnosis. Ischemic bow-
el disease is to be considered if arteriosclerosis and vascu-
lar obstruction are observed additionally to altered bowel
wall with submucosal edema.
Small Bowel Neoplasms
Small bowel neoplasms are rare and account for less than
5% of all gastrointestinal tumors. Adenocarcinoma, carci-
noid, lymphoma, GIST, and metastases constitute the ma-
lignant component. MR-enteroclysis has recently been re-
ported to be an effective diagnostic tool for the detection
of small bowel tumors, with a sensitivity, specificity, and
accuracy of 86, 98, and 97%, respectively [15]. The inci-
dence of small bowel cancer has risen in the past 30 years,
with the greatest increase for carcinoid tumors. Adeno-
carcinoma is the most common among the intestinal ma-
lignant neoplasms, followed by carcinoid tumors. After the
duodenum, the jejunum is the second most frequent loca-
tion for adenocarcinoma typically involving the ileum. The
imaging morphology for both adenocarcinoma and lym-
phoma (Fig. 3) is as previously described (see duodenal
neoplasms). Both may occur as complications of CD.
Neuroendocrine Tumors: Carcinoid
Carcinoid tumors are neuroendocrine neoplasms and ac-
count for approximately 2% of all gastrointestinal tumors.
They may be found along the entire gastrointestinal tract
Fig. 3. MR-enteroclysis study of a 63-year-old male with occult
bleeding at fecal occult blood test. T2-weighted SSFSE image af-
ter biphasic intraluminal contrast administration shows an ileal
segment of marked circumferential wall thickening with mass
effect preserving the lumen (arrow). The intraluminal surface is
irregular; contrast enhancement (not shown) was moderate. This
example shows the typical features of intestinal lymphoma
35
(85%) as well as in the pancreas and lung (10%). The
appendix (50%) and the ileum (~30%) are the most com-
mon primary locations. Up to 30% of intestinal carcinoids
are multifocal. Carcinoids have a tendency to metastasize
early to lymph nodes and the liver, even when they only
are 1-2 cm in size [16]. Therefore, distention of the bowel
lumen in MRE is required for better detection. Carcinoids
are typically hypervascularized and therefore are best
identified on contrast-enhanced T1-weighted fat saturated
GRE (gradient-recalled echo) sequences (Fig. 4). SSFSE
and SSFP sequence types depict these tumors less well as
slightly hyperintense or isointense to muscle and bowel
wall. Approximately half the tumors appear as a nodular
intraluminal mass, one third as focal circumferential wall
thickening, and 20% with both characteristics [17]. If only
wall thickening is present, carcinoids may easily be con-
founded with inflammatory disease. A desmoplastic reac-
tion in the adjacent mesentery occurs in up to 73% of the
cases of small tumors and may cause vascular engorge-
ment.
Gastrointestinal Stromal Tumors
Gastrointestinal stromal tumors represent 0.3-3% of all gas-
trointestinal tumors. They derive from the intestinal cells of
Cajal and originate most often from the stomach (~70%) or
the small bowel wall (20-30%). After appropriate immuno-
histochemical preparation, they can be shown to strongly
express the KIT protein (CD 117), which is a characteristic
feature of GISTs. These tumors are well delineated, non-in-
filtrative masses located in or arising from the intestinal
wall, typically extraluminally at the serosal side [18]. When
<5 cm in diameter, they are slightly heterogeneous and
mildly hyperintense on T2. Larger masses are heteroge-
neous,with a soft-tissue rim encompassing a necrotic cen-
ter. This rim is isointense or slightly hyperintense to muscle
on T2-weighted images and hypointense on T1-weighted
images, showing heterogeneous enhancement after contrast
administration. In the majority of cases, GISTs do not show
signs of infiltration, vessel encasement, or lymphatic
spread. The primary sites of metastases from malignant
GISTs are the liver and peritoneum. Treatment with tyro-
sine kinase inhibitors induces a loss of vascularization, and,
consequently, reduced contrast enhancement of the tumor
and its metastases [19], but no necessarily change in size.
Other Small Bowel Pathologies
For some small bowel pathologies, MRE has so far not
proven to be useful. For superficial mucosal pathologies
and angiodysplasia, MRI of the small bowel cannot be
recommended. The detection of acute ischemia and oc-
cult gastrointestinal bleeding is the domain of CT, main-
ly because of its availability in an emergency setting,
short examination time, and spatial resolution.
Post-operative or post-inflammatory small bowel ad-
hesions represent another important condition, formerly
investigated with conventional enteroclysis using manual
36
a b
compression and mobilization of the bowel loops. This
manipulation is not practicable in MRI due to the limited
access to the patient inside the magnet. To date, no liter-
ature is available to support the usefulness of MRE for
small bowel adhesions.
References
1. Maglinte DD, Chernish SM, Kelvin FM et al (1992) Crohn
disease of the small intestine: accuracy and relevance of ente-
roclysis. Radiology 184:541-545
2. Bessette JR, Maglinte DD, Kelvin FM, Chernish SM (1989)
Primary malignant tumors in the small bowel: a comparison of
the small-bowel enema and conventional follow-through ex-
amination. AJR Am J Roentgenol 153:741-744
3. Lee SS, Kim AY, Yang SK et al (2009) Crohn disease of the
small bowel: comparison of CT enterography, MR enterogra-
phy, and small-bowel follow-through as diagnostic techniques.
Radiology 251:751-761
4. Kim KW, Ha HK (2004) MRI for small bowel diseases. Magn
Reson Imaging Clin N Am 12:637-650
5. Maglinte DDT, Gourtsoyiannis N, Rex D et al (2003) Classi-
fication of small bowel Crohn’s subtypes based on multi-
modality imaging. Radiol Clin North Am 41:285-303
6. Sempere GAJ, Sanjuan VM, Chulia EM et al (2005) MRI eval-
uation of inflammatory activity in Crohn’s disease. AJR Am J
Roentgenol 184:1829-1835
7. Maccioni F, Bruni A, Viscido A et al (2006) MR imaging in
patients with Crohn disease: value of T2- versus T1-weighted
gadolinium-enhanced MR sequences with use of an oral su-
perparamagnetic contrast agent. Radiology 238:517-530
8. Gourtsoyiannis N, Papanikolaou N, Grammatikakis J et al
(2004) Assessment of Crohn’s disease activity in the small
bowel with MR and conventional enteroclysis: preliminary
results. Eur Radiol 14:1017-1024
Karin A. Herrmann
Fig. 4 a, b. T1-weighted contrast-enhanced
coronal GRE image after MRE with biphasic
intraluminal contrast depicts a 1.5-cm U-
shaped intraluminal mass (arrow) in the ileum
with increased enhancement (a; zoomed in b)
which after surgical removal proved to be car-
cinoid. No mesenteric desmoplastic reaction
is seen
9. Prassopoulos P, Papanikolaou N, Grammatikakis J et al (2003)
MR enteroclysis imaging of Crohn disease. Radiographics 21
Spec No:S161-172
10. Punwani S, Rodriguez-Justo M, Bainbridge A et al (2009) Mur-
al inflammation in Crohn disease: location-matched histologic
validation of MR imaging features. Radiology 252:712-720
11. Tillack C, Seiderer J, Brand S et al (2008) Correlation of mag-
netic resonance enteroclysis (MRE) and wireless capsule en-
doscopy (CE) in the diagnosis of small bowel lesions in
Crohn’s disease. Inflamm Bowel Dis 14:1219-1228
12. Seiderer J, Herrmann K, Diepolder H et al (2007) Double-bal-
loon enteroscopy versus magnetic resonance enteroclysis in di-
agnosing suspected small-bowel Crohn’s disease: results of a
pilot study. Scand J Gastroenterol 42:1376-1385
13. Herrmann KA, Michaely HJ, Zech CJ et al (2006) Internal fis-
tulas in Crohn disease: magnetic resonance enteroclysis. Ab-
dom Imaging 31:675-687
14. Herrmann KA, Michaely HJ, Seiderer J et al (2006) The “star-
sign” in magnetic resonance enteroclysis: a characteristic find-
ing of internal fistulae in Crohn’s disease. Scand J Gastroen-
terol 41:239-241
15. Masselli G, Polettini E, Casciani E et al (2009) Small-bowel
neoplasms: prospective evaluation of MR enteroclysis. Radiol-
ogy 251:743-50
16. Horton KM, Kamel I, Hofmann L, Fishman EK (2004) Carci-
noid tumors of the small bowel: a multi-technique imaging ap-
proach. AJR Am J Roentgenol 182:559-567
17. Schmid-Tannwald C, Zech CJ, Panteleon A et al (2009). Char-
acteristic imaging features of carcinoid tumors of the small
bowel in MR enteroclysis. Radiologe 49:242-245
18. Burkill GJ, Badran M, Al-Muderis O et al (2003) Malignant
Gastrointestinal Stromal Tumor: Distribution imaging features
and pattern of metastatic spread. Radiology 226:527-532
19. Schlemmer M, Sourbron SP, Schinwald N et al (2009) Perfu-
sion patterns of metastatic gastrointestinal stromal tumor
lesions under specific molecular therapy. Eur J Radiol 27:
278-284
 IDKD 2010-2013

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

Jaap Stoker1, Richard M. Gore2
1 Department of Radiology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
2 Department of Radiology, North Shore University Health System, Evanston Hospital, Evanston, IL, USA

Introduction of ulcerative colitis are beneath the spatial resolution of

CT. With progressive disease, submucosal edema pro-
Infectious and inflammatory diseases of the colon and rec- ducing a “target sign” may be seen. Severe mucosal ul-
tum are common disorders that are becoming increasingly ceration can denude certain portions of the colonic wall,
prevalent. Diverticulitis, appendicitis, and the inflammato- leading to inflammatory pseudopolyps (Fig. 1), which,
ry and infectious colitides are among the commonest caus- when sufficiently large, can be visualized on CT. Mural
es of acute abdominal pain. These patients frequently pre- thickening and luminal narrowing are the CT hallmarks
sent with non-specific symptoms. Since in most cases of of the subacute and chronic ulcerative colitides. Mural
acute abdominal pain clinical and laboratory assessments thinning, unsuspected perforations, and pneumatosis can
cannot confidently identify a specific etiology, cross-sec- be detected on CT in patients with toxic megacolon. In
tional imaging plays an important part in the work-up and this regard, CT can be quite helpful in determining the ur-
management of these patients. Ultrasound (US) and com- gency of surgery in patients with stable abdominal films
puted tomography (CT) are the primary means of evaluat- yet a deteriorating clinical course [1-3].
ing acute abdominal pain, as both modalities provide in- In chronic ulcerative colitis, the muscularis mucosa
sight into the pathological changes in the colon wall, becomes markedly hypertrophied, often by a factor of 40.
serosa, surrounding mesentery, and peritoneum. US is a Forceful contraction of this hypertrophied longitudinal
readily available, real-time technique with reasonable ac- muscle may pull the mucosa away from the submucosa,
curacy that uses no ionizing radiation. CT provides rapid
assessment with high accuracy, optimal field of view, and
good reproducibility. Magnetic resonance imaging (MRI)
in patients with acute abdominal pain offers great clinical
potential by virtue of its global imaging perspective, ex-
quisite soft-tissue resolution, and high accuracy without
the need for ionizing radiation. In patients with perirectal
and perianal inflammatory conditions, MRI is the pre-
ferred imaging technique, as the combination of high in-
trinsic contrast resolution and large field of view provides
detailed information on the presence and extent of disease.
This chapter describes the imaging features of these
infectious and inflammatory disorders, provides differen-
tial diagnostic guidelines, and presents the advantages
and disadvantages of the various imaging modalities in
the context of optimizing patient management.

Ulcerative Colitis
Ulcerative colitis is characterized pathologically by ex-
tensive confluent and circumferential ulceration and by
diffuse inflammation of the mucosa. The disease charac-
Fig. 1. Acute ulcerative colitis. CT demonstrates deep ulcerations
teristically begins in the rectum and extends proximally (arrows) of the fluid-filled rectosigmoid. Inflammatory pseudo-
in a contiguous fashion to involve part or all of the colon. polyps appear as residual islands of inflamed mucosa that protrude
The pathological changes found in the very early stages above the denuded colonic surface
38
producing diffuse or segmental narrowing of the lumen.
The contraction also causes shortening of the colon. The
submucosa becomes thickened due to the deposition of
fat or, in acute and subacute cases, edema. Submucosal
thickening further contributes to narrowing of the lumen.
Additionally, in acute and in chronic ulcerative colitis, the
lamina propria is thickened due to round-cell infiltration.
On CT, these mural changes produce a “target” or
“double-halo” appearance when axially imaged. The lu-
men is surrounded by a ring of soft-tissue density (mu-
cosa, lamina propria, hypertrophied muscularis mu-
cosae), then by a low-density ring (fatty infiltration of the
submucosa), and in turn by a ring of soft-tissue density
(muscularis propria). This mural stratification is not spe-
cific and can also be seen in Crohn’s disease, infectious
enterocolitis, pseudomembranous colitis, ischemic and
radiation enterocolitides, mesenteric venous thrombosis,
bowel edema, and graft-versus-host disease [1-3].
There are certain CT findings that can help differenti-
ate granulomatous from ulcerative colitis. Mural stratifi-
cation, i.e., the ability to visualize individual layers of
bowel wall, is seen in 61% of patients with chronic ul-
cerative colitis but only in 8% of patients with chronic
granulomatous colitis. Also, mean colon wall thickness in
chronic ulcerative colitis is 7.8 mm, significantly less
than that observed in Crohn’s colitis (11 mm). Finally, the
outer contour of the thickened colonic wall is smooth and
regular in 95% of ulcerative colitis patients while serosal
and outer mural irregularities are present in 80% of pa-
tients with granulomatous colitis [1-3].
Rectal narrowing and widening of the presacral space
are hallmarks of chronic ulcerative colitis. CT depicts the
anatomical alterations that underlie these rather dramatic
morphological changes. The rectal lumen is narrowed
due to the previously described mural thickening that at-
tends chronic ulcerative colitis. As a result, the rectum
has a target appearance on axial scans, which should not
be mistaken for the external anal sphincter, mucosal pro-
lapse, or the levator ani muscles. The increase in the pre-
sacral space is caused by proliferation of the perirectal
fat, which on CT is characterized by an increased num-
ber of nodular and streaky soft-tissue densities and an ab-
normal attenuation value, 10-20 HU higher than the nor-
mal extraperitoneal or mesenteric fat. These fatty changes
relate to a number of factors, including ex vacuo re-
placement by fat of the void produced by rectal lumen
narrowing and lipodystrophy resulting from an influx of
inflammatory cells and edema. Edematous adipose tissue
and enlarged lymph nodes are often observed in the
perirectal region at the time of abdominoperineal resec-
tions in patients with chronic ulcerative colitis [1-3].
Crohn’s Disease
Crohn’s disease most commonly affects the terminal
ileum and proximal colon. The acute, active phase of
Crohn’s disease is characterized by focal inflammation,
Jaap Stoker, Richard M. Gore
aphthoid ulceration with adjacent cobblestoning, an often
transmural inflammatory reaction with lymphoid aggre-
gates and granuloma formation, fissures, fistulas, and si-
nus tracts. The chronic and resolving phase of this disor-
der is associated with fibrosis and stricture formation.
The presence and extent of Crohn’s disease can be de-
termined by US, CT, or MRI. The accuracy of each of
these examinations is comparable but each technique has
its strength and limitations [4]. When Crohn’s disease is
limited to the mucosa, the CT scan is often normal. Al-
though inflammatory and post-inflammatory pseudo-
polyps may be identified on CT, the assessment of the
mucosa is best reserved for barium studies and colono-
scopy, which are more direct and sensitive. Crohn’s dis-
ease is manifested on CT by bowel wall thickening of 1-
2 cm. This thickening, which occurs in up to 83% of pa-
tients, is most frequently observed in the terminal ileum,
but other portions of the small bowel, colon, duodenum,
stomach, and esophagus may be similarly affected [1-3].
During the acute, non-cicatrizing phase of Crohn’s dis-
ease, the small bowel and colon maintain mural stratifi-
cation and often have a target or double-halo appearance.
As in ulcerative colitis, there is a soft-tissue density ring
(corresponding to mucosa), which is surrounded by a
low-density ring with an attenuation near that of water or
fat (corresponding to submucosal edema or fat infiltra-
tion, respectively), which in turn is surrounded by a high-
er density ring (muscularis propria). Inflamed mucosa
and serosa may show significant contrast enhancement
following bolus intravenous contrast administration, and
the intensity of enhancement correlates with the clinical
activity of the disease [1-3].
CT demonstration of mural stratification, i.e., the abil-
ity to visualize distinct mucosal, submucosal, and mus-
cularis propria layers, indicates that transmural fibrosis
has not occurred and that medical therapy may be suc-
cessful in ameliorating lumen compromise. Additionally,
prior to the onset of fibrosis, the edema and inflammation
of the bowel wall responsible for mural thickening and lu-
men obstruction are reversible to some extent. A modest
decrease in wall thickness often produces a dramatic in-
crease in lumen cross-sectional area as well as resolution
of the patient’s obstructive symptoms.
Loss of mural stratification is indicative of transmural
fibrosis [1-3]. In the background of long-standing
Crohn’s disease, this is typically visualized on CT as ho-
mogeneous attenuation in the affected bowel wall. If this
finding occurs against a background of good levels of
intravascular contrast medium and thin-section recon-
structions, then the fibrosis is most likely irreversible. In
these patients, anti-inflammatory agents may not provide
a significant reduction in bowel wall thickness. If these
segments become sufficiently narrow, surgery or stricturo-
plasty will be necessary to relieve the obstruction.
In a patient with Crohn’s disease, the palpation of an ab-
dominal mass or the separation of bowel loops as seen on
a barium study evokes a large differential diagnosis: ab-
scess, phlegmon, “creeping fat” or fibrofatty proliferation
Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases 39

of the mesentery, bowel wall thickening, and enlarged

mesenteric lymph nodes. Each of these disorders has sig-
nificantly different prognostic and therapeutic implications.
This diagnostic dilemma is further complicated by the fact
that many of these patients are receiving immunosuppres-
sive therapy, which can mask other signs and symptoms.
CT can readily differentiate the extraluminal manifes-
tations of Crohn’s disease. Fibrofatty proliferation, also
known as “creeping fat” of the mesentery, is the most
common cause of bowel-loop separation seen on barium
studies in patients with Crohn’s disease. On CT, the sharp
interface between bowel and mesentery is lost and the at-
tenuation value of the fat is elevated by 20-60 HU due to
the influx of inflammatory cells and fluid. Mesenteric
adenopathy, with lymph nodes ranging in size between
3 and 8 mm, may also be present. If these lymph nodes
are larger than 1 cm, the presence of lymphoma or carci-
noma, both of which occur with greater frequency in
Crohn’s disease, must be excluded [1-3].
Contrast-enhanced CT scans often show hypervascu-
larity of the involved mesentery, manifesting as vascular
dilatation, tortuosity, prominence, and wide spacing of
the vasa recta. These distinctive vascular changes have Fig. 2. Acute Crohn’s colitis. Sagittal reformatted image from a CT
been called the “comb sign” and its identification should enterography study reveals marked mural thickening of the de-
scending colon with markedly prominent vasa rectae
suggest active disease. In addition, it may be useful in dif-
ferentiating Crohn’s disease from lymphoma or metas-
tases, which tend to be hypovascular lesions.
With the advent of novel biological therapeutic agents mulative radiation dose of multiple CT examinations is
such as infliximab, CT enterography has become an im- substantial [6]. Accordingly, accurate technique without
portant technique for evaluating small bowel disease ac- ionizing radiation is preferable. For small bowel imaging
tivity in patients with Crohn’s disease, because of its ac- in IBD, there is a considerable body of evidence showing
curacy and non-invasive nature. Since colonic involve- the comparable accuracy of US, CT, scintigraphy, and
ment is common in patients with inflammatory bowel dis- MRI [4]. The lack of ionizing radiation exposure thus fa-
ease (IBD), the capability of CT enterography in evaluat- vors the use of either US or MRI (Fig. 3). US gives real-
ing colorectal involvement (Fig. 2) is being intensively time information on peristalsis in addition to the mor-
studied, and the preliminary results are encouraging [5]. phological features. While contrast-enhanced US corre-
Since patients with Crohn’s disease are often young lates well with colonoscopic disease activity [7, 8], this
and typically will require multiple examinations, the cu- technique is operator-dependent. The major advantages

a b

Fig. 3 a, b. Acute Crohn’s colitis of the descending colon shows a thickened de-
scending colon at (a) ultrasound and (b) MR enterography (coronal T2-weight-
ed turbo spin-echo). The arrow indicates the thickened descending colon
40
of MRI over US are its unlimited field of view and good
reproducibility; however, its lack of availability and high
cost may be prohibitive.
Optical colonoscopy can directly visualize the colon,
but the impetus for cross-sectional imaging studies of the
colon in patients with IBD has lagged behind imaging
evaluation of the small bowel. Indeed, there is a large body
of data concerning MR enterography and MR enteroclysis,
but very few studies have evaluated colonic disease [9, 10].
In MR evaluations of the colon, lumen distention is
mandatory [11, 12] to achieve optimal imaging results.
However, the experience with dedicated examinations of
the colon in IBD is limited and conflicting [13]. While
use of the bright lumen technique is preferred, in a series
of 15 healthy volunteers and 23 patients with known IBD
dark lumen MR colonography resulted in high sensitivi-
ty (87%) for identifying segmental IBD changes [14].
However, no biopsies were obtained of endoscopically
normal mucosa, which might have influenced the results.
Several studies have evaluated the efficacy of MRI in
determining small bowel disease activity in Crohn’s dis-
ease, with bowel wall thickening and bowel wall en-
hancement employed as markers of disease activity. Al-
though severe disease activity is well depicted, limited
disease activity and non-active disease are less reliably
demonstrated [15].
Data regarding the assessment of colonic disease activ-
ity are sparse and partly conflicting. In one study, increased
contrast enhancement of the bowel wall was significantly
related to colonoscopically active disease in patients with
IBD [16]. In another series, combining increased contrast
enhancement of the bowel wall with other MRI features
(bowel wall thickening, presence of mesenteric lymph
nodes, loss of the normal haustral pattern), the sensitivity
was poor (32%) but the specificity was good (88%) [17].
The combination of MR enterography with a water-based
enema seems to be a valuable approach to assess disease
activity in patients with established Crohn’s disease [18].
Data comparing MRI with other techniques are sparse, nor
is it possible to recommend the routine use of water-based
enema or other colon distention techniques for evaluating
the colon in patients with IBD.
Pseudomembranous Colitis
Pseudomembranous colitis is being encountered with in-
creasing frequency as a nosocomial infection complicating
antibiotic therapy [19]. This potentially life-threatening dis-
order is caused by overgrowth of Clostridium difficile. The
bacteria release a cytotoxic enterotoxin that causes ulcera-
tion of the colonic mucosa and the formation of 2-3 mm
pseudomembranes consisting of fibrin, mucus, sloughed
epithelial cells, and leukocytes. Mild cases may demon-
strate only mucosal irregularity and nodularity, with the for-
mation of small plaques that cannot be detected radiologi-
cally whereas in advanced cases there is thickening of the
haustral folds, a shaggy wall contour, and mucosal plaques.
Jaap Stoker, Richard M. Gore
Fig. 4. Pseudomembranous colitis. Coronal reformatted CT image
shows marked mural thickening and submucosal edema of the dis-
tal colon. Note the ascites
CT shows a pancolitis with mural thickening that may
be irregular or polypoid and characterized by a shaggy
endoluminal contour. The thickened wall (Fig. 4), which
is usually 1.6-1.8 cm, is a result of submucosal edema.
Mucosal and serosal enhancement is seen following in-
travenous contrast administration. The haustra are also
thickened and edematous, producing the “accordion pat-
tern” highly suggestive of pseudomembranous colitis.
This pattern, which is the result of contrast trapped be-
tween thickened haustral folds aligned in a parallel fash-
ion, can sometimes simulate deep ulcerations or fissures.
Pericolic stranding, ascites, pleural effusions, and sub-
cutaneous edema are other ancillary CT findings. Com-
plications of untreated pseudomembranous colitis in-
clude toxic megacolon and intestinal perforation with
subsequent peritonitis. CT is also useful in monitoring
the response to medical therapy with oral vancomycin
and metronidazole [20, 21].
AIDS-Related Colitis
Cytomegalovirus and cryptosporidiosis are common
pathogens seen in patients with acquired immunodefi-
ciency syndrome (AIDS). Patients with CD4 lymphocyte
counts of ≤200 mm–3 are at greatest risk for these infec-
tions. Typically, the cecum and proximal ascending colon
are involved by these organisms; however, a pancolitis
with continuous lesions may occur as well. CT shows
mural thickening of the involved segments of colon, with
low attenuation in the region of the submucosa due to ede-
ma as well as pericolonic fluid and stranding of the adja-
cent fat. Pneumatosis and ascites have also been described
[21]. AIDS-related colitis is being encountered with de-
creasing incidence in industrialized countries due to the
efficacy of highly active anti-retroviral therapy (HAART).
Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases
Typhlitis
Typhlitis (neutropenic colitis) is a potentially fatal in-
fection of the cecum and ascending colon caused by en-
teric pathogens in patients with severe immunosuppres-
sion. It is most frequently seen in patients with acute
leukemia receiving chemotherapy but also occurs in the
setting of AIDS, aplastic anemia, multiple myeloma,
and bone marrow transplantation. Bacteria, viruses, and
fungi penetrate the damaged cecal mucosa and are able
to proliferate due to the profound neutropenia. Edema
and inflammation involve the cecum, ascending colon,
and, occasionally, the ileum. Fever, abdominal pain,
nausea, and diarrhea are presenting symptoms. Prompt
diagnosis and supportive therapy with intensive anti-
biotics and fluids are required to prevent transmural
necrosis and perforation. Surgical resection is indicated
in patients with transmural necrosis, intramural perfora-
tion, abscess, or uncontrolled sepsis and gastrointestinal
hemorrhage.
Due to the inherent risks in these critically ill patients
of bowel perforation during barium enema administration
and colonoscopies, CT is the study of choice in typhlitis.
CT demonstrates circumferential mural thickening
(1-3 cm) of the cecum, low-density areas within the colonic
wall secondary to edema, pericolonic inflammation and
fluid, and, in severe cases, pneumatosis. Clinically, CT is
used to monitor the decrease in mural thickness with
therapy and to detect subtle pneumoperitoneum in cases
of silent perforation or necrosis [22, 23].
Graft-Versus-Host Disease
Colonic complications of graft-versus-host disease
(GVHD) are common. In this disorder, mature donor
lymphocytes attack recipient tissue in bone marrow trans-
plantation patients. Acute GVHD may manifest with pro-
fuse diarrhea, nausea, vomiting, intestinal hemorrhage,
and cramping abdominal pain. Transmural inflammation
is common but frank perforation is unusual.
Initially, CT shows diffuse mural thickening of the
colon with submucosal edema and increased intraluminal
secretions. In chronic GVHD, submucosal fat may be
rapidly deposited and is readily identified on CT [20].
Differential Diagnosis of the Colitides
The differentiation of granulomatous colitis and ulcera-
tive colitis is important in terms of medical management,
surgical options, and prognosis. This distinction can usu-
ally be made on the basis of colonoscopy, with biopsy
histology, double-contrast barium enema, disease distri-
bution, and clinical course. CT can occasionally help
distinguish these disorders by demonstrating differences
in mural thickness, wall density, distribution of colonic
41
involvement, and the presence or absence of small bow-
el disease, abscess, fistula, and fibrofatty mesenteric
proliferation.
Idiopathic IBD must also be differentiated from the in-
fectious colitides. Although there is considerable overlap in
the CT findings of these disorders, there are certain differ-
entiating features [24]. For example, the presence of ascites
is more suggestive of an acute rather than a chronic cause
of colonic inflammation. Peritoneal fluid is commonly
found in the acute colitides, particularly pseudomembra-
nous, infectious, and ischemic colitis, but not in chronic
IBD. Ascites is only infrequently seen in patients with
acute IBD. Submucosal fat deposition, seen on CT, is pri-
marily found in the subacute and chromic colitides, usual-
ly ulcerative colitis, but not in acute disease.
Appendicitis
Acute appendicitis is the most common abdominal surgi-
cal emergency, affecting 250,000 individuals in the USA
annually. The lifetime risk of developing acute appen-
dicitis is 8.6% for men and 6.7% for women. Radiologists
play a critical role in evaluating patients with suspected
appendicitis and in minimizing its complications by con-
firming or excluding the diagnosis in atypical cases. They
also can reduce the number of misdiagnoses and negative
laparotomies, provide a correct alternate diagnosis, and
manage appendiceal abscesses and postoperative compli-
cations. For example, one study reported a drop in the
negative appendectomy rate from 24 to 3% after the
widespread use of CT [25].
Of the cross-sectional imaging tests, CT has been the
most extensively studied in terms of accuracy, impact on
patient management, length of hospital stay, and cost sav-
ings. Contrast-enhanced helical CT has a sensitivity,
specificity, and accuracy of >95% in the diagnosis of
acute appendicitis [26-28]. US, when carried out with the
graded compression technique, has been shown to be a
valuable approach as well (Fig. 5) [29], although head-to-
head comparative studies show that CT has a significant
higher accuracy than US [30].
On CT, the abnormal appendix presents as a slightly
distended, fluid-filled (Fig. 6) or collapsed structure ap-
proximately 0.5-2 cm in diameter. The fluid level within
the lumen is >2.6 mm [31]. As a result of the inflamma-
tory hyperemia, the wall of the diseased appendix shows
hyperenhancement during the arterial phase of contrast
administration. The wall is circumferentially and asym-
metrically thickened (usually 1-3 mm). Periappendiceal
inflammation, the hallmark of appendicitis, is character-
ized by increased hazy density or linear fat stranding in
the adjacent mesoappendix, by fluid-containing abscesses,
and by ill-defined, heterogeneous soft-tissue densities
representing a phlegmon [26-28]. There may be sec-
ondary inflammatory and edematous changes – with
thickening of the wall of the adjacent ileum and cecum –
that may mimic primary ileocolic inflammatory disease.
42
Fig. 5 Acute appendicitis. Ultrasound shows a thickened, fluid-
filled appendix, surrounding infiltration and free fluid
Fig. 6. Acute appendicitis. Coronal reformatted CT image shows a
thickened, fluid-filled appendix (arrow) along the lateral aspect of
the right psoas muscle
CT after intravenous contrast medium administration
is the preferred technique, as it facilitates identification
of the inflamed appendix and suggests alternative diag-
noses. In patients with (imminent) renal insufficiency, a
non-contrast CT can be performed. The diagnosis of
acute appendicitis on non-contrast CT scans requires the
detection of a thickened appendix (diameter >6 mm) with
associated inflammatory changes in the periappendiceal
fat or abnormal thickening of the right lateroconal fascia,
with or without a calcified appendicolith.
The addition of coronal and sagittal reformatted im-
ages increases diagnostic confidence by virtue of the
more reliable demonstration of the entire appendix, sur-
rounding fat and lymph nodes, and periappendiceal in-
fection and inflammation [32]. Confidence in image in-
terpretation also improves with the use of thinner recon-
struction sections [33].
Jaap Stoker, Richard M. Gore
The combination of right lower quadrant inflammation,
phlegmon, and an abscess adjacent to the cecum is sug-
gestive but not diagnostic of appendicitis. Indeed, if an ab-
normal appendix or an appendicolith is not shown, the dif-
ferential diagnosis also must include Crohn’s disease, ce-
cal diverticulitis, ileal diverticulitis, perforated cecal or ap-
pendiceal carcinoma, and pelvic inflammatory disease.
Abscesses may be found in locations distant from the
cecum because of the length and position of the appendix
and the patterns of fluid migration in the peritoneal cavity.
It should be noted that (depending on referral patterns)
the majority (60-70%) of patients with suspected appen-
dicitis who are referred for cross-sectional imaging do
not have this disease. Instead, most of these patients have
benign, self-limited gastrointestinal disorders such as vi-
ral gastroenteritis. CT and US often can suggest a spe-
cific alternate diagnosis [26-28]. Adnexal cysts, masses,
salpingitis, and tubo-ovarian abscesses are readily shown
on US. Ureteral calculi and pyelonephritis can be detect-
ed on CT and US. Enlarged lymph nodes in the right low-
er quadrant suggest mesenteric adenitis or infectious
ileitis; mural thickening of the terminal ileum can be seen
in Crohn’s disease or infectious ileitis.
Although CT has a higher accuracy than US, the latter
technique can be used initially, to reduce radiation expo-
sure. Only those patients with a negative or inconclusive
examination should proceed to CT [34]. Another approach
is to use MRI instead of CT, which combines the benefits
of a lack of ionizing radiation exposure with high-contrast
resolution cross-sectional imaging [35, 36]. Initial reports
on MRI in diagnosing acute appendicitis are encouraging,
particularly concerning pregnant women (Fig. 7) [37, 38].
Fig. 7. Acute appendicitis. Coronal HASTE (half-Fourier acquisi-
tion single-shot turbo spin-echo) in a 28-year-old woman at
18 weeks gestation, clinically suspected of having appendicitis but
in whom ultrasound was non-diagnostic. MRI shows a thickened
retrocecal appendix (arrow) with increased signal intensity and
minimal infiltration of the surrounding fat. The MRI diagnosis of
appendicitis was confirmed at surgery
Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases
Scan times are short, namely 10-15 min. At present, how-
ever, MRI is not widely used in the Emergency Depart-
ment for the diagnostic work-up of patients with acute ab-
dominal pain, due to a lack of availability and expertise
and the uncertainly as to its cost-effectiveness. Further
studies should be directed at better defining the role of
MRI in acute abdominal pain, especially its role compared
to US and CT. Nonetheless, in pregnant women with an
inconclusive US, MRI is currently the preferred imaging
technique.
Diverticulitis
It is estimated that 10-25% of individuals with diverticu-
losis will suffer from episodes of peridiverticular inflam-
mation during their lifetime. In the USA, this complica-
tion accounts for approximately 200,000 hospitalizations
and a health-care expenditure of four billion dollars an-
nually. Among the patients who are hospitalized, 10-20%
require emergency surgery [39]. Clinical signs indicative
of diverticulitis are inaccurate, although the combination
of direct tenderness only in the left lower quadrant, the
absence of vomiting, and an elevated C-reactive protein
level is suggestive in 25% of these patients (Laméris et
al., personal communication).
Inflammatory change in the pericolic fat (Fig. 8) is
the hallmark of diverticulitis on CT and is seen in 98%
of patients with the disease. The extent of the inflam-
matory reaction is related to the size of the perforation,
degree of bacterial contamination, and the host re-
sponse. Mild cases may manifest as areas of slight in-
Fig. 8. Acute diverticulitis. Coronal reformatted CT image shows
mural thickening of the sigmoid colon, inflammatory changes, and
a gas bubble (arrow) in the sigmoid mesocolon
43
crease in the density of the fat adjacent to the involved
colon or as fine linear stranding with small fluid col-
lections or bubbles of extraluminal air. In sigmoid di-
verticulitis, the fluid is typically decompressed into the
inferior interfascial plane. Due to the hypervascularity
of the inflamed area, contrast-enhanced CT scans often
reveal engorged mesenteric vessels in the involved peri-
colic fat. Pericolic heterogeneous soft-tissue densities
representing phlegmons and partially loculated fluid
collections indicating abscess are seen in more severe
cases. The abscess cavities usually contain air bubbles
or air-fluid levels. They develop within the sigmoid
mesocolon or are sealed off by the sigmoid colon and
adjacent small bowel loops. Less commonly, they may
form in the groin, flank, thigh, psoas muscle, sub-
phrenic space, or liver [40].
On CT, diverticula are seen at the site of perforation
or adjacent to it in about 80% of cases. They appear as
small outpouchings of air, contrast, or fecal material
projecting through the colonic wall. Symmetrical mural
thickening of the involved colon of approximately
4-10 mm is found in about 70% of cases; however, if
there is marked muscular hypertrophy, the wall of the
colon can measure up to 2-3 cm in thickness. CT can
also demonstrate intramural abscesses and fistula, and
is helpful in patients with suspected colovesical fistu-
las. In the latter case, a pericolic inflammatory mass
involves the bladder wall; the presence of intraluminal
gas confirms the diagnosis.
CT has a reported sensitivity of up to 98% in the di-
agnosis of diverticulitis [40]. Additionally, it can demon-
strate disease extent, such as abscess and peritonitis re-
mote from the colon, and can guide percutaneous abscess
drainage. The diagnosis of other pathological conditions
that may clinically simulate diverticulitis can also be
achieved with CT.
Although the accuracy of US in most prospective
studies is not inferior to that of CT [41], it has its limi-
tations and a recent large cohort study demonstrated the
superiority of CT. Moreover, CT is more accurate in in-
dicating alternative diagnoses [41] and provides a better
overview of disease extent, which is important for clini-
cal management (Hinchey classification). An initial
study demonstrated that MRI is also accurate in diag-
nosing diverticulitis [42], but further studies are needed
to evaluate its role. The major advantage of MRI in
middle-aged patients with diverticulitis is avoidance of
intravenous contrast medium and thus of contrast-
induced nephropathy.
Epiploic Appendagitis
Primary epiploic appendagitis is a relatively uncommon
condition that results from acute ischemia and inflamma-
tion of the appendices epiploicae. This disorder is often
associated with torsion and infarction of these appendices
and can simulate diverticulitis if it occurs in the sigmoid
44
Fig. 9. Epiploic appendagitis at the junction of the descending colon
and sigmoid colon. A fatty density mass is surrounded by increased
attenuation. Thrombosed vessels (arrow) can be seen in the central
portion of the epiploic appendage
and appendicitis if located in the proximal colon. A char-
acteristic appearance of a small, round, or oval fat-
containing mass with an associated inflammatory reac-
tion of the pericolic fat is seen on US and CT (Fig. 9)
[43]. The thrombosed vessels within the affected
appendage epiploica can sometimes be visualized.
Epiploic appendagitis is a self-limited process with clin-
ical resolution in a few days. Follow-up CT examination
may show total resolution, shrinkage, and eventual calcifi-
cation of the inflamed and infarcted epiploic appendix.
Acute Abdominal Pain
Inflammatory conditions of the colon are the most fre-
quent causes of acute abdominal pain requiring urgent
treatment (e.g., appendicitis, diverticulitis). In the major-
ity of patients with acute abdominal pain, the clinical di-
agnosis is unclear or incorrect such that imaging is
mandatory [44]. In this setting, imaging tests give differ-
ent results when used in a setting of a specific disease
(e.g., acute appendicitis) than when referring to all pa-
tients with acute abdominal pain (a myriad of diagnoses).
Thus, papers reporting on a specific diagnosis are less in-
formative than those relatively scarce papers reporting on
patients with acute abdominal pain in general. Plain ab-
dominal radiographs have a low accuracy in this setting
and result in management changes in only 4% of the pa-
tients [45, 46].
In the majority of patients, further imaging is warrant-
ed after plain radiography. US was shown to increase the
Jaap Stoker, Richard M. Gore
diagnostic accuracy of clinical evaluation (from 70 to
83%) and resulted in changes in management in 22% of
patients [47-49]. CT has a significant impact on the di-
agnosis, as shown in one series in which the accuracy im-
proved from 71% in the pre-CT clinical diagnosis to 93%
in the post-CT diagnosis, with a change in management
of 46% [50]. Also, the level of confidence increases sig-
nificantly with CT [51]. In two randomized controlled tri-
als from the same institution, CT was studied: (1) within
24 h versus routine workup by plain X-ray or (2) when
considered necessary, US, CT, or fluoroscopy and CT
within 1 h versus routine workup. The first study demon-
strated significantly fewer deaths (0 vs. 7; p=0.014) [52].
Length of hospital stay was not significantly different but
with CT there was a significant overstatement of serious
diagnoses. In the second study, routine CT significantly
improved diagnostic certainty, but there were no other
significant differences to routine workup in patient man-
agement [53]. The overall inter-observer agreement of
abdominal CT is good [54], and for specific urgent diag-
noses it is very good (e.g., for appendicitis, diverticulitis,
and bowel obstruction the values are 0.84, 0.90, and 0.81,
respectively).
A cohort study evaluated 11 diagnostic strategies of
clinical diagnosis, plain radiography (supine abdominal
X-ray and upright chest X-ray), US, and CT in 1021 pa-
tients with acute abdominal pain [34]. Clinical diagnosis
(performed by surgical residents) had a sensitivity of
88% and a specificity of 41% for urgent diagnoses. Plain
radiography did not contribute to a higher sensitivity or
specificity for urgent cases of acute abdominal pain. Re-
garding the clinical diagnosis, US reduced the number of
false-positive urgent diagnoses to 85%, but at the ex-
pense of a lower sensitivity (70%, thus missing 30% of
the urgent conditions). CT had the highest sensitivity
(89%) and specificity (77%) as a single imaging strate-
gy. The best strategy for the detection of urgent diag-
noses was an initial US, reserving CT for patients with
negative or inconclusive US examinations [34]. This
strategy led to a sensitivity for urgent diagnosis of 94%.
Initial US reduced CT use and associated radiation bur-
den by 51% compared to CT in all patients. Strategies
driven by body mass index, age, or location of the pain
had a lower sensitivity for urgent diagnoses than
achieved with this conditional strategy. The use of MRI
in acute abdominal pain primarily has been studied in
acute appendicitis and diverticulitis [36].
Perianal Fistulas
Perianal fistulas mostly occur either as the result of fis-
tulous disease originating from the anal glands near the
anal crypts (cryptoglandular hypothesis) or in patients
with Crohn’s disease [55]. Infection of the anal glands
may result in abscess formation. This is a relatively com-
mon condition, with a prevalence of approximately
0.01%, predominantly affecting young adults. The course
Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases 45

of the fistula track may be simple and superficial or com-
plicated. The latter may be intersphincteric (through the
internal anal sphincter and then downward through the
intersphincteric space) or trans-sphincteric (transversing
not only the internal sphincter but also the external
sphincter or puborectis muscle) or have a supralevator ex-
tent or an extrasphincteric extension to the rectum, with-
out involvement of the anal sphincter. These complicated
tracks need detailed imaging for proper therapy, as inad-
equate treatment may lead to recurrent disease. The sur-
geon must be aware both of the presence and number of
tracks and of their extent, the location of the internal
opening, and the presence of abscesses.
Pre-operative evaluation of perianal fistulas may in-
clude physical examination, examination under anesthe-
sia (EUA), endoscopic ultrasound (EUS), or MRI. Phys-
ical examination has significant shortcomings, especially
in patients with recurrent disease. While EUA can be
used for determining disease extent, immediately fol-
lowed by treatment, it has limitations and disadvantages,
mostly related to probing. Firstly, not all fistulas have an
external opening that can be probed, and probing may
miss secondary tracks. It is well-recognized that missed
extensions are the commonest cause of recurrence, which
reaches 25% in some series. Forceful probing may lead
to perforation of the levator plate, worsening the extent of
the disease. Patients with recurrent disease are most like-
ly to harbor missed disease but are also the most difficult
to assess. Digital palpation frequently cannot distinguish
between scarring due to repeated surgery and induration
due to an underlying extension.
EUS, which can be enhanced by hydrogen peroxide in-
stillation within the track, may be used to determine disease
extent. Although initial reports were encouraging, later
studies have been less sanguine especially when EUS was
compared to MRI. This discrepancy probably relates to
operator expertise since EUS is highly operator-dependent.
Insufficient penetration beyond the external sphincter,
especially with high-frequency transducers, limits the abil-
ity of EUS to resolve ischioanal and supralevator sepsis,
with the result that it may miss extensions from the prima-
ry tract. This technique is impressive in demonstrating the
internal opening whereas infection is distinguished only
with difficulty from postoperative fibrosis; however, hydro-
gen peroxide instillation facilitates this differentiation. Low
simple tracks presumably can be identified by EUS as
accurately as by MRI, but the latter is definitely superior in
cases involving complex or high tracks [56].
MRI has been proven to provide the most compre-
hensive assessment of patients with perianal fistulas, fa-
cilitating accurate identification of tracks and exten-
sions as well as abscesses. MRI examination for peri-
anal fistulas should include T2-weighted sequences in
multiple planes, a fat-saturation sequence, and a contrast-
enhanced (fat-saturation) T1-weighted sequence [57].
Tracks (Figs. 10, 11) are identified on T2 as hyper-

a b

Fig. 10 a, b. Crohn’s disease and perianal fistu-

las. a Axial T2-weighted turbo spin-echo
demonstrates multiple tracks (arrowheads),
intersphincteric and outside the anal sphinc-
ter. On the right, an abscess (A) is seen in the
ischioanal space. b Coronal T2-weighted tur-
bo spin-echo demonstrates a trans-sphincteric
track in the ischioanal space (arrowhead) and
extension of the abscess (A) in the levator ani
at both sides and in the rectal wall (arrow)

a b

Fig. 11 a, b. Cryptoglandular perianal fistula.

a Coronal T2-weighted turbo spin-echo with
endoanal coil shows a fibrous track (arrow-
head) extending through the left external
sphincter. Normal sphincter anatomy is seen
on the right. E External sphincter, I internal
sphincter, PR puborectal muscle, LA levator
ani plate. b Axial T2-weighted turbo spin-
echo demonstrates the hyperintense track
(arrow), surrounded by fibrous tissue (arrow-
head), along the left external sphincter (E)
46
intense longitudinal structures, often with a hypointense,
fibrous wall. Collateral inflammation is often appreciat-
ed on fat saturation sequences. After the administration
of intravenous contrast medium, the lining of the wall
will enhance. Non-enhancing fluid can be identified in
the center of the track or the track can be completely
obliterated by granulation tissue. In the latter case, there
is complete enhancement of the part of the track that is
hyperintense on T2 sequences (Fig. 11). Abscesses are
readily appreciated on fat saturation sequences, al-
though some small fluid collections may be more diffi-
cult to identify. The use of external phased array coils
may result in limitations in detecting superficial exten-
sions and difficulty in locating the precise level of the
internal opening. In such cases, endoluminal MRI may
provide more information (Fig. 11).
A prospective triple-blinded comparison of the accura-
cy of anal endosonography (AES), pelvic MRI, and sur-
gical EUA in perianal Crohn’s disease showed that AES
correctly classified fistulas in 91% of the cases, compared
with 87% for pelvic MRI and 91% for surgical evaluation
[58]. A combination of any two of the three modalities in-
creased the accuracy to 100%. Another study, in which
MRI and AES were compared with surgical findings,
showed MRI to be superior to AES in fistula classifica-
tion, with sensitivities of 84 vs. 60% for the two modali-
ties, and specificities of 68 and 21%, respectively.
Several studies have indicated a positive effect of pre-
operative MRI on patient outcome. In one study, the ther-
apeutic effect of MRI before EUA was 21.1% [59]. Ad-
ditionally it was shown that disease recurrence after
surgery could be reduced by about 75% if surgery was
guided by the MRI findings.
The differential diagnosis of perianal fistulas primarily
concerns fistulas originating from skin appendages: acne
conglobata, suppurative hidranitis, and pilonidal sinus.
The first two disorders are easy to recognize clinically, but
this is more difficult with pilonidal sinus. Imaging can be
used to differentiate between perianal fistula and pilonidal
sinus. In a study in seven patients with pilonidal sinus and
14 sex- and age-matched individuals with perianal fistu-
las, these conditions could be readily discriminated by the
absence in the former of intersphincteric sepsis or an en-
teric opening [60]. Osteomyelitis of the pelvis or femur
may give rise to abscesses and tracks that extend to the
anorectal region, whereas osteomyelitis is a rare finding in
perianal fistulas due to Crohn’s disease. Differentiating
these two conditions is usually not difficult, as the pre-
dominant disease localization (either extensive bone mar-
row edema or extensive tracks with intersphincteric ex-
tension and internal opening) will establish the diagnosis.
References
1. Gore RM, Laufer I, Berlin JW (2008) Ulcerative and granulo-
matous colitis: idiopathic inflammatory bowel disease. In: Gore
RM, Levine MS (eds) Textbook of gastrointestinal radiology,
3rd edn. Saunders, Philadelphia, pp 1071-1109
Jaap Stoker, Richard M. Gore
2. Markose G, Ng CS, Freeman AH (2003) The impact of helical
computed tomography on the diagnosis of unsuspected inflam-
matory bowel disease in the large bowel. Eur Radiol 13:107-113
3. Furukawa A, Saotome T, Yamasaki M et al (2004) Cross-sec-
tional imaging in Crohn disease. Radiographics 24:689-702
4. Horsthuis K, Bipat S, Bennink R, Stoker J (2008) Inflammato-
ry bowel disease diagnosed with US, MR, scintigraphy, and
CT: Meta analysis of prospective studies. Radiology 247:64-79
5. Johnson KT, Hara AK, Johnson CD (2009) Evaluation of col-
itis: usefulness of CT enterography technique. Emerg Radiol
16:277-282
6. Desmond AN, O’Regan K, Curran C et al (2008) Crohn’s dis-
ease: factors associated with exposure to high levels of diag-
nostic radiation. Gut 57:1524-1529
7. Ripolles T, Martinez MJ, Pardes JM et al (2009) Crohn dis-
ease: Correlation of findings at contrast-enhanced US with
severity at endoscopy. Radiology 253:241-248
8. Migaleddu V, Scanu AM, Quaia E et al (2009) Contrast-en-
hanced ultrasonographic evaluation of inflammatory activity
in Crohn’s disease. Gastroenterology 137:43-61
9. Ziech M, Stoker J (2010) MRI of the small bowel: enterogra-
phy. In: Stoker J (ed) Magnetic resonance imaging of the gastro-
intestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 117-134
10. Papanikolaou N, Gourtsoyianni S (2010) MRI of the small
bowel: enteroclysis. In: Stoker J (ed) Magnetic resonance
imaging of the gastrointestinal tract. Springer-Verlag, Berlin,
Heidelberg, pp 135-148
11. Rimola J. Rodriguez S, Gracia-Bosch O et al (2009) Role of
3.0-T colonography in the evaluation of inflammatory bowel
disease. Radiographics 29:701-719
12. Ajaj W, Lauenstein TC, Langhorst J et al (2005) Small bowel
hydro-MR imaging for optimized ileocecal distension in
Crohn’s disease: should an additional rectal enema filling be
performed? J Magn Reson Imaging 22:92-100
13. Zijta F, Stoker J (2010) Magnetic resonance imaging of the
colon (Colonography): Results. In: Stoker J (ed) Magnetic res-
onance imaging of the gastrointestinal tract. Springer-Verlag,
Berlin, Heidelberg, pp 185-204
14. Ajaj WM, Lauenstein TC, Pelster G et al (2005) Magnetic res-
onance colonography for the detection of inflammatory dis-
eases of the large bowel: quantifying the inflammatory activi-
ty. Gut 54:257-263
15. Horsthuis K, Bipat S, Stokkers PC, Stoker J (2009) Magnetic
resonance imaging for evaluation of disease activity in Crohn’s
disease: a systematic review. Eur Radiol 19:1450-1460
16. Röttgen R, Herzog H, Lopez-Häninnen E, Felix R (2006) Bow-
el wall enhancement in magnetic resonance colonography for
assessing activity in Crohn’s disease. Clin Imaging 30:27-31
17. Langhorst J, Kühle CA, Ajaj W et al (2007) MR colonography
without bowel purgation for the assessment of inflammatory
bowel diseases: diagnostic accuracy and patient acceptance.
Inflamm Bowel Dis 13:1001-1008
18. Rimola J, Rodríguez S, García Bosch O et al (2009) Magnet-
ic resonance for assessment of disease activity and severity in
Crohn disease. Gut 58:1113-1120
19. Kelly CP, LaMont JT (2008) Clostridium difficile – more dif-
ficult than ever. N Engl J Med 359:1932-1940
20. Horton KM, Corl FM, Fishman EK (2000) CT evaluation of
the colon: inflammatory disease. Radiographics 20:399-418
21. Turner DR, Markose G, Arends MJ et al (2003) Unusual caus-
es of colonic wall thickening on computed tomography. Clin
Radiol 58:191-200
22. Gluecker TM, Williamson EE, Fletcher JG et al (2003) Dis-
eases of the cecum: a CT pictorial review. Eur Radiol 13 Sup-
pl 4:L51-61
23. Kirkpatrick ID, Greenberg HM (2003) Gastrointestinal com-
plications in the neutropenic patient: characterization and dif-
ferentiation with abdominal CT. Radiology 226:668-674
24. Thoeni RF, Cello JP (2006) CT imaging of colitis. Radi-
ology 240:623-638
Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases
25. Raman SS, Osuagwu FC, Kadell B et al (2008) Effect of CT
on false positive diagnosis of appendicitis and perforation. N
Engl J Med 358:972-973
26. Pinto Leite N, Pereira JM, Cunha R et al (2005) CT evaluation
of appendicitis and its complications: imaging techniques and
key diagnostic findings. AJR Am J Roentgenol 185:406-417
27. Daly CP, Cohan RH, Francis IR et al (2005) Incidence of acute
appendicitis in patients with equivocal CT findings. AJR Am
J Roentgenol 184:1813-1820
28. Paulson EK, Harris JP, Jaffe TA et al (2005) Acute appen-
dicitis: added diagnostic value of coronal reformations from
isotropic voxels at multi-detector row CT. Radiology 235:
879-885
29. Puylaert JB, Rutgers PH, Lalisang RI et al (1987) A prospec-
tive study of ultrasonography in the diagnosis of appendicitis.
N Engl J Med 317:666-669
30. van Randen A, Bipat S, Zwinderman AH et al (2008) Acute
appendicitis: meta-analysis of diagnostic performance of CT
and graded compression US related to prevalence of disease.
Radiology 249:97-106
31. Moteki T, Ohya N, Horikoshi H (2009) Prospective examina-
tion of patients suspected of having appendicitis using new
computed tomography criteria including “maximum depth of
intraluminal appendiceal fluid greater than 2.6 mm. J Comput
Assist Tomogr 33:383-389
32. Kim YJ, Kim J-E, Kim HS, Hwang HY (2009) MDCT with
coronal reconstruction: clinical benefit in evaluation of sus-
pected acute appendicitis in pediatric patients. AJR Am J
Roentgenol 192:150-152
33. Johnson PT, Horton KM, Kawamoto S et al (2009) MDCT
for suspected appendicitis: effect of reconstruction section
thickness on diagnostic accuracy, rat of appendiceal visual-
ization, and reader confidence using axial images. AJR Am J
Roentgenol 192:893-901
34. Laméris W, van Randen A, van Es HW et al (2009) Imaging
strategies for detection of urgent conditions in patients with acute
abdominal pain: diagnostic accuracy study. BMJ 338:b2431
35. Lee KS, Pedrosa I (2010) Magnetic resonance imaging of
acute conditions of the gastrointestinal tract. In: Stoker J (ed)
Magnetic resonance imaging of the gastrointestinal tract.
Springer-Verlag, Berlin, Heidelberg, pp 283-314
36. Stoker J (2008) Magnetic resonance imaging and the acute ab-
domen. Br J Surg 95:1193-1194
37. Oto A, Ernst RD, Ghulmiyyah LM et al (2009) MR imaging
in the triage of pregnant patients with acute abdominal and
pelvic pain. Abdom Imaging 34:243-250
38. Cobben L, Groot I, Kingma L (2009) A simple MRI protocol
in patients with clinically suspected appendicitis: results in
138 patients and effect on outcome of appendectomy. Eur Ra-
diol 19:1175-1183
39. Humes DJ, Solyamani-Dodaran M, Fleming KM et al (2009)
A population-based study of perforated diverticular disease in-
cidence and associated mortality. Gastroenterology 136:1198-
1205
40. Kircher MF, Rhea JT, Kihiczak D, Novelline RA (2002)
Frequency, sensitivity, and specificity of individual signs
of diverticulitis on thin-section helical CT with colonic contrast
material: experience with 312 cases. AJR Am J Roentgenol
178:1313-1318
41. Laméris W, van Randen A, Bipat S et al (2008) Graded com-
pression ultrasonography and computed tomography in acute
47
colonic diverticulitis: meta-analysis of test accuracy. Eur Ra-
diol 18:2498-2511
42. Heverhagen JT, Sitter H, Zielke A, Klose KJ (2008) Prospec-
tive evaluation of the value of magnetic resonance imaging in
suspected acute sigmoid diverticulitis. Dis Colon Rectum
51:1810-1815
43. Singh AK, Gervais DA, Hahn PF et al (2005) Acute epiploic
appendagitis and its mimics. Radiographics 25:1521-1534
44. Stoker J, van Randen A, Laméris W, Boermeester MA (2009)
Imaging in acute abdominal pain. Radiology 253:31-46
45. MacKersie AB, Lane MJ, Gerhardt RT et al (2005) Nontrau-
matic acute abdominal pain: unenhanced helical CT com-
pared with three view acute abdominal series. Radiology
237:114-122
46. Kellow ZS, MacInnes M, Kurzencwyg D et al (2008) The role
of abdominal radiography in the evaluation of the nontrauma
emergency patient. Radiology 248:887-893
47. Walsh PF, Crawford D, Crossling FT et al (1990) The value of
immediate ultrasound in acute abdominal conditions: a critical
appraisal. Clin Radiol 42:47-49
48. Allemann F, Cassina P, Rothlin M, Largiader F (1999) Ultra-
sound scans done by surgeons for patients with acute abdom-
inal pain: a prospective study. Eur J Surg 165:966-970
49. Dhillon S, Halligan S, Goh V et al (2002) The therapeutic im-
pact of abdominal ultrasound in patients with acute abdominal
symptoms. Clin Radiol 57:268-271
50. Tsushima Y, Yamada S, Aoki J et al (2002) Effect of contrast-
enhanced computed tomography on diagnosis and manage-
ment of acute abdomen in adults. Clin Radiol 57:507-513
51. Rosen MP, Sands DZ, Longmaid HE, 3rd et al (2000) Impact
of abdominal CT on the management of patients presenting to
the emergency department with acute abdominal pain. AJR
Am J Roentgenol 174:1391-1396
52. Ng CS, Watson CJ, Palmer CR et al (2002) Evaluation of ear-
ly abdominopelvic computed tomography in patients with
acute abdominal pain of unknown cause: prospective ran-
domised study. BMJ 325:1387
53. Sala E, Watson CJ, Beadsmoore C et al (2007) A randomized,
controlled trial of routine early abdominal computed tomogra-
phy in patients presenting with nonspecific acute abdominal
pain. Clin Radiol 62:961-969
54. van Randen A, Lameris W, Nio CY et al (2009) Inter-observ-
er agreement for abdominal CT in unselected patients with
acute abdominal pain. Eur Radiol 19:1394-1407
55. Halligan S, Stoker J (2006) Imaging fistula-in-ano. Radiology
239:18-33
56. Ziech M, Felt-Bersma R, Stoker J (2009) Imaging of perianal
fistulas. Clin Gastroenterol Hepatol 7:1037-1045
57. Horsthuis K, Stoker J (2004) MRI of perianal Crohn’s disease.
AJR Am J Roentgenol 183:1309-1315
58. Schwartz DA, Wiersema MJ, Dudiak KM et al (2001) A com-
parison of endoscopic ultrasound, magnetic resonance imag-
ing, and exam under anesthesia for evaluation of Crohn’s peri-
anal fistulas. Gastroenterology 121:1064-1072
59. Buchanan G, Halligan S, Williams A et al (2002) Effect of
MRI on clinical outcome of recurrent fistula-in-ano. Lancet
360:1661-1662
60. Taylor SA, Halligan S, Bartram CI (2003) Pilonidal sinus dis-
ease: MR imaging distinction from fistula in ano. Radiology
226:662-667

CT Colonography: Updated
Daniel C. Johnson1, Michael Macari2
1 Department of Radiology, Mayo Clinic, Scottsdale, AZ, USA
2 Department of Radiology, New York University Langone School of Medicine, New York, NY, USA
Computed tomography (CT) colonography has been in
development for more than a decade, with hundreds of
articles now published on its performance and technical
capabilities. With the conclusion and publication of the
National CT Colonography trial [1] and endorsement of
the technique for screening by a multi-society task force
(including the American Cancer Society, American Col-
lege of Radiology, US Multi-society Task Force on Colo-
rectal Cancer) [2], the clinical validation of CT colono-
graphy in the prepared colon has been completed. This
chapter highlights the most important current issues for
CT colonography.
Patient acceptance of routine colorectal screening, in-
cluding CT colonography, remains a major barrier. In
2000, only 43% of US adults age 50 or older had under-
gone a sigmoidoscopy or colonoscopy within the previous
10 years or had used a fecal occult blood home test kit
within the preceding year [3]. The major disincentive for
patients considering CT colonography as a screening op-
tion is the laxative purgation (the same as that required for
colonoscopy) [4]. Advantages include the lack of required
sedation and intravenous line placement for CT colono-
graphy, a quick return to work following the examination,
and no need to inconvenience others for transportation to
and from the exam. The risk of perforation at CT colono-
graphy is considerably less than at colonoscopy. Further-
more, the examination only requires two breath-holds on
the CT scanner (in the supine and prone positions), with
the completion of most average examinations in 10 min,
which may help reassure hesitant patients. Still, the reali-
ty of a full bowel preparation, an enema tip, and full (al-
though brief) colonic insufflation is likely to delay the de-
cision for screening for some people.
The performance of CT colonography has undergone
exhaustive testing. The Pickhardt trial demonstrated a
sensitivity similar to colonoscopy [5], but concerns were
raised that community practices might not be able to
achieve these results. The National CT Colonography tri-
al (ACRIN 6664) studied 2531 individuals nationally
across 15 centers, including academic and private prac-
tices. The findings of this trial were similar to those of
the Pickhardt trial and have reassured many groups [1].
Radiologist’s training and testing were required for the
IDKD 2010-2013
ACRIN trial. Although some participants required more
training than others, all of them received a passing score
of 90% for easy and moderately difficult to detect lesions
[6]. The ACRIN trial also insisted on strict adherence to
protocol requirements, including stool tagging regimens,
mechanical insufflation of the colon, and thin-section and
low-dose CT techniques [7]. It is clear that meticulous at-
tention to all aspects of the examination is required to
achieve optimal results.
Extracolonic abnormalities are common in patients of
screening age [8-11]. A pragmatic approach to these find-
ings is needed; for example, radiologists should recom-
mend follow-up studies for those patients with findings
most likely to be of clinical significance. Patients (and
clinicians) will be grateful if additional testing is mini-
mized; for those that need addition studies, the recom-
mended optimal follow-up should be included in the
report.
It is unfortunate that the risk associated with the low
radiation dose required for CT colonography has been
misunderstood. The standard dose at CT colonography is
about half of the dose used for a standard body CT ex-
amination. This results in an average dose of approxi-
mately 5 mSv. The real risk of this exposure is unknown,
but the Health Physics Society has stated that for doses
in this range the risks for the development of radiation-
induced cancer are too small to measure or are non-
existent [12]. Even if a very small risk is assumed from
radiation exposure at CT, it must be balanced against the
risk of developing colon cancer and of other alternative
procedures. The risk of perforation (1:1000) and death
(1:17,000) at colonoscopy are real and can be measured
[13], but as a society there is a consensus that these risks
are outweighed by the risk of developing colon cancer
(about 1 in 13 without screening) [14].
Maintaining high-quality interpretations is a responsi-
bility that each individual, each practice, and our special-
ty should assume. The American College of Radiology
has established a national CT colonography database
within the National Radiology Data Registry (NRDR)
[15]. Selected process and outcome metrics can be quick-
ly entered on-line and compared to national benchmarks.
These measures include process metrics related to the CT
CT Colonography: Updated
technique and the adequacy of patient preparation, and
outcome metrics related to colon perforation, true-posi-
tive and false-positive rates for large (≥1 cm) polyps, and
the prevalence of significant extracolonic findings. Prac-
tices seriously interested in providing the best care should
be encouraged to participate in this data registry and
manage their practice such that benchmark metrics are
achieved.
A spirit of cooperation between radiologists and gas-
troenterologists is needed for optimal patient care. Guide-
lines will need to be jointly developed for the proper use
of colonography and colonoscopy, and for processes to
efficiently transfer patients with polyps to colonoscopy.
Those practices that are able to do this effectively will of-
fer patients a service of high value – and will likely find
themselves very busy.
In summary, CT colonography has completed its clin-
ical validation and is now ready for widespread clinical
application. Radiologists committed to performing the
examination to the highest quality must obtain the edu-
cation and equipment needed. We must focus our efforts
on the best in patient care, and ignore the political dis-
tractions that will come. We have an obligation to educate
referring physicians on the correct use of the technique.
Collaborations with gastroenterologists to ensure same-
day polypectomy for selected patient will enhance patient
care. Extracolonic findings must be vigilantly and prop-
erly reported so that only those patients with highly sig-
nificant lesions are recommended for additional follow-
up testing. Lastly, we should be committed to ongoing
quality measures to both improve and maintain the high-
est standards of care. Radiology has another exciting op-
portunity to serve the public, and to potentially help re-
duce the incidence of a common cancer killer.
References
1. Johnson CD, Chen MH, Toledano A et al (2008) Accuracy of
CT colonography for detection of large adenomas and cancer.
N Engl J Med 359:1207-1217
49
2. Levin B, Lieberman DA, McFarland B et al (2008) Screening
and surveillance for the early detection of colorectal cancer
and adenomatous polyps, 2008: a joint guideline from the
American Cancer Society, the US Multi-Society Task Force on
Colorectal Cancer, and the American College of Radiology.
Gastroenterology 134:1570-1595
3. Colorectal (Colon) Cancer. http://cdcgov/cancer/colorectal/
statistics/screening_rateshtm
4. Beebe TJ, Johnson CD, Stoner S et al (2007) Assessing atti-
tudes toward laxative preparation in colorectal cancer screen-
ing and effects on future testing: potential receptivity to com-
puted tomographic colonography. Mayo Clinic Proceedings
82:666-671
5. Pickhardt PJ, Choi JR, Hwang I et al (2003) Computed tomo-
graphic virtual colonoscopy to screen for colorectal neoplasia
in asymptomatic adults. New Eng J Med 349:2191-2200
6. Fletcher JG, Johnson CD, Toledano A et al (2005) ACRIN
6664: Lessions for CT colonography (CTC) training and cer-
tification. Radiological Society of North America Scientific
Assembly and Annual Meeting Program, Chicago, IL
7. Johnson CD, Chen MH, Toledano A et al. The National CT
Colonography Trial Protocol, ACRIN 6664. http://wwwacrinorg/
Portals/0/Protocols/6664/Protocol-ACRIN%206664%20
Amendment%201,%207706pdf
8. Gluecker TM, Johnson CD, Wilson LA et al (2003) Extra-
colonic findings at CT colonography: evaluation of prevalence
and cost in a screening population. Gastroenterology 124:911-
916
9. Hara AK, Johnson CD, MacCarty RL, Welch TJ (2000) Inci-
dental extracolonic findings at CT colonography. Radiology
215:353-357
10. Hellstrom M, Svensson MH, Lasson A (2004) Extracolonic
and incidental findings on CT colonography (virtual
colonoscopy). AJR Am J Roentgenol 182:631-638
11. Rajapaksa RC, Macari M, Bini EJ (2004) Prevalence and im-
pact of extracolonic findings in patients undergoing CT
colonography. Journal of Clin Gastroenterol 38:767-771
12. Radiation risk in perspective (2004) Position Statement of the
Health Physics Society
13. Waye JD, Kahn O, Auerbach ME (1996) Complications of
colonoscopy and flexible sigmoidoscopy. Gastrointest Endosc
Clin N Am 6:342-377
14. Lifetime Probability of Developing or Dying From Cancer.
http://wwwcancerorg/docroot/CRI/content/CRI_2_6x_Life-
time_Probability_of_Developing_or_Dying_From_
Cancerasp?sitearea=&level=
15. National Radiology Data Registry. https://nrdracrorg/portal/
Nrdr/Main/pageaspx

Imaging of Diffuse and Inflammatory Liver Diseases
Pablo R. Ros1, Rendon C. Nelson2
1 Department of Radiology, University Hospitals Health System, Case Western Reserve University, Cleveland, OH, USA
2 Department of Radiology, Duke University, Durham, NC, USA
Introduction
The category of diffuse liver diseases includes a variety
of disorders that typically involve the liver in a non-focal
fashion. It is important to note, however, that even with-
in this group there may be focal abnormalities, repre-
senting unusual expression of a disease that typically has
a diffuse manifestation. At the same time there may be fo-
cal neoplasms, some of which are benign and others ma-
lignant. With recent advances in cross sectional imaging,
the detection, characterization, and follow-up of diffuse
liver disease has been greatly facilitated.
This chapter is divided along the traditional classifica-
tion of cirrhosis, vascular disorders, congenital, metabol-
ic and storage, and neoplastic diseases. In addition, dif-
fuse and focal inflammatory/infectious diseases are dis-
cussed.
Fibrous Tissue Deposition (Cirrhosis)
Background
The preliminary events that lead to cirrhosis involve a
mechanism of injury whereby there is repeated exposure
of some noxious agent to the liver, resulting in hepatocyte
injury and/or death. Etiological noxious agents include al-
cohol, viral infection (specifically hepatitis B and hepati-
tis C), non-alcoholic steatohepatitis, autoimmune disor-
ders (such as primary sclerosing cholangitis and primary
biliary cirrhosis), and toxic agents (such as aflatoxin and
iron, specifically, primary hemochromatosis). Pathologi-
cally, the liver attempts to repair itself from the injury by
regenerating hepatocytes and depositing collagen. With
repeated exposure to the noxious agent, there is ongoing
hepatocyte regeneration and collagen deposition, resulting
in the formation of regenerative nodules and, eventually,
the formation of fibrous scars, respectively. Note that fi-
brous tissue deposition, specifically in the form of colla-
gen, can be treated and is reversible to a point. However,
with more severe or profound degrees of collagen and fi-
brous tissue deposition, this process is no longer re-
versible and is henceforth referred to as cirrhosis. Con-
IDKD 2010-2013
siderable research has been devoted to the non-invasive
quantification of fibrosis using ultrasound (US) and mag-
netic resonance imaging (MRI). Some of these techniques
use elastography, in which a mechanical impulse is used
to push or distort the liver followed by measurement of the
amount of liver movement. A stiffer, more fibrotic liver
will move less than a soft, less fibrotic liver.
Regenerative Nodules
Regenerative nodules are classified as micro-regenerative
or macro-regenerative. Micro-regenerative nodules mea-
sure <3 mm in diameter while macro-regenerative nod-
ules are those ≥3 mm. Macro-regenerative but not micro-
regenerative nodules are large enough to be visualized
with US, computed tomography (CT), or MRI. Further-
more, in 10-20% of patients (not 10-20% of regenerative
nodules), iron deposition occurs within the cytoplasm of
the regenerating hepatocytes, leading to the formation of
so-called siderotic nodules. This is in contradistinction to
transfusional hemosiderosis, in which excess iron is
phagocytized by the reticuloendothelial cells or Kuppffer
cells, not the hepatocytes. Although somewhat controver-
sial, there is evidence that intracellular iron has a toxic ef-
fect that may increase the risk of hepatocellular carcino-
ma (HCC) development later in life. Since 7-10% of pa-
tients with cirrhosis develop HCC and 10-20% of pa-
tients with cirrhosis develop siderotic nodules, there may
well be a correlation between the two entities.
Confluent Hepatic Fibrosis
The coalescence of fibrous tissue leads to the formation
of a large fibrous scar in the liver; this is referred to as
confluent hepatic fibrosis. It typically occurs in the mid-
dle of the liver, specifically, the anterior segment of the
right hepatic lobe (segments V and VIII) and the medial
segments of the left hepatic lobe (segments IVa and IVb).
The large fibrous scar emanates from the more central
portion of the liver and is associated with peripheral cap-
sular atrophy and retraction, which at times can be pro-
found. The enhancement pattern of confluent hepatic fi-
brosis on CT is similar to that of fibrous tissue elsewhere;
Imaging of Diffuse and Inflammatory Liver Diseases 51

that is, the scar is usually hypoattenuating during the un-
enhanced state, hypo- to iso-enhancing during the late he-
patic arterial phase, hypo- to iso-enhancing during the
portal venous phase, and iso-to hyper-enhancing during
the equilibrium phase (Fig. 1). The hyper-enhancement
phenomenon noted during the equilibrium phase occurs
as a result of slow accumulation and then slow wash-out
of contrast material within fibrous tissue. On MRI, the
appearance of confluent hepatic fibrosis may be different
than that of fibrous tissue elsewhere in the body, such as
ligaments and tendons. Due to the high water content
within this fibrous scar, hepatic fibrosis is typically hypo-
intense on T1-weighted images and hyper-intense on T2-
weighted images. Whereas the signal-intensity character-
istics parallel those of hepatic neoplasms, the location,
shape, and capsular retraction aid in distinguishing the
two entities. Following Gd-chelate administration, the
scan reveals slow wash-in and slow wash-out, similar to
the pattern obtained with iodinated agents on CT.
Hepatic Morphology in Cirrhosis
The morphology and shape of the liver typically change
in patients as they develop progressive fibrosis. Although
there are variations from patient to patient and from dis-
ease to disease, the most typical pattern is atrophy of the
right hepatic lobe and hypertrophy of the left hepatic and
caudate lobes. As a result of right hepatic lobar atrophy,
there is a decrease in the angle of the gallbladder fossa
with the horizontal. Furthermore, caudate lobe hypertro-
phy can be quantified with US, CT, or MRI by measur-
ing the ratio between the size of the caudate lobe and
right hepatic lobe. The transverse measurement of the
caudate lobe medially to the bifurcation of the right por-
tal vein, divided by the transverse dimension of the entire
right hepatic and caudate lobes, shows low sensitivity but
high specificity for cirrhosis when exceeding 0.60-0.65.
Nodularity of the capsular surface of the liver is also a
feature of cirrhosis. The appreciation of nodularity, how-
ever, is related to the size of the regenerative nodules. With
smaller regenerative nodules, as may occur in alcoholic
cirrhosis, it may be difficult to detect capsular nodularity.
Among the disorders characterized by the development of
large regenerative nodules, such as primary sclerosing
cholangitis, there may be deep nodularity. Furthermore,
well-defined clefts can be identified in the capsular sur-
face, particularly on the gastric side of the lateral segment
of the left hepatic lobe and the renal side of the posterior
segment of the right hepatic lobe. Lastly, there tends to be
prominence of the hepatic fissures, specifically, the gall-
bladder fossa and the fissure for the falciform ligament.
Portal Hypertension
Signs of portal hypertension in patients with cirrhosis in-
clude splenomegaly and/or porto-systemic shunts. A
spleen with a volume exceeding 250 mL or with a longi-
tudinal dimension >12 cm or an anteroposterior dimen-
sion >9 cm is considered enlarged. In some patients with
splenomegaly, focal iron deposits can be appreciated
within the parenchyma of the spleen. These are seen on
MRI as hypointense foci on either T1-weighted gradient

a b

Fig. 1 a-d. Confluent hepatic

fibrosis in a 55-year-old
man with cirrhosis. Axial
images through the liver
during the a unenhanced c d
state, b late hepatic arterial
phase, c portal venous
phase, and d equilibrium
phase. There is a wedge-
shaped area in the sup-
capsular portion of the me-
dial segment of the left he-
patic lobe that demonstrates
hypoattenuation pre-con-
trast and slow wash-in and
slow wash-out post-con-
trast. There is also associat-
ed capsular retraction
52
echo or T2-weighted pulse sequences and are referred to
as Gamna-Gandy bodies.
Porto-systemic shunts can occur either intrahepatical-
ly or extrahepatically. Intrahepatic porto-systemic shunt-
ing is a result of high-pressure, low-volume arterial blood
that mixes at the level of the hepatic sinusoids with low-
pressure, high-volume venous blood. In the cirrhotic liv-
er, arterial flow is increased and portal venous flow is
decreased, thereby increasing the prevalence of these
shunts. During the arterial phase of imaging, there is typ-
ically a parenchymal blush that is associated with early
portal venous enhancement.
Extrahepatic porto-systemic shunts are also seen in
the setting of cirrhosis. The most common is a sponta-
neous spleno-renal shunt whereby venous blood from the
spleen is shunted to the left renal vein via the left inferi-
or adrenal vein. In this setting, a prominent collateral
vein is usually seen draining directly into an enlarged
left renal vein. Porto-systemic varices are also common
in patients with cirrhosis and portal hypertension. These
include esophageal varices, splenic hilar varices, peri-
gastric varices, and peripancreatic varices. In some pa-
tients, profound shunting of blood from the splenic vein
retrograde into the inferior mesenteric vein precludes the
development of splenomegaly.
Ancillary Findings in Cirrhosis
Other findings that often occur in the setting of cirrhosis
include ascites, mesenteric edema, and bowel wall ede-
ma, particularly in the small bowel. Some of these
changes are due to portal hypertension although they may
be exacerbated by hypoproteinemia.
a b
c d
Pablo R. Ros, Rendon C. Nelson
Vascular Disorders
Arterial-Portal Shunts
Although some arterial-portal shunts appear to arise spon-
taneously, many of them are due to prior liver intervention,
such as a biopsy, or in the setting of cirrhosis, as noted
above. Furthermore, they are often associated with other
vascular malformations, such as focal nodular hyperplasia
and cavernous hemangiomas. On contrast-enhanced CT or
MRI, findings include a focal or wedge-shaped parenchy-
mal blush, a large hepatic artery, early portal venous en-
hancement, and a possible increase in the luminal diame-
ter of the portal vein (Fig. 2). Most of these arterial portal
shunts are subclinical and do not require intervention.
Perfusion Abnormalities
Perfusion abnormalities are common following the ad-
ministration of contrast agents. They are seen as focal,
often web-shaped, areas of hyperenhancement that are
most pronounced during the late hepatic arterial phase,
hence the term transient hyperattenuation defect
(THAD) for CT and transient hyperintensity defect
(THID) for MRI. Although the etiology of these defects
is not always apparent, many of them are due to alter-
ations in portal venous flow. For example, they are
often visualized in the setting of liver tumors (either
primary or metastatic) that block flow from a branch of
the portal vein. They can also be seen when there are
aberrant veins draining directly into the liver, leaving
parenchyma that receives venous blood from the
Fig. 2 a-d. Arterial-portal
shunt in a 49-year-old man
with hepatitis C. Axial im-
ages through the liver during
the late hepatic arterial
phase demonstrate a wedge-
shape area of focal paren-
chymal hyperenhancement
in the left hepatic lobe (a), a
large hepatic artery (b) as
well as early and vivid en-
hancement of a large peri-
umbilical collateral vein (c,
d). These findings are con-
sistent with an intrahepatic
arterial-portal shunt
Imaging of Diffuse and Inflammatory Liver Diseases
aberrant source only and not from the portal vein.
These perfusion abnormalities typically occur:
a) in the subcapsular area;
b) in the periligamentous area (about both the fissure for the
falciform ligament and the ligamentus venosum), due to
aberrant internal mammary or periumbilical veins;
c) about the porta hepatis, due to aberrant gastric veins
(right more common than left);
d) about the gallbladder fossa, due to aberrant cholecys-
tic veins.
On contrast-enhanced CT or MRI, they appear as tran-
sient areas of hyperenhancement during the late hepatic
arterial phase and are often web-shaped and subcapsular.
There is usually not an appreciable abnormality in this
area on the unenhanced images or during the portal ve-
nous or equilibrium phases.
Portal Vein Thrombosis
In portal vein thrombosis, the thrombus can be either
bland or malignant and either occlusive or non-occlu-
sive. Bland thrombosis typically occurs in the setting of
trauma, cirrhosis, following an orthotopic liver trans-
plant at the end-to-end anastomosis, in certain hyper-
coagulable states, and in 25% of patients with Budd-
Chiari syndrome. If the thrombus is not occlusive, an
eccentric intraluminal filling defect is identified, which
often resolves with anti-coagulant therapy. In this case,
there is no cavernous transformation (Fig. 3). With oc-
clusive portal vein thrombosis, however, the imaging
findings are different. Early on, the portal vein demon-
strates a non-enhancing intraluminal filling defect that
may distend the vein and increase the luminal diameter.
Furthermore, there is often enhancement of the wall of
the vein via the vaso vasorum. Over time, typically in the
range of 3-6 weeks, the thrombus undergoes retraction,
with the development of cavernous transformation, main-
ly via collaterals that develop in the vaso vasorum itself.
Malignant portal vein thrombosis typically occurs in pa-
tients that have HCC, in the setting of either cirrhosis or,
less commonly, non-cirrhosis. Interestingly, although
metastases to the liver commonly obstruct small branches
a b
Fig. 3 a, b. Non-occlusive
portal vein thrombosis in a
39-year-old woman with a
long history of taking birth
control pills. a Axial T2-
weighted and b post-con-
trast (late hepatic arterial
phase) images through the
liver demonstrate an eccen-
tric intraluminal defect in
the main portal vein
53
of the portal vein, they are uncommonly associated with
portal vein thrombosis, either bland or malignant. On CT
and MRI, occlusive malignant portal vein thrombosis typ-
ically enlarges the portal vein, the luminal diameter of
which may exceed 23 mm. Furthermore, the thrombus it-
self may enhance and this enhancement can be quite vivid
during the late hepatic arterial phase. With Doppler US,
sampling of the thrombus itself reveals low-resistance arte-
rial wave forms that often flow in the hepatofugal direction.
In patients with cirrhosis and HCC who are anticipating
liver transplantation, it is important to determine whether
portal vein thrombosis is bland or malignant. If the throm-
bosis is indeed malignant, the patient cannot be considered
a candidate for transplantation. The best way to confirm the
diagnosis is to percutaneously biopsy the intraluminal
thrombus itself under direct real-time US guidance.
Budd-Chiari Syndrome
Budd-Chiari syndrome may occur when there is obstruc-
tion to the outflow of blood from the hepatic veins. In
Western countries, such as the USA, the majority of these
cases (70%) are idiopathic. In the Orient, however, they
are commonly due to congenital webs that occur within
the hepatic veins themselves. Other disorders that are as-
sociated with Budd-Chiari syndrome include trauma,
pregnancy, certain hypercoagulable states, and malignant
hepatic vein thrombosis. The latter is commonly associ-
ated with primary tumors of the liver, right adrenal gland,
and right kidney, specifically HCC, adrenal cortical car-
cinoma, and renal cell carcinoma, respectively.
Budd-Chiari syndrome can occur with occlusion of
one, two, or all three hepatic veins or with occlusion of
the suprahepatic inferior vena cava. Furthermore, the
imaging findings may differ depending upon whether the
disease is acute or chronic. With acute Budd-Chiari syn-
drome, there may be one or more intraluminal thrombi,
most commonly identified by Doppler US. With chronic
Budd-Chiari syndrome, however, the hepatic veins are
small and often difficult to identify, although tortuous
intrahepatic collateral veins or shunts may be apparent. In
this setting, shunts may develop from one hepatic vein
54
a b
c d
that is obstructed to another hepatic vein that is not.
Shunting can also occur from a hepatic vein that is ob-
structed to the hepatic vein in the caudate lobe. Finally,
shunting may occur from a hepatic vein that is obstruc-
ted to the portal vein, one of the reasons why 25% of pa-
tients with Budd-Chiari syndrome develop portal vein
thrombosis. Furthermore, chronic occlusion of the hepat-
ic vein can result in significant enlargement of the cau-
date lobe and atrophy of the peripheral portion of the
right and left hepatic lobes. At times, large intrahepatic
collateral veins can be seen shunting blood to the hepat-
ic vein in the caudate lobe. Furthermore, on T1-weighted
MRI, the caudate lobe is often hyperintense. Following
contrast administration on either CT or MRI, there is of-
ten differential enhancement of the central and peripher-
al portions of the liver (Fig. 4); that is, early on, the cen-
tral portion of the liver hyperenhances but the periphery
does not. Later on, there is a flip-flop phenomenon in
which the central portion washes out and the peripheral
portion accumulates contrast media. Over time, profound
atrophy of the peripheral parenchyma can be seen. Pa-
tients commonly have and first clinically present with as-
cites, which develops shortly after hepatic vein occlusion.
In suprahepatic inferior vena caval obstruction, the he-
patic vein in the caudate lobe cannot be used as a conduit
to shunt blood from the hepatic veins to the inferior vena
cava. As a result, there is neither hypertrophy of the cau-
date lobe nor development of large intrahepatic collater-
al veins. It is important to note that, in the setting of
Budd-Chiari syndrome, the liver is often swollen, which
Pablo R. Ros, Rendon C. Nelson
Fig. 4 a-d. Axial images
through the liver during the
portal venous phase demon-
strate intense enhancement
of the central portion of the
liver and hypoenhancement
of the peripheral portion
(especially on a and b). The
hepatic veins are not visual-
ized. There is a cleft in the
medial aspect of both the
right and left hepatic lobes
(c, d) consistent with chron-
ic Budd-Chiari syndrome
and peripheral parenchymal
atrophy
may narrow the lumen of the intrahepatic inferior vena
cava. Although this is not the cause of Budd-Chiari syn-
drome, it certainly exacerbates the condition.
Passive Hepatic Congestion
Right-sided heart failure can result in the delayed
drainage of blood from the liver into the inferior vena ca-
va and right atrium. Images through the lower chest typ-
ically demonstrate either cardiac enlargement or a large
pericardial effusion, although in the setting of restrictive
pericarditis the heart may be normal in size. The key
finding with right-sided congestive heart failure or pas-
sive hepatic congestion is enlarged and distended hepat-
ic veins and inferior vena cava. Secondary signs include
reflux of contrast material from the superior vena cava in-
to the hepatic veins, although this is occasionally seen in
normal patients as well. In addition, there may be a mot-
tled enhancement pattern in the liver that is more pro-
nounced peripherally and more apparent during the late
hepatic arterial phase. This pattern, referred to as the
“nutmeg liver”, may not be apparent during the portal
venous and/or equilibrium phases. Over time, the liver
can become fibrotic, at which point it may have all of the
manifestations of cirrhosis and portal hypertension.
Macro-regenerative Nodules
In a small percentage of patients with outflow obstruction
of the hepatic veins, large regenerative nodules develop
Imaging of Diffuse and Inflammatory Liver Diseases
that tend to hyperenhance during the late hepatic arterial
phase. These nodules have been associated with both
Budd-Chiari syndrome and passive hepatic congestion.
They typically are iso-enhancing during the portal venous
and equilibrium phases and demonstrate only slow, peri-
odic growth over time.
Hepatic Veno-occlusive Disease
Hepatic veno-occlusive disease (VOD) typically occurs
in bone marrow transplant recipients who have under-
gone total body irradiation. The definition of VOD is pro-
gressive non-thrombotic occlusion of the hepatic venules.
Although some reports have noted sporadic reversal or
hepatofugal flow in the portal vein with VOD, there are
no reliable imaging findings for this diagnosis. As a re-
sult, VOD can only be reliably diagnosed by microscopic
examination of biopsy tissue.
Peliosis Hepatis
Peliosis hepatis is a rare entity in which the hepatic sinu-
soids throughout the liver dilate, resulting in numerous
blood-filled lacunar spaces ranging in size from 1 to 3 mm.
Similar lacunar spaces can also occur in the spleen, lymph
nodes, bone marrow, and lungs. Although the cause of
peliosis is poorly understood, it is believed to be due to
outflow obstruction of the sinusoid. It typically occurs in
patients who use anabolic steroids, corticosteroids, ta-
moxifen, or birth control pills; following cardiac or renal
transplantation; with chronic debilitating diseases, such as
tuberculosis, malignancy or AIDS; in association with di-
abetes, sprue, or Hodgkin’s disease; or in patients exposed
to arsenic or polyvinyl chloride. Imaging shows numerous
small cystic lesions that demonstrate an enhancement pat-
tern similar to that of the blood pool.
Hepatic Infarction
Parenchymal infarction in the liver is relatively uncom-
mon for two reasons:
1. the liver has a dual blood supply;
2. the hepatocytes are relatively insensitive to hypoxia.
It has been seen, however, in patients with shock, sep-
sis, eclampsia, sickle cell disease or trait, or arteritis; in
those who have taken birth control pills; and in those who
have suffered arterial embolic events such as those due to
endocarditis, rheumatic heart disease, trauma, intra-arterial
chemotherapy, or iatrogenic tumor embolization. With
imaging, infarcted parenchyma may be hypoechoic on US:
anechoic bile lakes may be visualized as necrosis pro-
gresses. On contrast-enhanced CT, there is a wedge-shaped
subcapsular region of hypoenhancement that later on may
contain bile lakes and gas bubbles. On MRI, the edema of
infarction is typically hypointense on T1-weighted images
and hyperintense on T2-weighted images; as with CT,
hypoenhancement and bile lakes are seen as well.
55
Metabolic and Storage Diseases
Steatosis
Hepatic steatosis results from a variety of abnormal
processes, including the increased production or mobi-
lization of fatty acids (e.g., obesity, steroid use) or the de-
creased hepatic clearance of fatty acids due to hepatocel-
lular injury (e.g., alcoholic liver disease, viral hepatitis).
Histopathologically, the hallmark of all forms of fatty liv-
er is the accumulation of fat globules within hepatocytes.
The distribution of steatosis can be variable, ranging from
focal, to regional, to diffuse. Diffuse steatosis is common
and estimated to occur in approximately 30% of obese
patients. Patients with steatosis are usually asymptomatic
although some may present with right upper quadrant
pain or abnormal liver function parameters.
The histopathological findings of non-alcohol-related
liver steatosis, also known as non-alcoholic fatty liver
disease (NAFLD), vary from steatosis alone to steatosis
with inflammation, necrosis, and fibrosis. Non-alcoholic
steatohepatitis (NASH), with or without cirrhosis, is po-
sitioned at the most severe end of the NAFLD spectrum.
The histopathological findings of NASH include steato-
sis (predominately macrovesicular), mixed lobular in-
flammation, and hepatocellular ballooning. Unlike
steatosis alone, NASH may progress to cirrhosis.
Diffuse fatty change is easily identified on CT. The at-
tenuation value of normal liver is usually ~8 HU greater
than that of spleen on non-contrast CT images. In patients
with fatty change, however, an abnormally decreased
density will be demonstrated, typically 10 and 25 HU less
than the spleen on non-contrast and contrast-enhanced
CT images, respectively. The diagnosis of hepatic steato-
sis is more reliably made on non-contrast images. Un-
doubtedly, the most sensitive technique to detect fatty
change of the liver is the use of in-phase and out-phase
gradient echo MRI pulse sequences (Fig. 5).
Hepatic fatty change is, however, not always uniform
but can instead present as a focal area of steatosis in an
otherwise normal liver (focal steatosis) or as subtotal fat-
ty change with sparing of certain areas (focal sparing).
On imaging, several features allow the correct identifica-
tion of focal fatty change or focal spared areas:
1. the typical periligamentous and periportal location;
2. lack of mass effect;
3. sharply angulated boundaries of the involved area;
4. non-spherical shape;
5. absence of vascular displacement or distortion;
6. lobar or segmental distribution.
Iron Overload
Iron overload states may arise from hemochromatosis,
with the preferential accumulation of iron within hepato-
cytes, or hemosiderosis, in which iron is deposited in
Kupffer cells.
56 Pablo R. Ros, Rendon C. Nelson

a b c

Fig. 5 a-c. Diffuse fatty liver in a 41-year-old female presenting with epigastric pain. a Axial CECT image demonstrates diffuse low attenu-
ation of the liver without displacement of the hepatic vessels. b In- and c out-of-phase T1-weighted images show significant signal drop
in the liver on the out-of-phase images

Primary Hemochromatosis thalassemia major, sideroblastic anemia, pyruvate kinase

deficiency, chronic liver disease), the excess iron is
Hereditary or primary hemochromatosis is an autosomal processed and accumulates in organs containing reticu-
recessive disorder of iron metabolism that is character- loendothelial cells, including liver, spleen, and bone
ized by the abnormal absorption of iron from the gut, marrow.
with subsequent excessive deposition of iron into hepato- Diffuse, increased attenuation of the liver and spleen is
cytes, pancreatic acinar cells, myocardium, joints, en- seen on CT (Fig. 6). On MRI, the extrahepatic signal in-
docrine glands, and skin. In addition, cells of the reticulo- tensity changes in the spleen and bone marrow allow pri-
endothelial system (RES) in patients with primary hemo- mary hemochromatosis to be distinguished from hemo-
chromatosis are abnormal and unable to store processed siderosis. Although, in general, the clinical significance of
iron effectively. Consequently, these patients do not transfusional iron overload states is negligible, patients
accumulate iron into the RES. The clinical findings of with chronic hemosiderosis can develop symptoms similar
cirrhosis and its complications (portal hypertension, to those of the primary form as well as cirrhosis and HCC.
development of HCC) predominate in patients with long-
lasting disease. Wilson’s Disease
On CT, excessive iron storage within hepatocytes will
result in an overall increased density. However, this CT Wilson’s disease, also known as hepatolenticular degen-
appearance of a hyperdense liver is non-specific, since eration, is a rare autosomal recessive abnormality of cop-
similar features can be seen with gold deposition and in per metabolism that is characterized by the accumulation
Wilson’s disease, type IV glycogen storage disease, and
following amiodarone administration. The use of non-
contrast CT in patients with suspected hemochromatosis
is important because excessive iron cannot be detected in
the setting of enhancing parenchyma.
MRI is far more specific than any other imaging
modality for the characterization of iron overload, due to
the magnetic susceptibility effect of iron. The superpara-
magnetic effect of accumulated iron in the hepatocytes
results in a significant reduction of signal intensity on
T2-weighted images. Comparison of the signal intensity
of liver with that of paraspinal muscles provides a useful
internal control. HCC, seen in 35% of patients with ad-
vanced hemochromatosis, is usually easily detected on
T1- and T2-weighted images due to the background of
decreased signal intensity of the liver.

Hemosiderosis
Fig. 6. Hemosiderosis in a 45-year-old female with long history of
In patients with hemosiderosis or siderosis, due to transfu- sickle cell anemia requiring multiple transfusions. Axial non-
sional iron overload states or dyserythropoiesis (e.g., enhanced CT image demonstrates increased attenuation of the liver
Imaging of Diffuse and Inflammatory Liver Diseases
of toxic levels of copper in the brain, cornea (Kayser-
Fleischer rings), and liver, the latter due to impaired bil-
iary excretion. Hepatic deposition of copper, predomi-
nantly seen in periportal areas and along the hepatic si-
nusoids, evokes an inflammatory reaction resulting in
acute hepatitis with fatty change. Subsequently, chronic
hepatitis may result in liver fibrosis and, eventually,
macronodular cirrhosis.
Due to the high atomic number of copper, a hyper-
dense liver may be seen on unenhanced CT scans. How-
ever, this finding is not universally present; instead, usu-
ally only non-specific signs, such as hepatomegaly, fatty
change, and in advanced cases, cirrhosis, are observed.
During the early stage of the disease, and due to the para-
magnetism of ionic copper, MRI can be valuable as it
demonstrates focal copper depositions, such as multiple
nodular lesions, typically appearing hyperintense and
hypointense on T1- and T2-weighted images, respectively.
Amyloidosis
In amyloidosis, fibrils of protein-mucopolysaccharide
complexes are deposited throughout the body. The dis-
ease is classified based on the biochemical composition
of the amyloid fibrils. Primary amyloidosis is due to the
deposition of immunoglobin light chains and is associat-
ed with multiple myeloma and monoclonal gammopathy.
Secondary amyloidosis results from the deposition of
amyloid A protein and is associated with chronic infec-
tion, rheumatoid arthritis, and malignancies. Exceeded
only by the spleen and kidney, the liver is the third most
common solid organ prone to amyloid deposition.
Hepatic amyloidosis has a non-specific imaging ap-
pearance, with the most common finding being diffuse
hepatomegaly. CT sporadically demonstrates focal areas
of low attenuation within the liver, corresponding to sites
of amyloid deposition (amyloid pseudotumor).
Neoplastic Diseases
Metastatic Disease
Neoplastic infiltration due to diffuse metastatic disease
can occur with many primary tumors. Melanoma, malig-
nant neuroendocrine tumors, pancreatic adenocarcinoma,
breast carcinoma, and colonic adenocarcinoma are some
of the more commonly encountered causes of diffuse he-
patic metastatic disease.
The CT appearances of hepatic metastases depend on
the vascularity of the lesions compared with the normal
liver parenchyma. Diffuse metastatic involvement may
produce only subtle imaging findings and be detectable
only through indirect features, such as diffuse parenchy-
mal heterogeneity, vascular and architectural distortion,
or alterations of the liver contour. The latter, particularly
seen in patients with treated breast cancer metastases, has
been described as the “pseudocirrhosis” sign (Fig. 7).
57
Fig. 7. Diffuse metastatic breast cancer (pseudocirrhosis pattern) in
a 43-year-old woman treated for metastatic breast cancer. Axial
contrast-enhanced CT image demonstrate several small low-density
lesions in the liver and a nodular contour of the liver due to hepatic
capsular retraction
Lymphoma
Lymphoma can infiltrate the liver both primarily and sec-
ondarily. Primary lymphoma of the liver is extremely
rare. Conversely, the liver is often secondarily involved in
Hodgkin’s and in non-Hodgkin’s lymphoma. Typically,
the liver parenchyma is diffusely infiltrated with micro-
scopic nests of neoplastic cells, without significant ar-
chitectural distortion. Consequently, lymphomatous in-
volvement is difficult to detect by imaging alone. Asso-
ciated abnormalities, such as splenomegaly and lym-
phadenopathy, may narrow the differential diagnosis.
Diffuse Infectious and Inflammatory Diseases
Fungal Infections
Hepatosplenic fungal infection is a clinical manifestation of
disseminated fungal disease in patients with hematological
malignancies or compromise of the immune system. The
reported prevalence of fungal dissemination ranges from 20
to 40%. Most hepatic fungal microabscesses occur in
leukemia patients and are caused by Candida albicans.
Candidiasis
Candida albicans in the liver may evoke little or no in-
flammatory reaction, cause a superlative response, or oc-
casionally produce granulomas. The typical histological
pattern of hepatic candidiasis is characterized by micro-
abscesses, with the yeast or pseudohyphal forms of the
fungus in the center of the lesion and a surrounding area
of necrosis and polymorphonuclear infiltrate.
At contrast-enhanced CT, fungal microabscesses usu-
ally appear as multiple, round, discrete areas of low at-
tenuation, generally ranging in size from 2 to 20 mm
(Fig. 8). These microabscesses usually enhance centrally
58
Fig. 8. Hepatic candidiasis in a 66-year-old female with leukemia
who presented with abnormal liver function tests. Axial contrast-
enhanced CT image demonstrates multiple, small, low-attenuation
lesions distributed throughout the liver and spleen. Splenomegaly
and bilateral pleural effusions are also seen
after intravenous administration of contrast medium, al-
though peripheral enhancement may occur as well.
At MRI, the untreated nodules are rounded lesions
<1 cm in diameter that are minimally hypointense on T1-
weighted and gadolinium-enhanced images and markedly
hyperintense on T2-weighted images. After treatment, the
lesions appear mildly to moderately hyperintense on T1-
and T2-weighted images and demonstrate enhancement
on gadolinium-enhanced images. A dark ring is usually
seen around these lesions with all sequences. Completely
treated lesions are minimally hypointense on T1-weighted
images, isointense to mildly hyperintense on T2-weighted
images, moderately hypointense on early gadolinium-
enhanced images, and minimally hypointense on delayed
gadolinium-enhanced images. MRI is superior to CT and
US in the detection of these fungal foci.
Granulomatous Diseases
Granulomatous hepatitis is associated with numerous
conditions, most commonly, sarcoidosis, tuberculosis,
and histoplasmosis. Hepatic granulomas usually appear
as discrete, sharply defined nodules consisting of aggre-
gates of epithelioid cells by a rim of mononuclear cells,
predominantly lymphocytes.
Sarcoidosis
Sarcoidosis is a multi-system disorder of unknown patho-
genesis, characterized by non-caseating granulomas. Al-
though it may involve almost any organ in the body, pul-
monary sarcoidosis is the most common. Sarcoidosis of
the liver is also relatively frequently seen, but the granu-
lomas are usually not macroscopically detectable and
thus may not produce focal abnormalities on imaging
studies. Classically, the granulomas develop in a peri-
portal location, resulting in periportal fibrosis, cirrhosis,
and eventually portal hypertension.
Pablo R. Ros, Rendon C. Nelson
Hepatic contrast-enhanced CT may typically reveal
multiple, diffuse, small, low-density areas in both the liv-
er and spleen. MRI features of hepatic sarcoidosis are al-
so non-specific and include organomegaly, multiple le-
sions of low signal intensity relative to background
parenchyma with all sequences, increased periportal sig-
nal, irregularity of the portal and hepatic vein branches,
and patchy areas of heterogeneous signal.
Tuberculosis
Tuberculosis is one of the most common infectious dis-
eases worldwide. Generally, tuberculosis of the liver pre-
sents as either a miliary form or a local form, which is fur-
ther subdivided into nodular tuberculosis (i.e., tuberculous
abscess and tuberculoma) and tubular or hepatobiliary tu-
berculosis (i.e., tuberculosis involving the intrahepatic
ducts). Hepatic miliary tuberculosis is most common and
is reported to occur in 50-80% of all patients with terminal
pulmonary tuberculosis. Miliary tuberculosis is usually not
detected at imaging. Hepatomegaly may be the only radio-
logical abnormality. In the healing stage of tuberculosis,
CT may show diffuse hepatic calcifications (approximate-
ly 50% of cases). Reported CT findings of nodular tuber-
culosis are non-specific and include hypoattenuating le-
sions both before and after intravenous administration. At
MRI, the lesions are hypointense on T1-weighted images
and hypo- to iso-intense on T2-weighted images. Tubercu-
losis lesion differently enhance after gadolinium adminis-
tration. Given these rather non-specific findings with all
imaging techniques, percutaneous liver biopsy is necessary.
Histoplasmosis
Histoplasmosis is the most common cause of fungal in-
fection in the Ohio River Valley of the USA. Fortunately,
99% of patients exposed to histoplasmosis develop only
subclinical infections. Liver involvement is common in
disseminated histoplasmosis, which usually originates in
the lungs. The most common hepatic findings include
portal lymphohistiocytotic inflammation and discrete,
well-delineated granulomas. In patients with healed
histoplasmosis, the presence of small, punctate calcifica-
tions scattered throughout the liver and spleen is typical
but a non-specific finding.
Parasitic Infections
Schistosomiasis
Schistosoma japonicum, S. hematobium, and S. mansoni
are the three most important species that infect humans.
The schistosomas live in the bowel lumen and lay eggs in
the mesenteric veins. The eggs may then embolize to the
portal vein. The eggs themselves do not survive and subse-
quently calcify. Chronic infections with either S. japonicum
or S. mansoni result in the formation of cirrhosis and the
risk of development of HCC. Histologically, schistosomiasis
Imaging of Diffuse and Inflammatory Liver Diseases
is characterized by white, pinhead-sized granulomas scat-
tered throughout the liver. At the center of each granuloma
is a schistosome egg. In severe infections, the surface of the
liver shows granulomatous involvement and widespread fi-
brous portal enlargement (“pipe-stem” fibrosis).
At CT, the most pathognomonic pattern is the presence
of calcified septa, usually aligned perpendicular to the liv-
er capsule (“tortoise shell” or “turtle back” appearance).
Viral Infections
Viral Hepatitis
Acute viral hepatitis is a systemic infection that affects the
liver and is usually caused by one of five viral agents: he-
patitis A virus, hepatitis B virus (HBV), hepatitis C virus,
the HBV-associated delta agent or hepatitis D virus, and
hepatitis E virus. A vast array of other viruses may also
produce hepatitis, including herpes viruses, yellow fever
virus, rubella virus, Coxsackie virus, and adenovirus. The
diagnosis of acute hepatitis is usually based on serological,
virological, and clinical findings. Probably the most im-
portant role of radiology in patients with acute hepatitis is
to help rule out other diseases that produce similar clinical
and biochemical abnormalities, such as extrahepatic
cholestasis, diffuse metastatic disease, and cirrhosis.
At CT and MRI, the findings in acute viral hepatitis are
non-specific and include hepatomegaly and periportal ede-
ma. At CT, heterogeneous enhancement and well-defined
regions of low attenuation may be present. At MRI, peri-
portal edema appears as areas of high signal intensity on
T2-weighted images. Involved areas may be normal or
demonstrate decreased signal intensity on T1-weighted im-
ages and increased signal intensity on T2-weighted images.
There is also impaired uptake of liver-specific agents.
Extrahepatic findings in patients with severe acute hepati-
tis include thickening of the gallbladder wall due to edema
and, infrequently, ascites. On CT and MRI, the features of
chronic hepatitis resemble those of early-stage liver cir-
rhosis. Periportal lymphadenopathy may be the sole de-
tectable abnormality in both acute and chronic hepatitis.
HIV Infection
The liver and biliary tracts are frequent sites of involve-
ment during the course of HIV infection. Co-infection
with HBV and hepatitis C virus is particularly common
due to the shared means of transmission of these viruses
with HIV. AIDS-related cholangiopathy is the newest
common manifestation. At CT, inflammation of the gall-
bladder or biliary tree manifests as mural thickening or
abnormal contrast enhancement. Magnetic resonance
cholangiopancreatography is more sensitive and specific
than US or CT in depicting the mural irregularity of the
extrahepatic ducts, which results from the exuberant
periductal inflammation, focal mucosal ulcers, and the
interstitial edema found in AIDS-related cholangitis.
59
Uncommon Hepatic Infections
Cat-scratch Disease
Cat-scratch disease is an infection that affects immuno-
competent children or adolescents. It is caused by Bar-
tonella henselae, a gram-negative bacillus that is usual-
ly introduced by the scratch of a cat. Cat-scratch disease
takes many forms, from regional lymphadenitis to dis-
seminated infection. The typical clinical manifestation is
painful lymphadenopathy proximal to the site of inocu-
lation. Disseminated infection is seen in 5-10% of cases.
In the abdomen, multiple granulomas ranging from 3 to
30 mm may form in the liver and spleen, with or with-
out hepatosplenmegaly. Histopathological findings in-
clude vascular proliferative lesions (peliosis hepatis) and
necrotizing granulomatous lesions.
At unenhanced CT, the lesions are hypoattenuating rel-
ative to normal parenchyma. Three different patterns at
contrast-enhanced CT have been described:
1. persistent hypoattenuation relative to the liver;
2. isoattenuation relative to surrounding tissues;
3. marginal enhancement.
There are only a few MRI studies of cat-scratch dis-
ease. The lesions appear as nodules of low signal intensi-
ty on T1-weighted images and high signal intensity on
T2-weighted images. Peripheral enhancement may be
seen on gadolinium-enhanced T1-weighted images.
Bacillary Angiomatosis
Bacillary angiomatosis is also a manifestation of infec-
tion by Bartonella henselae, the same organism that
causes cat-scratch disease, but in immunocompromised
patients. The disease is characterized by localized areas
of vascular proliferation that may affect the skin, airway,
mucous membranes, visceral organs, bone, and brain.
Contrast-enhanced CT may demonstrate multiple diffuse
low- or high-attenuation lesions less than 1 cm that are
scattered throughout the hepatic parenchyma. Ascites,
mild periportal edema, and intrahepatic biliary ductal
dilatation may occur. These imaging features are non-
specific and must be distinguished, especially in AIDS
patients, from hepatic abscesses related to other bacteri-
al, viral, or fungal infections, AIDS-related lymphoma,
Kaposi sarcoma, and, less commonly, disseminated
Pneumocystis carinii infection.
Focal Infections
Bacterial (Pyogenic) Abscess
Pyogenic abscess, although uncommon in the antibiotic
era, is still challenging clinically since its presentation is
quite variable, from profound septicemia to chronic, in-
dolent symptoms.
60
Enhanced CT can reliably diagnose over 90% of he-
patic pyogenic abscesses, revealing two main patterns:
multiple microabscesses (disseminated or clustered) and
large macroabscesses. By virtue of its good spatial and
contrast resolution, CT is the single best method for de-
tecting hepatic abscess, with a sensitivity as high as 97%.
On CT scans, abscesses appear as generally rounded
masses that are hypodense on both contrast and non-
contrast scans. Central gas, as air bubbles or an airfluid
level, is a specific sign, but it is present in less than 20%
of cases. A thick, enhancing, peripheral rim is also noted.
At MRI, air within the abscess appears as a signal
void, thus making it more difficult to differentiate from
calcifications. However, the shape and location (air-fluid
level) of the abscess should enable the correct diagnosis.
After the administration of gadolinium-DTPA, abscesses
typically show rim enhancement (the “double target”
sign). Small lesions (<1 cm) may enhance homogeneously,
mimicking hemangiomas. Percutaneous, image-guided
aspiration followed by drainage is the method of choice
for definitive diagnosis and treatment, with success
achieved in >90% of cases.
Amebic Abscess
Hepatic abscess is the most common extraintestinal man-
ifestation of amebiasis, affecting approximately 10% of
patients with the disease. Although rare in the continen-
tal USA, 10% of the world’s population is infected with
Entamoeba histolytica. Clinically, patients with amebic
abscess are more acutely ill than those with pyogenic ab-
scess, with high fever and right upper quadrant pain. Di-
agnosis is made by positive serological amebic titers, al-
though they have false-negative rates of almost 20%.
Extrahepatic extensions to the chest wall, pleura, or
adjacent viscera are well demonstrated on CT. Percuta-
neous catheter drainage of an amebic abscess is rarely
necessary due to the effectiveness of amebicidal therapy.
Occasionally, percutaneous drainage is needed in large,
symptomatic abscesses with poor response to medical
therapy, suspected bacterial superinfection, and threaten-
ing intrapericardial rupture. The CT appearance of ame-
bic abscess is variable and non-specific. The lesions are
a b
Pablo R. Ros, Rendon C. Nelson
usually peripheral, round or oval areas of low attenuation
(10-20 HU). A peripheral rim of slightly higher attenua-
tion can be seen on non-contrast scans and shows
marked enhancement after the administration of contrast
material.
On MRI, amebic liver abscesses are spherical and usu-
ally solitary lesions with a hyperintense center on T2-
weighted images and a hypointense center on T1-weighted
images. The abscess wall is thick; on gadolinium en-
hanced images, the enhancement pattern is similar to that
of pyogenic abscess.
Echinococcal Disease
Hydatid disease has two main forms affecting humans,
resulting from infection with either Echinococcus granu-
losus or Echinococcus multilocularis or alveolaris. These
infections have well-defined and different geographical
distributions. The pathological and imaging findings dif-
fer dramatically between these parasites.
On CT scans, E. granulosus-infected sites are seen as
unilocular or multilocular, well-defined cysts with ei-
ther thick or thin walls. Daughter cysts are usually de-
tected as areas of lower attenuation than the mother cyst
and are usually in the periphery of the lesion. Daughter
cysts can also float free in the lumen of the mother cyst;
altering the patient’s position may change the position
of these cysts, confirming the diagnosis of echinococcal
disease. Curvilinear ring-like calcification is also a
common feature.
On MRI studies, the appearance of the cyst compo-
nent of echinococcal cysts is similar to that of other
cysts, with long T1 and T2 relaxation times. However,
MRI best demonstrates the pericyst, matrix and hydatid
sand (debris consisting of freed scolices), and the daugh-
ter cysts. The pericyst usually has low signal intensity on
T1- and T2-weighted images, because of its fibrous
component. This rim and a multiloculated or multicystic
appearance are distinctive features. The hydatid matrix
appears hypointense on T1-weighted images and
markedly hyperintense on T2-weighted images. When
present, daughter cysts are hypointense relative to the
matrix on both T1- and T2-weighted images (Fig. 9).
Fig. 9 a, b. Echinococcus gran-
ulosus cyst in a 28-year-old
man with right upper quad-
rant pain. a Axial T2-weight-
ed image demonstrates a
large cystic mass in the right
lobe of the liver which is sur-
rounded by a hypointense
rim and contains more hyper-
intense smaller cysts in its
periphery. b On the axial T1-
weighted image, the hypo-
intense rim is well visualized
and the peripheral cysts are
hypointense relative to the
center of the lesion
Imaging of Diffuse and Inflammatory Liver Diseases
Floating membranes have low signal intensities on T1-
and T2-weighted images.
E. multiloculars appear as a solid large mass or masses,
with minimal to no enhancement after intravenous admin-
istration of contrast material and possible small punctate
calcification.
Suggested Reading
General
Boll DT, Merkle EM (2009) Diffuse liver disease: Strategies for
hepatic CT and MR imaging. Radiographics 29:1591-1614
Cohen EI, Wilck EJ, Shapiro RS (2006) Hepatic imaging in the
21st century. Semin Liver Dis 26:363-372
Foley WD, Jochem RJ (1991) Computed tomography: Focal and
diffuse liver disease. Radiol Clin North Am 29:1213-1233
Gourtsoyiannis NC, Ros PR (eds) (2005) Radiologic-pathologic
correlations from head to toe: Understanding the manifesta-
tions of disease, Springer-Verlag, Berlin, Heidelberg
Mergo PJ, Ros PR (1998) Imaging of diffuse liver disease. Radiol
Clin North Amer 36:365-375
Mergo PJ, Ros PR, Buetow PC, Buck JL (1994) Diffuse diseases
of the liver: Radiologic-pathologic correlation. Radiographics
14:1291-1307
Mitchell DG (1992) Focal manifestations of diffuse liver disease at
MR imaging. Radiology 185:1-11
Mortele KJ, Ros PR (2001) Imaging of diffuse liver disease. Semin
Liver Dis 21:195-212
Rofsky NM, Fleishaker H (1995) CT and MRI of diffuse liver dis-
ease. Semin Ultrasound, CT, and MRI 16:16-33
Ros PR (1997) MR imaging of the liver. In: Magn Reson Imag Clin
North Amer, Vol 5, WB Saunders, Philadelphia, PA
Ros PR (1998) Hepatic imaging. In: Radiol Clin North Amer, WB
Saunders, Philadelphia, PA
Ros PR (2002) Hepatic imaging and intervention. In: Clinics in
Liver Disease, WB Saunders, Philadelphia, PA
Ros PR, Mortele K (2006) CT/MRI of the Abdomen and Pelvis,
Lippincott, Williams and Wilkins, Philadelphia, PA
Fibrous Tissue Deposition (Cirrhosis)
Aguirre DA, Behling CA, Alpert E et al (2006) Liver fibrosis:
Noninvasive diagnosis with double contrast material-enhanced
MR imaging. Radiology 239:425-437
Awaya H, Mitchell DG, Kamishima T et al (2002) Cirrhosis: Mod-
ified caudate-right lobe ratio. Radiology 224:769-774
Blachar A, Federle MP, Brancatelli G (2001) Primary biliary cir-
rhosis: Clinical, pathologic, and helical CT findings in 53 pa-
tients. Radiology 220:329-336
Bryce TJ, Yeh BJ, Qayyum A et al (2003) CT signs of hepatofugal
venous flow in patients with cirrhosis. AJR Am J Roentgenol
181:1629-1633
Cobbold JF, Wylezinska M, Cunningham C et al (2006) Non-inva-
sive evaluation of hepatic fibrosis using magnetic resonance
imaging and ultrasound techniques. Gut 55:1670-1672
Colli A, Fraquelli M, Andreoletti M et al (2003) Severe liver fi-
brosis or cirrhosis: Accuracy of US for detection: Analysis of
300 cases. Radiology 227:89-94
Day CP (2006) From fat to inflammation. Gastroenterology 130:
207-210
Dodd GD III, Baron RL, Oliver JH III et al (1999) End-stage pri-
mary sclerosing cholangitis: CT findings of hepatic morphol-
ogy in 36 patients. Radiology 211:357-362
Dodd GD, Oliver JH, Federle MP et al (1993) Spectrum of imag-
ing findings in hepatic cirrhosis: Pathologic correlation in 500
complete hepatectomy specimens. Radiology 189:421
61
Faria SC, Ganesan K, Mwangi I et al (2009) MR imaging of
liver fibrosis: Current state of the art. Radiographics 29:
1615-1635
Friedman SL, Rockey DC, Bissell DM (2007) Hepatic fibrosis
2006: Report of the third AASLD single topic conference. He-
patology 45:242-249
Hussain S, Reinhold C, Mitchell DG (2009) Cirrhosis and lesion
characterization at MR imaging. Radiographics 29:1637-
1652
Ito K, Mitchell DG, Gabata T et al (1999) Expanded gallbladder
fossa: Simple MR imaging sign of cirrhosis. Radiology 211:
723-726
Johnson KJ, Olliff JF, Olliff SP (1998) The presence and signifi-
cance of lymphadenopathy detected by CT in primary scleros-
ing cholangitis. Br J Radiol 71:1279-1282
Kang HK, Jeong YY, Choi JH et al (2002) Three-dimensional mul-
ti-detector row CT portal venography in the evaluation of por-
tosystemic collateral vessels in liver cirrhosis. Radiographics
22:1053-1061
Kim M-J, Mitchell DG, Ito K (2000) Portosystemic collaterals of
the upper abdomen: Review of anatomy and demonstration on
MR imaging. Abdom Imaging 25:462-470
Kim YJ, Raman SS, Yu NC et al (2007) Esophageal varices in cir-
rhotic patients: Evaluation with liver CT. AJR Am J Roentgenol
188:139-144
Lee V (2006) Can MR imaging replace liver biopsy for the diag-
nosis of early fibrosis? Radiology 239:309-310
Lipson JA, Qayyum A, Avrin DE et al (2005) CT and MRI of
hepatic contour abnormalities. AJR Am J Roentgenol 184:
75-81
Mortele KJ, Ros PR (2002) MR Imaging in chronic hepatitis and
cirrhosis. Semin in Ultrasound, CT, and MRI 23:79-100
Ohtomo K, Baron RL, Dodd GD et al (1993) Confluent hepatic fi-
brosis in advanced cirrhosis: Appearance at CT. Radiology
188:31-35
Revelon G, Rashid A, Kawamoto S et al (1999) Primary sclerosing
cholangitis MR imaging findings with pathologic correlation.
AJR Am J Roentgenol 173:1037-1042
Rouviere O, Yin M, Dresner MA et al (2006) MR elastography of
the liver: Preliminary results. Radiology 240:440-673
Vascular Disorders
Abu-Yousef MM, Milam SG, Farner RM (1990) Pulsatile portal
vein flow: A sign of tricuspid regurgitation on duplex Doppler
sonography. AJR Am J Roentgenol 155:785-788
Arai K, Matsui O, Takashima T et al (1988) Focal spared areas in
fatty liver caused by regional decreased portal flow. AJR Am J
Roentgenol 151:300-302
Bargallo X, Gilabert R, Nicolau C et al (2003) Sonography of the
caudate vein: Value in diagnosing Budd-Chiari syndrome. AJR
Am J Roentgenol 181:1641-1645
Bradbury MS, Kavanaugh PV, Bechtold RE et al (2002) Mesen-
teric venous thrombosis: Diagnosis and noninvasive imaging.
Radiographics 22:527-541
Brancatelli G, Federle MP, Grazioli L et al (2002) Large regenera-
tive nodules in Budd-Chiari syndrome and other vascular dis-
orders of the liver: CT and MR findings with clinicopatholog-
ic correlation. AJR Am J Roentgenol 178:877-883
Camera L, Mainenti PP, Di Giacomo A et al (2006) Triphasic he-
lical CT in Budd-Chiari syndrome: Patterns of enhancement
in acute, subacute and chronic disease. Clin Radiol 61:
331-337
Cazals-Hatem D, Vilfrain V, Genin P et al (2003) Arterial and
portal circulation and parenchymal changes in Budd-Chiari
syndrome: A study in 17 explanted livers. Hepatology 37:
510-519
Colegrande S, Centi N, Galdiero R et al (2007) Transient hepatic
intensity differences: Part 1. Those associated with focal le-
sions. AJR Am J Roentgenol 188:154-159
62
Colegrande S, Centi N, Galdiero R et al (2007) Transient hepatic
intensity differences: Part 2. Those not associated with focal
lesions. AJR Am J Roentgenol 188:160-166
Erden A (2007) Budd-Chiari syndrome: A review of imaging find-
ings. Eur J Radiol 61:44-56
Erden A, Erden I, Krayalcin S et al (2002) Budd-Chiari syndrome:
Evaluation with multiphase contrast-enhanced three-
dimensional MR angiography. AJR Am J Roentgenol 179:
1287-1292
Gallego C, Velasco M, Marcuello P et al (2002) Congenital and ac-
quired anomalies of the portal venous system. Radiographics
22:141-159
Gore RM, Mathieu DG, White EM et al (1994) Passive hepatic con-
gestion: Cross-sectional imaging features. AJR Am J Roentgenol
162:71-75
Harter LP, Gross BH, St Hilaire J et al (1982) CT and sonographic
appearance of hepatic vein obstruction. AJR Am J Roentgenol
139:176-178
Iannaccone R, Federle MP, Brancatelli G et al (2006) Peliosis he-
patis: Spectrum of imaging findings. AJR Am J Roentgenol
187:W43-W52
Kane R, Eustace S (1995) Diagnosis of Budd-Chiari syndrome:
Comparison between sonography and MR angiography. Radio-
logy 195:117-121
Kim HJ, Kim AY, Kim TK et al (2005) Transient hepatic attenua-
tion differences in focal hepatic lesions: Dynamic CT features.
AJR Am J Roentgenol 184:83-90
Mathieu D, Vasile N, Menu Y et al (1987) Budd-Chiari syndrome:
Dynamic CT. Radiology 165:409-413
Moore EH, Russell LA, Klein JS et al (1995) Bacillary an-
giomatosis in patients with AIDS: multiorgan imaging find-
ings. Radiology 197:67-72
Pliskin M (1975) Peliosis hepatis. Radiology 114:29-30
Rossi S, Rosa L, Ravetta V et al (2006) Contrast-enhanced versus
conventional and color Doppler sonography for the detection
of thrombosis of the portal and hepatic venous systems. AJR
Am J Roentgenol 186:763-773
Sandrasegaran K, Hawes DR, Matthew G (2005) Hepatic peliosis
(bacillary angiomatosis) in AIDS: CT findings. Abdom Imag-
ing 30:738-740
Metabolic and Storage Diseases
Kawamori Y, Matsui O, Takahashi S et al (1996) Focal hepatic fat
infiltration in the posterior edge of the medial segment associ-
ated with aberrant gastric venous drainage: CT, US, and MR
findings. J Comput Assist Tomogr 17:590-595
Ko S, Lee T, Ng S et al (1998) Unusual liver MR findings of Wil-
son’s disease in an asymptomatic 2-year-old girl. Abdom
Imaging 23:56-59
Marmoloya G, Karlins NL, Petrelli M, McCullough A (1990) Un-
usual computed tomography findings in hepatic amyloidosis.
Clin Imaging 14:248
Queiroz-Andrade M, Blasbalg R, Ortega CD et al (2009) MR
imaging findings of iron overload. Radiographics 29:
1575-1589
Siegelman ES, Mitchell DG, Semelka RC (1996) Abdominal iron
deposition: metabolism, MR findings, and clinical importance.
Radiology 199:13-22
Pablo R. Ros, Rendon C. Nelson
Infectious Diseases
Akhan O, Ozmen MN, Dincer A et al (1996) Liver hydatid disease:
long-term results of percutaneous treatment. Radiology
198:259-264
Balci NC, Sirvanci M (2002) MR imaging of infective liver le-
sions. Magn Reson Imaging Clin N Am 10:121-135
Cheung H, Lai YM, Loke TK et al (1996) The imaging diagnosis of
hepatic schistosomiasis japonicum sequelae. Clin Radiol 51:51-
55
Czermak BV, Unsinn KM, Gotwald T et al (2001) Echinococcus
multilocularis revisited. AJR Am J Roentgenol 176:1207-1212
Danon O, Duval-Arnould M, Osman Z et al (2002) Hepatic and
splenic involvement in cat-scratch disease: imaging features.
Abdom Imaging 25:182-183
Doyle DJ, Hanbidge AE, O’Malley ME (2006) Imaging of hepatic
infections. Clin Radiol 61(9):737-748
Giorgio A, Tarantino L, Mariniello N et al (1995) Pyogenic liver
abscesses: 13 years of experience in percutaneous needle aspi-
ration with US guidance. Radiology 195:122-124
Hickey N, McNulty JG, Osborne H, Finucane J (1999) Acute he-
patobiliary tuberculosis: A report of two cases and a review of
the literature. Eur Radiol 9:886-889
Jeffrey RB, Tolentino CS, Chang FC et al (1998) CT of small pyo-
genic hepatic abscesses: the cluster sign. AJR Am J Roentgenol
151:487-489
Kalovidouris A, Gouliamos A, Vlachos L et al (1994) MRI of ab-
dominal hydatid disease. Abdom Imaging 19:489-494
Lee TY, Wan YL, Tsai CC (1994) Gas-containing liver abscess: Ra-
diological findings and clinical significance. Abdom Imaging
19:46-52
Mendez RJ, Schiebler ML, Outwater EK et al (1994) Hepatic ab-
scesses: MR imaging findings. Radiology 190:431-436
Monzawz S, Ohtomo K, Oba H et al (1994) Septa in the liver of
patients with chronic hepatic schistosomiasis japonica: MR
appearance. AJR Am J Roentgenol 162:1347-1351
Mortele KJ, Segatto E, Ros PR (2004) The infected liver: radio-
logic-pathologic correlation. Radiographics 24:937-955
Pastakia B, Shawker TH, Thaler M et al (1988) Hepatosplenic can-
didiasis: wheels within wheels. Radiology 166:417-444
Ralls PW (2002) Inflammatory disease of the liver. Clin Liver Dis
6:203-225
Ralls PW, Henley DS, Colletti PM et al (1987) Amebic liver ab-
scess: MR imaging. Radiology 165:801-804
Rappaport DC, Cumming WA, Ros PR (1991) Disseminated he-
patic and splenic lesions in cat-scratch disease: Imaging fea-
tures. AJR Am J Roentgenol 156:1227-1228
Ros PR, Sobin LH (1994) Amyloidosis: The same cat, with differ-
ent stripes. Radiology 190:14-15
Ros PR, Ertuk SM (2007) Focal hepatic infections. In: Gore RM,
Levine MS (eds) Textbook of Gastrointestinal Radiology, 3rd
edn. WB Saunders, Philadelphia, PA, pp 1663-1684
Sakai T, Maeda M, Takabatake M et al (1995) MR imaging of he-
patosplenic sarcoidosis. Radiat Med 13:39-41
Taourel P, Marty-Ane B, Charasset S et al (1993) Hydatid cyst of
the liver: comparison of CT and MRI. J Comput Assist Tomo-
gr 17:80-85
Van Allen RJ, Katz MD, Johnson MB et al (1992) Uncomplicated
amebic liver abscess: prospective evaluation of percutaneous
therapeutic aspiration. Radiology 183:827-830

Focal Liver Lesions
Wolfgang Schima1, Richard Baron2
1 Department of Radiology, KH Göttlicher Heiland, Wien, Austria
2 Department of Radiology, University of Chicago, Chicago, IL, USA
Introduction
State-of-the-art multidetector computed tomography
(MDCT) and magnetic resonance (MR) imaging tech-
nologies offer detailed insights into the liver’s anatomy
and the pathophysiology of liver disease, such that imag-
ing has become the pacemaker in the development of new
therapeutic techniques. Understanding different imaging
techniques and the diagnostic potential of different
modalities, including contrast utilization, is essential to
optimize patient diagnoses. In the current environment of
cost containment, the most appropriate modality should
be chosen to answer the clinical question. Ultrasonogra-
phy (US) is widely available, non-invasive, and the least
expensive, but it is limited by low sensitivity and speci-
ficity unless US contrast agents are used. Instead, contrast-
enhanced CT has emerged as the modality of choice for
routine liver imaging while MR imaging is used primarily
as a problem-solving technique for liver evaluation when
CT or US results are equivocal or if high concern exists
for malignancy in certain high-risk populations. This
chapter highlights the imaging of hepatic focal liver le-
sions, focusing on MDCT and MR imaging, but includ-
ing ultrasonography in selected cases. It aims to provide
readers with an understanding of how to optimize their
CT and MR imaging protocols and to familiarize them
with the different CT and MR imaging appearances of
focal liver lesions. The value of liver-specific MR contrast
agents for non-invasive characterization of focal lesions
is highlighted as well.
MDCT Imaging Techniques
Organ imaging in multiple planes rather than in single se-
quential slices is the concept underlying MDCT. Thus, in
systems consisting of 64 or more detectors the entire liv-
er can be scanned within 1-4 s using a sub-millimeter de-
tector configuration to achieve high-quality multiplanar
reconstructions [1]. When viewed axially, reconstructed
sections of 2.5-4 mm (overlap 0.5-1 mm) are preferred –
because thinner slices do not improve lesion conspicuity
– but at the expense of increased image noise [2] and
IDKD 2010-2013
substantially decreased specificity [3]. The amount of
contrast material administered may depend on the pa-
tient’s weight, but is typically 42-45 g of iodine (i.e., 120-
150 mL of contrast at 300 mg iodine/mL, equivalent to
110 mL of contrast at 400 mg/mL). The total amount of
iodine given determines the quality of portal-venous-
phase imaging, with a goal of increasing liver attenuation
through contrast enhancement by 50 HU [4]. To achieve
good arterial-phase imaging, the flow rate of the contrast
material is crucial; a rate of 4-5 mL/s is recommended
[5]. Others have used a variable amount of contrast ma-
terial (e.g., 2 mL/kg body weight) and a fixed injection
duration of 30 s, which means that the injection rate
varies according to the patient’s weight.
Timing of image acquisition in relation to contrast ma-
terial administration depends on whether one needs to im-
age during the early arterial phase (for arterial anatomy on-
ly), late arterial phase (for hypervascular tumor detec-
tion/characterization), or portal-venous phase (in some
cases of follow-up imaging and hypovascular tumor detec-
tion). Routinely, late-arterial-phase imaging (with a delay
consisting of the aortic transit time plus 15-18 s) [6, 7] and
a portal-venous scan (20-30 s interscan delay or with a
fixed delay of ~60-70 s) are acquired, although the actual
combination of imaging phases will depend on the indi-
vidual indication [8]. Automated methods of measuring ar-
terial (aortic) enhancement on CT (often termed bolus
tracking), rather than fixed scan times, can be helpful [9].
Magnetic Resonance Imaging Technique
Unenhanced T1- and T2-weighted pulse sequences and
contrast-enhanced sequences are mandatory in MR exam-
ination of the liver. Specific pulse sequences vary by manu-
facturer, and their use depends on patient compliance and
the clinical question being addressed. In- and opposed-
phase (or out-of-phase) T1-weighted imaging is recom-
mended to allow for maximal tumor detection and in the
characterization of fatty tumors and steatosis/non-steatosis.
T2-weighted pulse sequences with fat suppression provide
better lesion contrast than non-fat-suppressed sequences.
Diffusion-weighted imaging, which is helpful for lesion
64
detection, is now available on most scanners. In general,
dynamic imaging with extracellular gadolinium-based
contrast agents is required for lesion characterization, the
detection of tumors in cirrhosis, evaluation of the tumor
response to therapy, and the detection of marginal tumor
recurrences following tumor ablation.
Two types of liver-specific MR contrast agents are
available. Following their intravenous injection, hepato-
cyte-specific agents (mangafodipir: Teslascan, GE Health-
care, Norway; gadobenate: MultiHance, Bracco, Italy; and
gadoxetic acid: Primovist or Eovist, Bayer-Schering,
Germany) are taken up by hepatocytes and provide T1
enhancement of liver tissue. These compounds are used to
improve the detection of metastases and to characterize
lesions [10-13]. The reticuloendothelial agent ferumoxide
(Endorem, Guerbet), a superparamagnetic iron oxide
(SPIO), is phagocytosed by Kupffer cells after intravenous
infusion and leads to a drop in the T2 signal intensity of
liver tissue but not of most parenchymal masses.
Benign Hepatic Lesions
Cyst
Cysts are common, usually asymptomatic liver lesions
with an incidence in the population of 5-14%. At US, he-
patic cysts are anechoic, with an imperceptible wall and
increased acoustic enhancement behind the cyst. On CT
scans, a hepatic cyst appears as a well-circumscribed, ho-
mogeneous mass with an attenuation value similar to that
of water (<15 HU). Cysts lack any mural thickening or
nodularity and do not show contrast enhancement after
the administration of intravenous contrast material. Small
lesions may appear to have higher attenuation measure-
ments because of partial-volume averaging. Occasional-
ly, unenhanced scans will suggest the diagnosis of small
cysts if they are well visualized as hypodense lesions,
whereas very small metastases are usually not discernible
on unenhanced scans. In most instances, however, these
a b
Fig. 1 a-c. Hemangioma type 1. a Unenhanced CT shows a small hypodense lesion adjacent to the falciform ligament (arrow). b Contrast-
enhanced CT in the arterial phase shows rapid and complete enhancement of the hemangioma, which persists in the venous phase (c). The
attenuation of the hemangioma in the enhanced phases is similar to that of the aorta
Wolfgang Schima, Richard Baron
small lesions remain problematic for characterization at
CT. On MR imaging, cysts are well-defined, homoge-
neous lesions that are hypointense on T1-weighted im-
ages and markedly hyperintense on T2-weighted images.
The marked increased in T2 signal intensity of even very
small cysts can be very helpful to confirm the benign na-
ture of small lesions.
Hemangioma
Hemangioma is the most common benign liver tumor. Its
typical US characteristics include a sharply circum-
scribed, well-defined hyperechoic lesion with distal
acoustic enhancement. Small hemangiomas are usually
homogeneous in appearance whereas larger ones (>4 cm)
are frequently heterogeneous and do not demonstrate all
of the characteristic features of these lesions.
On unenhanced CT images, hemangiomas are well-
defined hypodense masses. On MR imaging, they are
hypointense on T1-weighted sequences and markedly
hyperintense on T2-weighted sequences. In addition, MR
imaging is useful in differentiating hemangiomas from
solid neoplasms based on the long T2 relaxation time
(= hyperintensity) of hemangioma compared with other
hepatic masses [14, 15]. A relatively long T2 echo time
(>140 ms) will demonstrate the presence of a homoge-
neously “light-bulb-bright” lesion, which is characteris-
tic of a benign lesion, either cyst or hemangioma. The
exceptions to this would include cystic metastases, and
gastrointestinal stromal tumor (GIST) and neuro-
endocrine tumor metastases.
Hemangiomas show three distinctive patterns of en-
hancement at CT/MR imaging [16], with the common
characteristic feature that areas of lesion enhancement
closely follow the enhancement characteristics of blood
pooling elsewhere [17]. Small lesions (up to ~2 cm) may
show immediate and complete filling in the arterial phase,
with sustained enhancement in the venous and delayed
phases (type I, also termed flash filling) [18] (Fig. 1). The
most common enhancement pattern is one of peripheral
c
Focal Liver Lesions 65

a b c

Fig. 2 a-d. Hemangioma type III. a T2-weighted turbo spin-echo (TSE) MR image shows a
very hyperintense lesion in the right lobe. b-d Dynamic gadolinium-enhanced T1-weighted
gradient-recalled echo (GRE) images: b arterial, c venous, and d equilibrium phases show
peripheral nodular enhancement with progressive centripetal fill-in. d In the equilibrium-
phase, after 5 min, there is no complete fill-in

nodular discontinuous enhancement that progresses with
increased fill-in over time (type II). Larger lesions (>5
cm) or lesions with central thrombosis/fibrosis may lack
central fill-in (type III) (Fig. 2). With SPIO agents, the
blood pooling effect, with accumulation of contrast mate-
rial in the sinuses, may lead to the prolonged enhancement
of hemangiomas on T1 images and signal intensity loss on
T2 images, despite the lack of Kupffer cells in these le-
sions. This imaging feature helps in the differentiation be-
tween hemangiomas and metastases [19].
Recent studies have shown that non-contrast diffusion-
weighted imaging may help to differentiate between he-
mangioma and solid lesions, as the apparent diffusion co-
efficient of hemangiomas is higher than that of solid le-
sions [20].
Focal Nodular Hyperplasia
This benign lesion is usually of no clinical consequence
other than the confusion it causes when incidentally de-
tected during abdominal imaging examinations. The
sonographic appearance of focal nodular hyperplasia
(FNH) is non-specific; the lesion may be isoechoic, or
slightly hypoechoic [21] to liver, while in patients with
diffuse hepatic steatosis it is always hypoechoic. One
characteristic feature is the presence of a central scar,
seen in approximately two-thirds of large lesions but in
only one-third of small lesions (<3 cm) [22]. The central
scar is most often hyperintense on T2-weighted images,
with a comma-shaped or spoke-wheel appearance. This is
a key differentiating feature from fibrolamellar hepato-
cellular carcinoma, in which the central scar is predomi-
nately of low signal intensity on T2-weighted MR se-
quences. The use of color/power Doppler US may
demonstrate blood vessels within the scar [23].
On unenhanced CT, FNH are isodense or minimally
hypodense and are sometimes detectable only by the
mass effect on adjacent vessels. On unenhanced MR im-
ages, FNH often has a signal intensity similar to that of
hepatic parenchyma but usually slightly different on ei-
ther T1- or T2-weighted images (Fig. 3). Due to the
prominent arterial vascular supply, FNH undergoes
marked homogeneous enhancement during the arterial
phase of contrast-enhanced CT/MR imaging, with rapid
wash-out of contrast to isodensity/isointensity on venous-
phase images [22]. The central scar often enhances on de-
layed scans [21], a feature typical of a fibrous compo-
nent. One key aspect is that, other than the scar, these le-
sions tend to be very homogeneous in appearance.
With liver-specific MR contrast agents, FNH shows
enhancement on delayed images after the administration
of hepatobiliary contrast agents (such as mangafodipir
(Fig. 3) and signal loss after the administration of reticulo-
endothelial agents [24]. This difference is particularly
helpful for the differentiation between hypervascular
metastases (which do not accumulate liver-specific
agents) or hepatic adenoma and incidentally encountered
FNH [18]. The influence of oral contraceptives on the
growth of FNH is still discussed controversially. Studies
with serial imaging have shown FNH growth during
follow-up to be rare (3-11%) [25, 26] and not stimulated
by oral contraceptives [26].
66 Wolfgang Schima, Richard Baron

a b c

Fig. 3 a-d. Focal nodular hyperplasia (FNH). a The lesion (arrow) is isointense on T1-weight-
ed imaging, with a small central scar. b On T2, the lesion is also isointense (arrow); the cen-
tral T2-weighted bright scar is better discernible. Mangafodipir-enhanced T1-weighted im-
ages in the axial (c) and coronal (d) planes show homogeneous uptake of the liver-specific
agent, typical for FNH (arrow). The central spoke-wheel scar is nicely depicted

Hepatocellular Adenoma trast-enhanced CT or MR, adenomas usually show

This is an uncommon benign neoplasm of low malignant
potential [26]. Unlike FNH, the relationship between oral
contraceptives and the development of hepatocellular
adenoma (HCA) is well established. Histologically, HCA
is composed of cells resembling normal hepatocytes but
lacking bile ducts. This is a key feature in histologically
differentiating between adenoma and FNH [26]. Adeno-
mas are typically hypervascular and, when large, often
heterogeneous due to the presence of fat, necrosis, or he-
morrhage [26, 27] (Fig. 4). These tumors often contain
intratumoral fat, and, as T1-weighted chemical-shift MR
imaging is sensitive for detecting fat, this technique can
be very helpful in characterizing adenomas. The second
typical feature of adenoma is its propensity for sponta-
neous hemorrhage. If no tumor capsule is present, such
hemorrhage may lead to spontaneous rupture, with hy-
potension and even death.
The CT and MR imaging appearances of HCA may be
non-specific. On unenhanced CT images, the lesion may
be hypodense due to the presence of fat or necrosis. On
T1- and T2- weighted MR images, HCA are non-specif-
ic in appearance, being usually mildly hypointense on T1
and isointense or mildly hyperintense on T2. If intracel-
lular lipid is present within the lesion, it will often appear
hyperintense on T1 in-phase images, with a drop in sig-
nal intensity on opposed-phase images [26]. The presence
of intratumoral fat helps to narrow the differential diag-
nosis, as hemangioma can be excluded and metastases
and FNH only very rarely contain fat. On dynamic con-
marked arterial-phase enhancement with a rapid transi-
tion to either isoattenuating or hypoattenuating/intense to
hepatic parenchyma on portal-venous-phase imaging.
Hepatic adenomas may show signal loss after the admin-
istration of reticuloendothelial agents and delayed en-
hancement after hepatobiliary MR contrast agents. How-
ever, they will not show sustained enhancement on de-
layed imaging to degrees greater than normal liver
parenchyma after gadobenate administration, which helps
to differentiate FNH from adenoma [28].
Biliary Hamartomas (von Meyenburg Complex)
Bile duct hamartomas are congenital malformations of
the ductal plate but they have no connections with the
bile ducts. These lesions are of no clinical significance
but are incidentally encountered in patients undergoing
abdominal imaging examinations. They appear as small
cystic lesions of round, oval, or irregular shape, found ei-
ther in the periportal region or diffusely spread through-
out the liver (Fig. 5). They are usually <10 mm in diam-
eter and lack contrast enhancement, but occasional pe-
ripheral rim enhancement may simulate small hypo-
vascular metastases [27].
Hepatic Abscess and Hydatid Cyst
Abscess appearances can vary depending on etiology.
Peribiliary abscesses tend to be small and scattered adja-
cent to the biliary tree; hematogenous distribution via the
Focal Liver Lesions 67

a b c

Fig. 4 a-d. Hepatocellular adenoma. a T1-weighted in-phase GRE image demonstrates a very
large mass in a young woman. The mass is inhomogeneous and shows bright spots, sug-
gestive of hemorrhage (asterisk). b There is a typical drop in signal intensity on the op-
posed-phase image, indicative of intratumoral fat (arrows), whereas the hemorrhage
(methemoglobin, asterisk) does not lose signal intensity. c T2-weighted TSE sequence con-
firms the presence of hemorrhage (asterisk). Intratumoral fat and hemorrhage are typical
for adenoma. d The gadolinium-enhanced image shows moderate and inhomogeneous en-
hancement. In a large, very inhomogeneous adenoma, malignant degeneration cannot be
ruled out radiologically. However, at surgery this lesion was shown to be an adenoma with
central hemorrhage

a b

Fig. 5 a, b. Biliary hamartomas (von

Meyenburg complex). Coronal (a)
and axial (b) T2-weighted images
show multiple small cystic lesions
of different sizes and shapes (“star-
ry sky” appearance), typical for bil-
iary hamartomas

hepatic artery or portal vein in appendicitis or divertic-
ulitis tends to lead to larger lesions. US reveals a cystic
lesion with internal echoes. On CT, hepatic abscess ap-
pears as a hypodense lesion with a capsule that may show
enhancement. The cluster sign may be noted when multi-
ple abscesses are present as focal clusters of lesions [29].
The CT appearance of hepatic abscess is non-specific
and can be mimicked by cystic or necrotic metastases.
Thus, clinical information and laboratory values play a
key role in guiding the radiological diagnosis. Although
seen in only a small minority of patients, the presence of
central gas is highly specific for abscess (Fig. 6). On T1-
weighted MR images, hepatic abscesses are hypointense
relative to liver parenchyma whereas on T2-weighted se-
quences they are markedly hyperintense and often sur-
rounded by a local area of slight T2 hyperintensity, rep-
resenting perilesional edema (Fig. 6).
Amoebic liver abscess has a non-specific appearance
on CT, but is usually seen as a solitary, hypodense lesion
with an enhancing wall that may be smooth or nodular,
often associated with an incomplete rim of edema. With
MR imaging, the lesions are hypointense on T1-weighted
images and heterogeneously hyperintense on T2-weighted
images [30].
On CT scan, involvement of the liver by Echinococcus
granulosus (hydatid cyst) manifests as unilocular or
68
a b
Fig. 6 a-c. a Typical large subcapsular abscess with an air-fluid level and a pleural empyema. b In another patient, CT shows a small, thick-
walled abscess after pancreatic surgery. c T2-weighted image of the same patient as in (b) shows the thick indistinct wall of the abscess
and peripheral edema
multilocular cysts, with thin or thick walls and calcifica-
tions and usually accompanied by daughter cysts. The lat-
ter are seen as smaller cysts with septations at the mar-
gins of or inside the mother cyst. This appearance is
therefore quite different from that of a “usual” multicys-
tic tumor. On MR imaging, a hypointense rim on T1- and
T2-weighted images and a multiloculated appearance are
considered to be important diagnostic features.
Malignant Primary Tumors
Hepatocellular Carcinoma
Hepatocellular carcinoma (HCC) is the most common
primary liver cancer worldwide and is particularly
prominent in Asian and Mediterranean populations. In
European countries, HCC occurs mostly in patients with
chronic liver disease (hepatitis B or C, liver cirrhosis, or
hemochromatosis). These tumors consist of abnormal
hepatocytes arranged in a typical trabecular, sinusoidal
pattern. They may be solitary, multifocal, or diffusely
infiltrating.
The imaging appearances of HCC can vary dramati-
cally, but generally can be separated into those based on
early versus late presentation. Early presentation is typi-
cal of patients with chronic liver disease who undergo
a b
Wolfgang Schima, Richard Baron
c
imaging-based screening for HCC or frequent liver
imaging due to complications of their chronic disease. In
these patients, HCC lesions are typically small (<3 cm)
and homogeneous in appearance. In non-cirrhotic pa-
tients, the disease is usually asymptomatic; thus, by the
time symptoms occur and imaging is necessarily per-
formed, the lesion is very large and usually heteroge-
neous in appearance.
The US presentation of HCC is quite variable, with
isointensity, hypointensity, or hyperintensity. Smaller le-
sions are typically homogeneous and larger ones hetero-
geneous. A surrounding fibrous capsule is often present
and relatively characteristic for HCC, appearing as a hy-
poechoic rim surrounding the lesion.
On unenhanced CT images, most HCCs are hypo-
dense. The presence of intratumoral fat can result in the
lowered CT attenuation of these tumors; this finding is
characteristic of primary hepatocellular tumors. Due to
their predominant arterial supply, small HCCs enhance
vividly in the arterial phase of hepatic contrast en-
hancement, becoming isoattenuating or hypoattenuat-
ing with hepatic parenchyma in the portal-venous
phase of enhancement (so-called wash-out). On de-
layed images, most HCC lesions are hypodense with
surrounding liver (Fig. 7).
There have been several studies addressing the re-
quired phases of scanning for optimal HCC detection
Fig. 7 a, b. Detection of hepatocellular carcino-
ma with late-arterial-phase multidetector CT.
a In the late-arterial phase, MDCT shows a
small HCC that was not visible on unen-
hanced scan. b The delayed-phase scan
reveals wash-out of the lesion, which is now
slightly hypodense. The combination of
arterial hypervascularity and wash-out is a
very specific sign of malignancy
Focal Liver Lesions
a b
Fig. 8 a-c. Large HCC with tumor thrombus in the inferior vena cava. a Arterial-phase MDCT shows a large HCC with peripheral irregular
rim enhancement. The tumor has grown into the right hepatic vein and the inferior vena cava (IVC). b Arterial- and c venous-phase images
in the coronal plane better demonstrate the extension of the tumor thrombus into the IVC above the diaphragm
and characterization. Arterial-phase imaging is the most
sensitive for the detection of small lesions; as highest
visibility is achieved in the late arterial phase, to allow
time for contrast diffusion into the tumor parenchyma,
there is no need for early-arterial-phase imaging [6, 31,
32]. A venous phase is always necessary for tumor de-
tection and the assessment of venous structures (Fig. 8)
as well as other abdominal organs. For HCC detection,
the delayed phase can visualize a few lesions that would
otherwise go unnoticed [33] and also is very helpful in
differentiating HCC from benign enhancing lesions by
demonstrating tumor wash-out greater than liver
parenchyma [34]. Unenhanced images are important to
document siderotic nodules as different from arterial
enhancing lesions and to detect intratumoral fat. How-
ever, to reduce the radiation dose to the patient, these
images should be obtained only intermittently during
serial imaging examinations. Nonetheless, they are
mandatory in the follow-up after chemoembolization or
tumor ablation and when hemorrhage is suspected. In
summary, a three- to four-phase MDCT protocol is rec-
ommended by most centers to optimally detect and
characterize HCC.
The presence of focal hypervascularity in the arterial
phase may lead to false-positive results [35]: transient fo-
cal enhancement of liver parenchyma during arterial
phase enhancement, often termed transient hepatic atten-
uation differences (THAD), can be caused by a multitude
of factors. In cirrhotic patients, transient focal enhance-
ment is most often due to arterial-portal shunting, result-
ing in inappropriately early focal areas of portal-venous
distribution enhancement in the liver. These usually are
peripheral, often wedge shaped, and not round. Sub-
capsular lesions that do not show a substantial mass
effect or round nature should be evaluated carefully
before a diagnosis of HCC is concluded. The combi-
nation of hyperdensity on arterial-phase images com-
bined with wash-out to hypodensity on venous-phase or
delayed-phase images, although not sensitive (33%), is a
very specific (100%) feature for the presence of HCC
[36] (Fig. 7). Diffusely infiltrating and small HCCs in
69
c
cirrhosis may be difficult to detect with CT. Larger HCC
lesions typically have a different appearance and are
visualized as a mosaic, due to hemorrhage and fibrosis.
Also, about 10% of small HCCs can appear hypodense to
liver; these are generally thought to be well-differentiated
lesions.
Typical MR imaging findings of larger HCC consist
of a fibrous capsule, central scar, intratumoral septa,
daughter nodules, and tumor thrombus [37]. In addition,
there is often a somewhat organized internal, mosaic-
like pattern that is seen on CT as well as on MR imag-
ing [38]. While most large HCCs are hyperintense on
T2-weighted sequences, small lesions (<2 cm) are often
isointense but may also be hypointense. On T1-weighted
sequences, HCC has variable signal intensity relative to
hepatic parenchyma. A tumor capsule may be seen on
T1-weighted and, less commonly, on T2-weighted im-
ages as hypointense (Fig. 9). Conventional gadolinium
contrast imaging in HCC parallels that described for CT,
with characteristic early peak contrast enhancement and
delayed-phase tumor wash-out. These enhancement fea-
tures are useful in differentiating HCC from heman-
gioma, which generally shows early peripheral en-
hancement, marked peak enhancement >2 min after
contrast injection, and marked pooling of contrast on
delayed images. Dynamic gadolinium-enhanced MR
imaging has been found to be superior to MR with liver-
specific contrast agent in terms of HCC detection
[39, 40], because hypervascularity is the key feature
marking the transition of dysplastic nodules into early
HCC [41]. However, double-contrast MR imaging with
sequential administration of gadolinium and reticulo-
endothelial contrast agents has been found to be the
most sensitive and specific method to detect HCC and
to differentiate between early HCC and dysplastic
nodules [42, 43]. HCC may show enhancement on
delayed images after the administration of hepatobiliary
MR contrast agent (Fig. 9). However, such enhancement
is not specific for HCC and can be seen with other
primary hepatocellular tumors, such as dysplastic
nodules, FNH, and adenoma.
70 Wolfgang Schima, Richard Baron

a b c

Fig. 9 a-d. Hepatocellular carcinoma. MR imaging with mangafodipir and gadolinium. a Ax-
ial T1-weighted GRE shows an isointense mass with a pseudocapsule and a small hyper-
intense hemorrhage. b The lesion is only minimally hyperintense on T2-weighted images.
c The mangafodipir-enhanced T1-weighted GRE image shows enhancement of this well-
differentiated HCC. The enhancement does not help in the differential diagnosis of hepato-
cellular lesions. d The gadolinium-enhanced sequence in the arterial phase shows typical
hypervascularity. The key features guiding the correct diagnosis in this hypervascular lesion
are the presence of a pseudocapsule (very rare in FNH or adenoma) and cirrhosis (see
enlarged left and caudate lobes)

Fibrolamellar HCC retention of conventional CT and MR contrast agents.

Fibrolamellar HCC (FL-HCC) is a less aggressive tu-
mor with a better prognosis than HCC. It consists of
malignant hepatocytes separated into cords by fibrous
strands. On CT, FL-HCC appears as a large, well-
defined vascular mass with a lobulated surface and
often a central scar; calcifications are identified in up to
70% of patients [44, 45]. On MR imaging, FL-HCC is
hypointense on T1-weighted images and hyperintense
on T2-weighted images, with the central scar being
hypointense on both sequences (Fig. 10). This is in con-
trast to the scar in FNH, which is most often hyper-
intense on T2-weighted images. The fibrous central
zones of both FNH and FL-HCC will show delayed
Contrast enhancement patterns in FL-HCC are almost
always heterogeneous whereas in FNH they are almost
always homogeneous.
Cholangiocellular Carcinoma
Cholangiocellular carcinoma (CCC) is the second most
common primary malignancy of the liver. Intrahepatic
CCC originates from the intralobular bile ducts, unlike
hilar CCC, which arises from a main hepatic duct or from
the bifurcation. Intrahepatic CCC presents as a large
mass, because the tumor does not cause symptoms in the
early stages [46] (Fig. 11). According to its growth char-
acteristics, CCC is classified as mass-forming, periductal-

a b

Fig. 10 a, b. Fibrolamellar hepatocellular carcinoma. a CT during arterial phase shows a typical heterogeneously enhancing mass in the left
lobe (arrows), with a low-attenuation central fibrous scar containing calcifications (arrowheads). b MR T2-weighted shows a large mass
in the left lobe (arrows) with a heterogeneous appearance and mild to moderately increased signal intensity. The fibrous central scar is of
very low signal intensity (arrowheads)
Focal Liver Lesions
a b
Fig. 11 a, b. Cholangiocellu-
lar carcinoma. Contrast-en-
hanced CT in the arterial
(a) and venous (b) phases
demonstrates a large hypo-
vascular mass with some
calcifications. Capsular re-
traction is quite often seen
in peripheral CCC
infiltrative, or intraductal-growing, with the mass-form-
ing type being most common in intrahepatic CCC [46].
At CT and MR imaging, the lesions tend to be hypodense
on unenhanced CT and hypointense on T1-weighted im-
ages, with peripheral enhancement at dynamic contrast-
enhanced studies [47]. Delayed-phase CT/MR imaging
(after 5-15 min) may show enhancement homogeneously
or in the center of the lesion due to the rich fibrous stro-
ma, which is suggestive of the diagnosis of CCC [48].
Periductal-infiltrative CCC causes early segmental dilata-
tion of the bile ducts at a stage when the tumor itself may
be difficult to discern [47].
Rare Primary Liver Tumors
Biliary cystadenoma/cystadenocarcinomas
These tumors have a similar appearance and morphology
as their mucinous counterparts in the pancreas and are
seen predominantly in women. Even when benign, they
have a propensity for malignant degeneration, and any
such tumor should be considered malignant. These
unilocular or multilocular cystic masses have a typical
anechoic and hypoechoic US appearance; on CT, their
contents have near-water attenuation, with peripheral
soft-tissue nodularity and traversing septations (Fig. 12).
Fig. 12. Biliary cystadenoma. T2-weighted MR shows a mass of
very high signal intensity (arrow) that is mostly homogeneous in
appearance, with the exception of a few thin internal septations
(arrowheads)
71
The greater the presence of papillary excrescences, soft-
tissue nodularity, or septations, the more likely it is that
the lesion is malignant [49]. However, this is a moot point
as it has been shown that benign lesions can undergo ma-
lignant degeneration. The cystic areas at T1-weighted
imaging are of variable signal intensity, including hyper-
intense to liver, presumably due to proteinaceous content.
Coarse calcifications can be seen at US and CT in both
cystadenoma and cystadenocarcinoma and therefore are
not a helpful differentiating feature.
Hepatic Angiosarcoma
This rare tumor has a strong association with carcino-
gens such as vinyl chloride and Thorotrast and is also
seen in patients with hemochromatosis. However, the
majority of hepatic angiosarcoma patients have no
known exposure to toxic agents. Pathologically, angio-
sarcoma may appear as a large solitary mass or with
multiple tumor nodules of varying size. In addition, the
tumors may contain vascular channels that create sinus-
oidal spaces; these can result in imaging findings in
some ways simulating those of hemangiomas. The imag-
ing appearances of angiosarcoma are most often non-
specific, with hypoattenuation on unenhanced CT, hypo-
intensity on T1 MR, and mild hyperintensity on T2
(although if prominent sinusoidal vascular spaces are
present pathologically, these can be of homogeneous
very high T2 signal). Following iodinated or gadolinium-
based contrast administration, most lesions show non-
specific heterogeneous enhancement. Potentially prob-
lematic, however, are those tumors with prominent sinu-
soidal vascular spaces, in that, albeit rarely, the CT or
MR contrast enhancement characteristics of some an-
giosarcomas can simulate those of benign hemangioma.
The high MR T2 signal in such lesions further com-
pounds this problem. In most such cases, however, care-
ful observation will reveal that tumoral enhancement
does not follow the characteristics of blood pooling at all
phases or that there are other features, such as innumer-
able lesions, that make the diagnosis of hemangioma
unlikely [50, 51].
72
Epithelioid Hemangioendothelioma
An EHE is a rare tumor of vascular origin and is not to
be confused with infantile hemangioendothelioma,
which is a very different tumor. EHE is a primary liver
tumor characterized by the presence of multiple,
peripheral-based lesions that progressively become con-
fluent masses (Fig. 13). In addition to its unusual
peripheral liver distribution, a key and characteristic
feature of EHE is the presence of overlying capsular
retraction, attributed to fibrosis and scarring within the
tumor [52]. The CT attenuation or MR signal intensity
characteristics are non-specific and mimic those of
other tumors, although tumoral calcifications may occa-
sionally be seen. Contrast enhancement with CT or MR
gadolinium chelates often shows a central zone of de-
creased enhancement with marked enhancement periph-
erally and, in some cases, concentric zones of marked
enhancement (Fig. 13). The lesions often become con-
fluent and may grow large enough to replace nearly the
entire liver parenchyma.
Fig. 13. Epithelioid hemangioendothelioma. Contrast CT (portal-
venous phase) shows multiple, predominantly peripherally based
hypodense lesions, some of which have a laminated appearance
(arrows). Early development of capsular retraction is present, with
flattening of the capsule overlying some of the lesions (arrowheads)
a b
Wolfgang Schima, Richard Baron
Hepatic Lymphoma
This tumor is most often seen in patients with widespread
non-Hodgkin’s lymphoma or, rarely, in those with
Hodgkin’s disease. Although unusual, hepatic lymphoma
constitutes a primary liver tumor; it is usually associated
with an immunocompromised state, such as in AIDS or
post-transplantation with immunosuppression therapy.
The imaging appearances of these lesions are variable,
without any unique characteristics. CT, MR, or US imag-
ing shows focal lesions with an appearance similar to that
of many other neoplastic lesions. Diffuse infiltrating
forms can be difficult to detect regardless of the imaging
modality, although hepatomegaly may be present.
Hepatic Metastases
On US, metastases may appear hypoechoic, isoechoic, or
hyperechoic. Dynamic contrast-enhanced CT visualizes
most metastases as hypovascular and hypodense relative
to liver parenchyma on portal-venous phase (Fig. 14).
Hypervascular metastases are most commonly seen in pa-
tients with renal cell or neuroendocrine tumors, sarcomas,
or breast tumors (Fig. 14). These tumors are best visual-
ized in the arterial phase and may become isodense and
difficult to detect during the redistribution phase of en-
hancement. At MR imaging, metastases are usually hy-
pointense on T1-weighted images and hyperintense on T2-
weighted images [53] (Fig. 15). Peritumoral edema makes
the lesions appear larger on the T2, a finding that is very
suggestive of a malignant mass [54]. High signal intensi-
ty on T1-weighted sequences is typical for melanoma
metastases due to the paramagnetic nature of melanin.
Some lesions may have a central area of hyperintensity
(target sign) on T2-weighted images, which corresponds
to central necrosis. On dynamic contrast-enhanced MR
imaging, metastases demonstrate enhancement character-
istics similar to those described for these tumors on CT.
Metastases may demonstrate a hypointense rim compared
with the center of the lesion on delayed images (peripher-
al wash-out sign), which is very specific for malignancy.
Fig. 14 a, b. a Contrast-
enhanced MDCT in the
arterial phase demonstrates
several predominantly hyper-
vascular liver metastases
(arrows) of a neuro-
endocrine cancer of the pan-
creas. b Contrast-enhanced
MDCT in the venous phase
shows typical hypovascular
colorectal metastases
Focal Liver Lesions
a b
Fig. 15 a-c. Colorectal liver metastases at gadoxetate-enhanced MR imaging. a Unenhanced T1-weighted MR imaging shows two hypointense
lesions in segments 6/7 and 4. b The T2-weighted TSE pulse sequence shows the lesions to be moderately hyperintense. c The gadoxetate-
enhanced T1-weighted GRE image in the hepatobiliary phase shows two additional small subcapsular metastases (arrows) that were not
seen on unenhanced MR imaging or MDCT
The most recent studies have shown MR imaging to be
more sensitive than contrast-enhanced CT for the detec-
tion of hepatic metastases [10, 12, 13], and especially for
the detection of small lesions (Fig. 15).
Differential Diagnosis of Focal Liver Lesions
The approach to characterizing a focal liver lesion seen
on CT begins with a determination of its density. If the
lesion is of near-water density, homogeneous in charac-
ter, and has sharp margins, then a cyst should be consid-
ered and can be confirmed with an equilibrium-phase CT,
or US or even MR imaging (T2 bright and non-enhanc-
ing post-gadolinium).
If the lesion has some enhancement, then the aim of
the next analysis is to determine whether the enhance-
ment is peripheral and nodular, with the density of the en-
hancing portions of the lesions following the same gen-
eral levels of blood vessels in the arterial, venous, and de-
layed phases. In this case, a hemangioma may be diag-
nosed with high certainty.
Arterially enhancing lesions include FNH, HCA,
HCC, and metastases from neuroendocrine tumors,
melanoma, renal cell carcinoma, and breast cancer. In
general, HCC is considered in the setting of cirrhosis.
FNH is most likely in a young woman with a non-cir-
rhotic liver, if the lesion is homogeneous and near-iso-
dense/isointense on unenhanced CT/MR imaging, with a
central T2-weighted hyperintense scar. Thick irregular,
heterogeneous enhancement or the presence of peripher-
al wash-out at delayed phase suggests a malignant mass,
such as metastases, CCC, or even HCC.
References
1. Laghi A (2007) Multidetector CT (64 Slices) of the liver: ex-
amination techniques. Eur Radiol 17:675-683
2. Weg N, Scheer MR, Gabor MP (1998) Liver lesions: improved
detection with dual-detector-array CT and routine 2.5-mm thin
collimation. Radiology 209:417-426
73
c
3. Ichikawa T, Nakajima H, Nanbu A et al (2006) Effect of in-
jection rate of contrast material on CT of hepatocellular carci-
noma. AJR Am J Roentgenol 186:1413-1418
4. Foley WD, Hoffmann RG, Quiroz FA et al (1994) Hepatic he-
lical CT: contrast material injection protocol. Radiology
192:367-371
5. Kim T, Murakami T, Takahashi S et al (1998) Effects of injec-
tion rates of contrast material on arterial phase hepatic CT.
AJR Am J Roentgenol 171:429-432
6. Schima W, Hammerstingl R, Catalano C et al (2006) Quadru-
ple-phase MDCT of the liver in patients with suspected hepa-
tocellular carcinoma: effect of contrast material flow rate. AJR
Am J Roentgenol 186:1571-1579
7. Sultana S, Awai K, Nakayama Y et al (2007) Hypervascular he-
patocellular carcinomas: bolus tracking with a 40-detector CT
scanner to time arterial phase imaging. Radiology 243:140-147
8. Oliver JH, Baron RL (1996) Helical biphasic contrast-en-
hanced CT of the liver: technique, indications, interpretations,
and pitfalls. Radiology 201:1-14
9. Mehnert F, Pereira PL, Trubenbach J et al (2001) Biphasic spi-
ral CT of the liver: automatic bolus tracking or time delay? Eur
Radiol 11:427-431
10. Oudkerk M, Torres CG, Song B et al (2002) Characterization
of liver lesions with mangafodipir trisodium-enhanced MR
imaging: multicenter study comparing MR and dual-phase spi-
ral CT. Radiology 223:517-524
11. Scharitzer M, Schima W, Schober E et al (2005) Characterization
of hepatocellular tumors: value of mangafodipirenhanced mag-
netic resonance imaging. J Comput Assist Tomogr 29:181-190
12. Ward J, Robinson PJ, Guthrie JA et al (2005) Liver metastases
in candidates for hepatic resection: comparison of helical CT
and gadolinium- and SPIO-enhanced imaging. Radiology
237:170-180
13. Hammerstingl R, Huppertz A, Breuer J et al (2008) Diagnos-
tic efficacy of gadoxetic acid (Primovist)-enhanced MRI and
spiral CT for a therapeutic strategy: comparison with intra-
operative and histopathologic findings in focal liver lesions.
Eur Radiol 18:457-467
14. Schima W, Saini S, Echeverri JA et al (1997) T2-weighted MR
imaging for characterization of focal liver lesions: convention-
al spin-echo vs fast spin-echo. Radiology 202:389-393
15. Farraher SW, Jara H, Chang KJ et al (2006) Differentiation of
hepatocellular carcinoma and hepatic metastasis from cysts and
hemangiomas with calculated T2 relaxation times and the T1/T2
relaxation times ratio. J Magn Reson Imaging 24:1333-1341
16. Semelka RC, Brown ED, Ascher SM et al (1994) Hepatic he-
mangiomas: a multi-institutional study of appearance on T2-
weighted and serial gadolinium-enhanced gradient-echo MR
images. Radiology 192:401-406
74
17. Kim T, Federle MP, Baron RL et al (2001) Discrimination of
small hepatic hemangiomas from hypervascular malignant tu-
mors smaller than 3 cm with three-phase helical CT. Radiolo-
gy 219:699-706
18. Ba-Ssalamah A, Uffmann M, Saini S et al (2009) Clinical val-
ue of MRI liver-specific contrast agents: a tailored examina-
tion for a confident non-invasive diagnosis of focal liver le-
sions. Eur Radiol 19:342-357
19. Grangier C, Tourniaire J, Mentha G et al (1994) Enhancement
of liver hemangiomas on T1-weighted MR SE images by su-
perparamagnetic iron oxide particles. J Comput Assist Tomo-
gr 18:888-896
20. Vossen JA, Buijs M, Liapi E et al (2008) Receiver operating
characteristic analysis of diffusion-weighted magnetic resonance
imaging in differentiating hepatic hemangioma from other hyper-
vascular liver lesions. J Comput Assist Tomogr 32:750-756
21. Kehagias D, Moulopoulos L, Antoniou A et al (2001) Focal
nodular hyperplasia: imaging findings. Eur Radiol 11:202-212
22. Brancatelli G, Federle MP, Grazioli L et al (2001) Focal nodu-
lar hyperplasia: CT findings with emphasis on multiphasic he-
lical CT in 78 patients. Radiology 219:61-68
23. Uggowitzer MM, Kugler C, Mischinger HJ et al (1999) Echo-
enhanced Doppler sonography of focal nodular hyperplasia of
the liver. J Ultrasound Med 18:445-451; quiz 453-444
24. Ba-Ssalamah A, Schima W, Schmook MT et al (2002) Atypi-
cal focal nodular hyperplasia of the liver: imaging features of
nonspecific and liver-specific MR contrast agents. AJR Am J
Roentgenol 179:1447-1456
25. Leconte I, Van Beers BE, Lacrosse M et al (2000) Focal nodu-
lar hyperplasia: natural course observed with CT and MRI. J
Comput Assist Tomogr 24:61-66
26. Mathieu D, Kobeiter H, Maison P et al (2000) Oral contracep-
tive use and focal nodular hyperplasia of the liver. Gastroen-
terology 118:560-564
27. Prasad SR, Sahani DV, Mino-Kenudson M et al (2008) Benign
hepatic neoplasms: an update on cross-sectional imaging spec-
trum. J Computer Assist Tomogr 32:829-840
28. Grazioli L, Morana G, Kirchin MA, Schneider G (2005) Ac-
curate differentiation of focal nodular hyperplasia from hepat-
ic adenoma at gadobenate dimeglumine-enhanced MR imag-
ing: prospective study. Radiology 236:166-177
29. Jeffrey RB, Jr, Tolentino CS, Chang FC, Federle MP (1988)
CT of small pyogenic hepatic abscesses: the cluster sign. AJR
Am J Roentgenol 151:487-489
30. Barreda R, Ros PR (1992) Diagnostic imaging of liver ab-
scess. Crit Rev Diagn Imaging 33:29-58
31. Laghi A, Iannaccone R, Rossi P et al (2003) Hepatocellular car-
cinoma: detection with triple-phase multi-detector row helical
CT in patients with chronic hepatitis. Radiology 226:543-549
32. Ichikawa T, Kitamura T, Nakajima H et al (2002) Hypervascu-
lar hepatocellular carcinoma: can double arterial phase imag-
ing with multidetector CT improve tumor depiction in the cir-
rhotic liver? AJR Am J Roentgenol 179:751-758
33. Monzawa S, Ichikawa T, Nakajima H et al (2007) Dynamic CT
for detecting small hepatocellular carcinoma: usefulness of de-
layed phase imaging. AJR Am J Roentgenol 188:147-153
34. Iannacone R, Laghi A, Catalano C et al (2005) Hepatocellular
carcinoma: Role of unenhanced and delayed-phase multi-de-
tector row helical CT in patients with cirrhosis. Radiology
234:460-467
35. Baron RL, Brancatelli G (2004) Computed tomographic imaging
of hepatocellular carcinoma. Gastroenterology 127:S133-143
36. Forner A, Vilana R, Ayuso C et al (2008) Diagnosis of hepat-
ic nodules 20 mm or smaller in cirrhosis: Prospective valida-
Wolfgang Schima, Richard Baron
tion of the noninvasive diagnostic criteria for hepatocellular
carcinoma. Hepatology (Baltimore, Md) 47:97-104
37. Tublin ME, Dodd GD, Baron RL (1997) Benign and malignant
portal vein thrombosis: differentiation by CT characteristics.
AJR Am J Roentgenol 168:719-723
38. Stevens WR, Gulino SP, Batts KP et al (1996) Mosaic pattern
of hepatocellular carcinoma: histologic basis for a characteris-
tic CT appearance. J Comput Assist Tomogr 20:337-342
39. Pauleit D, Textor J, Bachmann R et al (2001) Hepatocellular
carcinoma: detection with gadolinium- and ferumoxides-
enhanced MR imaging of the liver. Radiology 222:73-80
40. Tang Y, Yamashita Y, Arakawa A et al (1999) Detection of he-
patocellular carcinoma arising in cirrhotic livers: Comparison
of gadolinium- and ferumoxides-enhanced MR imaging. AJR
Am J Roentgenol 172:1547-1554
41. Kitao A, Zen Y, Matsui O et al (2009) Hepatocarcinogenesis:
multistep changes of drainage vessels at CT during arterial
portography and hepatic arteriography-radiologic-pathologic
correlation. Radiology 252:605-614
42. Ward J, Guthrie JA, Scott DJ et al (2000) Hepatocellular car-
cinoma in the cirrhotic liver: double-contrast MR imaging for
diagnosis. Radiology 216:154-162
43. Lee DH, Kim SH, Lee JM et al (2009) Diagnostic performance
of multidetector row computed tomography, superparamagnet-
ic iron oxide-enhanced magnetic resonance imaging, and dual-
contrast magnetic resonance imaging in predicting the appro-
priateness of a transplant recipient based on milan criteria:
correlation with histopathological findings. Invest Radiol
44:311-321
44. Ichikawa T, Federle MP, Grazioli L et al (1999) Fibrolamellar
hepatocellular carcinoma: imaging and pathologic findings in
31 recent cases. Radiology 213:352-361
45. Ichikawa T, Federle MP, Grazioli L, Marsh W (2000) Fibro-
lamellar hepatocellular carcinoma: pre- and posttherapy eval-
uation with CT and MR imaging. Radiology 217:145-151
46. Lim JH (2003) Cholangiocarcinoma: morphologic classifica-
tion according to growth pattern and imaging findings. AJR
Am J Roentgenol 181:819-827
47. Han JK, Choi BI, Kim AY et al (2002) Cholangiocarcinoma:
pictorial essay of CT and cholangiographic findings. Radio-
graphics 22:173-187
48. Lee WJ, Lim HK, Jang KM et al (2001) Radiologic spectrum of
cholangiocarcinoma: emphasis on unusual manifestations and
differential diagnoses. Radiographics 21 Spec No:S97-S116
49. Buetow PC, Buck JL, Pantongrag-Brown L et al (1995) Bil-
iary cystadenoma and cystadenocarcinoma: clinical-imaging-
pathologic correlations with emphasis on the importance of
ovarian stroma. Radiology 196:805-810
50. Peterson MS, Baron RL, Rankin SC (2000) Hepatic angiosar-
coma: findings on multiphasic contrast-enhanced helical CT
do not mimic hepatic hemangioma. AJR Am J Roentgenol
175:165-170
51. Koyama T, Fletcher JG, Johnson CD et al (2002) Primary he-
patic angiosarcoma: findings at CT and MR imaging. Radio-
logy 222:667-673
52. Miller WJ, Dodd GD 3rd, Federle MP, Baron RL (1992) Ep-
ithelioid hemangioendothelioma of the liver: imaging findings
with pathologic correlation. AJR Am J Roentgenol 159:53-57
53. Schima W, Kulinna C, Langenberger H, Ba-Ssalamah A
(2005) Liver metastases of colorectal cancer: US, CT or MR?
Cancer Imaging 5 Spec No A:S149-156
54. Lee MJ, Saini S, Compton CC, Malt RA (1991) MR demon-
stration of edema adjacent to a liver metastasis: pathologic cor-
relation. AJR Am J Roentgenol 157:499-501

Imaging Diseases of the Gallbladder and Bile Ducts
Angela D. Levy1, Celso Matos2
1 Department of Radiology, Georgetown University Hospital, Washington, DC, USA
2 Department of Radiology, University Clinics of Brussels, Erasme Hospital, Brussels, Belgium
Introduction
Patients with gallbladder and biliary disease may present
with complaints of right upper quadrant or mid-epigastric
pain, fever, jaundice, pruritis, nausea, vomiting, or they
may be asymptomatic, with only laboratory abnormali-
ties. Ultrasound, multidetector computed tomography
(MDCT), and magnetic resonance imaging (MRI) are
used for the non-invasive evaluation of patients with
signs and symptoms of gallbladder and biliary disease.
New MRI techniques, such as functional magnetic reso-
nance cholangiography (MRC), have improved visualiza-
tion of the bile ducts. In many instances, non-invasive
imaging will establish the diagnosis prior to endoscopic,
radiological, surgical intervention. The differential diag-
nosis for these patients is broad and includes infectious,
non-infectious inflammatory, neoplastic, and congenital
disorders of the liver, gallbladder, and bile ducts. This
chapter discusses functional MRI of the bile ducts, in-
cluding the use of functional MRC, and the patterned ap-
proach to the differential diagnosis of gallbladder and
bile-duct disease.
Functional Magnetic Resonance Imaging of the Bile Ducts
The bile ducts are generally investigated at MRI by us-
ing a heavily T2-weighted sequence. This approach is
completely non-invasive and is extremely sensitive for
the diagnosis of anatomical abnormalities of the biliary
tree. However, the assessment of biliary drainage re-
mains challenging with T2-weighted imaging. In T2-
weighted MRC, dynamic views of the Vaterian sphincter
complex can be obtained in addition to functional infor-
mation when cholecystokinin stimulation is performed.
Unfortunately, cholecystokinin is associated with impor-
tant side effects (nausea, vomiting) and therefore is not
widely used. An alternative approach that can be used to
obtain functional information regarding the bile ducts re-
lies on the intravenous administration of paramagnetic
contrast agents. These compounds are taken up by the
hepatocytes and then transported to the bile, where they
cause T1 shortening as a result of their paramagnetic
IDKD 2010-2013
properties. The functional capabilities of MRI such as
those conferred by contrast-enhanced MRC have ex-
panded the clinical applications of magnetic resonance
cholangiography.
Technique, Indications, Results
T2-weighted imaging
The use of T2-weighted imaging on functional MRC al-
lows evaluation of the Vaterian sphincter. These types of
studies require a sequence with high temporal resolution,
as is the case with single-shot single-slice turbo spin echo
(TSE) T2-weighted sequences, which have an acquisition
time of about 2-3 s. Serial MRC images (we usually ob-
tain 20 consecutive dynamic views) are optimally ob-
tained in the coronal plane to evaluate dynamic changes
of the sphincteric segment of the pancreaticobiliary junc-
tion. This technique, also called kinematic magnetic res-
onance cholangiopancreatography (MRCP), allows nor-
mal physiological sphincter contractions to be distin-
guished from biliary stenosis in patients presenting with
bile duct dilatation without an evident cause of obstruc-
tion. In the study of Kim et al., failure to visualize
sphincteric relaxation had a sensitivity of 88% and a
specificity of 100% in the diagnosis of periampullary ob-
struction. Schmidt et al. reported high accuracy (similar
to that of endoscopic ultrasound) of the same technique
for the detection of choledocholithiasis. Kinematic
MRCP may also be used after the intravenous adminis-
tration of cholecystokinin. This hormone, which is
formed and secreted by the APUD cells of the duodenum,
induces bile flow by contracting the gallbladder and im-
posing an inhibitory effect on the sphincter of Oddi.
Therefore, a transient or persistent biliary dilatation may
be observed when dilatation of the bile-duct without an
evident cause is evaluated.
Contrast-Enhanced Functional MRC
This approach takes advantage of the biliary excretion of
paramagnetic contrast agents, allowing images to be ob-
tained with a high-resolution 3D gradient-recalled echo
76
(GRE) T1-weighted sequence. Three different hepatobil-
iary contrast agents are available for functional MRC:
mangafodipir-trisodium, gadobenate dimeglumine, and
gadolinium ethoxybenzyl diethylenetriaminepentaacetic
acid (Gd-EOB-DTPA). Proportions of hepatobiliary ex-
cretion are higher for mangafodipir-trisodium and for Gd-
EOB-DTPA (about 50%) than for gadobenate dimeglu-
mine. Unlike mangafodipir-trisodium, gadobenate dimeg-
lumine and Gd-EOB-DTPA allow evaluation of liver
parenchyma and angiographic renderings in the same ex-
amination. For both reasons (high proportion of biliary
excretion and vascular assessment), we currently use Gd-
EOB-DTPA. It should be emphasized that the degree of
biliary excretion of the contrast agent is dependent on liv-
er function. Tschirch et al. reported decreased visualiza-
tion or non-visualization of the biliary tree in a substan-
tial percentage of patients with liver cirrhosis. In patients
with normal liver function, good visualization of the bil-
iary tree is observed 20-30 min after intravenous admin-
istration of 0.025 mmol Gd-EOB-DTPA/kg body weight.
Due to the fact that contrast injection indications are gen-
erally based on the T2-weighted MRCP findings and be-
cause of possible signal extinction in the bile duct on T2-
weighted imaging and lack of visualization of the pancre-
atic duct, contrast-enhanced functional MRC is always
obtained after T2-weighted MRCP has been acquired.
a b
a b
Angela D. Levy, Celso Matos
Therefore, after T2-weighted MRCP, high-resolution fat-
suppressed 3D GRE T1-weighted sequences are obtained
without contrast and then again after intravenous admin-
istration of the contrast agent. First, dynamic axial views
are acquired, to evaluate the vessels and the liver
parenchyma in the arterial and portal-venous phases. Sub-
sequent scans are acquired in the coronal plane until vi-
sualization of the contrast agent in the gallbladder and
duodenum is noted. If additional delayed imaging is need-
ed, the transit time of the contrast agent into the biliary
tree and gallbladder may be determined; this estimate is
valuable in the diagnosis of partial obstruction of the bile
duct. Contrast-enhanced functional MRC increases the
diagnostic performance of conventional MRC in detect-
ing bile-duct leaks (Fig. 1), in assessing the patency of
biliary-enteric anastomoses (Fig. 2), and in diagnosing
biliary complications of liver transplantation (Fig. 3).
Bridges et al. reported that, in the evaluation of liver
transplants, delayed excretion (>45 min) was associated
with high-grade strictures. Moreover, neither the type of
anastomosis nor the presence of edema or ascites had any
influence on contrast-enhanced fMRC. When no excre-
tion is observed in the absence of biliary dilatation, trans-
plant rejection should be considered as a possible cause.
Other potential clinical scenarios include sphincter of
Oddi dysfunction and impaired patency of biliary stents.
Fig. 1 a, b. Bile duct leakage after cholecystec-
tomy. a T2-weighted MRCP shows a fluid-
filled collection adjacent to the extrahepatic
bile duct (arrows). b Contrast-enhanced func-
tional MRC confirms the leakage (arrows)
Fig. 2 a, b. Biliary-enteric anastomoses: as-
sessing patency. a T2-weighted MRCP
shows dilatation of the intrahepatic bile duct
(arrow). The anastomoses is not depicted. b
Contrast-enhanced functional MRC shows
the anastomoses and progressive filling of
the jejunal loop (arrow)
Imaging Diseases of the Gallbladder and Bile Ducts 77

a b

Fig. 3 a, b. Liver transplantation: assessment of

biliary anastomoses. a T2-weighted MRCP
shows dilatation of the intrahepatic bile ducts
upstream to the biliary-biliary anastomoses
(arrow). b Contrast-enhanced functional
MRC shows the patency of the anastomoses
(arrow)

Conclusion an unusual pseudotumoral chronic inflammatory condition

Major indications of functional MRC concern the assess-
ment of complications after bile-duct surgery (diagnosis
of bile duct leakage and determination of the patency of
biliary-enteric anastomoses). Contrast administration
should be considered after the biliary morphology has
been evaluated on T2-weighted MRC.
Patterned Approach to the Diagnosis of Gallbladder
and Bile-Duct Diseases
Gallbladder Wall Thickening
Focal or diffuse thickening of the gallbladder wall may be
caused by acute or chronic cholecystitis or non-inflamma-
tory conditions such as heart failure, cirrhosis, hepatitis,
hypoalbuminemia, renal failure, and human immunodefi-
ciency virus (HIV) infection. Gallbladder carcinoma may
also cause thickening of the gallbladder wall and should be
suggested when there are findings of a focal mass, lym-
phadenopathy, extension of the process to adjacent organs,
hepatic metastases, or biliary obstruction at the level of the
porta hepatis. Xanthogranulomatous cholecystitis (XGC) is
of the gallbladder that radiologically simulates gallbladder
carcinoma. There is a significant overlap in the CT features
of XGC and gallbladder carcinoma. Both entities may
demonstrate wall thickening, infiltration of the surrounding
fat, hepatic involvement, and lymphadenopathy. Adeno-
myomatous hyperplasia, also called adenomyomatosis, is a
more common tumor-like lesion of the gallbladder that may
produce focal, segmental, or diffuse mural thickening.
Sonographically, adenomyomatous hyperplasia is charac-
terized by focal or diffuse thickening of the gallbladder
wall, with echogenic foci and ring-down artifact emanating
from the wall. The echogenic foci represent bile salts, cho-
lesterol crystals, or small stones in Rokitansky-Aschoff si-
nuses. On MRI, Rokitansky-Aschoff sinuses are best visu-
alized on breath-hold T2-weighted sequences (Fig. 4). Con-
sequently, MRI can be useful to distinguish adenomyo-
matous hyperplasia from gallbladder carcinoma.
Gallbladder Polypoid Mass
Polypoid gallbladder masses are commonly demonstrated
on ultrasound as incidental findings when the gallbladder is
imaged sonographically. Polyps are estimated to be present
in approximately 3% of gallbladders. The differential diag-
nosis for a gallbladder polyp includes cholesterol polyp,

a b

Fig. 4 a, b. A 48-year-old man evaluated for

jaundice was found to have a benign stricture
in the distal common bile duct and adeno-
myomatous hyperplasia of the gallbladder.
a Longitudinal sonogram of the gallbladder
shows mural thickening along the anterior
gallbladder wall and echogenic foci in the
gallbladder wall that produce ring-down re-
verberation artifact. b Single-shot fast spin-
echo T2-weighted MR image with fat satura-
tion shows multiple hyperintense foci within
the thickened gallbladder wall, consistent
with adenomyomatous hyperplasia
78
adenoma, focal adenomyomatous hyperplasia, inflammato-
ry polyp, heterotopia, neurofibroma, carcinoma, carcinoid
tumor, lymphoma, and metastasis. The majority of gallblad-
der polyps are benign. Of these, the most common type is a
cholesterol polyp, which accounts for ~50% of all polypoid
lesions. Cholesterol polyps have no malignant potential. On
sonography, they are typically brightly echogenic, round or
slightly lobulated masses that do not produce acoustic shad-
owing. Larger cholesterol polyps are usually less echogenic
and may contain an aggregation of echogenic foci.
The management of gallbladder polyps is based on the
risk of malignancy, which increases for polyps >10 mm
in size and in patients over the age of 60. As the incidence
of malignancy in polyps >10 mm ranges from 37 to 88%,
patients with symptomatic polyps >10 mm are encour-
aged to undergo cholecystectomy while those with polyps
<10 mm should be followed periodically by ultrasound.
At sonography, careful attention should be paid to other
features that suggest malignancy: thickening or nodular-
ity of the gallbladder wall, evidence of hepatic invasion
such as an indistinct margin between the liver and gall-
bladder, biliary duct dilatation, and peripancreatic hepato-
duodenal ligament adenopathy. If there are any features
suggestive of malignancy, MDCT or MRI should be con-
sidered for further evaluation of the lesion.
Cystic Dilatation of the Extrahepatic Bile Duct
Mechanical biliary obstruction is the most common cause
of extrahepatic bile duct dilatation. Upon initial imaging,
an obstructive lesion should always be sought when bil-
iary dilatation is present. Once an obstructive lesion is ex-
cluded, congenital etiologies of bile duct dilatation should
be considered. Choledochal cysts, unlike obstructive di-
latation, generally have more focal extrahepatic bile duct
dilatation or are typically more expansive than what is
usually encountered in mechanical dilatation (Fig. 5).
Fig. 5. Choledochal cyst in a 3-year-old boy evaluated for a palpa-
ble right upper quadrant mass. MRCP shows marked extrahepatic
bile duct dilatation with minimal dilatation of the central intra-
hepatic ducts
Angela D. Levy, Celso Matos
These cysts may have normal intrahepatic bile ducts or
partial dilatation of the intrahepatic bile ducts in a non-
obstructive pattern. It may be more difficult to differen-
tiate a choledochal cyst with mild or fusiform dilatation
of the extrahepatic duct from a duct that is dilated sec-
ondary to an obstructing lesion. In these cases, MRCP
and/or endoscopic retrograde cholangiopancreatography
(ERCP) are useful to exclude an obstructing lesion and to
evaluate the pancreaticobiliary junction, an anomaly
which is commonly observed in patients with chole-
dochal cyst. Occasionally, pancreatic pseudocysts,
echinococcal (hydatid) cysts, and cystic biliary neo-
plasms, such as biliary cystadenoma or biliary cystade-
nocarcinoma, may occur in or around the porta hepatis,
simulating biliary dilatation and a choledochal cyst. The
appearance of rim-like calcification and enhancing sep-
tations or mural nodules should help in establishing the
diagnosis of a biliary cystadenoma or cystadenocarcinoma.
Likewise, echinococcal (hydatid) cysts generally have
evidence of inner membranes, daughter cysts, or rim-like
peripheral calcification. T2-weighted MRI of hydatid
cysts may show a fibrous capsule of low signal intensity
as well as membranes.
Cystic Dilatation of Intrahepatic Bile Ducts
Similar to dilatation of the extrahepatic duct, mechanical
biliary obstruction is the most common cause of intra-
hepatic bile duct dilatation. Intrahepatic biliary dilatation
due to mechanical obstruction is generally tubular and
lacks focal stricture formation. Caroli disease, recurrent
pyogenic cholangitis, polycystic liver disease, primary
sclerosing cholangitis, choledochal cyst, and peribiliary
cysts should be included in the differential diagnosis of
cystic intrahepatic biliary dilatation. Caroli disease is
suggested by focal or diffuse biliary dilatation that is cys-
tic or fusiform in character (Fig. 6). When diffuse in-
volvement is present, the bile ducts converge toward the
porta hepatis. Echogenic intraductal sludge or inflamma-
tory debris may be seen, as well as echogenic stones with
posterior acoustic shadowing. The most important differ-
ential diagnosis for patients with suspected Caroli disease
is recurrent pyogenic cholangitis, which is characterized
by biliary dilatation with intrahepatic stone formation.
The left hepatic lobe is more commonly involved than the
right in recurrent pyogenic cholangitis. Polycystic liver
disease may also mimic Caroli disease. However, in most
cases, the bile ducts in polycystic liver disease are intrin-
sically normal; only rarely will the cysts communicate
with the bile ducts.
Although intrahepatic bile duct dilatation is a feature
of primary sclerosing cholangitis, the dilatation is typi-
cally fusiform and isolated. In primary sclerosing cholan-
gitis, the degree and extent of duct dilatation is not as se-
vere as that in obstructive biliary dilatation, Caroli dis-
ease, or recurrent pyogenic cholangitis; instead, fibrosis,
stricture formation, and secondary cirrhosis are the ma-
jor features.
Imaging Diseases of the Gallbladder and Bile Ducts 79

a b

Fig. 6 a, b. Caroli disease in a 23-year-old man

who complained of abdominal pain. a Sagittal
sonogram of the liver shows multiple cystic
spaces in the posterior right lobe. b MRCP
shows high signal intensity within these cysts

Choledochal cyst should be considered in the differen-
tial diagnosis if there is both intrahepatic and extrahepat-
ic duct dilatation. Generally, in patients with choledochal
cyst the extrahepatic dilatation is more severe than the
intrahepatic dilatation.
Multiple peribiliary cysts in sequence may simulate
bile duct dilatation characterized by a beaded or saccular
appearance. Since the bile ducts adjacent to peribiliary
cysts are normal, correct diagnosis depends upon the vi-
sualization of a normal bile duct. Peribiliary cysts are
usually associated with hepatic diseases such as cirrhosis,
polycystic liver disease, portal hypertension, portal vein
obstruction, and metastatic disease.
(in or near the confluence of the right and left hepatic
ducts) may be secondary to hilar cholangiocarcinoma
(Klatskin tumor) (Fig. 7), inflammation, or vascular
impressions. Strictures in the mid-portion of the extra-
hepatic bile duct are commonly related to diseases of the
gallbladder, such as carcinoma, that have invaded the cys-
tic duct and hepatoduodenal ligament or to inflammatory
conditions, such as impaction of a stone in the cystic duct
(Mirrizi syndrome). Distal extrahepatic strictures may be
due to inflammatory or neoplastic diseases of the pan-
creas, primary carcinomas of the bile duct or ampulla,
sphincter of Oddi dysfunction, or, less commonly, infec-
tious papillitis such as seen in AIDS cholangiopathy.

Biliary Stricture
Focal narrowing or strictures in the biliary ducts may be
secondary to neoplasia, inflammation, trauma (iatrogenic
or non-iatrogenic), or mass effect from adjacent processes.
The location of the biliary stricture narrows the differen-
tial diagnosis. Strictures at the level of the porta hepatis
Conclusions
A patterned approach to differential diagnosis of gall-
bladder and biliary duct disease is useful and may be ap-
plied to the findings identified on all non-invasive imag-
ing techniques.

a b

Fig. 7 a, b. Hilar cholangiocarcinoma in a 40-

year-old man who complained of abdominal
pain. Coronal MRCP (a) and percutaneous
transhepatic cholangiogram (b) show a hilar
stricture and mild intrahepatic biliary di-
latation
80
Suggested Reading
Functional MR imaging of the bile duct
Bridges MD, May GR, Harnois DM (2004) Diagnosing biliary com-
plications of orthotopic liver transplantation with mangafodipir
trisodium-enhanced MR Cholangiography: comparison with
conventional MR Cholangiography. AJR 182:1497-1504
Fayad LM, Holland GA, Bergin D et al (2003) Functional MR
cholangiography of the gallbladder and biliary tree with con-
trast-enhanced magnetic resonance cholangiography. JMRI
18:449-460
Hottat N, Winant C, Metens T et al (2005) MR cholangiography
with manganese dipyridoxyl diphosphate in the evaluation of
biliary-enteric anastomoses: preliminary experience. AJR
184:1556-1562
Kim JH, Kim MJ, Park SI et al (2002) Using kinematic MRCP to
evaluate biliary dilatation. AJR 178:909-914
Schmidt S, Chevalier P, Novellas S et al (2007) Choledocholithia-
sis: repetitive thick slab single-shot projection MRCP versus
endoscopic ultrasonography. Eur Radiol 17:241-250
Tschirch FTC, Struwe A, Petrowsky et al (2008) Contrast-en-
hanced MR cholangiography with Gd-EOB-DTPA in patients
with liver cirrhosis: visualization of the biliary ducts in com-
parison with patients with normal liver parenchyma. Eur Ra-
diol 18:1577-1586
Van Hoe L, Gryspeerdt S, Vanbeckevoort D et al (1998) Normal
Vaterian sphincter complex: evaluation of morphology and
contractility with dynamic single-shot MRCP. AJR 170:1497-
1500
Patterned Approach to the Diagnosis of Gallbladder and
Bile Duct Diseases
Albores-Saavedra, J, Hensen De, Klimsta DS (2000) Tumors of the
gallbladder, extrahepatic bile ducts, and ampulla of vater: At-
las of tumor pathology. Fasc 27, ser 3. Armed Forces Institute
of Pathology, Washington, DC
Angela D. Levy, Celso Matos
Baron RL, Campbell WL, Dodd GD 3rd (1994) Peribiliary cysts
associated with severe liver disease: imaging-pathologic cor-
relation. AJR Am J Roentgenol 162:631-636
Baron RL, Tublin ME, Peterson MS (2002) Imaging the spectrum
of biliary tract disease. Radiol Clin North Am 40:1325-1354
Chun KA, Ha HK, Yu ES et al (1997) Xanthogranulomatous chole-
cystitis: CT features with emphasis on differentiation from
gallbladder carcinoma [see comments]. Radiology 203:93-97
Goodman ZD, Ishak K (1981) Xanthogranulomatous cholecystitis.
Am J Surg Pathol 5:653-659
Guy F, Cognet F, Dranssart M et al (2002) Caroli’s disease: mag-
netic resonance imaging features. Eur Radiol 12:2730-2736
Haradome H, Ichikawa T, Sou H et al (2003) The pearl necklace
sign: an imaging sign of adenomyomatosis of the gallbladder
at MR cholangiopancreatography. Radiology 227:80-88
Ishikawa O, Ohhigashi H, Imaoka S et al (1989) The difference in
malignancy between pedunculated and sessile polypoid lesions
of the gallbladder. Am J Gastroenterol 84:1386-1390
Levy AD, Murakata LA, Rohrmann Jr CA (2001) Gallbladder car-
cinoma: radiologic-pathologic correlation. Radiographics
21:295-314
Levy AD, Rohrmann Jr CA (2003) Biliary cystic disease. Curr
Probl Diagn Radiol 32(6):233-263
Levy AD, Rohrmann Jr CA, Murakata LA, Lonergan GJ (2002)
Caroli’s disease: radiologic spectrum with pathologic correla-
tion. AJR Am J Roentgenol 179:1053-1057
Raghavendra, BN, Subramanyam BR, Balthazar EJ et al (1983)
Sonography of adenomyomatosis of the gallbladder: radiolog-
ic-pathologic correlation. Radiology 146:747-752
Schulte SJ, Baron RL, Teefey SA et al (1990) CT of the extrahep-
atic bile ducts: wall thickness and contrast enhancement in
normal and abnormal ducts. AJR Am J Roentgenol 154:79-85
Sugiyama M, Atomi Y, Kuroda A et al (1995) Large cholesterol
polyps of the gallbladder: diagnosis by means of US and en-
doscopic US. Radiology 196:493-497
Yoshimitsu K, Honda H, Jimi M et al (1999) MR diagnosis of ade-
nomyomatosis of the gallbladder and differentiation from gall-
bladder carcinoma: importance of showing Rokitansky-
Aschoff sinuses. AJR Am J Roentgenol 172:1535-1540
 IDKD 2010-2013

Diseases of the Pancreas, I: Pancreatitis

Thomas Helmberger
Institute of Diagnostic and Interventional Radiology and Nuclear Medicine, Klinikum Bogenhausen, Munich, Germany

Introduction importance because of significant differences in progno-

The incidence of pancreatic inflammatory diseases in the
Western world is 110-240 cases/1 million individuals. In
about 80% of these cases, the underlying cause is an un-
detected gall stone disease or alcohol abuse. The former
is more prevalent in women and the latter in men, espe-
cially those between 30 and 50 years of age. Neverthe-
less, there are many other factors that can cause pancre-
atitis (Table 1).
The manifestations of pancreatitis range from mild,
self-limiting disease to severe, lethal forms in acute pan-
creatitis or permanent loss of exocrine and/or endocrine
function in chronic pancreatitis [1-5]. Clinically, it can be
difficult to differentiate cystic changes of the pancreas
from cystic tumors, and chronic pancreatitis (CP) from
pancreatic cancer due to similarities in imaging presenta-
tion. Nevertheless, the differential diagnosis is of ample
sis and management.
The differential diagnosis in terms of solid and cystic
tumors is discussed in the chapter by Thoeni (Diseases of
the Pancreas II: Tumors).
Acute Pancreatitis
Acute pancreatitis (AP) is defined as an inflammatory
process involving the gland that was normal prior to the
attack and is normal again once the derangements that
precipitated the attack have been corrected. AP is a com-
mon problem, with a rising incidence in the Western
world. In the USA, it currently accounts for about 200,000
admissions/year and about 10,000 deaths/year. The dis-
ease is most frequently seen in individuals in the fifth to
sixth decade of life, and the most common causes are

Table 1. Potential causes of pancreatitis

Metabolic Alcohol
Drugs (e.g., glucocorticosteroids, azathioprine, hydrochlorothiazide, furosemide, sulfonamides,
estrogens, pentamidine, didanosine (DDI), valproic acid)
Hypertriglyceridemia
Hyperalimentation with lipids
Hypercalcemia
Hyperparathyroidism
Mechanical/obstructive Gall stones
Tumors (carcinoma of the pancreas or common bile duct, tumor in the ampulla and duodenum,
metastases)
Sphincter of Oddi dysfunction
After upper gastrointestinal tract surgery or endoscopy
Following abdominal trauma
Duodenal obstruction/scar (e.g., after ulcer disease)
Infectious Mumps
Coxsackie virus infection
Ascaris infection, with a mechanical component
Vascular Panarteriitis nodosa
After cardiac or pulmonary surgery
Severe arteriosclerosis
Other Hereditary, idiopathic
Anomalies (pancreas divisum, annulare)
Autoimmune
Scorpion sting/gila monster bite
Unknown (microlithiasis??)
82 Thomas Helmberger

cholelithiasis (75%) and alcohol abuse (15%). Other caus-
es are listed in Table 1.
With respect to pathophysiology, the most likely fac-
tors for the sudden onset of AP are pancreatic hyper-
secretion, intra- and extravasation of pancreatic secre-
tions, and premature activation of pancreatic enzymes,
followed by autodigestion and necrosis of the pancreatic
gland and peripancreatic tissues.
AP can be divided clinically into mild and severe
forms, which are almost paralleled by the pathophysio-
logical findings of interstitial (edematous) and necrotiz-
ing forms [6, 7]. Mild AP (~50% of cases) is character-
ized by mild symptoms and transitory elevation of amy-
lase levels that recover rapidly without complications. In
general, the gland may be enlarged due to a moderate
Fig. 1. Computed tomography image of acute biliary pancreatitis.
Edematous enlargement of the pancreatic head and an exudation sur-
rounding the duodenum and along the pararenal fascia are seen. The
cause was a biliary stone trapped in the ampullary region (arrow)
edema and peripancreatic fluid collections can be pre-
sent. Nonetheless, in 30% of the cases, no morphological
changes can be appreciated. In cholelithiasis, segmental
pancreatitis, mainly of the pancreatic head, is seen in up
to 20% of cases (Fig. 1).
Ultrasound (US) may reveal a normal to mildly en-
larged gland with homogeneous (hypoechoic) echogenic-
ity, but sufficient visualization by US is possible only in
60-70% of cases. In contrast-enhanced computed tomo-
graphy (CT) and, while generally not necessary, magnetic
resonance imaging (MRI), the gland is diffusely enlarged
and a small amount of fluid is seen outlining the gland.
Imaging is needed to rule out other underlying conditions
that can be accompanied by hyperamylasemia, such as
bowel obstruction, bowel infarction, gangrenous chole-
cystitis, and perforated ulcers.
In mild to moderate progressing forms of AP, the con-
tour of the gland becomes shaggy. The appearance of the
parenchyma on CT and MRI may be heterogeneous, and
small intraglandular and/or retroperitoneal fluid collec-
tions adjacent to the organ may develop (Fig. 2).
The severe forms of pancreatitis are determined by a
delayed/absent response to conservative therapy or even
deterioration under therapy. Mortality at the latter stage
can be as high as 100%. Typical findings in severe
(necrotizing) AP are varying degrees of parenchymal
necrosis accompanied by progressive exudation, superin-
fection of necrotic tissue, hemorrhage, abscess forma-
tion, phlegmon (~inflammatory pannus), and vascular
erosion (Fig. 3).
Depending on the particular pathomorphological con-
dition, the pancreas and its surroundings present a rather
wide spectrum of imaging findings. In severe cases, US
imaging is often compromised by overlying gas, peripan-
creatic exudation, and phlegmonous changes. In necrosis,

a b

Fig. 2 a, b. Magnetic resonance imaging of acute pancreatitis after cholecystectomy and reconstruction of the extrahepatic common bile duct.
a T1-weighted gradient-echo image shows the slightly dilated side branches of the pancreatic duct (arrow). A small fluid rim on the ante-
rior renal fascia is seen. b The heavily T2-weighted image reveals fluid outlining the pancreatic gland and the slightly dilated pancreatic
duct. Note the moderate stenosis (arrow) of the common bile duct after surgical reconstruction
Diseases of the Pancreas, I: Pancreatitis 83

a b

Fig. 3 a, b. Acute severe pancreatitis. a Contrast-enhanced CT obtained at admission shows a fuzzy contour of the pancreatic gland togeth-
er with a peripancreatic exudation (arrows) during the venous phase. Note the hypo- and hyperdense hepatic lesions (arrowheads). b Con-
trol scan 10 days later revealed an almost normal gland with resorption of the peripancreatic fluid. However, an area with a lack of en-
hancement, representing focal necrosis (large arrow), was observed. In the liver, one lesion turned out to be a hemangioma (arrow), while
the other two lesions were small abscesses

the pancreatic appearance becomes increasingly hypoe-
choic, without differentiation of vital from necrotic tis-
sue. Therefore, US is generally used for second-line,
complementary imaging during patient follow-up in or-
der to detect fluid formations such as pseudocysts.
Parenchymal necrosis is best displayed on contrast-en-
hanced CT during at least the portal-venous phase. Char-
acteristic findings are patchy areas showing a lack of en-
hancement, (pseudo) fragmentation, and liquid necroses.
Additionally, increasing peripancreatic exudations dis-
secting along retroperitoneal fascia planes into the meso-
colon and the small bowel mesentery as well as peri-
pancreatic inflammatory tissue (phlegmon) and infected
areas are frequently seen. In <10% of cases, small amounts
of intraperitoneal fluid (ascites) are present, whereas large
volumes of intraperitoneal fluids are very rare.
According to the literature, there is no significant su-
periority of MRI over CT in the diagnosis of AP and its
related complications. The superior tissue resolution and
higher sensitivity to slightly edematous or necrotic
changes and to hemorrhage or fluid dissection of fat
planes favor MRI. However, these advantages are often
hampered by the impaired study conditions in severely ill
patients that may degrade the image quality. In patients
with severe AP who are administered iodinated contrast
agents, pancreatic flow can be reduced followed by an in-
creased rate of necrosis and mortality. While this would
seem to favor MRI as a staging tool in AP, this potential
complication has not been proved for the non-ionic con-
trast agents that are nowadays used almost exclusively in
CT [8-10].
The local inflammatory conditions are often compli-
cated by regional and systemic involvement induced by
autodigestion and activation of systemic inflammatory
mediators [11-17]. A rapid change in the local pancreat-
ic and overall abdominal situation demands an adequate
diagnostic and therapeutic regime to avoid a disastrous
Table 2. CT grading in acute pancreatitis (from [26, 27])
Grade CT findings
A Normal
B Focal (~20%), diffuse enlargement of the gland, irregu-
lar contour, inhomogeneous density
C Grade B + inflammation of the peripancreatic fat
D Small, mostly occasional fluid collections or phlegmon
E Two or more fluid collections, gas within the pancreas
or retroperitoneum
outcome. Several clinical and laboratory scoring systems
have been established to stage and predict the clinical
course of severe AP (Ranson’s score, APACHE II). How-
ever, these scoring systems are mainly dependent on sys-
temic alterations and are quite non-specific, as they fail
to address the local condition of the pancreas [18-25].
Balthazar et al. [26, 27] showed that contrast-enhanced
CT is the most helpful diagnostic modality to detect
complications that may necessitate medical, surgical, or
interventional management, and to predict outcome de-
pending on the local condition of the pancreas. The pro-
posed 5-grade scoring system (Tables 2, 3), by estimat-
ing the presence and degree of pancreatic and peripan-
creatic inflammation and fluid accumulation and by de-
tecting the presence and extent of pancreatic necrosis to-
gether with estimation of the lack of gland enhancement
(<30, 30-50, >50%), can be translated into a CT-severity
index (Table 3) that allows estimation of the complica-
tions (morbidity) and of mortality (Fig. 4). If >50% of
the pancreatic volume is necrotic, morbidity rises to al-
most 100%. Recently, a modified and simplified CT
severity index was proposed by Mortele et al. [28] that
more closely correlates with patient outcome measures
than is the case with the currently accepted CT severity
index (Table 3).
84 Thomas Helmberger

Table 3. “Classic” and modified CT severity index

CT severity index Points Modified CT severity index Points

Pancreatic inflammation Pancreatic inflammation

Normal pancreas 0 Normal pancreas 0
Focal or diffuse enlargement of the pancreas 1
Intrinsic pancreatic abnormalities with or without 2 Intrinsic pancreatic abnormalities with inflammatory 2
inflammatory changes in peripancreatic fat changes in peripancreatic fat
Single, ill-defined fluid collection or phlegmon 3 Pancreatic or peripancreatic fluid collection or 4
Two or more poorly defined collections or presence 4 peripancreatic fat necrosis
of gas in or adjacent to the pancreas
Pancreatic necrosis Pancreatic necrosis
None 0 None 0
≤30% 2 ≤30% 2
>30-50% 4 >30% 4
>50% 6 Extrapancreatic complications (one or more of 2
pleural effusion, ascites, vascular complications,
parenchymal complications, or gastrointestinal
tract involvement)

100 % In a septic patient, not-water-like fluid collections and

90
80
70
60
50 Mortality
40 Complications
30
20
10
0
0-3 4-6 7 - 10
Fig. 4. CT severity index related to the degree of necrosis of the
pancreatic parenchyma. Grade A = 0, B = 1, C = 2, D = 3, E = 4;
no necrosis = 0, 30% necrosis = 2, <50% necrosis = 4, >50%
necrosis = 6 (from [26,27])
About a fifth of the patients without necrotic changes
of the pancreatic gland will also develop local complica-
tions. Fluid collections are seen in up to 50% of patients
with AP. In about half of these patients, the collections
will resolve spontaneously within several weeks. In the
rest, however, the fluid collections will persist, eventual-
ly followed by encapsulation, superinfection (abscess), or
pseudocyst formation.
Complications
Pseudocysts are fluid collections with a noticeable capsule
that typically develop 4-5 weeks after the onset of AP. On
US, CT, and MRI, they have a cyst-like appearance, usu-
ally without septations. Since large cysts are prone to
complications (e.g., rupture, infection, hemorrhage, bil-
iary obstruction, or fistulization to the gastrointestinal
(GI) tract), cysts >5-7 cm in diameter should be treated by
percutaneous drainage or operative marsupialization.
rim enhancement on contrast-enhanced CT or MRI stud-
ies should be considered as abscesses until proven other-
wise. Gas, a characteristic sign of an infected fluid col-
lection, is detected in only 20% of patients with pancre-
atic abscesses. Percutaneous aspiration or drain place-
ment is the proper treatment.
In contrast to abscess formations, superinfected
necrotic areas of the pancreas are much more difficult to
treat. Due to the more solid consistency of the infected
necrosis, percutaneous drainage therapy is mostly frus-
trating; however, biopsy is often needed to prove the di-
agnosis. In most cases, either percutaneous, endoscopy-
guided necrosectomy, or surgical intervention has to be
considered.
Pseudoaneurysm formation and hemorrhage may re-
sult from the extravasated pancreatic enzymes that cause
vascular injury. They are typically late complications that
occur after several episodes of severe AP. While pseudo-
aneurysms are generally easily detected by any kind of
imaging modality, retroperitoneal hemorrhage is best
depicted by contrast-enhanced CT or unenhanced MRI.
Angiography with arterial embolization is the treatment of
choice and in general is superior to surgical therapy [29].
Groove Pancreatitis
First described by Becker in 1973, groove pancreatitis is a
rare, late complication occurring after several attacks of
AP. It is defined as an inflammatory reaction and fluid col-
lection located in the groove between the head of the pan-
creas, the duodenum, and the common bile duct. The an-
terior anlage of the pancreas seems to be mainly affected,
with duodenal stenosis and/or strictures of the common
duct in about 50% of the cases. Therefore, this disease may
mimic cancer of the pancreatic head, necessitating surgi-
Diseases of the Pancreas, I: Pancreatitis
cal exploration. Dynamic CT and MRI with delayed en-
hancement of collagen fibrous tissue during the late post-
equilibrium phase may reveal a potential soft-tissue mass
to be fibrosis and thus, in the absence of complications,
obviate the need for surgical exploration [30-33].
Autoimmune Pancreatitis
Autoimmune pancreatitis (AIP) is a relatively new syn-
drome of clinical and histological findings that was first
described by Yoshida in 1995 [34]. The condition has al-
so been described as lymphoplasmocytic sclerosing pan-
creatitis with cholangitis, non-alcoholic duct-destructive
chronic pancreatitis, and chronic sclerosing pancreatitis.
The features of AIP include hypergammaglobulinemia,
elevation of serum IgG4, IgG4-containing immune com-
plexes, and a number of other antibodies as antinuclear
antibodies, as well as antibodies against lactoferrin, car-
bonic anhydrase type II, and rheumatoid factors. Histo-
logically, there is fibrosis and a lymphoplasmacytic infil-
tration of the interlobular ducts. The majority of lympho-
cytes are CD8+ and CD4+, while B lymphocytes are less
frequent. In general, the diagnosis of AIP is established
by clinical signs, together with laboratory and morpho-
logical findings. An association with other autoimmune
diseases, such as Sjögren-syndrome, primary biliary cir-
rhosis, primary sclerosing cholangitis, Crohn’s disease or
ulcerating colitis, systemic lupus erythematosus, and
retroperitoneal fibrosis is found in a third of the cases.
At imaging, a focal (“mass-forming”) or diffuse
(“sausage-like”) enlargement of the pancreas may be pre-
sent. In contrast-enhanced studies, peripancreatic nodular
or rim-like enhancement can be appreciated. Focal AIP of
the pancreatic head that involves the pancreatic and dis-
tal common bile duct must be differentiated from pan-
creatic carcinoma, necessitating biopsy proof [35, 36].
In most patients, the symptoms as well as the labora-
tory and morphological abnormalities appear to respond
to steroid treatment [34, 37-42].
Chronic Pancreatitis
The hallmarks of chronic pancreatitis (CP) are a continu-
ing (aseptic) inflammation of the gland accompanied by ir-
reversible morphological and functional damage. The most
common reasons are chronic alcohol abuse (70%) and
cholelithiasis (20%), with rare cases arising from cystic fi-
brosis or idiopathically. Patients are typically in their 3rd to
4th decade of life and present with a history of epigastric
pain (95%), weight loss (95%), and signs of endocrine/
exocrine deficiency (diabetes mellitus 58%, malabsorption
syndrome and steatorrhea 80%). Acute exacerbations of
CP are accompanied by episodes of pain attacks that
may mimic an acute abdomen. With progressive destruc-
tion of the gland, CP may be painless after several years.
In 1.5-12% of cases it is complicated by pancreatic cancer.
85
Tumor markers such as CA 19-9 and CA-50 may be
elevated transiently and are non-specific. Laboratory
tests of secretin-creozyme and secretin-caerulein have a
high diagnostic accuracy except in early stages of the dis-
ease but are invasive and cumbersome for the patient.
However, these tests are of particular importance in the
diagnostically challenging, newly defined small-duct CP,
in which chronic inflammation occurs without ductal ab-
normalities.
In CP, the most characteristic findings are dilatation of
the pancreatic main duct and of the ductal side branches
(70-90%), small cystic changes, scattered glandular and
ductal calcifications (40-50%), and ductal protein plugs.
The grade and shape of the ductal dilatation may help to
differentiate chronic (benign) obstructions from malig-
nant occlusions: in CP, the contour of the pancreatic duct
and its side branches is commonly irregular (73%) while
this is true only in 15% of pancreatic malignancies. Ad-
ditionally, the duct usually accounts for <50% of the pan-
creatic anterior-posterior diameter in CP while the oppo-
site is true in pancreatic cancers (due to obstructive atro-
phy). In some cases, additional secretin-enhanced mag-
netic resonance cholangiopancreatography (MRCP) can
be helpful as it provides an improved display of the duct
system and allows assessment of the excretory capacity
of the pancreatic gland [43].
In CP, the gland may have a normal appearance in 15-
20% of patients, but most common is a diffuse (50%) or
focal (25%) enlargement that may arouse suspicion of a
neoplasm. With time, atrophy of the organ will occur in
10-50% of these patients. The variable appearance of CP
explains the shortcomings in establishing the diagnosis.
In the absence of gross morphological changes it is very
difficult to identify incipient forms of CP. Moreover,
morphological changes correlate very poorly with the
functional exocrine and endocrine deficits. Consequent-
ly, endoscopically guided (endoscopic US) or percuta-
neous biopsy may be necessary for the diagnosis.
Complications
These include consolidated cystic degeneration involving
dilatation of the cystic and side branches of the pancreat-
ic duct and peripancreatic pseudocyst formation (Fig. 5).
In addition, there may be obstruction of the common bile
duct secondary to fibrosis, splenic vein thrombosis (note:
upper GI tract bleeding may result from gastric varices in
the absence of esophageal varices!), and pancreatic fistu-
lae (communications between the pancreatic duct and ab-
dominal organs or the skin).
Diagnostic Challenge
In general, the diagnosis of inflammatory and tumorous
pancreatic conditions relies on imaging (US, CT, MRI)
and, more invasively, endoscopic retrograde cholan-
giopancreatography (ERCP). Imaging findings are deter-
mined by the macro-structural changes of the organ and
86 Thomas Helmberger

a b

Fig. 5 a, b. Chronic pancreatitis. Cystic degeneration of the pancreatic head (a) together with irregular dilatation of the pancreatic main duct
(b) is seen on MRI (fast spin-echo T2)

its surroundings. On plain films and US, the calcifica- Table 4. Ranson score based on clinical and laboratory signs at
tions in CP are readily depicted. Additionally, US is able admission and at 48-h follow-up (each sign = 1 point)
to display ductal dilatation, micro- and macrocystic At admission 48-h follow-up
changes, and the gland itself.
Age >55 years Hematocrit decrease >10%
Contrast-enhanced multidetector CT (MDCT) is well
WBC >16,000 Blood urea nitrogen increase
established in the assessment of ductal changes, calcifi- >5 mg/dL
cations, and alterations in the form and shape of the pan- Blood glucose >200 mg/dL Ca (serum) <8 mg/dL
creatic gland, as well as potential concomitant conditions Serum LDH >350 IU/L PO2 <60 mmHg
such as pseudocysts. In addition, multiplanar, curved re- SGOT (AST) >250 U/L Base deficit >4 meq/L
constructions yield high-resolution display of the entire Estimated fluid sequestration
gland and the anatomical course of the duct. Depending >600 mL
on the fibrotic changes in CP, contrast enhancement can Score Mortality (%)
be variable, whereas most ductal carcinomas show no or 0-2 <10
only minor enhancement during arterial-dominant- and 3-5 10-20
parenchymal-phase imaging. However, late enhancement >5 >50
can be seen on delayed imaging without substantial addi-
tional information [44].
In addition to MRI’s superior tissue resolution in the
differentiation of varying “qualities” of pancreatic missed diagnosis of carcinoma. However, if local or re-
parenchyma, using unenhanced and Gd-DTPA-enhanced gional lymph node enlargement, vascular encasement, or
T1-weighted (±fat suppression) and heavily T2-weighted remote metastases is displayed, the differential is ruled by
sequences, it optimally displays the pancreatic gland, the these secondary signs of malignancy, in which case the
pancreatic duct including the first-degree side branches, tumor must be staged correctly for further treatment strat-
and even small stones. Nevertheless, initial, minor ductal ification (Tables 3, 4). In ambiguous cases, biopsy or
changes are best seen on ERCP (Table 3), which in this even surgical exploration may be necessary.
respect is superior to all other imaging modalities. How- CP can cause a focal pancreatic mass indicative of a
ever, the clinical significance of these slight changes re- neoplasm. Moreover, it represents a major risk factor for
mains contentious, further compromised by potentially pancreatic cancer, with a 26-fold increased risk of devel-
“non-physiological” distention of the ducts due to the in- oping cancer, according to an international, multicenter
jected contrast material. cohort study [45]. Therefore, the differential between CP
Pancreatic cancer is the most serious complication of and pancreatic cancer remains challenging and under-
CP and is the major diagnostic challenge because the fo- lines the need for multiple diagnostic approaches. In one
cal enlargement of the gland induced by a fibrotic in- study, US, CT, MRI, and positron emission tomography
flammatory pseudotumor may be indistinguishable from (PET)/CT for pancreatic cancer were shown to have a
carcinoma. A comparison using state-of-the-art MDCT sensitivity of 76-83% and a specificity of 91-93% [46].
and MRI showed no difference in the detection rate of Nevertheless, the rate of incorrect diagnoses is as high as
pancreatic carcinoma, according to the recent literature. 25%. The use of various differential criteria (Table 5)
Nevertheless, the potential tumor-like appearance of CP may help to improve the overall diagnostic accuracy be-
accounts for the fact that it is still the major reason for a yond that achieved based solely on image interpretation.
Diseases of the Pancreas, I: Pancreatitis
Table 5. Differential criteria for chronic pancreatitis (CP) versus
pancreatic cancer (PCa)
CP PCa
History +++ – 9. Werner J, Schmidt J, Warshaw AL et al (1998) The relative
Duct Irregular Smooth
Duct/parenchyma <0.5 >0.5
Calcification +++ – mography and magnetic resonance imaging. Eur Radiol
Enhancement Diffuse Focal
Cysts +++ (+)
Lymph nodes (+) ++
Metastases – +++
Recently, the use of new methods and techniques, such
as oxygen insensitivity testing, have been described in con-
junction with conventional pathology studies of brush cy-
tology in facilitating discrimination between CP and pan-
creatic cancer [47, 48]. Nevertheless, to date, biopsy is the
most reliable diagnostic tool in ambiguous cases of pan-
creatic masses, with no significant difference whether the
biopsy was performed endoscopically or percutaneously,
while laparoscopic procedures are compromised by an in-
creased potential for tumor seeding and adverse events.
Pancreatitis in Children
Fortunately, pancreatitis in childhood is rare. The best
known causes for pediatric pancreatitis are traumas to the
pancreas (typically, a bicycle accident), the genetically
transmitted hereditary pancreatitis, and cystic fibrosis.
Nevertheless, in most of the cases the reason is unknown
(idiopathic pancreatitis), although microcalculi and pro-
tein plugs have been suggested. In general, the diagnosis
can be established based on the history, clinical findings,
and imaging findings, which are not different from those
described in adults [49-51].
References
1. Mitchell RM, Byrne MF, Baillie J (2003) Pancreatitis. Lancet
361:1447-1455
2. Glasbrenner B, Kahl S, Malfertheiner P (2002) Modern diag-
nostics of chronic pancreatitis. Eur J Gastroenterol Hepatol
14:935-941
3. Etemad B, Whitcomb DC (2001) Chronic pancreatitis: diag-
nosis, classification, and new genetic developments. Gastroen-
terology 120:682-707
4. Halonen KI, Leppaniemi AK, Puolakkainen PA et al (2000)
Severe acute pancreatitis: prognostic factors in 270 consecu-
tive patients. Pancreas 21:266-271
5. Bank S, Indaram A (1999) Causes of acute and recurrent pan-
creatitis. Clinical considerations and clues to diagnosis. Gas-
troenterol Clin North Am 28:571-89, viii
6. Cavallini G, Frulloni L, Bassi C et al (2004) Prospective mul-
ticentre survey on acute pancreatitis in Italy (ProInf-AISP): re-
sults on 1005 patients. Dig Liver Dis 36:205-211
7. Losanoff JE, Asparouhov OK, Jones JW (2001) Multiple fac-
tor scoring system for risk assessment of acute pancreatitis. J
Surg Res 101:73-78
87
8. Plock JA, Schmidt J, Anderson SE et al (2005) Contrast-en-
hanced computed tomography in acute pancreatitis: does con-
trast medium worsen its course due to impaired microcircula-
tion? Langenbecks Arch Surg 390:156-163
safety of MRI contrast agent in acute necrotizing pancreatitis.
Ann Surg 227:105-111
10. Robinson PJ, Sheridan MB (2000) Pancreatitis: computed to-
10:401-408
11. Pamuklar E, Semelka RC (2005) MR imaging of the pancreas.
Magn Reson Imaging Clin N Am 13:313-330
12. Laurens B, Leroy C, André A et al (2005) [Imaging of acute
pancreatitis]. J Radiol 86:733-46; quiz 746-747
13. Vaishali MD, Agarwal AK, Upadhyaya DN et al (2004) Mag-
netic resonance cholangiopancreatography in obstructive jaun-
dice. J Clin Gastroenterol 38:887-890
14. Arvanitakis M, Delhaye M, De Maertelaere V et al (2004) Com-
puted tomography and magnetic resonance imaging in the as-
sessment of acute pancreatitis. Gastroenterology 126:715-723
15. Akahane T, Kuriyama S, Matsumoto M et al (2003) Pancreat-
ic pleural effusion with a pancreaticopleural fistula diagnosed
by magnetic resonance cholangiopancreatography and cured
by somatostatin analogue treatment. Abdom Imaging 28:92-95
16. Sica GT, Miller FH, Rodriguez G et al (2002) Magnetic reso-
nance imaging in patients with pancreatitis: evaluation of sig-
nal intensity and enhancement changes. J Magn Reson Imag-
ing 15:275-284
17. Okai T, Fujii T, Ida M et al (2002) EUS and ERCP features of
nonalcoholic duct-destructive, mass-forming pancreatitis be-
fore and after treatment with prednisolone. Abdom Imaging
27:74-76
18. Taylor SL, Morgan DL, Denson KD et al (2005) A compari-
son of the Ranson, Glasgow, and APACHE II scoring systems
to a multiple organ system score in predicting patient outcome
in pancreatitis. Am J Surg 189:219-222
19. Mentula P, Kylänpää ML, Kemppainen E et al (2005) Early
prediction of organ failure by combined markers in patients
with acute pancreatitis. Br J Surg 92:68-75
20. Leung TK, Lee CM, Lin SY et al (2005) Balthazar computed
tomography severity index is superior to Ranson criteria and
APACHE II scoring system in predicting acute pancreatitis
outcome. World J Gastroenterol 11:6049-6052
21. Johnson CD, Toh SK, Campbell MJ (2004) Combination of
APACHE-II score and an obesity score (APACHE-O) for the
prediction of severe acute pancreatitis. Pancreatology 4:1-6
22. Gerlach H (2004) Risk management in patients with severe
acute pancreatitis. Crit Care 8:430-432
23. King NK, Powell JJ, Redhead D, Siriwardena AK (2003) A sim-
plified method for computed tomographic estimation of prog-
nosis in acute pancreatitis. Scand J Gastroenterol 38:433-436
24. Triester SL, Kowdley KV (2002) Prognostic factors in acute
pancreatitis. J Clin Gastroenterol 34:167-176
25. Sandberg AA, Borgstrom A (2002) Early prediction of severi-
ty in acute pancreatitis. Is this possible? Jop 3:116-125
26. Balthazar EJ (2002) Staging of acute pancreatitis. Radiol Clin
North Am 40:1199-1209
27. Balthazar EJ (2002) Acute pancreatitis: assessment of severity
with clinical and CT evaluation. Radiology 223:603-513
28. Mortele KJ, Wiesner W, Intriere L et al (2004) A modified CT
severity index for evaluating acute pancreatitis: improved cor-
relation with patient outcome. AJR Am J Roentgenol 183:
1261-1265
29. Mofidi R, Patil PV, Suttie SA, Parks RW (2009) Risk assess-
ment in acute pancreatitis. Br J Surg 96:137-150
30. Irie H, Honda H, Kuroiwa T et al (1998) MRI of groove pan-
creatitis. J Comput Assist Tomogr 22:651-5
31. Gabata T, Kadoya M, Terayama N et al (2003) Groove pancre-
atic carcinomas: radiological and pathological findings. Eur
Radiol 13:1679-1684
88
32. Shanbhogue AK, Fasih N, Surabhi VR et al (2009) A clinical
and radiologic review of uncommon types and causes of pan-
creatitis. Radiographics 29:1003-1026
33. Triantopoulou C, Dervenis C, Giannakou N et al (2009)
Groove pancreatitis: a diagnostic challenge. Eur Radiol
19:1736-1743
34. Okazaki K (2005) Autoimmune pancreatitis: etiology, patho-
genesis, clinical findings and treatment. The Japanese experi-
ence. Jop 6(1 Suppl):89-96
35. Takahashi N, Fletcher JG, Hough DM et al (2009) Autoim-
mune pancreatitis: differentiation from pancreatic carcinoma
and normal pancreas on the basis of enhancement characteris-
tics at dual-phase CT. AJR Am J Roentgenol 193:479-484
36. Weili L, Jiaguo W (2009) Education and imaging: Hepatobil-
iary and pancreatic: autoimmune pancreatitis. J Gastroenterol
Hepatol 24:1574
37. Sahani DV, Kalva SP, Farrell J et al (2004) Autoimmune pan-
creatitis: imaging features. Radiology 233:345-352
38. Farrell JJ, Garber J, Sahani D, Brugge WR (2004) EUS find-
ings in patients with autoimmune pancreatitis. Gastrointest
Endosc 60:927-936
39. Wakabayashi T, Kawaura Y, Satomura Y et al (2003) Clinical
and imaging features of autoimmune pancreatitis with focal
pancreatic swelling or mass formation: comparison with so-
called tumor-forming pancreatitis and pancreatic carcinoma.
Am J Gastroenterol 98:2679-2687
40. Ito K, Koike S, Matsunaga N (2001) MR imaging of pancre-
atic diseases. Eur J Radiol 38:78-93
41. Irie H, Honda H, Baba S et al (1998) Autoimmune pancreati-
tis: CT and MR characteristics. AJR Am J Roentgenol
170:1323-1327
Thomas Helmberger
42. Sahani DV, Sainani NI, Deshpande V et al (2009) Autoimmune
pancreatitis: disease evolution, staging, response assessment,
and CT features that predict response to corticosteroid thera-
py. Radiology 250:118-129
43. Sainani NI, Conwell DL (2009) Secretin-enhanced MRCP:
proceed with cautious optimism. Am J Gastroenterol
104:1787-1789
44. Yamada Y, Mori H, Matsumoto S et al (2009) Pancreatic ade-
nocarcinoma versus chronic pancreatitis: differentiation with
triple-phase helical CT. Abdom Imaging 22
45. Lowenfels AB, Maisonneuve P, Cavallini G et al (1994) Prog-
nosis of chronic pancreatitis: an international multicenter
study. International Pancreatitis Study Group. Am J Gastroen-
terol 89:1467-1471
46. Tang S, Huang G, Liu J et al (2009) Usefulness of (18)F-FDG
PET, combined FDG-PET/CT and EUS in diagnosing primary
pancreatic carcinoma: A meta-analysis. Eur J Radiol (in press)
47. Cho SG, Lee DH, Lee KY et al (2005) Differentiation of
chronic focal pancreatitis from pancreatic carcinoma by in vi-
vo proton magnetic resonance spectroscopy. J Comput Assist
Tomogr 29:163-9
48. van Kouwen MC, Jansen JB, van Goor H et al (2005) FDG-
PET is able to detect pancreatic carcinoma in chronic pancre-
atitis. Eur J Nucl Med Mol Imaging 32:399-404
49. Manfredi R, Lucidi V, Gui B et al (2002) Idiopathic chronic
pancreatitis in children: MR cholangiopancreatography after
secretin administration. Radiology 224:675-682
50. DeBanto JR, Goday PS, Pedroso MR et al (2002) Acute pan-
creatitis in children. Am J Gastroenterol 97:1726-1731
51. Levy MJ, Geenen JE (2001) Idiopathic acute recurrent pan-
creatitis. Am J Gastroenterol 96:2540-2555

Diseases of the Pancreas, II: Tumors
Ruedi F. Thoeni
Department of Radiology and Biomedical Imaging, University of California, San Francisco, CA, USA
Introduction
In the imaging of pancreatic disease and in assessment of
the etiology of jaundice, abdominal ultrasound (US) and
computed tomography (CT) traditionally have been em-
ployed [1]. These two methods are widely available and
have the advantages of their familiarity to radiologists and
clinicians and their non-invasiveness. With the introduction
of magnetic resonance imaging (MRI), magnetic resonance
cholangiopancreatography (MRCP), and endoscopic ultra-
sound (EUS), visualization of the pancreatic and biliary
ducts improved, allowing tumors to be more accurately
staged and safely sampled [2-12]. This led to a diminishing
role for endoscopic retrograde cholangiography in the diag-
nostic arena but its therapeutic use has remained unchal-
lenged. In recent years, technological advances with multi-
detector row CT (MDCT) imaging have improved the abil-
ity of CT to detect even small lesions in the pancreas and to
stage pancreatic tumors more accurately [13]. Microbubble
contrast enhancement and secretin stimulation have in-
creased the diagnostic acumen of US and MRI, respective-
ly, and may widen the utility of these techniques [14, 15].
Nevertheless, MDCT remains the primary tool in as-
sessing patients with suspected pancreatic disease, while
EUS and MRI are used as problem-solving modalities to
confirm suspected lesions not identified with CT, to find
additional lesions, and to obtain a definitive tissue diag-
nosis with EUS-guided tissue sampling. In recent years,
position emission tomography/CT (PET/CT) has been in-
creasingly employed in the assessment of patients with
suspected pancreatic tumors but its ultimate role still
needs further definition [16-21]. Also, somatostatin re-
ceptor scintigraphy has gained popularity in recent years
for neuroendocrine tumors [22, 23]. This discussion will
focus on diagnosing and staging the various pancreatic
neoplasms with CT and MRI, mentioning EUS, PET/CT,
and somatostatin receptor scintigraphy where appropriate.
Ductal Adenocarcinoma of the Pancreas
About 90% of all neoplasms of the pancreas are ductal
adenocarcinomas. Pancreatic adenocarcinoma is one of
IDKD 2010-2013
the leading causes of cancer death in the western world,
and the overall relative 5-year survival rate of only 5.5%
is dismal [24]. Late clinical presentation with advanced
disease and the aggressiveness of the tumor result in a
low rate of surgical intervention and overall poor out-
come. It is estimated that 42,470 men and women will be
diagnosed with cancer of the pancreas in 2009 and that
over 80% of these patients will die of the disease [25].
The tumor serum marker CA 19-9 is sensitive, although
not specific for the diagnosis of adenocarcinoma of the
pancreas. The treatment approach is based on whether the
tumor can or cannot be resected at presentation. There-
fore, imaging plays a crucial role in disease management.
The initial diagnosis of pancreatic tumor, particularly
if the patient presents with jaundice and the tumor is lo-
cated in the head of the pancreas, may be made by US.
The ultrasonographic signs of pancreatic carcinoma in-
clude a focal or diffuse pancreatic mass that is hypo-
echoic relative to normal gland parenchyma and dilata-
tion of the pancreatic duct without or with biliary duct
distention (double-duct sign). The accuracy of US for de-
tecting the level of bile duct obstruction varies greatly,
and ultrasonographic staging of pancreatic carcinoma is
inferior to that of CT. Ultrasonography often fails to pro-
vide an adequate examination of the entire gland, result-
ing in an overall decrease in the sensitivity of this tech-
nique. Some of these limitations are overcome by en-
dosonography, but tumors in the tail of the pancreas are
not always accessible by EUS.
For optimal evaluation of pancreatic neoplasms, MDCT
is the modality of choice. A triple-phase protocol is rec-
ommended that includes thin sections (0.625 or 1.25 mm)
through the abdomen, initially without intravenous contrast
followed by a rapidly delivered bolus of contrast material
(we use bolus tracking and 150 mL at 5 mL/s chased by
50 mL of saline). It is best to use a neutral oral contrast
agent (water or VoLumen, Bracco Diagnostics, USA) be-
cause it permits optimal assessment of tumor extension to
the stomach and/or duodenum and does not interfere with
the determination of vascular invasion. We recommend a
scan delay of 40-45 s (10-s delay from peak aortic enhance-
ment) for the late arterial or pancreatic phase and a scan
delay of 80 s for the hepatic phase. Rarely, an arterial phase
90 Ruedi F. Thoeni

at 20-25 s is performed if requested by surgery [26]. Arte-
rial involvement and tumor mass are best detected in the
pancreatic phase whereas the hepatic phase enables opti-
mal visualization of the liver, veins, and the entire ab-
domen in the search for liver metastases and peritoneal
seeding. One study demonstrated that a single-phase thin-
slice MDCT technique is sufficient for accurately assess-
ing the resectability of pancreatic adenocarcinoma [27].
Pancreatic adenocarcinoma arises from the pancreatic
duct. On MDCT, the tumor usually appears as a low-den-
sity mass, often associated with poorly defined margins
(Fig. 1) and pancreatic and/or bile duct dilatation. The
low-density central zone represents either hypovascular,
scirrhous tumor surrounded by normal parenchyma or in-
flammatory tissue caused by obstructive pancreatitis. Oc-
casionally, cystic degeneration is seen within the tumor
[28]. Neoplastic pancreatic duct obstruction frequently
produces a dilated duct as well as atrophy of the pancre-
atic parenchyma proximal to the neoplasm. Tumor ob-
struction of the main pancreatic duct can lead to rupture
of the side branches, resulting in the formation of
pseudocysts. Occasionally, a low-density mass cannot be
identified because the tumor is isodense to the surround-
ing normal parenchyma. In these cases, a dilated duct
with abrupt cut-off is often seen proximal to a small im-
perceptible tumor mass. Ancillary findings are local tu-
mor extension, including direct invasion of neighboring
organs such as the liver and the stomach, the arteries (loss
of fat planes surrounding celiac axis, superior mesenteric
artery, etc., vascular “cuffing”) and veins (tear-drop sign,
flattening, irregularity of margins, etc., of the portal vein,
superior mesenteric vein and its branches), and metastat-
ic disease to local lymph nodes, as well as spread to the
liver, peritoneum (often associated with ascites), and
more distant sites. The double-duct sign (dilatation of the
biliary and pancreatic ducts) occurs in <5% of patients
with pancreatic carcinoma. Biductal obstruction is a non-
specific sign and may also be seen in bile duct or am-
pullary carcinoma, pancreatitis, and ampullary stenosis.
For MRI, T1-weighted fat-suppressed sequences and
dynamic gadolinium-enhanced spoiled gradient echo
(SPGR) sequences are superior to T2-weighted se-
quences, as most pancreatic carcinomas have a signifi-
cant desmoplastic reaction that renders the tumor less
conspicuous on T2-weighted images. T1-weighted fat-
suppressed images using an early (arterial) gadolinium-
enhanced 3D vascular time-of-flight SPGR sequence pro-
vide optimal delineation of the tumor, particularly if it is
small and does not change the contour of the pancreas.
Diffusion-weighted imaging appears promising, especial-
ly for metastases to the liver. MRCP sequences consist-
ing of thin and thick axial and coronal sequences with
heavy T2-weighting and breath-holding are often added
to better assess the pancreatic and biliary ducts.
The CT imaging results for pancreatic carcinoma vary
widely, but with the current generation of scanners and
state-of-the-art scanning techniques a sensitivity of >90%
for detecting pancreatic carcinoma can be achieved [1, 27].
Small metastatic implants on the liver and peritoneum are
the lesions most likely to be missed by MDCT. MDCT
generally provides accurate information on vascular in-
volvement as long as a pancreatic protocol is observed; for
resectability, sensitivities of >80% have been obtained [1].
The positive predictive values for predicting unresectabili-
ty are much better than those for predicting resectability.
Presently, most studies show a slight advantage of MDCT

a b

Fig. 1 a, b. a Thin-section (1.25 mm) axial MDCT of a pancreas carcinoma in the pancreatic phase (~40 s). A low-attenuation mass is ap-
parent in the head of pancreas near the uncinate process (white arrows), with encasement of the replaced right hepatic artery originating
from the superior mesenteric artery (arrowhead). Retroperitoneal lymphadenopathy also is seen (black arrow). The mass is easily distin-
guished from adjacent normal pancreas. b Thin-section (1.25 mm) axial MDCT of a pancreas carcinoma in the hepatic phase (~80 s) in
the same patient. Note the teardrop-shaped superior mesenteric vein (black arrowhead) as a sign of venous encasement. The low-attenua-
tion pancreatic neoplasm (arrows) is less well seen
Diseases of the Pancreas, II: Tumors 91

over MR for detecting and staging pancreatic adenocarci- and has been shown to improve patient management before
nomas. A meta-analysis comparing CT, MRI, and US possible resection. In one PET/CT study, management was
demonstrated a sensitivity and specificity of 91 and 85% changed in 16% of patients with pancreatic cancers that
for helical CT and a sensitivity and specificity of 84 and were initially staged as being resectable [17]. In suspected
82% for MRI whereas the results for resectability were tumor recurrence, PET reliably detected local recurrence
similar [1]. For US, the sensitivity for diagnosing pancreas and was advantageous in diagnosing distant disease [18].
carcinoma and the specificity for determining resectabili-
ty are much lower. The advantage of MRI is thought to be
in the area of small tumors that do not alter the contour of Neuroendocrine Neoplasms of the Pancreas
the gland [10] and in detecting hepatic metastases. At pre-
sent, MRI appears to be a problem-solving modality. It Hyperfunctioning Neuroendocrine Neoplasms
should be considered in patients with suspected pancreatic Among hyperfunctioning or syndromic neuroendocrine
neoplasms in the presence of: (1) allergy to iodine contrast neoplams (NEN, formerly called islet cell tumors) of the
or other contraindications to iodine contrast administra- pancreas, insulinoma is the most common followed by
tion, (2) a MDCT scan showing focal enlargement of the gastrinoma, glucagonoma, VIPoma, and other rarely en-
pancreas but no definable mass, (3) a clinical history sug- countered secretory neoplasms. In functioning pancre-
gesting malignancy and MDCT images that are equivocal atic adenomas, the clinical diagnosis is based on clinical
or difficult to interpret, and (4) when distinction between data and laboratory tests that usually permit an accurate
chronic pancreatitis with focal enlargement and pancreatic diagnosis, with cross-sectional imaging used only for
cancer is needed. When choosing an imaging modality, one localizing the pancreatic neoplasm [31].
has to take into account that, today, MDCT of the pancreas Insulinomas, and especially extrapancreatic NENs,
requires a small fraction of the time needed for a complete that are small and located in the duodenal or gastric wall
MRI study of the pancreas. (Fig. 2) may be difficult to detect pre-operatively by any
False-positive MDCT diagnoses of pancreatic cancer
can occur, especially in patients with chronic pancreati-
tis; therefore, percutaneous aspiration biopsies are need- a
ed if non-operative treatment is planned. Fine-needle as-
piration biopsy of pancreatic cancer under CT guidance
is a frequently performed procedure and is associated
with severe pancreatitis in <3%. The sensitivity of percu-
taneous CT biopsies reaches 79%, with a positive predic-
tive value of 100% and a negative predictive value of
47% [7]. However, because of possible tumor seeding in
the needle tract, patients with potentially resectable tu-
mors (only 10% of all cases) who are acceptable candi-
dates for surgery should undergo exploratory surgery [7].
While EUS excels in detecting even small pancreatic
adenocarcinomas, reaching sensitivities as high as 97% [9],
and can be used in the differential diagnosis of pancreatic
tumors [29], it demonstrates poor sensitivity and specifici-
ty for diagnosing vascular involvement by the tumor [30].
The technique suffers from limited depth penetration. To- b
day, endoscopic biopsies often replace percutaneous CT
biopsies of the pancreas. They have a sensitivity of 80%
with a positive predictive value of 99% and a negative pre-
dictive value of 73%. They are particularly indicated when
CT is equivocal or negative despite a strong clinical suspi-
cion for tumor and when the lesion is <3 cm in size [7].
PET, and particularly PET/CT, has emerged as an im-
portant modality for effectively managing patients with
suspected pancreatic cancer. Nevertheless, more studies
are needed to demonstrate its true value and cost-effec-
tiveness since at least one study found no benefit over CT Fig. 2 a, b. a Thin-section (1.25 mm) axial MDCT of an ectopic
alone [16]. It was also reported that if helical CT was pos- neuroendocrine tumor of the pancreas (insulinoma) in the pancre-
itive for pancreas carcinoma, PET had a sensitivity of 92% atic phase. A small hypervascular mass (arrow) is seen at the junc-
tion of the second to third duodenum. b Thin-section (1.25 mm)
and a specificity of 68%; if CT was negative, the sensitiv- coronal MDCT of an ectopic neuroendocrine tumor of the pan-
ity of PET was 73% and the specificity 86% [19]. PET/CT creas (insulinoma) in the pancreatic phase, which clearly demon-
allows hot tracer spots to be to more precisely localized strates the mass in the duodenal wall (arrow)
92
of the radiographic techniques, and even intra-operative
US fails to detect some of these lesions. However, MD-
CT with 0.5-0.625 mm sections has improved the results.
These ectopic lesions are more likely to occur in patients
with multiple endocrine adenomatosis (MEA) or multiple
endocrine neoplasia (MEN). A combination of intra-
operative palpation and intra-operative US was found to
achieve the best results during surgery. Intra-operative
US is particular important in patients with multiple le-
sions and MEN.
On MDCT and MRI, functioning NENs generally
show intense enhancement in the arterial phase with rapid
washout in the portal venous phase. The most common
NEN, the insulinoma, usually is small (≤2 cm in diame-
ter) and seldom metastasizes (5-10% of cases). All other
NENs tend to be large and frequently have metastases
(60-65% of cases). The appearance of liver metastases in
patients with functioning NEN is similar to that of the
primary tumor.
The reported sensitivity of conventional CT for de-
tecting an insulinoma ranges from 28 to 79% with a
mean of 38%. It is slightly higher for gastrinomas, due
primarily to their larger size. A dual-phase MDCT pro-
tocol with thin sections improves the detection rate to
94% and reaches 100% if combined with EUS [32]. The
latter modality provides excellent results in the head of
the pancreas, but the results are less convincing for the
tail of the pancreas because of its distance from the
stomach. EUS usually allows the detection of even small
NENs and their precise location. Ectopic gastrinomas
may be missed by EUS but combining this technique
with somatostatin receptor scintigraphy (SSR or Octreo-
scan) increases the overall sensitivity for gastrinomas
[33]. The sensitivity of transabdominal US for detecting
insulinomas is low (mean of 46%) and therefore should
not be used for this purpose.
Functioning NENs of the pancreas are of low signal in-
tensity on T1-weighted images and of high signal inten-
sity on T2-weighted images [5]. Occasionally, an insuli-
noma is of dark signal intensity on T2-weighted se-
quences due to a fibrous stroma. In our study, the MRI
sensitivity for detecting functioning NENs of ≤2 cm in
diameter reached 85%, which is similar to the sensitivity
achieved by invasive procedures [5]. For gastrinomas, a
MRI sensitivity of up to 62% has been reported [34].
With present techniques, MRI should detect lesions
>2 cm with a sensitivity of over 85%. Therefore, MRI
with state-of-the-art equipment and optimal imaging
techniques appears to be a useful strategy for diagnosing
small pancreatic NENs. Nevertheless, contrast-enhanced
MDCT, with its superior spatial resolution and very thin
sections, currently surpasses MRI in diagnosing these
small neoplasms.
SSR with various derivatives of long-acting somato-
statin analogues [22] can be used for small gastrinomas,
somatostatinoma, glucagonoma, carcinoid, and VIPoma,
but insulinomas may be missed due to reduced receptor
expression [22, 33]. SSR with 111In-octreotide can
Ruedi F. Thoeni
often diagnose small lesions only suspected on CT or
MRI and detect metastases not diagnosed with other
modalities.
Non-hyperfunctioning Neuroendocrine Neoplasms
Non-hyperfunctioning or non-syndromic NENs are less
frequently encountered than insulinomas or gastrinomas,
representing 15-25% of these neoplasms [31]. While they
arise from the alpha or beta cells of the pancreas, these
neoplasms are hormonally quiescent (probably very min-
imal secretion) and often present as a mass with or with-
out jaundice or gastric outlet obstruction. These tumors
are usually located in the pancreatic head and can mea-
sure up to 20 cm in diameter. There may be solid and
necrotic components, with coarse calcifications present
in up to 25%. The mass is hypervascular with a late cap-
illary stain. The tumor does not encase vessels but in
80-100% there is malignant transformation, with liver
metastases and adenopathy. The cumulative 5-year sur-
vival is 52-58% [35]. The key features of non-function-
ing NENs are their large size, hypervascularity, and the
absence of vascular encasement. Results with CT and
MRI are similar.
Cystic Neoplasms of the Pancreas
Serous and Mucinous Cystic Neoplasms of the Pancreas
Cystic neoplasms of the pancreas are uncommon tumors
and account for <5% of pancreatic neoplasms. Pancre-
atic cystic neoplasms are classified into two categories:
(1) serous cystic (usually microcystic, occasionally
macrocystic: unilocular or oligocystic) neoplasms that
are benign; and (2) mucinous cystic (macrocystic) neo-
plasms that are potentially malignant or already malig-
nant at the time of diagnosis. The rare serous macro-
cystic variant is benign and exhibits radiological fea-
tures similar to those of mucinous cystadenoma. Serous
and mucinous cystic neoplasms, except for intraductal
papillary mucinous tumors (IPMT), do not communi-
cate with the pancreatic duct.
Serous cystic neoplasms of the pancreas are observed
in middle-aged and elderly women. This type of tumor
may not require surgical treatment whereas mucinous
cystic tumors should be resected because of their malig-
nant potential. Nevertheless, some surgeons prefer to re-
sect the serous type as well. In general, the patient’s age,
symptoms, and overall condition, the lesion’s location,
and its growth over time are factors that help to decide
whether surgery is needed [36]. Often, any cyst that in-
creases in size over time, any symptomatic cyst, and cysts
in older fit patients are selected for surgery. CT can ac-
complish pre-operative differentiation of the two types in
many cases. In serous cystic tumors, traditionally the di-
agnosis is made if the number of cysts within the tumor
is more than six and the diameters of the cysts <2 cm. A
Diseases of the Pancreas, II: Tumors
newer nomenclature prefers to call cysts ≤1 cm definite-
ly serous, those >1-2 cm equivocal, and those >2 cm def-
initely mucinous. Grossly, these serous tumors appear ei-
ther as solid tumors with innumerable tiny cysts or as
honeycombed cystic tumors, depending on the amount of
connective tissue. They have a lobulated margin (Fig. 3).
At times, it is difficult to visualize the cystic areas. Cal-
cifications in serous tumors are central in location. A cen-
tral enhancing scar may be present and is characteristic of
a serous tumor [28].
Mucinous cystic neoplasms of the pancreas (also
called cystadenomas and cystadenocarcinomas according
to the old nomenclature) have six or fewer cysts, the di-
ameters of the cysts measure >2 cm, a central enhancing
scar is rarely seen, and calcifications are peripheral [28].
The margins usually are smooth and metastatic disease
may be present at the time of diagnosis.
Based on the above-mentioned criteria, a correct diag-
nosis of a serous cystic pancreatic tumor can be made in
62% of patients by CT, in 74% by sonography, and in
84% using both modalities [37]. Overall, the results for
mucinous cystic tumors are inferior. Pancreatic pseudo-
cysts and cystic forms of islet cell tumors, ductal carci-
nomas, solid and papillary tumors, and lymphangioma of
the pancreas can be indistinguishable on CT from cystic
neoplasms. Thus, EUS needle biopsies of the lesions of-
ten are necessary [38].
Better definition of the internal architecture of these
cystic neoplasms is frequently obtained with MRI rather
than CT. MRI also demonstrates the presence of mucin,
seen as an area of increased signal intensity within the
cysts on T1-weighted sequences. Septa and wall thick-
ness of the lesions are well demonstrated by MRI but this
is not always true for calcifications. MRI is of great help
in distinguishing these cystic neoplasms from pseudo-
Fig. 3. Thin-section (1.25 mm) axial MDCT of a serous cystic neo-
plasm of the pancreas in the pancreatic phase. A lobulated and sep-
tated cystic mass is present in the tail of the pancreas (arrows). The
individual cysts are small and the septa barely perceptible
93
cysts of the pancreas, particularly if they are multiple.
Both MRCP and MDCT with curved planar reconstruc-
tion can demonstrate the absence of a connection to the
main pancreatic duct.
Intraductal Papillary Mucinous Neoplasm of the Pancreas
A rare tumor that is considered a subtype of the mucinous
cystic neoplasms of the pancreas is the intraductal papil-
lary mucinous tumor of the pancreas (IPMN, formerly al-
so called ductectatic cystadenoma or ductectatic cystade-
nocarcinoma). IPMN can be classified as main duct,
branch duct (side-branch), or mixed type depending on
the site and extent of involvement [39]. The cystic
changes always demonstrate a connection to the pancre-
atic duct, which is a diagnostic feature that can be seen
on MDCT and MRI. The branch duct tumor consists of
cystic dilation of the side branches of the pancreatic duct,
usually in the uncinate process. These ducts are lined
with atypical, hyperplastic, or clearly malignant epitheli-
um. In the late stages, the tumor nodules of the ducts pro-
duce copious mucinous secretions that fill the entire duct.
Since extension into the parenchyma and beyond occurs
relatively late in branch duct IPMN and overall malignant
degeneration is rare, the overall prognosis is good. In 25-
44% of resected specimens of the other two types, ma-
lignancy is present. Resection is therefore the treatment
of choice in these patients.
CT shows markedly dilated ducts and cystic-appearing
structures filled with mucinous material, which has a
slightly higher attenuation than that of water. MRI seems
to have a slight advantage over CT because it can visu-
alize mucin within the cysts as well as the internal
architecture of the lesion, including a solid mass and
mural nodules (which are signs of malignancy) slightly
better than CT. EUS also is well suited to detect mural
nodules.
Solid Pseudopapillary Epithelial Neoplasm of the Pancreas
Solid pseudopapillary epithelial neoplasms, previously
called solid and cystic tumors of the pancreas, are rare tu-
mors. They are seen almost exclusively in young women
and are located mostly in the tail of the pancreas. This
neoplasm is characterized by a solid peripheral area of tu-
mor and a central zone of degeneration consisting of he-
morrhage and cystic spaces filled with necrotic debris
that can be visualized by CT (Fig. 4) and MRI. On imag-
ing, these tumors appear as sharply defined, heteroge-
neous, large cystic pancreatic masses with solid compo-
nents. They usually are benign but in older women they
may be malignant [40]. Calcifications may be present in
the capsule or in the inner portion of the mass. EUS also
can be helpful in visualizing the nodules and the internal
architecture of these masses.
94
Fig. 4. Thin-section (1.25 mm) axial MDCT of a solid pseudopap-
illary epithelial neoplasm of the pancreas in the pancreatic phase.
A large enhancing mass (arrows) with partial necrosis is seen
in the tail of the pancreas. The origin from the pancreas is clearly
visualized on this image
Follow-Up for Cystic Neoplasms of the Pancreas
Small lesions (≤3 cm) that are asymptomatic, show no
signs of malignancy, and have negative fine-needle aspi-
ration biopsy results can be followed every 6 months for
one year and then yearly for a total of 4 years. If the cyst
becomes symptomatic, increases in size during observa-
tion, shows malignant features including a thick wall,
multiple irregular septation, and nodules, and/or has in-
creased CEA or CA 19.9, positive cytology, or mucin in
the aspirate, it should be surgically removed. Often, cysts
detected in the elderly and fit patient are removed re-
gardless of the features because of the increased inci-
dence of malignancy in these lesions.
References
1. Bipat S, Phoa SS, van Delden OM et al (2005) Ultrasonogra-
phy, computed tomography and magnetic resonance imaging
for diagnosis and determining resectability of pancreatic ade-
nocarcinoma: a meta-analysis. J Comput Assist Tomogr 29:
438-445
2. Midwinter MJ, Beveridge CJ, Wilsdon JB et al (1999) Corre-
lation between spiral computed tomography, endoscopic ultra-
sonography and findings at operation in pancreatic and am-
pullary tumours. British J Surg 86:189-193
3. McLean AM, Fairclough PD (2005) Endoscopic ultrasound in
the localisation of pancreatic islet cell tumours. Best Pract Res
Clin Endocrinol Metab 19:177-193
4. Canto MI, Goggins M, Yeo CJ et al (2004) Screening for pan-
creatic neoplasia in high-risk individuals: an EUS-based ap-
proach. Clin Gastroenterol Hepatol 2:606-621
5. Thoeni RF, Mueller-Lisse UG, Chan R et al (2000) Detection
of small, functional islet cell tumors in the pancreas: Selection
of MRI Imaging sequences for optimal sensitivity. Radiology
214:483-490
Ruedi F. Thoeni
6. Semelka RC, Kroeker MA, Shoenut JP et al (1991) Pancreat-
ic disease: prospective comparison of CT, ERCP, and 1.5 T MR
imaging with dynamic Gadolinium enhancement and fat sup-
pression. Radiology 181:785-791
7. Volmar KE, Vollmer RT, Jowell PS et al (2005) Pancreatic
FNA in 1000 cases: a comparison of imaging modalities. Gas-
trointest Endosc 61:854-861
8. Ardengh JC, de Paulo GA, Ferrari AP (2004) EUS-guided
FNA in the diagnosis of pancreatic neuroendocrine tumors be-
fore surgery. Gastrointest Endosc 60:378-384
9. Maguchi H (2004) The roles of endoscopic ultrasonography in
the diagnosis of pancreatic tumors. J Hepatobiliary Pancreat
Surg 11:1-3
10. Vachiranubhap B, Kim YH, Balci NC et al (2009) Magnetic
resonance imaging of adenocarcinoma of the pancreas. Top
Magn Reson Imaging 20:3-9
11. Manfredi R, Graziani R, Motton M et al (2009) Main pancre-
atic duct intraductal papillary mucinous neoplasms: accuracy
of MR imaging in differentiation between benign and malig-
nant tumors compared with histopathologic analysis. Radiolo-
gy 253:106-115
12. Ku YM, Shin SS, Lee CH et al (2009) Magnetic resonance
imaging of cystic and endocrine pancreatic neoplasms. Top
Magn Reson Imaging 20:11-18
13. Brennan DD, Zamboni GA, Raptopoulos VD et al (2007)
Comprehensive preoperative assessment of pancreatic adeno-
carcinoma with 64-section volumetric CT. Radiographics
27:1653-1666.
14. D’Onofrio M, Zamboni G, Faccioli N et al (2007) Ultrasono-
graphy of the pancreas. 4. Contrast-enhanced imaging. Abdom
Imaging 32:171-181
15. Akisik MF, Sandrasegaran K, Aisen AA et al (2006) Dynamic
secretin-enhanced MR cholangiopancreatography. Radio-
graphics 26:665-677
16. Lytras D, Connor S, Bosonnet L et al (2005) Positron emission
tomography does not add to computed tomography for the di-
agnosis and staging of pancreatic cancer. Dig Surg 22:55-61
17. Heinrich S, Goerres GW, Schafer M et al (2005) Positron
emission tomography/computed tomography influences on the
management of resectable pancreatic cancer and its cost-effec-
tiveness. Ann Surg 242:235-243
18. Ruf J, Lopez Hanninen E, Oettle H et al (2005) Detection of
recurrent pancreatic cancer: comparison of FDG-PET with
CT/MRI. Pancreatology 5:266-272
19. Orlando LA, Kulasingam SL, Matchar DB (2004) Meta-analy-
sis: the detection of pancreatic malignancy with positron emis-
sion tomography. Aliment Pharmacol Ther 20:1063-1070
20. Lee TY, Kim MH, Park do H et al (2009) Utility of 18F-FDG
PET/CT for differentiation of autoimmune pancreatitis with
atypical pancreatic imaging findings from pancreatic cancer.
AJR Am J Roentgenol 193:343-348
21. Farma JM, Santillan AA, Melis M et al (2008) PET/CT fusion
scan enhances CT staging in patients with pancreatic neo-
plasms. Ann Surg Oncol 15:2465-2471
22. Virgolini I, Traub-Weidinger T, Decristoforo C (2005) Nuclear
medicine in the detection and management of pancreatic islet-
cell tumours. Best Pract Res Clin Endocrinol Metab 19:213-227
23. Buchmann I, Henze M, Engelbrecht S et al (2007) Compari-
son of 68Ga-DOTATOC PET and 111In-DTPAOC (Octre-
oscan) SPECT in patients with neuroendocrine tumours. Eur J
Nucl Med Mol Imaging 34:1617-1726
24. Horner MJ, Ries LAG, Krapcho M et al (eds) posted to the SEER
web site (2009) SEER Cancer Statistics Review, 1975-2006, Na-
tional Cancer Institute. Bethesda, MD, based on November 2008
SEER data submission, http://seer.cancer.gov/csr/1975_2006
25. Cancer Facts & Figures (2009) American Cancer Society
(ACS), Atlanta, Georgia
26. Fletcher JG, Wiersema MJ, Farrell MA et al (2003) Pancreat-
ic malignancy: value of arterial, pancreatic, and hepatic phase
imaging with multi-detector row CT. Radiology 229:81-90
Diseases of the Pancreas, II: Tumors
27. Imbriaco M, Megibow AJ, Ragozzino A et al (2005) Value of
the single-phase technique in MDCT assessment of pancreat-
ic tumors. AJR Am J Roentgenol 184:1111-1117
28. Sahani DV, Kadavigere R, Saokar A et al (2005) Cystic pan-
creatic lesions: a simple imaging-based classification system
for guiding management. Radiographics 25:1471-1484
29. Morris-Stiff G, Webster P, Frost B et al (2009) Endoscopic
ultrasound reliably identifies chronic pancreatitis when
other imaging modalities have been non-diagnostic. JOP 10:
280-283
30. Aslanian H, Salem R, Lee J et al (2005) EUS diagnosis of vas-
cular invasion in pancreatic cancer: surgical and histologic cor-
relates. Am J Gastroenterol 100:1381-1385
31. Thoeni RF (2009) Imaging of endocrine tumors. In: Heiken JP
(ed) Pancreatic cancer. Contemporary issues in cancer imag-
ing. Cambridge University Press, Cambridge, pp 104-129
32. Gouya H, Vignaux O, Augui J et al (2003) CT, EUS combined
protocol for preoperative evaluation of pancreatic insulinoma.
AJR Am J Roentgenol 181:987-992
33. T. Zimmer, U. Stolzel, M. Bader et al (1996) Endoscopic ul-
trasonography and somatostatin receptor scintigraphy in the
preoperative localisation of insulinomas and gastrinomas. Gut
39:562-568
95
34. Pisegna JR, Doppman JL, Norton JA et al (1993) Prospective
comparative study of the ability of MR imaging and other
imaging modalities to localize tumors in patients with
Zollinger-Ellison syndrome. Dig Dis Sci 38:1318-1320
35. H. Liang H, Wang P, Wang XN et al (2004) Management of non-
functioning islet cell tumors. World J Gastroenterol 10:1806-1809
36. Spinelli KS, Fromwiller TE, Daniel RA et al (2004) Cystic pan-
creatic neoplasms: Observe or operate. Ann Surg 239:651-659
37. Procacci C, Graziani R, Bicego E et al (1997) Serous cystade-
noma of the pancreas: report of 30 cases with emphasis on the
imaging findings. J Comput Assist Tomogr 21:373-382
38. Belsley NA, Pitman MB, Lauwers GY et al (2008) Serous cys-
tadenoma of the pancreas: limitations and pitfalls of endo-
scopic ultrasound-guided fine-needle aspiration biopsy. Can-
cer 114:102-110
39. Ogawa H, Itoh S, Ikeda M et al (2008) Intraductal papillary
mucinous neoplasm of the pancreas: assessment of the likeli-
hood of invasiveness with multisection CT. Radiology
248:876-886
40. Lee JH, Yu JS, Kim H et al (2008) Solid pseudopapillary car-
cinoma of the pancreas: differentiation from benign solid
pseudopapillary tumour using CT and MRI. Clin Radiol
63:1006-1014

Adrenal Imaging and Intervention
William W. Mayo-Smith1, Isaac R. Francis2
1 Department of Radiology, Warren Alpert School of Medicine Brown University, Rhode Island Hospital, Providence, RI, USA
2 Department of Radiology, University of Michigan, Ann Arbor, MI, USA
Introduction
The objectives of this chapter are: (1) to describe the dif-
ferent work-ups for adrenal masses, depending on clinical
scenario, (2) define adrenal incidentaloma, (3) discuss the
relative risk factors for benign and malignant adrenal
masses, (4) describe the imaging techniques to differenti-
ate benign from malignant adrenal masses, and (5) discuss
the recommended medical work-up of an adrenal mass.
The adrenal gland is a complex organ that is made up
of the catecholamine-producing medulla and the steroid-
producing cortex. It is a common site of primary tumors
(functional and non-functional) and of metastases. The
optimal work-up for an adrenal mass depends on the pa-
tient’s clinical scenario and whether detection or charac-
terization is the primary concern. In general, it is useful
to separate adrenal work-ups into one of three algorithms:
1. Detection of an adrenal tumor in a patient with a
known biochemical abnormality.
2. Staging of a patient with a known primary neoplasm.
3. Characterization of an incidental adrenal mass detect-
ed on cross-sectional imaging.
This is a relevant topic as there has been an over 20-
fold increase in medical literature reports on this subject
in the past two decades.
Detection of Biochemically Active Adrenal Tumor
Biochemically active adrenal neoplasms originate in the
adrenal cortex, in which case they produce an excess of
either glucocorticoids, aldosterone, or androgens, or in the
adrenal medulla, thus producing an excess of cate-
cholamines. Cushing’s syndrome results from an over-
production of cortisol by the adrenal cortex and approxi-
mately 80% of these cases are due to stimulation of the
adrenal glands by a pituitary adenoma. A primary adren-
al cortical tumor is seen in 20% of patients with Cushing’s
syndrome and <1% have ectopic production of ACTH by
a non-pituitary neoplasm. The work-up of patients pre-
senting with Cushing’s syndrome involves a dexametha-
sone suppression test, pituitary magnetic resonance imag-
ing (MRI) to look for a pituitary adenoma, and an adrenal
IDKD 2010-2013
computed tomography (CT) study depending on the
suspected source of ACTH production. If pituitary and
adrenal neoplasms are ruled out and an ectopic source of
hormone secretion is suspected, then a chest and abdomi-
nal CT should be performed to search for it.
Conn’s syndrome can result from either adrenal hyper-
plasia or an adrenal cortical tumor producing elevated
levels of aldosterone, leading to increased sodium reten-
tion, hypertension, and potassium wasting. The diagnosis
is suspected in a hypertensive patient with low serum
potassium and is confirmed by measuring the ratio of
serum aldosterone to renin levels. When the diagnosis is
suspected based on biochemical assays, CT scans using
3-mm collimation targeted to the adrenals is useful to dif-
ferentiate a small adrenal neoplasm from bilateral hyper-
plasia. If findings are equivocal on CT, then adrenal
venous sampling to localize and lateralize the site of
elevated aldosterone production should be performed.
Pheochromocytomas originate from the adrenal medul-
la and produce an excess of catecholamines, causing hy-
pertension. These tumors are solitary and occur sporadi-
cally in 90% of cases. If the diagnosis is suspected, the
most appropriate first-line test is the measurement of
plasma metanephrines. If these are equivocal, urinary
metanephrines can be measured. Once the diagnosis has
been established biochemically, the primary role of the ra-
diologist is to determine the site of origin of the pheo-
chromocytoma. Over 95% of pheochromocytomas origi-
nate in the adrenals; therefore, a non-contrast-enhanced
CT examination of the adrenals is usually sufficient. MRI
of pheochromocytoma typically demonstrates a T2 hyper-
intense mass, although the finding is non-specific, as
pheochromocytomas can also have intermediate signal in-
tensity on T2-weighted images, thus simulating adrenal
cortical carcinoma; also, other adrenal lesions can be T2
hyperintense (adrenal cysts and lymphangiomas). MRI can
also detect extra-adrenal paragangliomas along the sympa-
thetic chain. While metaiodobenzyl-guanidine (MIBG)
scintigraphy has high specificity (>95%) for the diagnosis
of pheochromocytoma, its sensitivity is only 77-90%.
Recent studies have suggested that MIBG scintigraphy
should be used selectively and only in patients with famil-
ial or hereditary disorders, in the detection of metastatic
Adrenal Imaging and Intervention
disease, and in patients with biochemical evidence for
pheochromocytoma and negative CT or MRI. These stud-
ies also concluded that MIBG scintigraphy does not offer
any added advantage in patients with biochemical evi-
dence for a pheochromocytoma, no hereditary or familial
diseases, and a unilateral adrenal mass detected on CT or
MRI [1, 2].
The standard treatment of a biochemically active
adrenal tumor is open or laparoscopic resection. More re-
cently, non-invasive techniques have been described, in-
cluding selective arterial embolization, percutaneous in-
jection of acetic acid, and radiofrequency ablation.
Staging Patients with Known Carcinoma
Evaluation of the adrenal gland in the oncology patient is
complicated because the gland is a frequent site of metas-
tases, but benign adrenal adenomas are also common (de-
tected in 2-5% of autopsy series). Thus, the presence of
an adrenal mass does not necessarily implicate metas-
tases. The role of cross-sectional imaging in the oncolo-
gy patient is to detect enlargement of the adrenal gland
and characterize the enlargement as either benign or ma-
lignant. More recently, PET imaging has facilitated the
staging of neoplasms because adrenal metastases tend to
demonstrate increased activity, having a greater uptake
relative to the liver, while most benign adenomas do not.
More recent studies have confirmed the high sensitivity
of PET/CT in detecting malignant lesions but the speci-
ficity is lower (87-97%). This loss of specificity is attrib-
utable to a small number of adenomas and other benign
lesions that mimic malignant lesions [3, 4].
Depending on the primary tumor, CT or PET/CT is a
useful first-line exam to stage a known neoplasm. If the pa-
tient demonstrates multiple sites of metastatic disease, then
evaluation of an adrenal mass is not important. If the adren-
al mass is the only abnormality, further evaluation is re-
quired to differentiate an adenoma from a metastatic focus.
Currently, there are two main criteria (anatomical and
physiological) used to differentiate benign adenomas from
malignant adrenal masses: (1) the intracellular lipid con-
tent of the adrenal mass, which represents the anatomical
difference between adenomas and metastases, and (2) dif-
ferences in vascular enhancement patterns, which repre-
sent the physiological difference. Approximately 80% of
benign adenomas have abundant intracytoplasmic lipid in
the adrenal cortex and thus are of low density on unen-
hanced CT or show signal drop-off on out-of-phase chem-
ical shift MRI (CSMRI). Conversely, most metastases
have little intracytoplasmic lipid and thus do not have a
low density on non-contrast CT. At a threshold of 10 HU,
CT has a 71% sensitivity and 98% specificity for charac-
terizing adrenal adenomas. While a low HU is useful to
characterize lipid-rich adenomas, it is estimated that up to
20% of adenomas do not contain sufficient lipid to be of
low density on unenhanced CT [5-7]. More recently, Bae
et al. showed that a histogram analysis of adrenal masses
97
(evaluating microscopic levels of lipid on a pixel by pixel
basis) is useful to differentiate adenomas from metastases
on non-contrast- and contrast-enhanced CT [8]. Accord-
ing to their findings, if an adrenal mass has >10% nega-
tive pixels, it is diagnostic of an adenoma (on either non-
contrast- or contrast-enhanced CT). However, due to dif-
fering results in more recent studies, this approach is still
considered as “research” and its use in clinical practice re-
mains limited.
The physiological difference in perfusion between ade-
nomas and metastases can be used to differentiate these
entities. Adenomas enhance rapidly with intravenous
contrast (iodinated CT contrast or MR gadolinium
chelates) and also have rapid washout. Metastases also
enhance vigorously with dynamic contrast but the
washout of contrast is more prolonged than in adenomas.
This difference in contrast washout has been exploited to
further differentiate benign from malignant adrenal le-
sions by comparing pre-contrast HU values with dynam-
ic and 15-min delayed HU values [9, 10]. Absolute per-
cent washout (APW) values are calculated by the formu-
la: (HU at dynamic CT – HU at 15-min delayed CT)/(HU
at dynamic CT – HU at non-contrast CT) × 100. A value
≥60% is diagnostic of an adenoma. Relative percent
washout (RPW) is used when a non-contrast CT value is
not available and the dynamic enhanced values are com-
pared to 15-min delayed scans. RPW is calculated by the
formula: (HU dynamic CT – HU 15-min delayed
CT)/HU dynamic CT × 100, and a value >40% is diag-
nostic of adenoma.
Adenomas can be differentiated from metastases using
CSMRI if the patient has a non-diagnostic CT, is allergic
to iodinated contrast, or in young patients, in whom radi-
ation exposure is an issue [11, 12]. Most adrenal adeno-
mas contain sufficient intracellular lipid and lose signal
on the out-of-phase image compared to the spleen. Visu-
al analysis is adequate in most cases to make this obser-
vation, but quantitative methods, such as the signal in-
tensity index, may also be useful [13, 14].
If the CT, MRI, or PET findings are equivocal, adren-
al biopsy using CT guidance should be performed, par-
ticularly to stage a lung carcinoma in patient who has no
other sites of metastatic disease, as this may determine
whether surgical resection is a therapeutic option. The
role of adrenal biopsy has evolved in the last few years;
in addition to the above indication of an indeterminate
adrenal mass, adrenal biopsy can also be used to confirm
metastatic disease to the adrenal glands in patients with
suspected solitary adrenal metastasis. CT-guided biopsy
has been shown to be safe, with a diagnostic yield of 83-
96% and a 3% complication rate [15].
Evaluation of an Incidentally Discovered Adrenal Mass
As the indications for abdominal imaging (particularly
CT) continue to increase, so does the detection of the
incidental adrenal mass-given the high prevalence of
98
adenomas in the general population (3-7%) [16-18]. In
general, the overwhelming majority of incidentally dis-
covered adrenal masses (incidentalomas) are benign in a
patient with no known malignancy [19]. An adrenal inci-
dentaloma is defined as “an unsuspected and asympto-
matic mass (≥1 cm) detected on imaging exams obtained
for other purposes”.
Risk factors for an incidental adrenal mass being ma-
lignant include lesion size, change in size, and their
occurrence in a patient with a history of malignancy.
For patients with no history of malignancy, most small
(<4 cm) incidentally discovered adrenal masses are be-
nign, and an extensive and costly imaging workup is
usually not justified. An additional biochemical workup
to exclude functioning adrenal cortical or medullary le-
sions is recommended by endocrinologists; however,
these functioning lesions present very rarely as “inci-
dental” adrenal masses, such that this expensive workup
is not done routinely [20].
If a mass of any size has the typical features of a be-
nign lesion, such as a cyst or myelolipoma, no additional
workup or follow-up imaging is needed. For adrenal
lesions <4 cm with non-diagnostic imaging features, if
prior imaging is available and the lesion has been stable
for at least 1 year, then it can be deemed benign and there
is no need for additional imaging follow-up. But if the
lesion is enlarging, then it may be prudent to proceed
to adrenal biopsy or resection. If there are no prior com-
parison CT or MRI exams and patient has no hystory of
malignancy, and if the lesion has benign imaging fea-
tures (homogeneous, smooth margins), a diagnosis of a
benign lesion may be made and one can consider a
follow-up with an unenhanced CT or CSMRI exam in
12 months. However, if there are suspicious imaging
features on contrast-enhanced CT, then one should
proceed with an unenhanced CT or CSMRI; if these
exams do not confirm that the lesion is a lipid-rich
adenoma, an adrenal protocol CT with washout calcula-
tions is recommended. If the lesion does not have the
imaging and washout features of a benign lesion, then a
biopsy may be appropriate.
In patients with a prior history of cancer and an adren-
al masses of any size, if imaging features on contrast-
enhanced are not diagnostic for a benign lesion and there
is no prior imaging, one can consider an unenhanced CT
or CSMRI or PET imaging. If the lesion does not behave
like a typical adenoma, then one should proceed to
adrenal CT with washout. If the lesion does not show
either washout features of an adenoma or the findings of
an adenoma on PET imaging, then a biopsy should be
considered.
In patients with no history of cancer and an adrenal
mass >4 cm in size, resection is an option; but if there is
a history of prior cancer, then a PET scan or a biopsy is
recommended [21].
William W. Mayo-Smith, Isaac R. Francis
References
1. Miskulin J, Shulkin BL, Doherty GM et al (2003) Ispreopera-
tive iodine 123 metaiodobenzylguanidine scintigraphy rou-
tinely necessary before initial adrenalectomy for pheochromo-
cytoma? Surgery 134:918-923
2. Greenblatt DY, Shenker Y, Chen H (2008) The utility of
metaiodobenzylguanidine (MIBG) scintigraphy in patients
with pheochromocytoma. Ann Surg Oncol 15:900-905
3. Boland GW, Blake MA, Holalkere NS, Hahn PF (2009)
PET/CT for the characterization of adrenal masses in patients
with cancer: qualitative versus quantitative accuracy in 150
consecutive patients. AJR Am J Roentgenol 192:956-962
4. Groussin L, Bonardel G, Silvéra S et al (2009) 18F-Fluoro-
deoxyglucose positron emission tomography for the diagnosis
of adrenocortical tumors: a prospective study in 77 operated
patients. J Clin Endocrinol Metab 94:1713-1722
5. Lee MJ, Hahn PF, Papanicolaou N et al (1991) Benign and ma-
lignant adrenal masses: CT distinction with attenuation coeffi-
cients, size, and observer analysis. Radiology 179:415-418
6. Korobkin, M, Brodeur FJ, Francis IR et al (1998) CT time-at-
tenuation washout curves of adrenal adenomas and nonadeno-
mas. AJR Am J Roentgenol 170:747-752
7. Boland GW, Lee MJ, Gazelle GS et al (1998) Characterization
of adrenal masses using unenhanced CT: an analysis of the CT
literature. AJR Am J Roentgenol 171:201-204
8. Bae KT, Fuangtharnthip P, Prasad SR et al (2003) Adrenal
masses: CT characterization with histogram analysis method.
Radiology 228:735-742
9. Korobkin M, Giordano TJ, Brodeur FJ (1996) Adrenal adeno-
mas: relationship between histologic lipid and CT and MR
findings. Radiology 200:743-747
10. Caoili EM, Korobkin M, Francis IR et al (2002) Adrenal
masses: characterization with combined unenhanced and de-
layed enhanced CT. Radiology 222:629-633
11. Tsushima Y, Ishizaka H, Matsumoto M (1993) Adrenal mass-
es: differentiation with chemical shift, fast low-angle shot MR
imaging. Radiology 186:705
12. Israel GM, Korobkin M, Wang C et al (2004) Comparison of
unenhanced CT and chemical shift MRI in evaluating lipid-
rich adrenal adenomas. AJR Am J Roentgenol 183:215-219
13. Fujiyoshi F, Nakajo M, Fukukura Y, Tsuchimochi S (2003)
Characterization of adrenal tumors by chemical shift fast low-
angle shot MR imaging: comparison of four methods of quan-
titative evaluation. AJR Am J Roentgenol 180:1649-1657
14. Mayo-Smith WW, Lee MJ, McNicholas MM et al (1995) Char-
acterization of adrenal masses (5 cm) by use of chemical shift
MR imaging: observer performance versus quantitative mea-
sures. AJR Am J Roentgenol 165:91-95
15. Silverman SG, Mueller PR, Pinkney LP et al (1993) Predictive
value of image-guided adrenal biopsy: analysis of results of
101 biopsies. Radiology 187:715-718
16. Grumbach MM, Biller BM, Braunstein GD et al (2003) Man-
agement of the clinically inapparent adrenal mass (“inciden-
taloma”). Ann Intern Med 138:424-429
17. Young WF (2007) Clinical practice. The incidentally discov-
ered adrenal mass. N Engl J Med 356:601-610
18. Choyke PL (2006) ACR Appropriateness Criteria on inciden-
tally discovered adrenal mass. J Am Coll Radiol 3:498-504
19. Song JH, Chaudhry FS, Mayo-Smith WW (2008) The inci-
dental adrenal mass on CT: Prevalence of adrenal disease in
1049 consecutive adrenal masses in patients with no known
malignancy. AJR Am J Roentgenol 190:1163-1168
20. NIH state-of-the-science statement on management of the clin-
ically inapparent adrenal mass (“incidentaloma”) (2002) NIH
Consens State Sci Statements 19:1-25
21. Boland GW, Blake MA, Hahn PF, Mayo-Smith WW (2008)
Imaging characterization of adrenal incidentalomas: princi-
ples, techniques and algorithms. Radiology 249:756-775

Renal Tumors
Richard H. Cohan1, Ronald J. Zagoria2
1 University of Michigan, Ann Arbor, MI, USA
2 Wake Forest University, Winston-Salem, NC, USA
Introduction
This chapter reviews recent advances in the imaging of
renal masses, primarily using computed tomography
(CT) and magnetic resonance imaging (MRI). The focus
is on developments in the use of imaging to differentiate
benign from malignant renal lesions and to assess pa-
tients following treatment for a renal mass.
Ultrasonography
Although non-contrast ultrasound evaluates the internal
morphology of cystic lesions with more detail than CT, it
is not as sensitive in detecting renal masses as CT or
MRI. Further, ultrasound is limited in its ability to char-
acterize detected masses. Most investigators consider
ultrasound to be diagnostically definitive only when it
identifies a renal mass as a simple cyst. The majority of
radiologists recommend that complex cysts and solid
masses detected on ultrasound be evaluated further with
CT or MRI.
CT and MRI Techniques
Adequate CT evaluation of a patient with a known or sus-
pected renal mass requires that at least two series be ob-
tained: unenhanced and delayed-enhanced images, with
the latter obtained at least 100 s after initiation of contrast
material injection [1, 2]. Similarly, MRI examinations
should include T1-weighted sequences obtained before
and after gadolinium administration [3]. Additional MRI
sequences or techniques that are often helpful include T2-
weighted, fat suppression, chemical shift, and diffusion-
weighted imaging.
Small Renal Masses
Most renal masses detected with CT and some masses
detected with MRI cannot be characterized due to their
small size. For CT, this is felt to be true for renal masses
IDKD 2010-2013
<1-1.5 cm in diameter. Given that small renal masses are
very common, further imaging assessment of all of these
lesions is not feasible, even though a few will be cancers.
Recently, Patel and colleagues reported that the subjec-
tive impressions of radiologists are frequently quite accu-
rate in identifying which small renal masses are cysts and
therefore which lesions may not require follow-up [4].
Accordingly, we recommend that follow-up imaging of
small renal masses be performed in only a few circum-
stances: (a) when the imaging suggests to a radiologist
that a lesion may not be a simple cyst (often due to lesion
heterogeneity), (b) when a lesion is detected in a young
patient (<40 years), or (c) when a new small mass is seen
in a patient at risk for renal malignancy (such as in pa-
tients with von Hippel Lindau, hereditary papillary renal
cell cancer, or hereditary leiomyomatosis-renal cancer
syndromes) [5]. In these instances, further evaluation
with immediate MRI can be performed. Often, long-term
follow-up (for up to 5 years) will be necessary. Follow-
up imaging can be performed safely in most patients and
does not usually result in a change in tumor stage or prog-
nosis, even if the mass being followed is later found to be
a small cancer. Most small renal cancers grow very slow-
ly, with a mean growth rate of 0.4-0.5 cm/year [6, 7].
Therefore, on a 6-month follow-up examination, the typical
growth of a renal cancer is near the range of CT measure-
ment error (~2-3 mm). While the only generally known
exceptions are the small aggressive renal type II papillary
cancers seen in patients with hereditary leiomyomatosis-
renal cancer syndrome [5], a recent study has demonstrated
that some other small (<4 cm) renal cancers can invade the
renal capsule or even metastasize distantly [8].
Cystic Renal Masses
In 1986, Bosniak presented a CT cyst classification system
[9] that categorized cysts based on their likelihood of their
being malignant. This system has undergone a number of
revisions, with the most recent version published in 2005
[10]. According to Bosniak, lesions with no or minimal
complexity on CT, classified as category I or II lesions,
are nearly always benign and do not require follow-up.
100
Only a tiny percentage of these contains malignant cells
[11]. Category II lesions require imaging follow-up with
immediate MRI, and/or with surveillance imaging, since
a few of these are malignant. When these follow-up
studies are interpreted, it must be remembered that both
benign and malignant lesions can enlarge over time [11].
To further confuse the issue, some benign and malignant
lesions can decrease in size over time [7]. Therefore, it is
important to assess cystic lesions for increasing com-
plexity, such as the development of new wall thickening
or nodularity on follow-up studies [11]. These features
are the best predictors of malignancy. More complex cat-
egory III and IV cysts should be treated (in appropriate
candidates), since many of these are cancers and imaging
distinction between benign and malignant category III
and IV lesions is not possible.
When the Bosniak system is used with MRI, about
80% of cystic masses demonstrate similar complexity as
seen on CT, but one in five appears more complex,
demonstrating additional septations, wall or septal thick-
ening, or subtle enhancement [12]. This is due, in part, to
the ability of MRI to detect some internal cyst features
not visible with CT, but also to the fact that MRI is more
prone to artifacts.
Angiomyolipomas
While nearly all angiomyolipomas (AMLs) are echogenic
on ultrasound examinations, so are many small renal can-
cers. In a few series, acoustic shadowing has been iden-
tified posterior to some AMLs but not posterior to renal
cancers; however, at many institutions this finding alone
is not accepted as conclusive evidence of an AML [13].
Thus, echogenic masses generally are evaluated further
with CT or MRI to determine whether macroscopic fat is
present in the mass. If macroscopic fat is present, then the
mass can be diagnosed as an AML (with case reportable
exceptions, see [14]). Otherwise, the mass is highly likely
to be a renal cell carcinoma.
On CT, the presence within a renal mass of even small
areas measuring –10 HU or less is considered diagnostic
of macroscopic fat and of an AML [15]. On MRI, such
fat typically is of high T1 and T2 signal intensity and loses
signal with fat suppression. On opposed-phase chemical-
shift imaging, there is a characteristic “India ink” artifact
at fat-water interfaces within the AML and between the
AML and adjacent tissue [16].
On occasion, AMLs can be very exophytic and diffi-
cult to differentiate from perinephric liposarcomas. This
is an important distinction, given the very different treat-
ments and prognoses for these two neoplasms. While per-
cutaneous biopsy is often helpful in making the distinc-
tion, imaging features have also been identified that can
be used to facilitate differentiation. Fatty perinephric
masses are more likely to be exophytic AMLs if they con-
tain large vessels or vessels that can be seen to extend to
the renal cortex, or if they are associated with a renal
Richard H. Cohan, Ronald J. Zagoria
parenchymal defect [17, 18]. A mass that contains calci-
fication is more likely to be a liposarcoma than an AML.
Some AMLs do not contain identifiable macroscopic
fat, including up to one-third of such tumors in patients
with tuberous sclerosis. Although these tumors cannot be
differentiated from other renal neoplasms on gross in-
spection, a number of studies have assessed the ability of
imaging to differentiate minimal-fat-containing AMLs
from other solid renal masses.
A few studies have attempted to identify small foci of
fat in minimal-fat-containing AMLs in the hope that this
will permit these lesions to be correctly identified. The
authors of these series have studied unenhanced CT at-
tenuation [19], analyzed CT histograms [20], counted fat
attenuation pixels [21], and quantitatively assessed fat on
MRI [22]. Unfortunately, the results have been mixed. In
some series, minimal-fat-containing AMLs have con-
tained more measurable fat than renal cancers, while in
other studies this has not been confirmed. A series by
Catalano et al. [20] comparing AMLs to clear cell carci-
nomas is particularly concerning. The authors found that
many clear cell renal cancers actually contained more
measurable fat than did minimal-fat-containing AMLs.
Additional MRI features independent of fat detection
have also been evaluated. It has been suggested that a di-
agnosis of minimal-fat-containing AML should be con-
sidered, albeit not definitively, if a solid mass demon-
strates low signal intensity on T2-weighted sequences and
if it has only mild enhancement after gadolinium-based
contrast material injection. These MRI characteristics are
not seen in the majority of renal cancers.
Differentiating Non-Fat-Containing Renal Tumors
Renal masses that do not contain recognizable fat cannot be
distinguished consistently from each other. In the past, this
was not felt to be an important problem, as it was believed
that the vast majority of such lesions were renal cancers.
Accordingly, all patients with non-fat-containing solid
renal masses were referred for treatment. It has become
apparent, however, that a sizeable minority of small solid
renal masses are benign, a fact that has led some to
recommend pre-treatment biopsy for all small renal
tumors. Recent studies suggested that about one in five
solid renal masses measuring <4 cm is benign [23, 24]. The
frequency of benign tumors is even greater for renal le-
sions measuring <1 cm, with nearly half being benign [24].
Differentiating Benign from Malignant Renal Neoplasms
Benign solid renal masses that are encountered routinely
include the previously discussed minimal-fat-containing
AMLs and oncocytomas. While some oncocytomas con-
tain central scars that can be detected on imaging studies
and oncocytomas also may demonstrate a spoke-wheel
vascular pattern at CT or MR arteriography, these fea-
tures are not diagnostic. Necrosis in renal cancers and
Renal Tumors
scars in oncocytomas, when they are present, are indis-
tinguishable from one another. Furthermore, most onco-
cytomas evaluated on CT do not contain identifiable cen-
tral scars [25]. Many renal cancers demonstrate a spoke-
wheel arterial pattern and since renal cancer is much
more common, most tumors that have this appearance are
malignant.
One cannot rely on an assessment of growth rate to
distinguish oncocytomas from malignant renal neo-
plasms, because small renal cancers and many onco-
cytomas enlarge slowly and to a similar degree over time.
In a meta-analysis reported by Chawla et al., there was no
significant difference in the growth rates between these
two different neoplasms [26].
Traditionally, percutaneous biopsy differentiation of
oncocytomas from some renal cancers based upon histo-
logical appearance was difficult, if not impossible. How-
ever, several stains have recently been used in conjunc-
tion with an assessment of tumor morphology to facili-
tate this distinction, such as Hale colloidal iron, cyto-
keratin-7, and vimentin immunohistochemical stains.
While several researchers have suggested that in many
cases these stains permit the distinction of these two
types of tumors [27, 28], the accuracy of biopsy for this
distinction remains controversial.
There are a variety of additional benign renal neo-
plasms that likewise have a non-specific appearance.
These include the common papillary adenomas and
renomedullary interstitial cell tumors (generally measur-
ing <1 cm and encountered in 40-50% of adults in autop-
sy series) as well as very rare lesions, such as metanephric
neoplasms, hemangiomas and lymphangiomas, leio-
myomas, juxtaglomerular tumors, mixed epithelial and
stromal tumors, and cystic nephromas [29].
Differentiating among Renal Cancer Subtypes
As is the case with benign versus malignant renal lesions,
one also cannot rely on differences in growth rates to dis-
tinguish among cell types of renal cancers. In a recent se-
ries, neither the initial size of a detected renal cancer nor
the cell type of that cancer predicted the likelihood that the
cancer would be more or less likely to grow quickly [7].
Nonetheless, there are some morphological differences
among the different types of renal cancers. Papillary tu-
mors can often be differentiated from other renal cancer
cell types on MRI, as many papillary cancers demon-
strate characteristic T2 signal hypointensity [30, 31]. In
contrast, only a small minority of clear cell carcinomas
are T2 hypointense [31]. Additionally, on contrast-
enhanced CT or MRI, both papillary subtypes, i.e., the
better-prognosis type 1 tumors and the more aggressive
type II tumors, tend to demonstrate more well-defined
margins, more homogeneity, and less enhancement than
do other renal cancer cell types [30, 32]. In one recent
study, for example, the mean percent increase of T1
signal intensity after gadolinium-based contrast material
administration during the corticomedullary phase was
101
over 200 for 75 clear cell, but only 110 for 10 chromo-
phobe, and 32 for 28 papillary cancers [32].
A number of investigators have used diffusion-weighted
MRI in the evaluation of renal masses [33]. Preliminary
studies have demonstrated differences in diffusion for
different types of renal masses. Not surprisingly, dif-
fusion is least restricted in simple renal cysts and most
restricted in the more cellular renal neoplasms (such as
papillary cancers and angiomyolipomas) [33]. Diffusion-
weighted MRI may be able to play a role in renal mass
characterization, particularly in patients who cannot re-
ceive gadolinium-based contrast material [33]. However,
diffusion-weighted imaging has a number of problems.
Firstly, there is overlap in the degree of restricted diffu-
sion identified in benign and malignant renal lesions.
Secondly, there are many technical issues, with measured
apparent diffusion coefficients varying among scanner
brands and protocols.
Pre-operative Imaging and Treatment of Renal Cancers
CT and MRI are very accurate in their ability to localize
and stage renal cancers prior to treatment. Pre-treatment
imaging studies should be used to assess renal cancer
size, location (with respect to the renal poles, renal sinus,
and collecting system), evidence of gross perinephric
extension, involvement of the ipsilateral renal vein or
inferior vena cava (including a description of the extent
of involvement), the presence or absence of enlarged
lymph nodes, and of any distant metastases. Either tech-
nique can also be used for anatomical definition prior
to anticipated surgery (including location and number of
renal vessels).
Treatment choice depends upon the pre-operative
imaging characteristics of each tumor and the patient’s
condition. Although percutaneous thermal ablation tech-
niques are being performed with increasing frequency,
they are still most often reserved for patients who are not
good operative candidates, who have renal insufficiency,
or who have multiple renal neoplasms. Based on early
studies, the cure rate for percutaneous thermal ablation is
comparable to that for surgery, but studies evaluating
long-term oncological efficacy have yet to be completed.
Masses most amenable to percutaneous ablation include
smaller lesions (<4 cm) and lesions for which there is no
radiological evidence of advanced disease (N0, M0) [34].
Masses in almost any location in the kidney now can be
treated with equal efficacy with ablation. Precautions to
protect adjacent organs, such as the bowel, should be
used with ablation when the mass to be treated is closer
than 2 cm from these structures.
Currently, many urologists have shifted away from use
of open total and partial nephrectomy to laparoscopic
procedures, because, as is the case with percutaneous
thermal ablation, laparoscopic nephrectomy is associated
with considerably less patient perioperative morbidity
and a shorter length of hospital stay [34].
102
Imaging after Renal Mass Treatment
Both CT and MRI, often performed using unenhanced
arterial phase imaging, and delayed enhanced imaging, are
used widely to image patients after treatment of renal
cancers. Recent studies have described the imaging ap-
pearance of masses treated with radiofrequency ablation
[35, 36]. It is important to be aware of the normal imaging
appearance of the ablated mass following successful abla-
tion in order to avoid confusion with tumor recurrence.
After successful renal mass ablation, there is initial expan-
sion of the ablation site. The ablation bed then typically
decreases in size, but rarely disappears entirely. Other typ-
ical “normal” findings include fat invagination between
the ablation bed and normal renal parenchyma and a peri-
lesional halo, such that the ablation cavity mimics an AML.
Recurrent or residual tumor should be suspected when
the ablation bed increases in size or when areas of en-
hancement are identified. The latter are usually nodular
and crescentic and are often located at the interface be-
tween the ablation bed and adjacent renal parenchyma
[37, 38]. Close-interval follow-up (for example, at 1, 3,
6, and 12 months) should be performed after ablation,
since most tumor recurrences are detected within the first
3 months of the procedure [38].
The imaging appearance of patients treated with open
and laparoscopic partial nephrectomy has also been de-
scribed. Hemostatic devices such as Surgicel (Johnson &
Johnson, USA) and renal bolsters, when inserted after
partial nephrectomy, can be mistaken for areas of infection
or recurrent tumor. While recurrent tumors often typically
occur in the surgical bed, in adjacent lymph nodes, or with
distant metastatic disease, patients treated laparoscopically
may demonstrate a different pattern of recurrence, with
implants growing along the laparoscopic port sites or
elsewhere in the mesentery or peritoneum [39].
References
1. Cohan RH, Sherman LS, Korobkin M et al (1995) Renal mass-
es: assessment of corticomedullary-phase and nephrographic-
phase CT scans. Radiology 196:445-451
2. Birnbaum BA, Jacobs JE (1996) Multiphasic renal CT: com-
parison of renal mass enhancement during the corticomedullary
and nephrographic phases. Radiology 200:753-758
3. Hecht EM, Israel GM, Krinsky GA et al (2004) Renal masses:
quantitative analysis of enhancement with signal intensity
measurements versus qualitative analysis of enhancement with
image subtraction for diagnosing malignancy at MR imaging.
Radiology 232:373-378
4. Patel NS, Poder L, Wang ZJ et al (2009) The characterization
of small hypoattenuating renal masses on contrast enhanced
CT. Clin Imag 33:295-300
5. Choyke PL (2003) Imaging of hereditary renal cancer. Radiol
Clin North Am 41:1037-1051
6. Bosniak MA, Birnbaum BA, Krinsky GA, Waisman J (1995)
Small renal parenchymal neoplasms: further observations on
growth. Radiology 197:589-597
7. Zhang J, Kang SK, Wang L et al (2009) Distribution of renal
tumor growth rates determined by using serial volumetric CT
measurements. Radiology 250:137-144
Richard H. Cohan, Ronald J. Zagoria
8. Pahernik S, Ziegler S, Roos F et al (2007) Small renal tumors:
correlation of clinical and pathological features with tumor
size. J Urol 178:414-417
9. Bosniak MA (1986) The current radiological approach to re-
nal cysts. Radiology 158:1-10
10. Israel GM, Bosniak MA (2005) How I do it: evaluating renal
masses. Radiology 236:441-450
11. Gabr AH, Gdor Y, Roberts WW, Wolf JS (2008) Radiographic
surveillance of minimally and moderately complex renal cysts.
Br J Urol Int 103:1116-1119
12. Israel GM, Hindman N, Bosniak MA (2004) Evaluation of
cystic renal masses: comparison of CT and MR imaging by us-
ing the Bosniak classification system. Radiology 231:365-371
13. Farrelly C, Delaney H, McDermott R, Malone D (2008) Do all
non-calcified echogenic renal lesions found on ultrasound need
further evaluation with CT? Abdominal Imaging 33:44-47
14. Helenon O, Chretien Y, Paraf F et al (1993) Renal cell carcinoma
containing fat: demonstration with CT. Radiology 188:429-430
15. Simpson E, Patel U (2006) Diagnosis of angiomyolipoma us-
ing computed tomography-region of interest ≤10 HU or 4 ad-
jacent pixels ≤10 HU are recommended as the diagnostic
thresholds. Clin Radiol 61:410-416
16. Israel GM, Hindman N, Hecht E, Krinsky G (2005) The use of
opposed-phase chemical shift MRI in the diagnosis of renal
angiomyolipomas. AJR Am J Roentgenol 194:1868-1872
17 Israel GM, Bosniak MA, Slywotzky CM et al (2002) CT dif-
ferentiation of large exophytic renal angiomyolipomas and
perirenal liposarcomas. AJR Am J Roentgenol 179:769-773
18. Ellingson JJ, Coakley FV, Joe BN et al (2008) Computed to-
mographic distinction of perirenal liposarcoma from exophyt-
ic angiomyolipoma: a feature analysis study. J Comput Assist
Tomogr 32:548-552
19. Jinzaki M, Silverman SG, Tanimoto A et al (2005) Angiomy-
olipomas that do not contain fat attenuation at unenhanced CT.
Radiology 234:311
20. Catalano OA, Samir AE, Sahani DV, Hahn PF (2008) Pixel dis-
tribution analysis: can it be used to distinguish clear cell car-
cinomas from angiomyolipomas with minimal fat? Radiology
247:738-746
21. Simpfendorfer C, Herts BR, Motta-Ramirez GA et al (2009)
Angiomyolipoma with minimal fat on MDCT: can counts of
negative attenuation pixels aid diagnosis? AJR Am J Roentgenol
192:438-443
22. Kim JK, Kim HS, Jang YJ et al (2006) Renal angiomyolipoma
with minimal fat: differentiation from other neoplasms at dou-
ble-echo chemical shift FLASH MR imaging. Radiology
239:274-280
23. Tuncali K, van Sonnenberg E, Shankar S et al (2004) Evalua-
tion of patients referred for percutaneous ablation of renal tu-
mors: importance of a preprocedural diagnosis. AJR Am J
Roentgenol 183:575-582
24. Frank I, Blute ML, Cheville JC et al (2003) Solid renal tumors:
an analysis of pathological features related to tumor size. J
Urol 2217-2220
25. Choudhary S, Rajesh A, Mayer NJ et al (2009) Renal oncocy-
toma: CT features cannot reliably distinguish oncocytoma
from other renal neoplasms. Clin Radiol 64:517-522
26. Chawla SN, Crispen PL, Hanlon AL et al (2006) The natural
history of observed enhancing renal masses: meta-analysis and
review of the world literature. J Urol 175:425-431
27. Shah RB, Bakshi N, Hafez KS et al (2005) Image-guided biop-
sy in the evaluation of renal mass lesions in contemporary uro-
logical practice: indications, adequacy, clinical impact, and lim-
itations of the pathological diagnosis. Hum Pathol 36:1309-1315
28. Hes O, Michal M, Kuroda N et al (2007) Vimentin reactivity
in renal oncocytoma: immunohistochemical study in 234 cas-
es. Arch Pathol Lab Med 131:1782-1788
29. Prasad S, Surabhi VR, Menias CO et al (2007) Benign renal
neoplasms in adults: cross-sectional imaging findings. AJR
Am J Roentgenol 190:158-164
Renal Tumors
30. Yamada T, Endo M, Tsuboi M et al (2008) Differentiation of
pathologic subtypes of papillary renal cell carcinoma on CT.
AJR Am J Roentgenol 191:1559-1563
31. Oliva MR, Glickman JN, Zou KH et al (2009) Renal cell car-
cinoma: T1 and T2 signal intensity characteristics of papillary
and clear cell types correlated with pathology. AJR Am J
Roentgenol 192:1524-1530
32. Sun MRM, Ngo L, Genega EM et al (2009)) Renal cell carci-
noma: dynamic contrast-enhanced MR imaging for differenti-
ation of tumor subtypes- correlation with pathologic findings.
Radiology 250:793-802
33. Taouli B, Thakur RK, Mannelli L et al (2009) Renal lesions:
characterization with diffusion-weighted imaging versus con-
trast-enhanced MR imaging. Radiology 251:398-407
34. Ng CS, Wood CG, Silverman PM et al (2008) Renal cell car-
cinoma: diagnosis, staging, and surveillance. AJR Am J
Roentgenol 191:1220-1232
103
35. Wile GE, Leyendecker JR, Krehbiel KA et al (2007) CT and
MR imaging after imaging-guided thermal ablation of renal
neoplasms. Radiographics 27:325-339
36. Davenport M, Caoili EM, Cohan RH et al (2009) MR and
CT characteristics of successfully ablated renal masses sta-
tus-post radiofrequency ablation. AJR Am J Roentgenol
192:1571-1578
37. Rutherford EE, Cast JEI, Breen DJ (2008) Immediate and
long-term CT appearances following radiofrequency ablation
of renal tumours. Clin Radiol 63:220-230
38. Kawamoto S, Solomon SB, Bluemke DA, Fishman EK (2009)
Computed tomography and magnetic resonance imaging ap-
pearance of renal neoplasms after radiofrequency ablation and
cryoablation. Semin Ultrasound CT MRI 30:67-77
39. Masterson TA, Russo P (2008) A case of port-site recurrence
and locoregional metastasis after laparoscopic partial nephrec-
tomy. Nat Clin Prac 5:345-349
 IDKD 2010-2013

Urinary Tract Obstruction and Infection

Parvati Ramchandani1, Julia R. Fielding2
1 Department of Radiology, University of Pennsylvania Medical Center, Philadelphia, PA, USA
2 Department of Radiology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Urinary Tract Obstruction as discussed below; however, some controversy exists as

to which imaging studies are best for investigating sus-
Imaging plays an important role in the evaluation of pa- pected ureteral obstruction. Nonetheless, regardless of
tients with acute and chronic urinary tract obstruction. composition, virtually all ureteral stones will have high at-
The cause of obstruction varies greatly with the patient tenuation values, making them readily detectable with CT.
population and the geographic locale. In a series report- Non-mineralized matrix stones and some drug-related
ing on percutaneous nephrostomy drainage, urinary ob- stones (protease inhibitors) may not be visible on CT im-
struction was related to calculus disease in 26% of pa- ages but these are rarely encountered in routine clinical
tients and to malignancy in 61%. Carcinoma of the blad- practice. A history of protease inhibitor therapy use is
der, cervix, and colon were the most common primaries required to deduce that the drugs are the cause of acute
in patients with malignancies and urinary obstruction [1]. flank pain and of the imaging findings of obstruction [10].
Acute ureteral obstruction is usually secondary to
urolithiasis, and patients commonly present to an acute
care facility with flank pain or acute renal colic. Non- Table 1. CT findings in acute renal colic
contrast helical computed tomography (CT) is often the Hoppe Katz Ahmad
first study performed in suspected acute obstruction, as it (2006) [2] (2000) [3] (2003) [4]
is a safe and rapidly performed examination and the N 1500 1000 233
accuracy for detecting ureteral stones, the most common Urinary stone 69% 62% 68%
cause of ureteral obstruction, exceeds that of other imag- Additional or alternative 71% 10% 12%
ing studies. Other causes of acute abdominal pain, such as diagnosis
appendicitis, leaking aortic aneurysm, and diverticulitis, Alternate diagnosis only 24% 7.50% X
can also be readily diagnosed and occur in 13-19% of Normal 7% 28% X
cases (Table 1) [2-4]. Non-contrast helical CT has an over- Pyelonephritis 3% 1% 1%
all accuracy of 97% for diagnosing ureteral stone disease, Renal cell carcinoma/ 2% 0.40% 1%
with a reported sensitivity of 97-100% and specificity of renal mass
94-96% [5-9]. This far exceeds the accuracy of intra- Cholecystitis 0.30% 0.30% 0.40%
venous urography (IVU) or ultrasound (US) (Table 2), Adnexal mass X 2% 2%

Table 2. Diagnostic performance for CT, US, and IVU in detection of ureteral stones
Lead author Year of publication N Stone + Test Sensitivity Specificity
Catalano 2002 181 82 CT 0.92 0.96
US/plain radiography 0.77 0.96
Boulay 1999 51 49 CT 1.0 0.96
Sheley 1999 180 87 CT 0.86 0.91
Sourtzis 1999 36 36 CT 1.0 1.0
IVU 0.66 1.0
Yilmaz 1998 97 64 CT 0.94 0.97
US 0.19 0.97
IVU 0.52 0.94
Smith 1996 210 100 CT 0.97 0.96
N, number of patients; Stone +, number of patients with ureteral stone
Urinary Tract Obstruction and Infection
The proper technique for performing non-contrast he-
lical CT to detect ureteral stone disease using a helical
scanner is detailed below. Imaging should be performed
from the top of the kidneys to the base of the bladder
without intravenous or oral contrast material and scans
should be obtained during a single breath-hold. A 16- to
64-slice MDCT with 1.5-mm collimation and a review of
5-mm contiguous images is appropriate. To reduce the ra-
diation dose, a variable mA is used for each slice based
on beam attenuation. In a recent report, the artificial ad-
dition of noise to a renal colic CT was used to show that
tube current could be diminished 75% without changing
the detection rate of stones >3 mm in diameter [11]. For
the obese patient, a fixed mA equal to his/her weight in
pounds will usually suffice. The expected dose is 30-40
mSv. Review of the images in cine mode on a worksta-
tion facilitates continuous identification of the ureter and
workflow. While 3D reconstructions are usually not nec-
essary, they may be occasionally useful to distinguish
intraureteral from extraurinary calcifications.
Due to the accuracy of CT and greater familiarity with
the examination by referring physicians, especially with-
in the emergency room, the number of renal colic CT
scans has dramatically increased. This alone is not worri-
some. More concerning is the number of patients who un-
dergo repeat CT scanning. In a study by Katz et al. pub-
lished in 2004, 4% of patients were found to have under-
gone three or more CT scans for renal stones during a
6-year period [12]. However, other reports indicate that
there has been no decrease in the rate of positive diagno-
sis of obstructing urinary stones on CT, validating the in-
creased frequency with which CT scans are obtained in
the setting of acute flank pain [13, 14]. The overall rate
of stone positivity on CT scanning in patients with sus-
pected renal colic was reported in several studies to be
60-66% [13, 14]. Additionally, as discussed above, CT is
the diagnostic test of choice to evaluate many acute ab-
dominal problems, and a decrease in the stone positivity
rate may not necessarily reflect overuse of CT [15]. Nev-
ertheless, it seems reasonable that a patient with known
history of stones and the appropriate clinical findings can
be treated conservatively and without imaging beyond an
abdominal radiograph.
In addition to direct visualization of the ureteral stone,
secondary signs of ureteral obstruction on non-contrast
CT scans include unilateral nephromegaly, perinephric
stranding, hydronephrosis, and periureteral stranding.
The combination of perinephric stranding and unilateral
hydronephrosis has a positive predictive value of 96% for
the presence of stone disease. The absence of both of
these signs has a negative predictive value of 93% for ex-
cluding stone disease. CT also gives information that de-
termines therapy. Stones that are f5 mm in size, of
smooth shape, and located within the distal one-third of
the ureter are likely to pass spontaneously [7].
The major pitfall in non-contrast helical CT evaluation
of the urinary tract for stone disease is difficulty in dis-
tinguishing pelvic phleboliths from ureteral calculi. The
105
presence of a tissue “rim” sign usually indicates that the
calcification is a stone rather than a phlebolith. Alterna-
tively, absence of the tissue rim sign or presence of a
“comet tail” sign strongly suggests that the calcification
is a phlebolith rather than a stone. In practice, the pres-
ence of two or more secondary signs of obstruction even
without clear visualization of a calcification within the
ureter indicates obstruction. If there is no history of re-
cent stone passage and the CT scan demonstrates find-
ings suggestive of obstruction, a contrast-enhanced study
of the upper tracts may be needed to exclude a urothelial
neoplasm, with additional cystoscopy to optimally evalu-
ate the bladder.
Intravenous urography is an alternative technique for
the detection of urinary tract obstruction. It was once the
imaging modality of choice in a patient who presented
with acute flank pain. However, the necessity for intra-
venous radiographic contrast administration and the
delay in obtaining relevant information about the site of
obstruction makes IVU a less desirable study than unen-
hanced CT in an acute pain setting. Extraurinary causes
of acute abdominal pain are not usually detectable with
IVU. The technique was also once considered to have a
role in the evaluation of pregnant patients with acute
flank pain, when the results of an ultrasonographic ex-
amination were negative or equivocal. However, even in
this patient population, CT may be the modality of
choice, as it provides a definitive and rapid diagnosis.
These advantages outweigh the slightly greater radiation
exposure [16, 17]. Acute colic in pregnant patients is dis-
cussed below. It is safe to state that IVU in the evaluation
of acute colic has been relegated to historical significance
in most patients.
Ultrasound, usually combined with an abdominal
radiograph, is an alternative method for evaluating the
obstructed or dilated urinary tract and is often the first
imaging procedure in patients who should avoid radia-
tion, such as pregnant women and children. Renal calculi
as small as 0.5 mm may be detectable under optimal
imaging conditions. For stones >5 mm in diameter, US
has been shown to have a sensitivity and specificity of
96% and nearly 100%, respectively [18]. US is excellent
for evaluation of the renal parenchyma and the collecting
system to the ureteropelvic junction but it is limited in its
evaluation of the ureter and of soft-tissue lesions within
the collecting system. The use of renal US in the evalua-
tion of suspected acute ureteral obstruction is also re-
stricted because dilatation often does not develop for
hours or even days. In these cases, US findings are nor-
mal in up to 50% of patients. US, either alone or com-
bined with conventional radiography, has been compared
with unenhanced CT. It has a much lower sensitivity:
24-77% [19-22] vs. 92-96% for CT. In Sheafor’s study, in
which CT and US were compared [19], CT depicted 22
of 23 ureteral calculi (sensitivity 96%) whereas US de-
picted 14 of 23 (sensitivity 61%). Differences in sensitiv-
ity were statistically significant (p = 0.02). The specifici-
ty for each technique was 100%. CT can give a rapid and
106
definitive diagnosis of urinary calculus disease as well as
other abdominal disorders with the same presentation.
US identification of ureteral jets within the urinary blad-
der lumen is helpful for assessing the presence of ob-
struction. One study showed an absent ureteral jet in 11
of 12 patients with high-grade obstruction and in 3 of 11
patients with low-grade obstruction [23]. The identifica-
tion of jets at the ureterovesical junctions indicates that
obstruction is incomplete, a finding that may be used to
guide therapy.
In diuresis renography, radionuclides are injected to
evaluate the urinary tract for obstruction. Since consider-
ably less anatomical detail is available with this test than
with other radiographic examinations, it is less useful in
the acute setting than for follow-up or the evaluation of
chronic urinary tract obstruction. Diuresis renography
does have the advantage of yielding objective data re-
garding the physiological significance of hydronephrosis
detected on imaging studies, and also allows evaluation
of the function of each kidney. Administration of a di-
uretic, usually furosemide, augments the standard
renogram and is useful in evaluating whether dilated uri-
nary systems are functionally obstructed or not.
Magnetic resonance urography (MRU) using rapid
scanning techniques, such as half acquisition turbo spin
echo or single-shot fast spin echo and 3D gradient echo
contrast-enhanced sequences, is used for evaluation of
the urinary tract. Following the administration of 250 mL
of normal saline and 10 mg of furosemide, the kidneys
and dilated ureters are very bright on T2-weighted images
and their stable position allows for clear imaging of the
level of obstruction (Table 3). Unfortunately, stones ap-
pear as signal voids and can be difficult to identify and
measure on MRU. Small calculi (which account for the
majority of symptomatic stones) are also difficult to de-
tect with this method. Also, urothelial lesions, blood
clots, and debris can mimic calculi.
In patients with renal impairment due to ureteral ob-
struction [24, 25], non-contrast CT was found to be the
best imaging modality to identify calculus causes of ob-
struction, while MRU was superior for identifying non-
calculus causes. In patients with normal renal function,
contrast-enhanced CT can identify the presence and
cause of hydronephrosis in nearly all cases [26]. MRU is
particularly helpful in delineating the anatomy in patients
with urinary diversion to bowel conduits [27].
An abdominal radiograph is a reasonable initial test in
patients who have a history of radiopaque calculi and
Table 3. Computed tomography urography protocol
No oral contrast, patient supine, 10 mg IV furosemide
Axial non-contrast enhanced abdomen
Inject intravenous contrast agent (100 mL of 350 mgI/mL)
Axial abdomen and pelvis at 75 s post-injection
Axial abdomen and pelvis at 5 min post-injection
Coronal reformatted images of 5-min delay exam
Review using bone and soft tissue windows
Parvati Ramchandani, Julia R. Fielding
Table 4. Magnetic resonance urography protocol
250 mL IV saline: begin 15 min before imaging
Patient supine, give 10 mg IV furosemide
Coronal HASTE/SSFSE scout
Axial abdomen HASTE/SSFSE
Axial IP/OOP gradient echo image
Coronal 3D TSE, HASTE or other fluid-bright sequence
If GFR <3.0 or serum Cr >3.0: STOP
If GFR >3.0 and suspicion for transitional cell carcinoma is high:
coronal 3D gradient echo image such as LAVA, VIBE with fat
saturation; 2-, and 5-min delay
acute flank pain that is similar to that of previous
episodes. In the absence of such a history, abdominal radio-
graphy may not be of value as an initial examination [9].
If a ureteral calculus is present on CT but not clearly
identifiable on the CT scout view, a conventional abdom-
inal radiograph may be useful – especially if the stone
is >4-5 mm in size and over 300 HU in density [28] – to
follow the progress of the stone for management purpos-
es. For radio-opaque calculi, confirmation of stone loca-
tion during conservative therapy is best performed using
plain films [29] (Table 4).
In dealing with the pregnant patient with flank pain,
fetal age and estimated radiation dose are of paramount
importance. Right hydronephrosis is commonly encoun-
tered, as the enlarging uterus turns slightly to the right
thus compressing the ureter. When an obstructing stone
is suspected in pregnant patients, US should be per-
formed first. Based on clinical findings, some urologists
will place a stent in a pregnant patient with severe hydro-
nephrosis. If more imaging information is needed from a
patient in the first trimester, a limited IVU using a plain
scout film followed by a 10-min post-infusion delayed
film yields the least radiation. After 20-24 weeks, IVU
becomes difficult to interpret because of the enlarging
uterus and the developing fetal skeleton, such that CT
should be considered [17]. The expected fetal dose is ap-
proximately 16 mSv, well below that expected to cause
developmental anomalies.
Urinary Tract Infection
Acute Pyelonephritis
This is usually an ascending infection from the bladder,
seen predominantly in females. Rarely, the source of in-
fection is hematogenous bacteremia. Diagnosis is usually
made on clinical grounds and with urine analysis. Imaging
may be needed to detect complications or the sequela of
pyelonephritis. When clinical pyelonephritis persists for
more than 3 days after suitable antibiotic therapy has been
initiated, imaging is recommended. In such cases, CT is the
imaging technique of choice to evaluate the kidneys for
possible complications of pyelonephritis, e.g., the develop-
ment of an abscess. CT is also the most sensitive and
Urinary Tract Obstruction and Infection
specific test for detecting the changes of acute pyelo-
nephritis and its complications. Typical CT findings of
pyelonephritis include unilateral nephromegaly, renal stri-
ations, wedge-shaped defects, and perinephric inflamma-
tory changes; detection of the latter usually requires con-
trast-enhanced images. Areas of liquefaction within the re-
nal parenchyma indicate the development of a renal ab-
scess. CT is more sensitive than US in detecting the de-
velopment of a renal abscess and in assessing its extent.
In males with a urinary tract infection (UTI) and/or
suspected pyelonephritis, clinical and imaging evaluation
for causes of UTI, such as epididymitis, orchitis, and pro-
statitis, may be helpful in management. Patients with a
neurogenic bladder secondary to a spinal cord injury pose
a difficult problem as the urine is usually colonized with
bacteria. Development of systemic symptoms should
prompt rapid imaging as these patients may not be sen-
sate to pain and a devastating abscess can develop quick-
ly [30]. Finally, in order to diminish radiation dose to
pregnant patients, US with power Doppler may be at-
tempted prior to CT to detect areas of aberrant blood
flow. This approach also has been shown to be useful in
children [31, 32].
Vesicoureteral reflux in children can lead to reflux
nephropathy, which may not be detected until they are
adults. The reflux of urine through the ducts of Bellini re-
sults in broad-based renal parenchymal scars that are cen-
tered over clubbed and blunted calyces. Changes of reflux
nephropathy occur preferentially in the poles of the kid-
neys, with the upper pole being affected most frequently.
The interpolar regions are almost always spared of scar-
ring in these patients.
Emphysematous Pyelonephritis
This life-threatening infection with a gas-producing or-
ganism may have a mortality of up to 90% without
prompt treatment, and nephrectomy is often required. In
diabetic patients, this infection is usually caused by a
strain of Escherichia coli. The diagnosis of emphysema-
tous pyelonephritis is made when gas is seen in the renal
parenchyma. CT is the most accurate technique for diag-
nosing emphysematous pyelonephritis and for differenti-
ating this entity from emphysematous pyelitis or peri-
nephric emphysematous infections. CT is also the most
accurate approach to differentiate localized gas-produc-
ing infections from diffuse emphysematous pyelonephri-
tis; the former can be successfully treated with percuta-
neous drainage in combination with systemic antibiotic
management.
Granulomatous Renal Infections
Tuberculosis, xanthogranulomatous pyelonephritis
(XGP), malacoplakia, and fungal infections can all af-
fect the urinary tract. Renal tuberculosis is usually
spread hematogenously from the lungs, seeding the
kidneys. Symptomatic renal tuberculosis results from
107
secondary, reactivation tuberculosis. Symptoms typical-
ly include hematuria and sterile pyuria. The earliest
signs of renal tuberculosis are, among others, focal pap-
illary necrosis and inflammation of the calyces. With
progression, areas of fibrosis and calcification may de-
velop. Long-standing tuberculosis may result in numer-
ous fibrotic strictures, ureteral wall thickening, hy-
dronephrosis, and autonephrectomy. Pyelonephritis re-
lated to Bacillus-Calmette-Guerin (BCG) therapy for
urothelial carcinoma has also been reported.
XGP, an inflammatory condition with a marked female
predominance, is associated with recurrent UTIs caused
by proteases or by E. coli infection. An infection-based
stone is seen in the majority of cases. The classic radio-
graphic triad consists of reniform enlargement of the
kidney, a renal stone, and markedly decreased or absent
renal function in the affected kidney. Localized XGP
occurs in 20% of patients and can mimic renal neoplasms
on imaging studies.
Both malacoplakia and fungal infections have non-
specific appearances on imaging, with diagnosis only be-
ing established by histological examination to exclude
neoplasm. Malacoplakia constitutes congregations of his-
tiocytes and is more commonly seen in the bladder and
ureter than in the kidney. The microscopic hallmark of
malacoplakia is the Michaelis-Gutman inclusion body,
which is seen within the abnormal histiocytes. When
malacoplakia involves the ureter or bladder, multiple sub-
mucosal masses are usually identified. Imaging findings
are non-specific and tissue is required for definitive di-
agnosis. Fungal infections are usually seen in immuno-
compromised patients, including diabetics. Debris, often
present within the renal collecting system, forms a
“hand-in-glove” filling defect of the contrast-opacified
calyces.
AIDS Nephropathy
Autoimmune deficiency syndrome (HIV/AIDS) nephro-
pathy comprises a variety of renal pathologies. Findings
are generally non-specific but patients with HIV infection,
renal failure, and hyperechoic nephromegaly likely have
AIDS nephropathy. These sonographic findings in an
AIDS patient usually indicate that the patient will develop
irreversible renal failure.
BK Virus Nephropathy
During the last decade, strong anti-rejection agents have
been used to improve renal allograft survival. The re-
sulting immunosuppression has allowed a generally be-
nign virus, known as BK, to become a deadly pathogen.
BK is present in the urine of 50% of transplant recipi-
ents 3 months post-operatively and may cause graft fail-
ure in 10%. The method of action of this small polyoma
virus is unknown. Patients with transplanted kidneys
who are BK-positive are treated with steroids. There is
no known cure [33].
108
Pyonephrosis
Pyonephrosis is a bacterial infection of the urine associ-
ated with ureteral obstruction. It is a medical emergency
as it has high morbidity if untreated. Pyonephrosis is best
diagnosed with US. Any febrile patient with hydro-
nephrosis should be suspected of also having pyonephrosis.
Findings that suggest pyonephrosis include echogenic
urine and the presence of debris within the hydro-
nephrotic calyces. Prompt drainage of the urinary tract
accompanied by systemic antibiotic administration is
essential in these patients. When obstruction and infection
are due to ureteral calculi, retrograde ureteral catheteriza-
tion and percutaneous nephrostomy are equally effective
in relieving the obstruction and infection; neither tech-
nique is superior to the other in promoting rapid drainage
or clinical defervescence [34].
Schistosomiasis
Schistosomiasis of the urinary tract is caused by infection
with Schistosoma hematobium. Infection is endemic in
Egypt but also occurs in Southern and sub-Saharan
Africa, South America, the Middle East, Southern China,
and in Southeast Asia. With the global nature of today’s
world, patients may be seen in developed western coun-
tries as well. This infection usually arises in the bladder
but can spread to the ureters and kidneys. Dystrophic cal-
cifications in the wall of the bladder and ureter are typi-
cal findings and are caused by calcification of the dead
parasitic ova. Typical radiographic findings include mur-
al calcifications, ureteral strictures, and vesicoureteral re-
flux. These patients have a markedly increased risk for
the development of squamous cell carcinoma as well as
transitional cell carcinoma of the urinary tract.
Post-therapeutic Changes
One of the most common causes of urinary obstruction is
pelvic cancer and its complications. Many patients with
gynecological and prostatic malignancies are treated with
surgery and radiation therapy, which may cause strictur-
ing of the ureter. The ureter may be transected during
hysterectomy or other pelvic surgery, leading to hydro-
nephrosis and urinoma. Large lymphoceles are un-
common with these surgeries but if present may com-
press the ureter or bladder. Ureteric obstruction sec-
ondary to stricture or fibrosis may occur, as well as fis-
tulae involving the bowel and urinary tract [35].
References
1. Farrell TA, Hicks MS (1997) A review of radiologically guid-
ed percutaneous nephrostomies in 303 patients. J Vasc Interv
Radiol 8:769-774
2. Hoppe H, Studer R, Kessler TM et al (2006) Alternate or ad-
ditional findings to stone disease on unenhanced computerized
tomography for acute flank pain can impact management. J
Urol 175:1725-1730; discussion 1730
Parvati Ramchandani, Julia R. Fielding
3. Katz DS, Scheer M, Lumerman JH et al (2000) Alternative or
additional diagnoses on unenhanced helical computed tomo-
graphy for suspected renal colic: experience with 1000 consec-
utive examinations. Urology 56:53-57
4. Ahmad NA, Ather MH, Rees J (2003) Incidental diagnosis of
diseases on un-enhanced helical computed tomography per-
formed for ureteric colic. BMC Urology 17:2
5. Sourtzis S, Thibeau JF, Damry N et al (1999) Radiologic inves-
tigation of renal colic: unenhanced helical CT compared with
excretory urography. AJR Am J Roentgenol 172:1491-1494
6. Yilmaz S, Sindel T, Arslan G et al (1998) Renal colic: com-
parison of spiral CT, US and IVU in the detection ureteral cal-
culi. Eur Radiol 8:212-217
7. Fielding JR, Silverman SG, Samuel S et al (1998) Unenhanced
helical CT of ureteral stones: a replacement for excretory urogra-
phy in planning treatment. AJR Am J Roentgenol 171:1051-1053
8. Catalano O, Nunziata A, Altei F et al (2002) Suspected ureteral
colic: primary helical CT versus selective helical CT after un-
enhanced radiography and sonography. AJR Am J Roentgenol
178:379-387
9. Levine JA, Neitlich J, Verga M et al (1997) Ureteral calculi in
patients with flank pain: correlation of plain radiography with
unenhanced helical CT. Radiology 204:27-31
10. Chan-Tack KM, Truffa MM, Struble KA, Birnkrant DB (2007)
Atazanavir-associated nephrolithiasis: cases from the US Food
and Drug Administration’s Adverse Event Reporting System.
AIDS 21:1215-1218
11. Ciaschini MW, Remer EM, Baker ME et al (2009) Urinary cal-
culi: Radiation dose reduction of 50% and 75% at CT - effect
on sensitivity. Radiology 251:105-111
12. Katz SI, Saluja S, Brink JA et al (2006) Dose associated with
unenhanced CT for suspected renal colic: impact of repetitive
studies. AJR Am J Roentgenol 186:1120-1124
13. Kirpalani A, Khalili K, Lee S, Haider MA (2005) Renal colic:
comparison of use and outcomes of unenhanced helical CT for
emergency investigation in 1998 and 2002. Radiology 236:
554-558
14. Ha M, MacDonald RD (2004) Impact of CT scan in patients
with first episode of suspected nephrolithiasis. J Emerg Med
27:225-231
15. Kenney PJ (2003) CT evaluation of urinary lithiasis. Radiol
Clin N Am 41: 979-999
16. LeRoy AJ (2006) Imaging of acute maternal diseases in preg-
nancy. Categorical course in diagnostic radiology at 92nd sci-
entific assembly and annual meeting, Chicago, Illinois, RSNA.
Genitourinary radiology syllabus 2006:271-279
17. Fielding JR, Washburn D (2005) Imaging the pregnant patient:
a uniform approach. J Women’s Imaging 7:16-21
18. Middleton WD, Dodds WJ, Lawson TL, Foley WD (1988) Re-
nal calculi: sensitivity for detection with US. Radiology
167:239-244
19. Sheafor DH, Hertzberg BS, Freed KS (2000) Nonenhanced He-
lical CT and US in the Emergency Evaluation of Patients with
Renal Colic: Prospective Comparison. Radiology 217:792-797
20. Catalano O, Nunziata A, Alei F et al (2002) Suspected ureteral
colic: primary helical CT versus selective helical CT after un-
enhanced radiography and sonography. AJR Am J Roentgenol
178:379-387
21. Fowler KAB, Locken JA, Duchesne JH, Williamson MR
(2002) US for detecting renal calculi with nonenhanced CT as
a reference standard. Radiology 222:109-113
22. Tublin ME, Dodd GD, Verdile VP (1994) Acute renal colic: di-
agnosis with duplex Doppler US. Radiology 193:697-701
23. Burge HJ, Middleton WD, McClennan BL, Hildebolt CF
(1991) Ureteral jets in healthy subjects and in patients with
unilateral ureteral calculi: comparison with color Doppler US.
Radiology 180:437-442
24. Shokeir AA, El-Diasty T, Eassa W et al (2004) Diagnosis of
ureteral obstruction in patients with compromised renal function:
the role of noninvasive imaging modalities. J Urol 171:2303-2306
Urinary Tract Obstruction and Infection
25. Shokeir AA, El-Diasty T, Eassa W et al (2004) Diagnosis of
noncalcareous hydronephrosis: role of magnetic resonance
urography and noncontrast computed tomography. Urology
63:225-229
26. El-Ghar MEA, Shokheir AA, El-Diasty T et al (2004) Contrast
enhanced spiral computerized tomography in patients with
chronic obstructive uropathy and normal serum creatinine: a
single session for anatomical and functional assessment. J
Urol 172:985-988
27. Zielonko J, Studniarek M, Markuszewski M (2003) MR Urog-
raphy of obstructive uropathy: diagnostic value of the method
in selected clinical groups. Eur Radiol 13:802-809
28. Zagoria RJ, Khatod EG, Chen MYM (2001) Abdominal radio-
graphy after CT reveals urinary calculi: a method to predict use-
fulness of abdominal radiography on the basis of size and CT
attenuation of calculi. AJR Am J Roentgenol 176:1117-1122
29. Assi Z, Platt JF, Francis IR et al (2000) Sensitivity of CT scout
radiography and abdominal radiography for revealing ureteral
calculi on helical CT: implications for radiologic follow-up.
AJR Am J Roentgenol 175:333-337
109
30. Rubenstein JN, Schaeffer AJ (2003) Managing complicated
urinary tract infections. The urologic view. Infect Dis Clin N
Am 17:333-351
31. Dacher J, Pfister C, Monroc M et al (1996) Power Doppler
sonographic pattern of acute pyelonephritis in children: com-
parison with CT. AJR Am J Roentgenol 166:1451-1455
32. Majd M, Nussbaum Blask AR, Markle BM et al (2001) Acute
pyelonephritis: comparison with Tc99m-DMSA SPECT, spiral
CT, MR imaging, and power Doppler US in an experimental
pig model. Radiology 218:101-108
33. Wiseman AC (2009) Polyomavirus nephropathy: A current
perspective and clinical considerations. Am J Kidney Dis
54:131-142
34. Pearle MS, Pierce HL, Miller GL et al (1996) Optimal method
of urgent decompression of the collecting system for obstruc-
tion and infection due to ureteral calculi. J Urol 160:1260-
1264
35. Yablon CM, Banner MP, Ramchandani P et al (2004) Compli-
cations of prostate cancer treatment: spectrum of imaging
findings. Radiographics 24:S181-S194
 IDKD 2010-2013

Benign Diseases of the Female Genital Tract

Caroline Reinhold1, Rahel A. Kubik-Huch2
1 McGill University Health Center, Montreal, Quebec, Canada
2 Institute of Radiology, Department of Medical Services, Kantonsspital Baden, Baden, Switzerland

Introduction
Endovaginal sonography (EVS) remains the procedure of
choice for the initial evaluation of benign diseases of the
female genital tract. When EVS findings are indetermi-
nate, further evaluation is typically performed with mag-
netic resonance imaging (MRI), due to its excellent soft-
tissue differentiation, multiplanar capabilities, and ab-
sence of ionizing radiation. MRI is thus well suited for
imaging women of reproductive age, particularly during
pregnancy. Accordingly, the technique has come to play
an increasing role in pelvimetry and, more recently, as an
adjunct to sonography for fetal imaging. MRI is used in
the pre-operative characterization of adnexal masses and
as a problem-solving tool in benign uterine disease (for
example, uterine malformations), adenomyosis, and to
select appropriate candidates for therapies such as myo-
mectomy and uterine embolization. The role of comput-
ed tomography (CT) is limited in the evaluation of benign
disease of the female pelvis and is usually employed in Fig. 1. Normal MR anatomy (sagittal T2-weighted image) of the
an emergency situation, such as in an acute abdomen uterus and cervix. Three zones are recognized in the uterus: the
caused by ovarian torsion or pelvic inflammatory disease. hyperintense endometrium, the hypointense junctional zone
(arrow), and the outer layer of the myometrium of intermediate
signal intensity. Four zones are distinguished in the cervix: the
hyperintense mucous within the endocervical canal, the cervical
Anatomy of the Female Genital Organs mucosa, the hypointense cervical stroma, and an additional layer of
smooth muscle. Scarring from C-section (arrowhead)
In women of reproductive age, the uterus is approxi-
mately 6-9 cm in length and varies in appearance ac-
cording to the menstrual cycle. The uterine zonal anato- 3. the hypointense cervical stroma surrounding the mucosa;
my is best depicted using sagittal T2-weighted images 4. an additional layer of intermediate signal intensity in
(Fig. 1). In the pre-menopausal woman, three distinct continuity with the uterine myometrium, representing
zones are recognized: smooth muscle.
1. the high-signal-intensity endometrium of varying In post-menopausal patients, the uterine corpus, but
thickness, depending on the menstrual cycle; not the cervix, regresses and decreases in size.
2. the hypointense junctional zone, anatomically corre- In pre-menopausal women, the mean ovarian volume is
sponding to the innermost layer of the myometrium; 10 ± 6 mL, with an upper limit of 22 mL [1]. In post-
3. the outer layer of the myometrium, which is of inter- menopausal women, the ovary atrophies and the follicles
mediate signal intensity. disappear over subsequent years. As a result of atrophy, the
Four zones are distinguished in the cervix by high- ovaries are reduced in volume (1-6 mL) and may be diffi-
resolution MRI: cult to visualize sonographically [1, 2]. An ovary is abnor-
1. the hyperintense mucous within the endocervical canal; mal if the volume is >8 mL or it is more than twice the vol-
2. the cervical mucosa of intermediate to high signal in- ume of the contralateral ovary. Small (<3 cm) anechoic
tensity; cysts can be seen in up to 15% of menopausal women [3].
Benign Diseases of the Female Genital Tract
Unlike sonography, MRI can detect the ovaries in almost
100% of patients [3]. On T2-weighted images, the cortex
and stroma of a pre-menopausal ovary is of low signal in-
tensity while the medulla is of higher signal intensity. In
pre-menopausal women, gadolinium enhancement of the
ovaries is less than that of the myometrium while in post-
menopausal women, it is equivalent [4, 5].
Magnetic Resonance Imaging Technique
If possible, patients should be scheduled for MRI in the
second half of the menstrual cycle, since the thickness of
the endometrial stripe increases during the follicular and
secretory phases, allowing better appreciation of the nor-
mal zonal anatomy of the uterus.
The objective of patient preparation is to obtain the best
possible image quality, while making the examination as
comfortable as possible for the patient. To minimize mo-
tion artifact induced by bowel peristalsis, patients are ad-
vised to fast for 6-8 h before the procedure. Unless con-
traindicated, the intravenous or intramuscular injection of
peristaltic inhibitors, i.e., glucagon or butyl-scopolamine,
is recommended to further decrease peristalsis artifacts.
An empty urinary bladder minimizes ghosting artifacts
from patient motion. In addition, it maintains the uterus in
a more caudal position in the pelvis, away from small
bowel loops, and assures that the normally visible fat
plane between the uterus and urinary bladder – an impor-
tant criteria to exclude tumor invasion of the bladder wall
in oncological patients – will not be obliterated.
Examinations are usually performed with the patient in
supine position and using a body-flex phased-array MRI
surface coil; pregnant patients in the third trimester may
also be placed in an oblique or lateral decubitus position
for greater comfort during imaging.
Depending on the clinical questions to be answered,
one or several fast sequences of the upper abdomen
should be performed, e.g., a cine steady-state free pre-
cession sequence (TrueFisp) in the coronal and axial
planes and a T1-weighted gradient echo sequence to
exclude ascites, enlarged retroperitoneal lymph nodes,
hydronephrosis, or renal agenesis in patients with con-
genital uterine malformations.
T1-weighted as well as axial STIR (short tau inversion
recovery) sequences are standard techniques in the as-
sessment of the small pelvis. The presence of a bright
mass on T1-weighted imaging requires an additional fat
suppressed T1-weighted sequence using chemical pre-
saturation to distinguish fat (e.g., in a mature teratoma of
the ovary, which shows signal loss) from blood, mucin, or
other proteinaceous material that remains of high signal
intensity. A pre-contrast fat-saturated T2-weighted se-
quence in the axial or sagittal plane is also helpful to de-
tect small endometriosis implants, e.g., in the cul-de-sac.
T2-weighted images are most important in the assess-
ment of the uterus, with sagittal sections best-suited to im-
age the uterus. Oblique coronal and axial slice orientations
111
that are in accordance with the long and short axes of the en-
dometrial cavity should be applied for the assessment of
uterine and cervical pathologies. It is recommended that
these oblique sequences are acquired under medical super-
vision, as the technician might not have sufficient anatomi-
cal knowledge to clearly identify the long axis of the uterine
body. The short-axis coronal oblique sequence (perpendicu-
lar to the long-axis of the endometrial cavity) is particularly
valuable for assessing localized endometrial pathology as
well as the thickness of the junctional zone. It is also valu-
able for determining the extent of a uterine septum. A T2-
weighted sequence performed parallel to the long axis of the
endometrial cavity is critical to characterize the external
uterine contour in patients with mullerian duct anomalies.
Since this series is so important in the classification of uter-
ine anomalies, it is best performed early in the examination,
prior to filling of the bladder, which with increasing disten-
tion often displaces the uterus. If the fundal contour is in-
adequately characterized on T2, then T1-weighted images
parallel to the long axis can facilitate characterization of the
external contour due to increased contrast between the
myometrial fundal contour and the overlying fat.
The uterus should be imaged using the smallest possi-
ble field of view (20-24 cm), with thin sections of 4-5 mm
and the largest possible matrix size appropriate to each
individual sequence. These imaging parameters provide
important anatomical detail, which becomes critical
when uterine anomalies and endometrial pathology are
imaged. In the imaging of large leiomyomas, the section
thickness and field-of-view (FOV) may need to be
adjusted accordingly. However, when the myometrial ori-
gin of a subserosal leiomyoma must be established, thin
sections at the level of the pedicle are frequently helpful.
We have recently added an axial echo-planar diffusion-
weighted MRI sequence of the small pelvis to our clini-
cal routine protocol. Diffusion-weighted imaging (DWI),
well established for intracranial imaging, is a functional
imaging technique whose contrast derives from the ran-
dom motion of water molecules within the extracellular
tissue. DWI is useful in pelvic female tumors, i.e., in the
differentiation of benign from malignant lesions, and has
a high sensitivity in the detection of iliac lymph nodes.
Gadolinium-enhanced imaging is not needed for most
benign conditions but can be useful in selected patholo-
gies. We routinely perform contrast-enhanced imaging in
patients with symptomatic leiomyomas scheduled for la-
paroscopic surgery or uterine artery embolization, in the
latter case with additional magnetic resonance angio-
graphy of the iliac arteries. Vascularity can help to pre-
dict response to treatment. In addition, the enhancement
pattern is one criterion that can be used to distinguish a
benign fibroid from a malignant leiomyosarcoma.
Congenital Uterine Anomalies
Buttram and Gibbons proposed a classification system in
1979 that was based on the degree of failure of normal de-
112
I Hypoplasia/agenesis
(a) Vaginal (b) Cervical (a) Communicating
(c) Fundal (d) Tubal (e) Combined (c) No cavity
V Septate
(a) Complete (b) Partial
Fig. 2. Classification system of mullerian duct anomalies according to the American Fertility Society [7]
velopment while taking into account similar clinical fea-
tures, reproductive outcomes, and management [6]. Mod-
ified by the American Fertility Society (now the American
Society of Reproductive Medicine) in 1988, the classifi-
cation system is the most widely accepted framework of
reporting uterovaginal anomalies [7] (Fig. 2). Uterine mal-
formations can be associated with subfertility, pregnancy
wastage, and menstrual disorders.
Class I: Mullerian agenesis and hypoplasia. In approxi-
mately 10% of uterine congenital anomalies, there is
some degree of early failure to form the mullerian ducts
prior to fusion [8]. Complete vaginal agenesis, or Mayer-
Rokitansky-Kuster-Hauser (MRKH) syndrome, is the
most common presentation of the class I anomalies. In
the group of patients with vaginal agenesis, 90% have as-
sociated uterine agenesis, while 10% present with a rudi-
mentary uterus. The ovaries are normal in appearance and
function, although they may be situated more cranially,
outside of the pelvis. Vaginal agenesis is best diagnosed
in the transverse plane with fatty and connective tissue in
the expected location of the vagina, between the urethra
and rectum.
Class II: unicornuate uterus. The unicornuate uterus ap-
pears elongated, curved, and is typically deviated to one
side [8]. The endometrium frequently takes on a “ba-
nana” or “bullet” shape. Unicornuate uteri without rudi-
mentary horns or those with non-cavitary rudimentary
horns require no treatment. Those with both non-
Caroline Reinhold, Rahel A. Kubik-Huch
II Unicornuate III Didelphus
(b) Non
Communicating IV Bicornuate
(d) No horn (a) Complete (b) Partial
VI Arcuate VII DES drug related
communicating and communicating rudimentary horns
are usually surgically resected. Hematometra and en-
dometriosis, as well as the potential for pregnancy with-
in the hypoplastic horn, may complicate the patient’s
course in this setting. The appearance of the rudimentary
horn is variable and depends on the presence of func-
tioning endometrial tissue and whether or not the horn is
obstructed (Fig. 3).
Class III: uterus didelphys. In uterus didelphys, two sep-
arate uterine horns (often widely divergent) and two sep-
arate cervices are visualized with MRI. The normal zon-
al anatomy is maintained within each hemiuterus, and the
two horns remain symmetrical in size.
Class IV: bicornuate uterus. In the bicornuate uterus, the
horns are symmetrical in size, with an intervening cleft
that extends to the internal cervical os in the “complete”
bicornuate uterus and which terminates more proximally
in the “incomplete” bicornuate uterus. The endocervical
canal may be solitary (bicornuate unicollis) or duplicated
(bicornuate bicollis). Communication of the endometrial
and/or endocervical segments must be present to diag-
nose this class of anomaly. Communication between the
segments allows differentiation of a bicornuate bicollis
uterus from a uterus didelphys. The uterus didelphys has
no communication between the endometrial segments. In
the bicornuate uterus, the fundal cleft on MRI is greater
than 1.0 cm. This feature distinguishes the bicornuate
from the septate uterus [8-11].
Benign Diseases of the Female Genital Tract
a b
Fig. 3 a, b. Uterine malformation: a axial and
b coronal T2-weighted MR images. A uni-
cornuate uterus with a rudimentary left horn
(arrows) is seen
Class V: septate uterus. In the septate uterus, the external
uterine contour can be convex, flat, or minimally con-
cave, with the fundal cleft always <1.0 cm [8-12]. The
fundal segment of the septum is isointense to myometri-
um in both partial and complete septa. In complete sep-
ta, the inferior segment of the septum is usually of low
signal intensity on T2-weighted images, corresponding to
the more fibrous component. This low-signal-intensity
band is frequently absent in partial septa, however, which
tend to be uniformly isointense to myometrium. It is im-
portant not to use the signal intensity of the septum to
classify this anomaly since septal composition can over-
lap between the septate and bicornuate uteri.
Benign Conditions of the Vagina, Cervix, and Uterus
Bartholin’s cysts are caused by retained secretions within
the vulvo-vaginal glands, mostly as a result of chronic in-
flammation or trauma. They are located in the posterolat-
eral parts of the lower vagina and vulva, whereas Naboth-
ian cysts are retention cysts of the cervical glands and
clefts. Gartner duct cysts represent remnants of the
caudal end of the Wolffian or mesonephric ducts and are
a b
Fig. 4 a, b. Multiple uterine leiomyomas.
a Sagittal T2-weighted and b contrast-
enhanced T1-weighted images. The multiple
lesions are sharply demarcated and present
as hypointense lesions on the T2-weighted
sequence. Contrast-enhancement is less pro-
nounced than in the adjacent normal my-
ometrium. A signal void is indicative of cal-
cification (arrow)
113
typically located in the anterolateral wall of the vagina,
above the level of the symphysis pubis.
Endometrial polyps are among the most common patho-
logical lesions of the uterine corpus. Patients with post-
menopausal bleeding and endometrial polyps usually un-
dergo endometrial sampling and polyp removal [13]. En-
dometrial polyps have a variable appearance at EVS but are
typically echogenic, with an intact overlying endometrium
or subendometrial halo. A vascular pedicle is usually iden-
tified at color/power Doppler imaging. On T2-weighted im-
ages, polyps present as masses that are slightly hypointense
or isointense relative to the normal endometrium. Large
polyps are frequently heterogeneous in signal intensity
[14, 15]. On T2-weighted sequences, the fibrous core is
seen as a hypointense area within the polyp. Endometrial
polyps show a variable degree of enhancement after
gadolinium administration, and the addition of gadolinium-
enhanced sequences significantly improves the detection
rate of endometrial polyps. Small polyps enhance early and
are well delineated against the hypointense endometrial
complex on early dynamic scans. In addition, a vascular
stalk can frequently be identified during the arterial phase.
Magnetic resonance imaging is useful for distinguishing
leiomyomas (Fig. 4) from other myometrial pathology and
114
solid pelvic masses, especially in patients with non-
diagnostic or equivocal ultrasound findings. A mass of
intermediate signal intensity on T1-weighted images, low
signal intensity on T2-weighted images, and splaying the
uterine serosa or myometrium allows the diagnosis of
leiomyoma to be made with confidence. The presence of
feeding vessels originating in the myometrium further
supports the uterine origin of the mass (Fig. 5). Howev-
er, if a mass is adjacent to the uterus and is of intermedi-
ate or high signal intensity relative to the myometrium on
T2-weighted images, the differential diagnosis includes
degenerated leiomyoma and extrauterine tumors (benign
and malignant). In these patients, the diagnosis of
leiomyoma should be reserved only for cases in which the
uterine origin of the mass is firmly established. Occa-
sionally, it may be difficult to distinguish a pedunculated
subserosal leiomyoma from an ovarian fibroma, since
both lesions may be hypointense on T2-weighted images.
This distinction is likely not significant as the latter is
rarely malignant [16]. Submucosal leiomyomas are usu-
ally distinguished from endometrial polyps by identifying
their myometrial origin and by their low signal intensity
on T2-weighted images.
a b
a b
Fig. 6 a-c. Focal adenomyosis: a sagittal T2-weighted and b, c axial T2-weighted STIR im-
ages. Enlarged uterus with diffusely thickened junctional zone with multiple hyperintense
foci indicative of adenomyoma. A small leiomyoma is also present (arrow)
Caroline Reinhold, Rahel A. Kubik-Huch
Leiomyomas must also be differentiated from uterine
adenomyosis (Fig. 6), although these conditions frequent-
ly coexist. Differentiating the two entities may be critical
because uterine-conserving therapy is established for
leiomyomas; whereas hysterectomy remains the definitive
treatment for debilitating adenomyosis. While MRI is ex-
tremely accurate in making this distinction, especially in
patients with diffuse adenomyosis, the imaging features of
focal adenomyosis or adenomyomas can overlap with
those of leiomyomas [17-19]. Imaging characteristics that
favor the diagnosis of focal adenomyosis include:
1. a lesion with poorly defined margins;
2. a lesion that is elliptical in shape extending along the
endometrium;
3. a lesion that has little mass effect upon the endometri-
um relative to its size;
4. a lesion with high signal intensity striations radiating
from the endometrium into the myometrium [17-19].
Cystic adenomyosis needs to be differentiated from a
leiomyoma with central hemorrhagic degeneration or a bi-
cornuate uterus with an obstructed horn.
Another entity that may mimic a leiomyoma or adeno-
myosis is a myometrial contraction. Myometrial contractions
Fig. 5 a, b. Subserosal leiomyoma: a axial T2-
weighted and b sagittal contrast-enhanced
T1-weighted sequences. A large, relatively
hyperintense mass is seen adjacent to the
uterus. The presence of feeding vessels (ar-
row) originating in the myometrium supports
the uterine origin of the mass. A smaller
hypointense intramural leiomyoma is seen in
the uterine fundus
c
Benign Diseases of the Female Genital Tract
are transient and usually change or resolve over the
course of the exam. Contractions image as hypoechoic or
low signal intensity lesions within the myometrium that
deform the endometrium while sparing the outer uterine
contour.
Malignant degeneration of a leiomyoma is a rare oc-
currence. Unfortunately, echogenicity or signal charac-
teristics do not reliably distinguish a benign leiomyoma
from a leiomyosarcoma. However, if a leiomyoma sud-
denly enlarges, especially after menopause, and/or has an
irregular or indistinct border, the possibility of sarcoma-
tous transformation should be raised.
Non-neoplastic Ovarian/Adnexal Masses
Follicular cysts are the most common benign ovarian
masses. Other non-neoplastic adnexal lesions include
hydrosalpinx, peritoneal inclusion cyst, endometrioma,
adnexal torsion, and tubo-ovarian abscess. These non-
neoplastic masses should be distinguished from benign
and malignant ovarian neoplasms.
Paraovarian cysts, also known as paratubal cysts,
account for 10-20% of adnexal masses. They are found
within the broad ligament or paraovarium and arise
from mesonephric (wolffian)/paramesonephric (mullerian)
structures or mesothelial inclusions.The hydatid cyst of
Morgani is the most common paramesonephric cyst and
arises from the fimbrial end of the fallopian tube [20]. On
imaging, paraovarian cysts have the typical appearance of
a cyst, although they can vary in size and contain com-
plex contents secondary to hemorrhage.
In pre-menopausal women, peritoneal fluid is pro-
duced by normally functioning ovaries and is reabsorbed
by the mesothelial cells of the peritoneal cavity. In
patients with adhesions composed of mesothelial and
a b
Fig. 7 a-c. Endometrioma: axial T1-weighted image with (a) and without (b) fat saturation;
c sagittal T2-weighted image. There is a large lesion of the right ovary with bright signal on
T1-weighted images with and without fat saturation and low signal intensity on T2-weighted
imaging. Small hyperintense follicular cysts adjacent to this lesion are seen on the sagittal
image
115
fibrous strands, either from previous surgery or sec-
ondary to pelvic inflammatory disease, endometriosis, or
trauma, the peritoneal fluid may not be reabsorbed; In-
stead, it accumulates and entraps the ovary, forming a
peritoneal inclusion cyst [21]. On imaging, peritoneal in-
clusion cysts are multiloculated cystic adnexal masses.
The diagnostic finding is the presence of an intact ovary
amid septations and fluid.
Hydrosalpinx is the result of obstruction of the fim-
briated end of the fallopian tube and dilatation of its am-
pullary and infundibular portions. The cause of obstruc-
tion includes pelvic inflammatory disease, endometriosis,
adjacent tumors, and adhesions from prior surgery. On
imaging, a hydrosalpinx appears as a fluid-filled tubular
structure with a somewhat folded configuration and well-
defined walls. It may contain folds that simulate an in-
complete thick septum.
The prevalence of endometriosis in all women is esti-
mated to be 5-10%. It is defined as the presence of func-
tioning endometrial glands and stroma outside the
uterus. The most common sites of involvement, in order
of decreasing incidence, are: the ovaries, cul-de-sac, pos-
terior uterine wall, uterosacral ligaments, anterior uterine
wall, and bladder dome. Other sites include the sigmoid
colon, fallopian tubes, and distal ureters. Endometrioma
is usually a term reserved for ovarian involvement. On
EVS, endometriomas typically appear as well-defined
unilocular or multilocular cystic masses with diffuse, ho-
mogeneously dispersed low-level internal echoes. A
fluid-fluid level and a thickened wall with calcifications
are sometimes present. On MRI, due to their hemor-
rhagic contents, endometriomas are typically high in sig-
nal intensity on T1-weighted images with and without fat
saturation. They demonstrate “shading” or a gradient of
low signal intensity on T2-weighted images, reflecting
the shortening of T2 due to blood (Fig. 7). Based on the
c
116
criteria of multiple adnexal, masses of very high signal
intensity on T1-weighted images or any mass with very
high signal intensity on T1-weighted images and low sig-
nal intensity (shading) on T2-weighted images, the over-
all sensitivity, specificity, and accuracy for diagnosing
endometriomas with MRI is 90, 98, and 96%, respec-
tively [22].
The MRI appearance of solid fibrotic endometriosis
has also been described: masses of intermediate signal
intensity studded with high-signal-intensity foci on
T1-weighted images, low-signal-intensity masses on
T2-weighted images, and enhancement following intra-
venous contrast [23]. Some fibrotic solid masses have
small foci of hyperintensity on T2-weighted images, re-
flecting embedded endometrial glands.
Neoplastic Ovarian Masses
Epithelial tumors are the most common histological type,
comprising 60% of all ovarian neoplasms and >85-90%
of ovarian malignancies. Serous tumors are the most
common ovarian epithelial tumors. On imaging, serous
cystadenomas are unilocular, thin-walled, large cystic
masses that may contain thin septations and occasionally
papillary projections [24]. Mucinous tumors are the sec-
ond most common epithelial tumor. On imaging, muci-
nous cystadenomas typically present as large cystic mass-
es (up to 30 cm) with multiple thin septations and low-
level echoes/T1-T2 shortening in the dependent portions
due to mucoid material. This results in the typical
“stained glass” appearance on MRI [25]. Papillary pro-
jections are less frequently seen than in serous cystade-
nomas. Rupture of the tumor capsule can result in
pseudomyxoma peritonei.
Mature cystic teratomas contain derivatives of at least
two of the three germ layers; ectoderm, mesoderm, and
endoderm. Ectodermal elements, however, tend to pre-
dominate, thus the term “dermoid cyst” was adopted. In
10% of the cases, mature cystic teratomas are bilateral.
They are usually asymptomatic except when complica-
tions, usually torsion, occur. Rupture is uncommon but its
a b
Fig. 8 a-c. Ovarian dermoid: a axial T2-weighted, b T1-weighted, and c T1-weighted fat suppressed images. A right adnexal mass with a flu-
id center and a rim that is hyperintense on the T1-weighted image; the mass shows signal loss on the fat suppressed image, indicating the
presence of fat
Caroline Reinhold, Rahel A. Kubik-Huch
occurrence can cause chemical peritonitis. In 1-2% of pa-
tients, usually in older women, there is malignant trans-
formation to squamous carcinoma.
On EVS, mature cystic teratomas can have variable ap-
pearances depending on the tumor contents; nonetheless,
several specific features have been described. A “der-
moid plug” or “Rokitansky” protuberance can be seen as
an echogenic mural nodule in a predominantly cystic
mass. It is composed of hair, teeth, and fat and causes
acoustic shadowing. The “dermoid mesh” is another spe-
cific sign; it refers to hair fibers that appear as linear
hyperechogenic interfaces floating in the cyst [26]. On
MRI, T1-weighted imaging with and without fat satura-
tion techniques can establish the presence of fat in a cys-
tic teratoma. The adipose tissue will have high signal in-
tensity on T1 without fat suppression and will show sig-
nal loss on fat-suppressed images, thus excluding the di-
agnosis of an endometrioma or hemorrhagic cyst (Fig. 8).
Chemical shift imaging using in- and out-of-phase imag-
ing is helpful in identifying tumors that have only a tiny
amount of fat [27].
Fibrothecomas account for only 1% of all ovarian
neoplasms; however, they are the most common solid be-
nign tumors affecting the ovary. They are derived from
stromal cells; because the histological appearance of fi-
bromas and thecomas overlap, the term “fibrothecoma”
is often applied to this spectrum of tumors. Like granu-
lose cell tumors, they are hormonally active. Pure theco-
mas are composed predominantly of theca cells and are
most common in peri-menopausal and post-menopausal
women. In 15% of patients, there is co-existing endome-
trial hyperplasia, and up to 30% of patients have en-
dometrial carcinoma. In contrast, pure fibromas are
composed predominantly of fibroblasts and are most
common in women under age 50, most of whom are
asymptomatic. Meigs syndrome, a combination of ovar-
ian fibroma, ascites, and right pleural effusion, is rare.
On EVS, fibrothecomas present as hypoechoic solid
masses with marked posterior sound attenuation as a re-
sult of homogeneous fibrous tissue. The MRI appear-
ance of fibromas and thecomas are similar: intermediate
signal intensity on T1-weighted images and very low
c
Benign Diseases of the Female Genital Tract
a Table 1. Reference values for magnetic resonance pelvimetry based
b graphy, with measurement of the various parameters on
Fig. 9 a, b. Ovarian fibrothecoma: a sagittal T2-weighted image and
b macroscopic specimen. The hypointense adnexal mass ventral to
the uterus was histologically confirmed as ovarian fibrothecoma.
A small submucosal and a larger subserosal leiomyoma are seen in
the uterus
signal intensity on T2-weighted images [28] (Fig. 9).
Due to these signal characteristics, fibrothecomas may
be difficult to distinguish from pedunculated fibroids or
Brenner tumors. Multiple imaging planes combined with
the observation of small follicular cysts surrounding the
mass should establish the diagnosis of a solid ovarian
neoplasm. The presence of a thickened endometrium
secondary to estrogen production favors the diagnosis.
Larger fibrothecomas may have high signal intensity
secondary to intratumoral edema, myxomatous change,
or cyst formation. Following intravenous gadolinium, fi-
brothecomas show little enhancement.
Magnetic Resonance Imaging in Obstetrics
There is no clinical or experimental evidence of terato-
genic or other adverse fetal effects from the use of MRI
in pregnancy. The technique is thus well suited in the
imaging of pregnant women for maternal, e.g., for the
117
on 100 women undergoing spontaneous vaginal delivery (from [30])
Reference values ± SD
(cm)
Obstetric conjugate 12.2 ± 0.9
Sagittal outlet 11.6 ± 1.0
Interspinous diameter 11.2 ± 0.8
Intertuberous diameter 12.1 ± 1.1
Transverse diameter 13.0 ± 0.9
work-up of colicky abdominal pain or suspected appen-
dicitis, or fetal reasons.
Magnetic resonance pelvimetry is indicated in preg-
nant women with a history of pelvic trauma, previous ce-
sarean section, or in women who desire a trial of labor
when the fetus is in breech presentation. The groundwork
in pelvimetry was conducted using conventional radio-
lateral and anterior-posterior views using various tech-
niques to correct the distortion due to differing distances
from the film. In recent years, X-ray pelvimetry has been
steadily replaced by MRI, which offers the benefit of ac-
curate measurement of the maternal pelvic dimensions
without exposure of the patient to ionizing radiation.
Magnetic resonance pelvimetric reference values in a
large study population, stratified by delivery modality,
have been established by a member of the author’s own
group [29, 30]. The results are shown in Table 1. In addi-
tion, it was demonstrated that the pelvimetric parameters
associated with the largest intra- and interobserver error
and intraindividual variability are the intertuberous dis-
tance and the sagittal outlet – a finding that must be tak-
en into careful account by obstetric decision-makers.
In conclusion, technical advances in ultrafast MRI
have revolutionized our ability to image the fetus. MRI is
most likely to soon become an important tool for in-
trauterine fetal imaging as an operator-independent sup-
plement to prenatal ultrasound [31, 32].
References
1. Cohen HL, Tice HM, Mandel FS (1990) Ovarian volumes mea-
sured by US: bigger than we think. Radiology 177:189-192
2. Goswamy RK, Campbell S, Royston JP et al (1988) Ovarian size
in postmenopausal women. Br J Obstet Gynaecol 95:795-801
3. Levine D, Gosink BB, Wolf SI et al (1992) Simple adnexal
cysts: the natural history in postmenopausal women. Radiology
184:653-659
4. Outwater EK, Talerman A, Duncon C (1996) Normal adnexa
uteri specimens: anatomic basis of MR imaging features. Ra-
diology 201:751-755
5. Outwater EK, Mitchell DG (1996) Normal ovaries and func-
tional cysts: MR appearance. Radiology 198:397-402
6. Buttram VC, Gibbons, WE (1979) Mullerian anomalies: a pro-
posed classification (an analysis of 144 cases). Fertil Steril
32:40-46
7. The American Fertility Society (1988) The American Fertility
Society classifications of adnexal adhesions, distal tubal ob-
118
struction, tubal occlusion secondary to tubal ligation, tubal
pregnancies, mullerian anomalies and intrauterine adhesions.
Fertil Steril 49:944-955
8. Trojano RN (2003) Magnetic resonance imaging of mullerian
duct anomalies of the uterus. Top Magn Reson Imaging 14:
269-279
9. Carrington BM, Hricak H, Nuruddin RN et al (1990) Mullerian
duct anomalies: MR imaging evaluation. Radiology 176:715-720
10. Pellerito JS, McCarthy SM, Doyle MB et al (1992) Diagnosis of
uterine anomalies: relative accuracy of MR imaging, endovaginal
ultrasound, and hysterosalpingography. Radiology 183:795-800
11. Junqueira BLP, Allen LM, Spitzer RF et al (2009) Mullerian
duct anomalies and mimics in children and adolescents: Cor-
relative intraoperative assessment with clinical imaging. Radio-
graphics 29:1085-1103
12. Homer HA, Li TC, Cooke ID (2000) The septate uterus: a re-
view of management and reproductive outcome. Fertil Steril
73:1-14
13. Sheth S, Hamper UM, Kurman RJ (1993) Thickened en-
dometrium in the postmenopausal woman: sonographic-patho-
logic correlation. Radiology 187:135-139
14. Pertl B, Lahousen M, Pieber D et al (1996) Stellenwert der
Sonograhie bei der Früherkennung des Endometriumkarzi-
noms. Gynäkologisch-geburtshilfliche Rundschau 36:14-20
15. Grasel RP, Outwater EK, Siegelman ES et al (2000) Endome-
trial polyps: MR imaging features and distinction from en-
dometrial carcinoma. Radiology 214:47-52
16. Hricak H, Tscholakoff D, Heinrichs L et al (1986) Uterine
leiomyomas: correlation of MR histopathologic findings, and
symptoms. Radiology 158:385-391
17. Reinhold C, McCarthy S, Bret PM et al (1996) Diffuse ade-
nomyosis: Comparison of endovaginal US and MR imaging
with histopathologic correlation. Radiology 199:151-158
18. Togashi K, Ozasa H, Konishi I et al (1989) Enlarged uterus:
Differentiation between adenomyosis and leiomyoma with MR
imaging. Radiology 171:531-534
19. Mark AS, Hricak H, Heinrichs LW et al (1987) Adenomyosis
and leiomyoma: Differential diagnosis with MR imaging. Ra-
diology 163:527-529
Caroline Reinhold, Rahel A. Kubik-Huch
20. Kurman RJ (1987) Blaustein’s pathology of the female genital
tract. 3rd edn. Springer-Verlag, New York
21. Levy AD, Arnaiz J, Shaw JC, Sobin LH (2008) From the
archives of the AFIP: Primary peritoneal tumors: Imaging fea-
tures with pathologic correlation. Radiographics 28:583-607
22. Togashi K, Nishimura K, Kimura I et al (1991) Endometrial
cysts: diagnosis with MR imaging. Radiology 180:73-78
23. Hottat N, Larrousse C, Anaf V et al (2009) Endometriosis:
Contribution of 3.0-T pelvic MR imaging in preoperative as-
sessment – initial results. Radiology 253:126-134
24. Thomassin-Naggara I, Bazot M, Darai E et al (2008) Epithe-
lial ovarian tumors: Value of dynamic contrast-enhanced MR
imaging and correlation with tumor angiogenesis. Radiology
248:148-159
25. Togashi K (2003) MR imaging of the ovaries: normal appear-
ance and benign disease. Radiol Clin N Am 41:799-811
26. Malde HM, Kedar RP, Chadha D et al (1992) Dermoid mesh: a
sonographic sign of ovarian teratoma. Letter Am J Roentgenol
159:1349-1350
27. Outwater EK, Siegelman ES, Hunt JL (2001) Ovarian ter-
atomas: tumor types and imaging characteristics. Radiographics
21:475-490
28. Troiano RN, Lazzarini KM, Scoutt LM et al (1997) Fibroma
and fibrothecoma of the ovary: MR imaging findings. Radiol-
ogy 204:795-798
29. Michel SC, Rake A, Treiber K et al (2002) MR obstetric pelvime-
try: effect of birthing position on pelvic bony dimensions. AJR
Am J Roentgenol 79:1063-1067 (Abstract and editorial comment
in Obstetrical&Gynecological Survey 2003; 54:238-239)
30. Keller TM, Rake A, Michel SCA et al (2003) Obstetric MR
pelvimetry: Reference values and evaluation of inter- and in-
traobserver error and intraindividual variability. Radiology
227:37-43
31. Keller TM, Rake A, Michel SCA et al (2004) MR assessment
of fetal lung development using lung volumes and signal in-
tensities. Eur Radiol 14:984-989
32. Huisman TA, Wisser J, Martin E et al (2002) Fetal magnetic
resonance imaging of the central nervous system. Eur Radiol
12:1952-1961

Malignant Diseases of the Female Genital Tract
Evis Sala1, Susan Ascher2
1 Department of Radiology, University of Cambridge and Addenbrookes Hospital, Hills Road, Cambridge, United Kingdom
2 Department of Radiology, Georgetown University Hospital, Washington, DC, USA
Introduction
Advances in cross-sectional imaging have led to an in-
creasingly important role for radiology in the manage-
ment of malignant gynecological conditions. A number
of imaging modalities can be used to evaluate malignant
diseases of the female pelvis, including ultrasound (US),
computed tomography (CT), magnetic resonance imag-
ing (MRI), and positron emission tomography/computed
tomography (PET/CT). These modalities have different
roles in screening, diagnosis, staging, treatment selection
and follow-up. The aim of this chapter is to review the
role of different techniques in the imaging of malignant
gynecological conditions. The emphasis is on the use of
MRI in the staging of endometrial and cervical cancer
following the revised FIGO (International Federation of
Gynecology and Obstetrics) criteria, implemented begin-
ning June 1, 2009 [1].
Ultrasound
The primary imaging modality in the initial assessment
of suspected gynecological pathology is US. It is used to
evaluate a suspected pelvic mass, characterize adnexal le-
sions, and identify endometrial abnormalities in the post-
menopausal patient. Transabdominal and transvaginal US
can assist in image-guided fine needle aspiration cytol-
ogy or biopsy and can also be used to guide placement of
brachytherapy seeds in the treatment of cervical and en-
dometrial cancer.
Magnetic Resonance Imaging
This is the imaging modality of choice in the staging of
uterine and cervical cancer and in the characterization of
adnexal lesions when the US findings are indeterminate.
The advantages of MRI include superb spatial and tissue
contrast resolution, the absence of ionizing radiation, its
multiplanar capability, and its fast techniques. However,
the optimization of MRI sequences and clinical proto-
cols, as outlined below, is crucial to ensure best results.
IDKD 2010-2013
Patient Preparation, Positioning, and Coil Selection
Patient preparation and positioning are very important
to obtain optimal imaging results. Patients are usually
instructed to fast for 4-6 h before the MRI examination
in order to limit artifacts due to small-bowel peristalsis.
An anti-peristaltic agent (hyoscine butylbromide or
glucagon) may be administered before imaging as an al-
ternative. Ideally, the patient is asked to empty the blad-
der prior to examination, as a full bladder may degrade
images due to ghosting and motion artifacts. Patients are
imaged in the supine position using a pelvic surface-
array multichannel coil; a cardiac coil usually offers the
best image quality [2].
Choice of Sequences and Imaging Planes
The basic imaging protocol for gynecological MRI con-
sists of T1-weighted images in the axial plane and T2-
weighted images in the axial, sagittal, and coronal planes.
T1-weighted axial images with a large field of view to
evaluate the entire pelvis and upper abdomen for lymph-
adenopathy as well as bone marrow changes are essential
in staging gynecological malignancies. High-resolution
axial oblique T2-weighted fast spin-echo (FSE) images
taken parallel to the short axis of the uterine corpus are
favored for the evaluation of primary tumor and depth of
myometrial invasion [3] in endometrial carcinoma,
whereas axial oblique T2-weighted FSE (parallel to the
short axis of the cervix) is crucial in assessing parame-
trial invasion in patients with cervical cancer.
Dynamic multiphase contrast-enhanced 3D T1-weight-
ed sequences through the uterus in the sagittal and axial
(parallel to the short axis of the uterine corpus) planes are
routinely used to improve staging accuracy in endometri-
al cancer [4, 5]. They are also useful for the evaluation of
complex adnexal lesions, as they may help differentiate
solid components or papillary projections from clots and
debris. Dynamic imaging is not necessary. Diffusion-
weighted imaging (DWI) may be useful in differentiating
benign from malignant endometrial lesions [6] and can
provide valuable information for pre-operative staging in
patient with endometrial and cervical carcinoma [7].
120
DWI can help in the evaluation of tumor response to ra-
diotherapy in patients with cervical cancer [8] and is use-
ful in the detection of peritoneal implants and metastatic
lymph nodes in patients with gynecological malignancies
[9]. DWI is also useful in detecting peritoneal implants in
patients with gynecological malignancies [10]. Ultra-
small particles of iron oxide (USPIO) have been shown
to improve the detection of lymph node metastases inde-
pendent of node size in patients with endometrial and
cervical cancer [10].
Computed Tomography
The role of CT in the imaging of malignant uterine
conditions is limited due to its poor soft-tissue contrast.
The main role of CT is in staging, treatment planning,
and follow-up of patients with ovarian cancer. However,
CT is important in the evaluation of other gynecological
malignancies by identifying enlarged lymph nodes and
distant metastases and detecting recurrent pelvic tumors.
Positron Emission Tomography/Computed Tomography
Patients with malignant gynecological conditions are in-
creasingly being evaluated with PET/CT. This modality is
very valuable in the detection of metastatic lymph nodes
as it has better sensitivity and specificity than MRI; it can
also differentiate tumor recurrence from radiation fibro-
sis. PET/CT is also very useful in evaluating recurrent tu-
mor prior to salvage therapy. Maximum standard uptake
values (SUV) at staging can predict survival in patients
with cervical carcinoma. However, it must be remem-
bered that in pre-menopausal patients physiological up-
take can be seen in the uterus, ovarian follicles, and cor-
pus luteum cysts. The uptake of 2-deoxy-2-[fluorine-
18]fluoro-d-glucose (FDG) can also be seen in certain
benign ovarian and uterine tumors as well as in inflam-
matory and infectious processes.
a b
Evis Sala, Susan Ascher
Endometrial Carcinoma
On US, endometrial carcinoma is seen as a thickened en-
dometrium (>5 mm in post-menopausal patients). On
sonohysterography, endometrial carcinoma may present
as an intrauterine polypoid mass or as an asymmetrical
thickening of the endometrium. It is, however, impossible
to distinguish between benign endometrial polyps, en-
dometrial hyperplasia, and endometrial carcinoma con-
fined to the endometrium using US alone. Therefore, en-
dometrial carcinomas are typically diagnosed at endome-
trial biopsy or dilatation and curettage, with MRI being
reserved to evaluate the extent of disease [11].
Imaging criteria for staging of endometrial cancer are
based on the TNM/FIGO (International Federation of
Obstetrics and Gynecology) classification. The overall
staging accuracy of MRI has been reported to be between
85 and 93% [4, 5, 12]. Routine use of dynamic intra-
venous contrast enhancement is necessary for state-of-
the-art MRI evaluation of endometrial carcinoma [4, 5].
Stage IA tumors involve <50% of the myometrial thick-
ness (Fig. 1). The presence of low-signal-intensity tumor
(equilibrium and later phases of enhancement) within the
outer myometrium or beyond indicates deep myometrial
invasion and thus stage IB disease. Erroneous MRI as-
sessment of the depth of myometrial invasion may occur
due to an indistinct zonal anatomy, the presence of co-
existent benign pathology (leiomyomas, adenomyosis),
tumor extension into the uterine cornu, or a polypoid
tumor distending the uterus so that rather than deep
myometrial infiltration there is a thin rim of myometrium
stretched over the tumor [2].
In stage II disease, the fibrocervical stroma is disrupt-
ed by high-signal-intensity tumor on T2-weighted im-
ages, together with the disruption of normal enhancement
of the cervical mucosa by low-signal-intensity tumor on
late dynamic contrast-enhanced MRI.
In stage III disease, tumor extends outside the uterus
but not outside the true pelvis. Stage IIIA is marked by
Fig. 1 a, b. Stage IA endometrial carcinoma.
Sagittal T2-weighted fast spin echo (FSE) (a)
and sagittal gadolinium-enhanced 3D T1-
weighted gradient-recalled echo (GRE) (b)
images show an endometrial carcinoma in-
vading the inner half of the myometrium. The
depth of myometrial invasion is better appre-
ciated on the gadolinium-enhanced 3D T1-
weighted GRE MR image (arrow)
Malignant Diseases of the Female Genital Tract
parametrial involvement, in which there is disruption of
the serosa with direct tumor extension into the surround-
ing parametrial fat. In stage IIIB disease, tumor extends
into the vagina and there is segmental loss of the low-
signal-intensity vaginal wall. In stage IIIC disease,
lymphadenopathy is present.
Tumor that extends beyond the true pelvis or invades the
bladder or rectum constitutes stage IV disease. A loss of the
low signal intensity of the bladder or rectal wall indicates
stage IVA disease whereas stage IVB disease includes dis-
tant metastasis, malignant ascites, or peritoneal deposits.
CT is valuable in detecting upper abdominal lymph-
adenopathy and distant metastases in patients with
advanced endometrial carcinoma. PET/CT is useful in
assessing nodal disease and distant metastases and has a
role in monitoring treatment response in these patients.
Cervical Carcinoma
The main role for imaging is staging a biopsy-proven cer-
vical carcinoma. MRI is the best single imaging investiga-
tion and can accurately determine tumor location (exophyt-
ic or endocervical) and size, the depth of stromal invasion,
and extension into the lower uterine segment [13, 14]. On
T1-weighted images, tumors are usually isointense with the
normal cervix and may not be visible. On T2-weighted
images, cervical cancer appears as a mass of intermediate
signal intensity and is easily distinguishable from the low
signal intensity of the cervical stroma. MRI is recommend-
ed in evaluating cervical carcinoma patients with clinical
stage IB disease or greater when the primary lesion is
>2 cm, because of the relatively high likelihood of parame-
trial invasion and/or lymph node metastases [13, 15]. The
staging accuracy of MRI ranges from 75 to 96%. The
reported sensitivity of MRI in the evaluation of parametrial
invasion is 69% and the specificity 93% [13, 15].
Although the FIGO system does not include radiology
in the staging of cervical cancer, the revised FIGO stag-
ing criteria for cervical carcinoma [1], implemented as of
June 1, 2009, encourage the use of imaging techniques,
if available, to assess important prognostic factors such as
parametrial and pelvic side wall invasion, tumor size, and
lymph node metastases. Imaging is therefore complimen-
tary to the clinical assessment. The FIGO committee
made three changes that impact on radiology vs MRI of
the pelvis.
1. The use of diagnostic imaging techniques to assess the
size of the primary tumor is now encouraged by FIGO
but is still not mandatory. MRI is highly accurate in
measuring tumor size, which can affect prognosis and
treatment.
2. Stage IIA has been subdivided: Stage IIA1 consists of
tumors without parametrial invasion that are ≤4 cm in
diameter and involve less than the upper two-thirds of
the vagina. Stage IIA2 tumors are without parametrial
invasion, >4 cm in diameter, and involve less than the
upper two-thirds of the vagina.
121
3. Examination under anesthesia (EUA), which includes
a cystoscopy and proctoscopy, is now recommended
rather than mandatory, as previously stated. MRI has
excellent imaging capabilities in assessing tumor in-
volvement of the bladder and of the rectosigmoid,
therefore avoiding an additional invasive, timely, and
costly EUA procedure. This is especially helpful in
clinical stage I and II disease, when the potential for
invasion and hence upgrading of the tumor staging af-
ter EUA is low.
The most important issue in the staging of cervical
cancer is to distinguish early disease (stage IIA1 and be-
low), which is treated with primary surgery, from ad-
vanced disease, which is treated with radiation, either
alone or in combination with chemotherapy. The excep-
tion is stage IB2, in which the lesion is >4 cm in diame-
ter and is treated as for advanced disease.
Stage IA and its subdivisions are defined as microinva-
sive tumors that cannot be reliably demonstrated on T2-
weighted images. However, dynamic contrast-enhanced
MRI may detect microinvasive disease as a focally en-
hanced area seen in the early dynamic phase. In distin-
guishing deep invasion (>3 mm) from superficial disease,
the accuracy of T2-weighted MRI, dynamic MRI, and con-
trast-enhanced T1-weighted imaging is 76, 98, and 63%,
respectively [16]. Stage IB is defined as clinically visible
lesions limited to the cervix uteri and is subdivided into
stage IB1, in which lesions are <4 cm in their greatest di-
mension, or stage IB2, in which lesions are >4 cm in their
greatest dimension. The carcinoma appears as a mass of
high signal intensity in contrast to the low signal intensity
of the cervical stroma on T2-weighted images.
Young women with stage IA or small IB (<2 cm) tu-
mors who wish to retain their fertility may be considered
for trachelectomy, in which the cervix is excised but the
uterine body and hence fertility is preserved. MRI is
highly accurate in predicting myometrial invasion, with a
sensitivity and specificity of 100 and 99% respectively.
For internal os involvement, the sensitivity of MRI is
90% and the specificity 98%.
Stage IIA is defined as a tumor that invades the upper
two-thirds of the vagina without parametrial invasion.
Segmental disruption of the hypointense vaginal wall is
demonstrated on T2-weighted images. When the tumor
extends beyond the uterus, with parametrial invasion,
it is defined as stage IIB. Spiculated irregular tumor/
parametrial interface, soft-tissue extension into the parame-
tria, or encasement of the peri-uterine vessels is required
to confidently diagnose parametrial invasion (Fig. 2).
MRI has a specificity and negative predictive value of 97
and 100%, respectively, in evaluating parametrial inva-
sion [3]. An important pitfall is the overestimation of
parametrial invasion on T2-weighted images of large
tumors (accuracy of 70%) compared to smaller tumors
(accuracy 96%). Large tumors can cause stromal edema
by tumor compression or inflammation – a fact that must
be considered when making treatment decisions in these
patients.
122
a b
In stage IIIA, the tumor involves the lower third of the
vagina without extending to the pelvic side wall (>3 mm
from pelvic side wall). When the tumor extends to the
pelvic side wall (pelvic musculature or iliac vessels) or
causes hydronephrosis, it is defined as stage IIIB.
If the tumor invades the bladder or rectal mucosa it is
stage IVA. There is segmental disruption of the low sig-
nal intensity of the bladder or rectal wall or segmental
thickening of the rectal wall. Prominent strands between
the tumor and the rectal wall may also indicate rectal in-
vasion. MRI can confidently exclude bladder or rectal in-
volvement, with a negative predictive value of 100%
[17]. Distant metastases define stage IVB disease. Al-
though pelvic lymph node metastases do not change the
FIGO stage, para-aortic or inguinal lymph node metas-
tases are also defined as stage IVB.
CT has a limited role in staging cervical cancer due
to its low accuracy in the detection of early parametrial
extension [18]. However, CT has a diagnostic accuracy
of approximately 90% in staging advanced cervical car-
cinoma and is very useful in evaluating the presence of
distant metastases [19]. PET/CT allows the identification
of involved nodes when CT findings are negative, result-
ing in a change in management in up to 25% of patients.
In the detection of recurrent cervical cancer, it has a
reported sensitivity, specificity, and accuracy of 90.3,
81.0, and 86.5, respectively. This is especially valuable to
exclude the presence of distant disease prior to pelvic
exenteresis [20].
Ovarian Carcinoma
Ultrasound enables the detection and characterization of
adnexal masses but has no role in staging. It can guide
biopsy of adnexal or peritoneal masses in patients
deemed unsuitable for primary surgery. CT is currently
the modality of choice in staging ovarian cancer and can
also be used to guide the biopsy of peritoneal or adnexal
disease. CT provides information on the primary tumor,
Evis Sala, Susan Ascher
Fig. 2 a, b. a Stage IIB cervical carcinoma.
Sagittal T2-weighted image demonstrates an
intermediate signal intensity mass centred on
the posterior fornix (white arrow). b Axial
oblique T2-weighted image demonstrates
parametrial invasion with spiculated irregu-
lar tumor to parametrium interface on the
right (black arrows)
the site and size of peritoneal deposits, and the presence
of enlarged lymph nodes and ascites (Fig. 3). This infor-
mation stratifies those patients with non-resectable dis-
ease, for whom neoadjuvant chemotherapy would be ben-
eficial, from those patients who should undergo primary
cytoreductive surgery. The primary ovarian tumor may be
seen as mixed solid/cystic tumors, which are often bilat-
eral, or as multilocular cystic lesions with thick internal
septations and solid mural or septal components. Assess-
ment can often be made as to whether the tumor invades
the pelvic side wall or rectosigmoid colon or bladder, and
associated complications, such as hydronephrosis and
bowel obstruction, can be identified. Peritoneal deposits
can be clearly identified; they are usually seen as discrete
enhancing soft-tissue nodules. Liver, lung, and renal
metastases and malignant pleural effusion indicate stage
IV disease. PET/CT is of value in cases of suspected re-
currence in which there is an increase in the level of the
tumor marker CA-125 but indeterminate findings on CT
or MRI [20].
Currently, the main role of MRI is in the characteriza-
tion of ovarian masses rather than the staging of histo-
logically proven ovarian cancer. MRI is very sensitive
(95%) in the detection of peritoneal metastases, which
show delayed enhancement on contrast-enhanced MRI
[21]. Gadolinium-enhanced MRI is comparable to la-
parotomy but superior to serum CA-125 levels in the de-
tection of residual or recurrent peritoneal and serosal im-
plants in women who have been treated for ovarian can-
cer [21, 22]. MRI plays a crucial role in the detection of
recurrent disease. It is important to realize that second-
look surgery is no longer routine and imaging diagnosis
of recurrence may obviate a second-look laparotomy
since secondary cytoreduction is only justified if resec-
tion is possible with no residual tumor. Imaging findings
that indicate non-resectable recurrent tumor are invasion
of the pelvic side wall, which should be suspected when
the primary tumor lies within 3 mm of the pelvic side
wall or when the iliac vessels are surrounded or distorted
by tumor.
Malignant Diseases of the Female Genital Tract
a b
c d
Fig. 3 a-d. Advanced ovarian carcinoma. Con-
trast-enhanced CT images show (a) a left
pleural effusion and a subcapsular liver de-
posit (arrow) and (b) peritoneal deposits along
the surface of the liver and gastrosplenic liga-
ment (arrows). There are also (c) a large omen-
tal cake and serosal deposits in the right para-
colic gutter (arrows), (d) bilateral solid cystic
ovarian masses and a large peritoneal deposit
in the Pouch of Douglas (arrows)
References
1. Pecorelli S (2009) Revised FIGO staging for carcinoma of the
vulva, cervix, and endometrium. Int J Gynaecol Obstet
105:103-104
2. Sala E, Wakely S, Senior E, Lomas D (2007) MRI of malig-
nant neoplasms of the uterine corpus and cervix. AJR Am J
Roentgenol 188:1577-1587
3. Shibutani O, Joja I, Shiraiwa M et al (1999) Endometrial car-
cinoma: efficacy of thin-section oblique axial MR images for
evaluating cervical invasion. Abdom Imaging 24:520-526
4. Manfredi R, Mirk P, Maresca G et al (2004) Local-regional
staging of endometrial carcinoma: role of MR imaging in sur-
gical planning. Radiology 231:372-378
5. Sala E, Crawford R, Senior E et al (2009) Added value of dy-
namic contrast-enhanced magnetic resonance imaging in pre-
dicting advanced stage disease in patients with endometrial
carcinoma. Int J Gynecol Cancer 19:141-146
6. Fujii S, Matsusue E, Kigawa J et al (2008) Diagnostic accura-
cy of the apparent diffusion coefficient in differentiating be-
nign from malignant uterine endometrial cavity lesions: initial
results. Eur Radiol 18:384-389
7. Shen SH, Chiou YY, Wang JH et al (2008) Diffusion-weight-
ed single-shot echo-planar imaging with parallel technique in
assessment of endometrial cancer. AJR Am J Roentgenol
190:481-488
8. Harry VN, Semple SI, Gilbert FJ, Parkin DE (2008) Diffusion-
weighted magnetic resonance imaging in the early detection of
response to chemoradiation in cervical cancer. Gynecol Oncol
111:213-220
123
9. Fujii S, Matsusue E, Kanasaki Y et al (2008) Detection of peri-
toneal dissemination in gynecological malignancy: evaluation
by diffusion-weighted MR imaging. Eur Radiol 18:18-23
10. Rockall AG, Sohaib SA, Harisinghani MG et al (2005) Diag-
nostic performance of nanoparticle-enhanced magnetic reso-
nance imaging in the diagnosis of lymph node metastases in
patients with endometrial and cervical cancer. J Clin Oncol
23:2813-2821
11. Ascher SM, Reinhold C (2002) Imaging of cancer of the en-
dometrium. Radiol Clin North Am 40:563-576
12. Hricak H, Rubinstein LV, Gherman GM, Karstaedt N (1991)
MR imaging evaluation of endometrial carcinoma: results of
an NCI cooperative study. Radiology 179:829-832
13. Nicolet V, Carignan L, Bourdon F, Prosmanne O (2000) MR
imaging of cervical carcinoma: a practical staging approach.
Radiographics 20:1539-1549
14. Okamoto Y, Tanaka YO, Nishida M et al (2003) MR imaging
of the uterine cervix: imaging-pathologic correlation. Radio-
graphics 23:425-445
15. Scheidler J, Heuck AF (2002) Imaging of cancer of the cervix.
Radiol Clin North Am 40:577-590,vii
16. Seki H, Takano T, Sakai K (2000) Value of dynamic MR imag-
ing in assessing endometrial carcinoma involvement of the
cervix. AJR Am J Roentgenol 175:171-176
17. Rockall AG, Ghosh S, Alexander-Sefre F et al (2006) Can
MRI rule out bladder and rectal invasion in cervical cancer
to help select patients for limited EUA? Gynecol Oncol
101:244-249
18. Hricak H, Gatsonis C, Chi DS et al (2005) Role of imaging
in pretreatment evaluation of early invasive cervical cancer:
124
results of the intergroup study American College of Radiology
Imaging Network 6651-Gynecologic Oncology Group 183. J
Clin Oncol 23:9329-9337
19. Mitchell DG, Snyder B, Coakley F et al (2006) Early invasive
cervical cancer: tumor delineation by magnetic resonance
imaging, computed tomography, and clinical examination, ver-
ified by pathologic results, in the ACRIN 6651/GOG 183 In-
tergroup Study. J Clin Oncol 24:5687-5694
20. Subhas N, Patel PV, Pannu HK et al (2005) Imaging of
pelvic malignancies with in-line FDG PET-CT: case exam-
Evis Sala, Susan Ascher
ples and common pitfalls of FDG PET. Radiographics
25:1031-1043
21. Ricke J, Sehouli J, Hach C et al (2003) Prospective evalua-
tion of contrast-enhanced MRI in the depiction of peritoneal
spread in primary or recurrent ovarian cancer. Eur Radiol
13:943-949
22. Low RN, Duggan B, Barone RM et al (2005) Treated ovarian
cancer: MR imaging, laparotomy reassessment, and serum
CA-125 values compared with clinical outcome at 1 year. Ra-
diology 235:918-926

Magnetic Resonance Imaging of Prostate Cancer
Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer
Department of Radiology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands
Introduction
With a total of 192,280 new cases predicted for 2009,
prostate cancer (PC) now accounts for 25% of all new
male cancers diagnosed in the USA [1]. Furthermore, in
their lifetime, one in six men will be clinically diagnosed
with PC, although many more will be found to have his-
tological evidence of PC at autopsy [2-4]. Presently, ap-
proximately one in ten men will die of PC [5, 6]. The
ever-aging population and more widespread use of the
blood prostate-specific antigen (PSA) test [7, 8], as well
as the tendency to apply lower cut-off levels for this test
[9], will further increase the diagnosis of PC [10].
An elevated PSA level, abnormal changes in PSA lev-
el and dynamics (such as PSA velocity or doubling time),
or an abnormal digital rectal examination are biological
indicators signaling an increased risk of PC. With the im-
provement and wider range of curative therapies, the de-
tection and subsequent exact localization of PC have be-
come increasingly important because of their influence
on treatment strategy [11, 12], in particular, laparoscopic
(robotic) radical prostatectomy and intensity-modulated
radiation therapy (IMRT) [13]. The urologist’s inability to
palpate the operating field during laparoscopic surgery
makes it even more crucial to precisely localize the can-
cer. Moreover, the urologist must determine whether the
cancer is near a neurovascular bundle since this affects
the decision of whether or not to perform nerve-sparing
prostatectomy [14]. IMRT also necessitates accurate PC
localization. While the prostate receives a standard dose
of radiation, a higher (boost) dose can be given to domi-
nant intraprostatic lesion(s) since it is those lesions that
regularly appear to be the sites of recurrent disease [15].
Furthermore, precision radiation dosimetry will decrease
radiation complications, particularly rectal wall toxicity
[16], thereby likely diminishing the development of post-
radiation rectal cancer [17].
In order to determine the optimal treatment for each
patient, it is necessary to thoroughly evaluate him and to
determine the cancer’s characteristics. In this regard, lab-
oratory values (PSA level and dynamics), the results of
the digital rectal examination (clinical staging), and
histopathological prostatic biopsy findings (Gleason
IDKD 2010-2013
score) are important aspects. Additionally, however, mag-
netic resonance imaging (MRI) can play an important
role in detecting and localizing those areas most reflec-
tive of the actual aggressiveness of the cancer. This di-
rectly influences patient assessment and may lead to im-
portant changes in treatment strategy, which can mean the
difference between treatment success and failure.
In the mid-1980s, the first prostate MRI examina-
tions were performed. Since that time MRI has evolved
from a promising technique into a mature imaging
modality for PC imaging [18, 19]. Beside anatomical in-
formation, MRI provides functional tissue-characteristic
information. Multiparametric MRI consists of a combi-
nation of anatomical T2-weighted imaging and func-
tional MRI techniques such as dynamic contrast-
enhanced MRI (DCE-MRI), diffusion-weighted imag-
ing (DWI), and 1H MR-spectroscopic imaging (MRSI).
Within a multiparametric MRI examination the relative
value of its component techniques differ. In addition to
T2-weighted MRI, which mainly assesses anatomy,
MRSI [20] can add specificity for PC detection, while
DCE-MRI [21] and DWI [22] are both very sensitive
and very specific.
The clinical challenges in the work-up of patients with
either suspected or proven PC include detection, local-
ization, TNM staging, determination of cancer aggres-
siveness, follow-up of patients in active surveillance pro-
tocols, and determination of the site and extent of cancer
recurrence after therapy.
This chapter reviews the MRI anatomy of the prostate
and the basic MRI techniques that can be applied in PC,
and discusses the clinical role of this imaging modality.
At the end of this chapter, three clinically applicable pro-
tocols are provided.
Magnetic Resonance Imaging: Anatomy
In order to effectively apply the various MRI techniques,
it is important to first understand the prostate’s normal
anatomy and its intrinsic age-related changes. The supe-
rior part of the prostate is called the base while its most
caudal part is referred to as the apex, analogous to the
126 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

a b
c

d e

Coronal

Sagittal Axial
Fig. 1 a-e. Normal prostate with signs of benign prostatic hyperplasia (BPH) as seen on sagittal (a, d), coronal (b), and axial (c, e) high-
resolution T2-weighted images. The peripheral zone is white, BPH is blue, the urethra is yellow, and the seminal vesicles are green

anatomy of the heart. The prostate consists of three
zones: (1) the peripheral zone, located posteriorly and
caudally at its middle portion; (2) the transition zone, lo-
cated interiorly, around the urethra; and (3) the central
zone, which is posterior and superior to the transition
zone. Ventral to the prostate is the anterior fibromuscular
stroma.
In aging, an important frequent change in prostate zon-
al anatomy occurs, namely, the transition zone becomes
hypertrophic (as in benign prostatic hyperplasia), thus
compressing the central gland. Consequently, most men
who are imaged for prostate cancer have only two identi-
fiable compartments in the prostate, the hyperplastic tran-
sition zone surrounded by the peripheral zone (Fig. 1).
Up to 70-80% of PCs are located in the peripheral
zone [23], with an overall analysis of these cancers show-
ing that they are homogeneously distributed across the
entire zone [24]. Additionally, over half of the prostates
examined contained two or more distinct cancer foci [25].
Nevertheless, while up to 20-52% of all PCs originate in
the transition zone, only a small percentage (3.6-25%) of
these cancers occur solely in that zone [24, 26], and many
such patients will have foci of concurrent peripheral-zone
cancer [23, 27, 28]. Thus, a solitary transition zone can-
cer is rare in the general PC population.
Magnetic Resonance Techniques and Their Role in
Detection and Localization
For evaluation of the prostate, anatomical (high-resolu-
tion) MRI can be combined with functional and meta-
bolic information. DWI, dynamic MRI, and MRSI pro-
vide information about the motion of free water mole-
cules and thus about cellular density (neo-)vasculariza-
tion and metabolism, respectively. These different types
of information can be combined into a multiparametric
MRI examination.
T2-Weighted Imaging
Compared to CT computed tomography (CT) scanning,
MRI has a high soft-tissue contrast resolution (Fig. 2).
The use of a disposable endorectal coil combined with
other external coils at 1.5 Tesla (T) increases the soft-
tissue contrast significantly and is now the accepted clin-
ical standard for MRI of the prostate, when information
about submillimeter extracapsular penetration is of clini-
cal importance [29]. A drawback is the extra time re-
quired for inserting and checking the position of the
endorectal coil as well as the substantial expense and
patient discomfort.
Magnetic Resonance Imaging of Prostate Cancer 127

a b

Fig. 2 a, b. A 60-year-old patient with a urinary

catheter, and stage 2b prostate cancer (PC) in
the left peripheral zone. a Axial MDCT does
not well-delineate the prostate and the tumor
is not visible. b T2-weighted axial MRI
shows good prostate delineation and tumor in
the left peripheral zone (arrow) without cap-
sular penetration

Fig. 3 a, b. a On axial T1-weighted MRI, the

post-biopsy hematoma (H (H) is white. b On the
T2-weighted image, both the hematoma (H (H),
benign prostatic hyperplasia ((B), and tumor
(*) are of low signal intensity

On MRI, PC typically appears as an area of low sig-
nal intensity within the brighter, healthy peripheral
zone, as seen using a T2-dominated sequence [30-32]
(Figs. 2b, 3). In the central gland, PC is not as clearly
discernible because the transition zone generally has
lower signal intensity than the peripheral zone and is
more inhomogeneous due to the architectural changes
induced by benign prostatic hyperplasia, which may
mimic PC. A recent study showed that a homogeneous-
ly low T2 signal intensity and lenticular shape were sig-
nificantly associated with the presence of transition-
zone [4, 33]. It was reported that, relative to muscle,
cancers with higher Gleason scores had lower signal in-
tensities than cancers with low Gleason scores [32].
However, the number of patients in that study was lim-
ited. In a comparison of T2-weighted MRI with prosta-
tectomy specimens, MRI attained 52-83% sensitivities
in PC localization, while specificities were somewhat
lower (46-88%).
A study that directly compared endorectal MRI with
digital rectal examination and transrectal ultrasound
(TRUS)-guided biopsy localization revealed significant
incremental value from MRI [34]. In patients subjected to
multiple prior negative TRUS-guided biopsies, anatomi-
cal MRI by means of T2-dominated acquisition plays an
important role. In this patient population, a sensitivity of
83% and positive predictive value of 50% for MRI have
been established [35].
Post-biopsy hemorrhage causes areas of low signal in-
tensity on T2-dominated sequences, thereby making PC
detection more difficult. However, it was shown recently
that the amount of hemorrhage was significantly lower in
areas of cancer than in healthy tissue [36].
Diffusion-Weighted Imaging
This non-invasive technique measures the fractional
anisotropy of water molecules within the prostate, ex-
pressed in apparent diffusion coefficient (ADC) map-
ping. Thereby, the movement of water molecules in can-
cer tissue has been shown to be more restricted, thus pro-
ducing lower ADC values [37, 38]. In a recent study of
38 patients who underwent DWI at 1.5 T with an en-
dorectal coil, the mean ADC values of regions of interest
128 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer
Fig. 4. Axial T2-weighted image obtained from a 70-year-old male
with PSA = 35 and six negative TRUS biopsies shows a homo-
geneous area of low signal intensity in the right ventral prostate
(arrows)
Fig. 5. Same patient as in Fig. 4. The ADC map (color coded) shows
restriction in the area suspicious for tumor on this T2-weighted im-
age (arrows)
within PC tissue were significantly lower than those with-
in healthy prostate tissue [39]. In preliminary studies, the
combination of this technique with MRSI [39] or T2-
weighted imaging [40] significantly improved localiza-
tion accuracy (Figs. 4, 5). This was confirmed in a more
recent study in 37 patients, in which sensitivity increased
from 51% for T2-weighted imaging to 71% when the lat-
ter was combined with DWI [41]. In a recent multipara-
metric analysis, DWI was the best-performing parameter
in localization [42]. Preliminary studies at 3 T have
shown promising results [43-45].
Dynamic Contrast-Enhanced MRI
The use of intravenous contrast agents further enhances
the localization accuracy of MRI. Endorectal DCE-
MRI, in which the contrast agent concentration is fol-
lowed over time [46], discriminates between healthy
prostate tissue and PC [47]. Early contrast enhancement
and high (relative) peak enhancement are the most ac-
curate predictors of PC of the peripheral zone, while
fast washout of contrast agent and high permeability of
the blood vessels (Fig. 6) are most sensitive for transi-
tion-zone PC [48, 49]. A recent study showed that the
area under the receiver operating curve (AUC) for lo-
calizing PC increased significantly, from 0.68 with reg-
ular anatomical MRI to 0.91 with contrast-enhanced
MRI. However, the limitations associated with the use
of contrast agents are the lack of a uniform threshold,
the low sensitivity for cancer, higher costs, and possible
adverse reactions, of which the most serious is nephro-
genic systemic fibrosis [50, 51].
1H Magnetic Resonance Spectroscopic Imaging
Additionally, MRSI (Fig. 7) can be added to the imag-
ing protocol to provide metabolic information based on
the tissue citrate, choline, and creatine concentrations,
and their relative ratios within the prostate. This is high-
ly informative since the ratio between choline and cit-
rate is markedly altered during the transformation from
healthy to malignant prostate cells [52, 53] and an in-
creasing choline+creatine/citrate ratio has been corre-
lated with higher Gleason scores [54]. Presently, three-
dimensional (3D) MRSI of the entire prostate can be
performed [55], thereby aiding in the diagnosis of
Fig. 6. Same patient as in Fig. 4. Ktrans-map obtained with dynam-
ic contrast-enhanced MRI shows pathological enhancement in area
suspicious for tumor on this T2-weighted image (arrows)
Magnetic Resonance Imaging of Prostate Cancer 129

c
b

Fig. 7 a-c. Images obtained from a 65-year-old male with stage T3a PC in the left peripheral zone. The T2-weighted image (a) shows the tu-
mor. MRSI of the right peripheral zone (b) shows low the choline and high citrate peak, whereas tumor in the left peripheral zone (c) shows
high choline

central-gland PC (Fig. 7). The addition of 3D MRSI to sonable detection rates of 25-55% [62, 63]. Moreover, di-
MRI increases localization accuracy, by raising the rect MRI-GB within the MRI scanner is technically feasi-
specificity to as high as 91% [56]. However, a limitation ble and can be performed on a routine basis. In patients
of MRSI is its low spatial resolution and cumbersome with one previous negative TRUS biopsy, transrectal MRI-
post-processing. Compared to systematic biopsy, PC GB performed at 1.5 T has produced promising cancer de-
localization by means of MRI and MRSI was found to tection rates of 38-56% [64-66]. Lesions >10 mm can suc-
be more sensitive (67 and 76% vs. 50%) but less spe- cessfully be biopsied using this approach [66].
cific (69 and 57% vs. 82%) than systematic biopsy [57]. A multiparametric MRI approach consisting of T2-
With whole-mount-section histopathology as the stan- weighted MRI, DWI, and DCE-MRI performed at 3 T has
dard of reference, 3D MRSI had a significantly larger a median MRI-guided biopsy time of just 35 min and can
AUC (0.80) in localizing cancer than obtained with T2- generate an average of four biopsy cores per patient, as
weighted MRI (0.68) [58]. The combination of T2- recently reported by Hambrock et al. (Fig. 8) [67]. Those
weighted imaging and MRSI information to clinical da-
ta yielded the highest accuracy (AUC 0.85) in predict-
ing the probability that a patient has insignificant PC
[59], which was significantly higher than that obtained
with clinical nomograms. A recent multi-institutional
American College of Radiology Imaging Network study
raised doubts on the additive value of MRSI over T2-
weighted imaging alone [60]. However, potential factors
resulting in this conclusion were the selected prostatec-
tomy population, the small size of the average cancer
focus, and the inclusion of health centers without any
previous MRSI experience.

MRI-Guided Biopsies

Random systematic TRUS-guided biopsy has relatively

low detection rates and is prone to sampling error [61].
MRI-guided prostate biopsy (MRI-GB) has the potential
to increase PC detection, as multiparametric MRI can tar-
get biopsies towards regions previously determined to be Fig. 8. Same patients as in Fig. 4. MRI-guided biopsy of the tumor-
suspicious for cancer. Indeed, MRI findings have been suspicious region (red) showed PC with a Gleason score of 4+3.
used to direct biopsies under TRUS guidance, with rea- The biopsy needle is highlighted in white
130 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

authors showed that a cancer detection rate of 59% can
be achieved in a large cohort of patients with more than
two previous negative TRUS biopsies [68]. In addition,
93% of the cancers found were clinically significant, thus
not contributing to the over-diagnosis of insignificant
cancers.
A limitation of MRI-GB is that multiparametric MRI
for tumor localization and MRI-guided biopsy need to be
performed in two different sessions, as image post-
processing and exact localization of the cancer demand
time. Another disadvantage is movement of the prostate
during the biopsy procedure [69]. A reduction of the
MRI-GB intervention time remains an important chal-
lenge, perhaps solvable by robotics. In the future, MRI-
based guidance might also be used in the focal treatment
of PC, such as in the form of brachytherapy or cryotherapy.
Prediction of Prostate Cancer Aggressiveness
Prostate cancer aggressiveness is pathologically graded
by the Gleason score, which consists of a combination of
the two most prevalent Gleason grades (range 1-5) based
on the architectural characteristics of PC tissue [70].
Biopsy specimens obtained from random TRUS guided-
biopsy are subject to sampling error in approximately
64% of the cases [71]; this results in incorrect Gleason
scores and thus incorrect patient risk stratification, which
in turn leads to under- or overtreatment [72].
Apart from a relationship between muscle-normalized
signal intensity on T2-weighted MRI and cancer Gleason
scores [73], a correlation between cancer visibility on
T2-weighted images and aggressiveness has been sug-
gested, with low-grade cancers being detected in 43%
and high-grade cancers in 79% of such cases [74]. More-
over, (choline+creatine)/citrate ratios, as determined by
MRSI, have been shown to correlate with the Gleason
score [75-77]. Preliminary results in the evaluation of
ADC as a marker of cancer aggressiveness are promis-
ing; ADC values were found to negatively correlate
(ρ = –0.497, p<0.0001) with peripheral-zone cancers and
Gleason scores. In another study, lesions in patients with
less aggressive cancers (PSA <10, T1 or T2a, Gleason
score <6) had significantly higher ADC values than le-
sions in patients with intermediate- to high-risk cancers
(PSA >10, T2b, Gleason score ≥7) (Fig. 9) [77-79].
Local (T) Staging
The application of MRI to determine whether PC is lo-
cally advanced remains controversial due to varying re-
sults across institutions. The most reliable MRI signs of
extracapsular extension are bulging of the prostate into
the periprostatic fat, obliteration of the recto-prostatic
angle, and asymmetry of the neurovascular bundles
(Fig. 10) [80]. Seminal vesicle invasion is usually easi-
ly detectable as areas of low signal intensity in the
brighter seminal fluid (Fig. 10).
Two meta-analyses on local staging by MRI at 1.5 T
reported combined maximum sensitivities and specifici-
ties of 71-74%, while sensitivity was 62-69% at a speci-
ficity of 80% [80-82]. Imaging in more than one plane as
well as the use of an endorectal coil resulted in a signif-
icantly better staging performance. A large study of 336
patients conducted by Cornud et al. found an overall sen-
sitivity, specificity, and positive and negative predictive
values of 40, 95, 79, and 76%, respectively [83]. High-
specificity MRI (in which only definite locally advanced
cases are excluded from curative therapy) is now the
optimal local staging method [84, 85].
The addition of MRI with an endorectal coil to clini-
cal data such as PSA, biopsy Gleason score, and Kattan
nomogram resulted in a significantly increased accuracy
of predicting disease stage, extracapsular extension, and

a b c

Fig. 9 a-c. Images from a 62-year-old patient with PSA = 9 and PC with a Gleason score of 3 but with a local component with a score of 4.
a Axial T2-weighted MRI shows low signal in almost the entire peripheral zone. b On the ADC map, the tumor can be delineated based
on its lower value (uninterrupted line) and on the presence of local areas with very low value (stippled areas). c Prostatectomy confirmed
the lower signal to be a Gleason 3 tumor and the very low signal component to be Gleason 4
Magnetic Resonance Imaging of Prostate Cancer 131

b d

Fig. 10 a-d. T2-weighted images from a 49-year-old male with stage T3B PC. a Coronal im-
age at basis shows seminal vesicle infiltration (*); b axial image at the mid-part shows
bulging (arrows); c axial image at the apex shows obliteration of the recto-prostatic angle
(arrow) and, for comparison, the normal recto-prostatic angle (drawn line); d schematic
drawing, as part of the structured report

seminal vesicle invasion [79, 86, 87]. Recently, it was
shown that the presence and the degree of pre-radiation
therapy extracapsular extension predicted by MRI was a
predictor of post-therapy outcomes [88].
The addition of DWI to T2-weighted imaging was re-
cently shown to be significantly better in establishing uri-
nary bladder wall invasion as well as seminal vesicle in-
vasion [89, 90]. Also, the addition of three-dimensional
MRSI to MRI improved staging accuracies, particularly
for less-experienced readers, and increased interobserver
agreement [91]. A drawback is the longer duration (by
approximately 15 min) of the examination.
Experience was found to be an important factor in dis-
ease staging [92]. However, the accuracy of a less-expe-
rienced reader could be increased by contrast-enhanced
examinations [93]. Likewise, MRI interpretation using
multiplanar cross-referencing significantly improved
staging accuracy compared with interpretation without
cross-referencing [94]. Interactive tutorials with direct
feedback were also shown to significantly increase the
accuracy of staging by less-experienced readers [95].
Imaging at higher magnetic field strengths (e.g., 3 T)
can achieve better image resolution. Although not yet
widely available for clinical work, local staging at 3 T
was shown in two studies to improve the sensitivities and
specificities of experienced readers to 80-88% and 94-
100%, respectively [96, 97]. In the current PSA era, this
higher resolution is mandatory as PC is detected at earli-
er stages. Likewise, if extracapsular extension is present,
it will most often be minimal.
Metastatic Disease (NM Staging)
Nodes
The prognosis of patients with PC is poorer if lymph
node metastases are present. The risk of lymph node
metastasis is currently determined (albeit inaccurately)
using nomograms [98, 99]. In patients with an elevated
risk for metastasis, additional examinations are required.
Today, the most commonly used imaging techniques for
detecting lymph node metastasis are multi-detector CT
scan (MDCT) and conventional MRI, with image inter-
pretation essentially based on lymph node size and shape
criteria. Although the criteria vary slightly [100], lymph
nodes with a short-axis diameter >8 mm for round lymph
nodes and >10 mm for oval ones are generally considered
to be malignant [101, 102]. Both MDCT [103] and MRI
[104] have a low sensitivity (36 and 39%, respectively)
for diagnosing PC lymph node metastases using these
size and shape criteria. In studies that have employed
thresholds as small as 6 mm [105], the specificity was
very high (95-100%) but the sensitivity was too low
(0-25%) to be useful in regular clinical practice for the
evaluation of metastatic lymph node disease [106]. Some
authors advocate restricting the application of these tech-
niques to high-risk patients (e.g., with PSA levels >20
ng/mL) in order for them to be cost-effective [107, 108].
Thus, supplementary, invasive diagnostic examinations in
the form of surgical pelvic lymph node dissection
(PLND) are still commonly performed.
132 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer
Magnetic resonance lymphography (MRL) using a
lymph-node-specific contrast agent (Combidex/Sinerem)
[109, 110] is an experimental technique that, compared to
PLND, has a high negative predictive value (>96%) for
the detection of lymph node metastasis in extended areas.
Importantly, its use can render PLND unnecessary in
negative cases [111].
Bone
Most metastatic bone lesions are sclerotic [112]; a 50%
change in bone mineral density is needed for metastat-
ic bone lesions to be visible on X-ray images [113]. The
most commonly used first-line diagnostic test to detect
or exclude bone metastases is technetium-99m-diphos-
phonate bone scintigraphy. However, this approach
lacks specificity, such that primary skeletal diseases
may generate false-positive findings. Conventional X-
ray examinations can be used to exclude false-positive
a b
findings on bone scintigraphy that have arisen due to
conditions such as trauma, degenerative joint disease,
and other chronic diseases. However, conventional X-
ray is too insensitive for the detection of metastatic bone
lesions. Lecouvet et al. evaluated the accuracy of bone
scintigraphy, targeted X-rays, and MRI in 66 patients
with prostate cancer, 41 of whom had bone metastases
(Fig. 11) [114]. Sensitivities were 46% for bone scinti-
graphy alone, 63% for bone scintigraphy and targeted
X-rays, and 100% for MRI; the corresponding specificities
were 32, 64, and 88%, respectively. Thus MRI was
significantly more sensitive than any other approach
( <0.001); furthermore, MRI limited to the pelvis and
(p
axial skeleton was shown to be sufficient, as the proba-
bility of finding metastases outside these locations in
the absence of metastases in the axial skeleton is negli-
gible. This is particularly the case in PC, which pre-
dominantly metastasizes to the spine and pelvis due to
the venous drainage routes [115, 116].
Fig. 11 a-c. Patient with bone metastasis from
PC. a T1-weighted image shows a small, low
signal intensity lesion (arrow) with high inten-
sity on DWI (b), consistent with metastasis.
The bone scan (c) and plain film were negative
Magnetic Resonance Imaging of Prostate Cancer
Recurrence
Currently, a rise in PSA level is the only major indicator
for PC recurrence after radical prostatectomy or radio-
therapy [117, 118]. The recurrence of PC is indicated by
PSA values >0.4 ng/mL following radical prostatectomy
and by a PSA value 2 ng/mL above the nadir value (after
the PSA bounce) following radiation therapy [119, 120].
Whenever such an elevation of PSA occurs, the main ob-
jective is to determine whether it is due exclusively to re-
current or residual disease in or outside the prostate,
since locally recurrent disease might still be cured with
adjuvant radiotherapy.
The use of DRE and TRUS for recurrence detection is
compromised because of the difficulty in distinguishing
between fibrotic changes and recurrent disease. TRUS-
guided biopsy has a low sensitivity and specificity with
respect to recurrence following prostatectomy or radio-
therapy [121, 122] whereas MRI can help in the detection
of recurrent disease. One of the major advantages of MRI
is that it can be used to evaluate the recurrence of PC at
either local or distant sites.
After external-beam radiation therapy, prostatic tissue
demonstrates diffusely low signal intensity on T2-weighted
images, with indistinct zonal anatomy [123]. The contrast
between benign, irradiated tissues and recurrent cancer is
a b
Fig. 12 a, b. A 69-year-old patient 2.5 years af-
ter external radiation beam therapy; his PSA
had risen to 2.1. a T2-weighted MRI shows
entirely low signal gland; no tumor can be
discriminated. b DCE-MRI shows enhance-
ment of the right half of the prostate (arrow).
Biopsy revealed Gleason 5+4 PC
a b
Fig. 13 a, b. Patient with PSA recurrence after
prostatectomy. a T1-weighted image shows
no lesion whereas (b) DWI shows enhance-
ment (red circle), which was due to recur-
rence of the tumor (Gleason 4+3)
133
therefore decreased, which makes the detection of recur-
rence more difficult. In one of the rare studies that used
radical prostatectomy specimens after radiotherapy as the
standard of reference, retrospectively assessed endorectal
MRI had a low to fair accuracy for the detection of local
post-radiotherapy recurrence (AUC 0.61-0.75), the pre-
diction of extracapsular extension (0.76-0.87), and evi-
dence of seminal vesicle invasion (0.70-0.76) [124].
Functional MRI techniques are of additional value in
the detection of disease recurrence following radiation
therapy (Fig. 12). On MRSI, the presence of three or
more suspicious voxels in a prostate half had a sensitivi-
ty of 89% and a specificity of 82% in the detection of lo-
cal post-radiotherapy recurrence (n=23) [125]. Further-
more, post-radiotherapy DCE-MRI was shown to have a
sensitivity of 70-74% and a specificity of 73-85% for re-
currence detection [126, 127]. The addition of DWI to
T2-weighted imaging (AUC 0.61) also improved the de-
tection of post-radiotherapy disease recurrence by 27%
(AUC 0.88) [128].
In a retrospective study evaluating the value of anatom-
ical T2-weighted MRI in the detection of post-prostatec-
tomy disease recurrence, the sensitivity was reported to be
95% and the specificity 100% [129]. Thus, also in the
evaluation of post-prostatectomy disease recurrence func-
tional MRI techniques are of additional value (Fig. 13). In
134 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer
this regard, DCE-MRI combined with anatomical T2-
weighted MRI was shown to improve sensitivity and
specificity [130, 131]. According to Sciarra et al. [132],
DCE-MRI is even better when used in combination with
MRSI, resulting in AUC values of 0.94-0.96 for the de-
tection of local recurrence compared to values of 0.81-
0.94 for either DCE-MRI or MRSI alone. One of the
limitations in the reported studies was the use of TRUS-
guided biopsies as the standard of reference [131].
Protocols
The European Society of Urogenital Radiology and the
Royal College of Surgeons (UK) are currently working
on a set of guidelines, with the first version to be pub-
lished at the end of 2011 in European Radiology.
Currently, three protocols can be recommended: one
for detection/localization and recurrence, one for staging,
and one for the assessment of nodal size and bone mar-
row. Unfortunately, despite the enormous clinical poten-
tial of Combidex/Sinerem has, due to the inability of the
pharmaceutical companies to provide convincing data to
the FDA and EMEA, this contrast agent was not ap-
proved for nodal imaging and further development has
been discontinued.
Detection, Localization, Recurrence
This is a fast (<30 min) protocol that does not involve the
use of an endorectal coil. The minimum requirements
should include images covering the entire prostate:
1. T1-weighted axial images, to detect post-biopsy
hematomas.
2. T2-weighted axial images and images in one other
plane: 4 mm at 1.5 T, 3 mm at 3 T. The resolution
should be at least 0.5 × 0.5 mm-0.7 × 0.7 mm at both
1.5 and 3 T.
3. DWI axial: 5 mm at 1.5 T, 4 mm at 3 T. The resolution
should be at least 1.5 × 1.5 mm-2.0 × 2.0 mm at 1.5 T
and 1.0 × 1.0 mm-1.5 × 1.5 mm at 3 T, with a b0 im-
age plus multiple b images allowing quantification.
a b
Fig. 14 a-c. This 45-year-old patient requested erectile-function-preserving surgery. a High-resolution T2-weighted image obtained with an
endorectal coil (3T) shows minimal capsular penetration close to the neurovascular bundle (detail in b, arrows). Prostatectomy was done,
saving the right and sacrificing the left neurovascular bundle. c Histopathology revealed submillimeter capsular extension, but negative re-
section margins. T Tumor, ECE extracapsular extension
The highest b value should be 1000 if ADC is calcu-
lated, otherwise it should be 1400 (with adequate sig-
nal to noise ratio).
4. Contrast-enhanced imaging, axial: 4 mm at 1.5 and 3
T. The resolution should be at least 1.0 × 1.0 mm at 1.5
T and 0.7 × 0.7 mm at 3 T. Quantitative or semi-quan-
titative DCE-MRI analysis does not comprise minimal
practice, but if available should be done as optimal
practice. The maximum temporal resolution should be
10 s following a single dose of contrast, with an injec-
tion rate of 3 mL/s. For DCE-MRI, imaging acquisition
should be continued for 5 min to detect washout.
Imaging can adequately be performed at 1.5 T. A
pelvic coil should always be used; bowel relaxants (Bus-
copan, glucagon) provide optimal imaging with fewer
motion artifacts.
Staging
This is a longer (45 min) protocol that allows determina-
tion of (minimal) capsular penetration (Fig. 14). Prefer-
ably, this exam should be done with an endorectal coil.
Minimum requirements should include images covering
the entire prostate:
1. T1-weighted axial, to detect hematomas.
2. T2-weighted axial and two other planes: 3 mm at 1.5
and 3 T. The resolution should be at least 0.3 × 0.3 mm -
0.7 × 0.7 mm at 1.5 T and 0.3 × 0.3 mm - 0.5 × 0.5 mm
at 3 T.
3. DWI axial: 5 mm at 1.5 T, 4 mm at 3 T. The resolution
should be at least 1.5 × 1.5 mm - 2.0 × 2.0 mm at 1.5 T
and 1.0 × 1.0 mm - 1.5 × 1.5 mm at 3 T, with a b0 im-
age plus multiple b images allowing quantification.
The highest b value should be 1000 if ADC is calcu-
lated, otherwise it should be 1400 (with adequate sig-
nal to noise ratio).
4. Contrast-enhanced imaging, axial: 3-4 mm at 1.5 T
and 3 mm at 3 T. The resolution should be at least
1.0 × 1.0 mm at 1.5 T and 0.7 × 0.7 mm at 3 T. Quan-
titative or semi-quantitative DCE-MRI analysis does
not comprise minimal practice, but if available should
be done as optimal practice. The maximum temporal
c
Magnetic Resonance Imaging of Prostate Cancer 135

resolution should be 10 s following a single dose of Interpretation, Structured Reporting

contrast, with an injection rate of 3 mL/s. For DCE-
MRI, imaging acquisition should be continued for 5
min to detect washout.
5. MRSI (optional).
Nodes and Bone
This 30-min protocol is used to assess nodal size and
bone marrow metastases.
1. T1 (SE or f/T SE)-weighted coronal of the lower lum-
bar spine plus pelvis: 3-mm slices.
2. 3D f/T SE T2-weighted coronal of the lower lumbar
spine plus pelvis; 1-mm isometric voxels.
3. DWI (b value 0 plus 600) coronal of the lower lumbar
spine plus pelvis; 2.5-3 mm isometric voxels.
4. T1 (SE or f/T SE)-weighted sagittal images of the cer-
vical and thoracic spine.
5. STIR or DWI sagittal images of the cervical and tho-
racic spine.
Due to a lack of standardization, the combined interpreta-
tion of anatomical and functional images is not possible.
Optimal results are obtained using an integrated comput-
er program that combines T2-weighted anatomical images
with DWI, DCE-MRI, and MRSI on one screen (Fig. 15).
Additionally, there is a dire need for a uniform scoring
system; for example, in which every modality allows the as-
signment of points on a 5-point scale regarding the impres-
sion of a lesion’s malignancy, analogous to the BIRADS
classification system in breast cancer imaging. Thus, a le-
sion score of 20/20 or 15/15 definitely indicates an inter-
mediate or highly aggressive tumor (Fig. 15). If the score
is <10/15 or <14/20, there may be a non-aggressive tumor
but other pathologies, such as prostatitis, are a possibility
(Fig. 16). This type of standardized scoring system allows
the quantification of lesion development and reliable com-
parison with a follow-up MRI examination (Fig. 16).

a b e

c d

Fig. 15 a-f. A 59-year-old patient with PSA = 12 and a Gleason 8 tumor of the right peripheral zone, stage T3a at prostatectomy. Screenshot
from computer monitor display. a DCE-MRI; b ADC-map (color coded); c choline image; d coronal T2-weighted image; e axial T2-weight-
ed image; f time concentration curve and H-spectrum of cursor. This patient’s score for all modalities was 5 points (20/20). Scale: 1 no tu-
mor, 5 definitely tumor
136 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

a b c

d e f

g
Fig. 16 a-g. This 55-year-old patient with PSA = 5, PC of stage T1 at digital rectal examina-
tion and Gleason 3+3 <5% on TRUS biopsy chose active surveillance. a T2-weighted im-
age shows no tumor: score 1/5; b DCE-MRI shows some asymmetrical enhancement with
curve 2: score 3/5; c ADC map shows only a small reduction: score 2/5. The total score of
6/15 argues for no or non-aggressive tumor. Follow-up MRI 1 year later: d T2-weighted im-
age at apex shows homogeneous, asymmetrical low signal: score 5/5; e DCE-MRI shows
pathological asymmetrical enhancement: score 5/5; f DWI shows an area with very low val-
ue: score 5/5. Based on these results, MRI-guided biopsy was performed. A Gleason 5+3
specimen was obtained on prostatectomy (g), which confirmed a Gleason 5+3 tumor

Conclusion cian in choosing the appropriate therapy based on PC

The use of MRI has led to improved localization of PC
compared to the results obtained with transrectal ultra-
sound. This has resulted in increased cancer detection
by targeting biopsy under MRI guidance. Furthermore,
radiotherapy planning and the guidance of minimally in-
vasive local therapies have become possible as a result
of adequate PC localization with MRI. Multiparametric
MRI using functional imaging has been shown to im-
prove the prediction of tumor aggression. Also, com-
pared to staging nomograms, high-resolution (endorec-
tal coil) MRI considerably improves the determination
of local tumor extension, thereby supporting the clini-
stage.
In case of a biochemical recurrence after radical treat-
ment, MRI can detect disease location with greater accu-
racy than obtained with DRE and TRUS. Finally, MRI of-
fers advantages over bone scintigraphy.
Nonetheless, among the different acquisition methods,
protocols, magnetic field strengths, and multimodality
techniques that are used, a consensus is needed regarding
dedicated MRI protocols for different specific clinical in-
dications. Implementation of protocols that on a larger
scale are highly similar if not identical will promote the
maturation of different multimodality techniques and im-
prove further development of supportive techniques such
Magnetic Resonance Imaging of Prostate Cancer
as computer-aided diagnosis. Furthermore, the use of
consensus protocols enables larger studies, the aims of
which include estimation of the true value of multipara-
metric MRI on a broader, more representative scale.
Take-Home Messages
• Currently, MRI is the most accurate imaging modality
in localizing and staging PC.
• Prostate cancer in the central gland of the prostate is
more difficult to detect and localize than PC in the pe-
ripheral zone of the gland.
• Multiparametric MRI comprises a combination of T2-
weighted imaging, DWI, MRSI, and DCE-MRI.
• The typical PC focus is of low signal intensity on T2-
weighted imaging and has a low apparent diffusion co-
efficient value, a high choline+creatine/citrate ratio,
and high contrast agent permeability and fast washout.
• As not all PC foci will exhibit all these features, mul-
tiparametric MRI is needed to obtain the highest lo-
calization accuracy.
• The apparent diffusion coefficient and choline+crea-
tine/citrate ratio are associated with the aggressiveness
of the cancer.
• Direct MRI-guided biopsy of a suspicious lesion in pa-
tients whose previous TRUS biopsies were negative is
feasible, highly accurate, and can be performed within
30 min by an experienced radiologist.
• For staging PC at 1.5 T, the use of an endorectal coil
is necessary; at 3 T, the endorectal coil is only needed
if knowledge of minimal capsular penetration is im-
portant for therapeutic decision-making.
• Conventional nodal staging with MRI is still based on
lymph node size and shape criteria; novel, more spe-
cific techniques still have to be approved by the FDA
and EMEA.
• For bone staging, MRI of the axial skeleton outper-
forms bone scintigraphy in detecting metastatic foci.
• Dynamic contrast-enhanced MRI may play an impor-
tant role in improving the detection of PC recurrence
after surgery or radiation therapy.
Acknowledgements
The authors would like to acknowledge Mrs. Louise
Dickinson from the Royal College of Surgeons and her
guideline group for her input in formulating minimal and
maximal requirements. The final guidelines will appear
by the end of 2010.
Suggested Reading
Anatomy
Coakley FV, Hricak H (2000) Radiological anatomy of the prostate
gland: a clinical approach. Radiol Clin N Amer 38:15-30
137
Hricak H, Dooms GC, McNeal et al (1987) MR imaging of the
prostate gland 148:51-58
McNeal JE (1988) Normal histology of the prostate. Amer J Surg
Pathol 12:619-633
Sommer FG, McNeal JE, Carrol CL (1986) MR depiction of zon-
al anatomy of the prostate at 1.5T. JCAT 1096:983-989
Villiers G, De Meereleer GO (2007) MRI anatomy of the prostate
and application of MRI in radiotherapy planning. Eur J Radiol
63:361-368
Prostate Cancer
Futterer JJ, Heijmink SWTP, Scheenen TWJ et al (2006) Prostate
cancer localization with dynamic contrast enhanced MR imag-
ing and proton MR spectroscopic Imaging. Radiology
241:449-458
Hambrock T, Somford DM, Hoeks C et al (2010) Magnetic reso-
nance imaging guided prostate biopsy in men with repeat neg-
ative biopsies and increased prostate specific antigen. J Urol
183:520-528
Heesakkers RAM, Hoevels AM, Jager GJ et al (2008) MRI with a
lymph-node-specific contrast agent as an alternative to CT scan
and lymph-node dissection in patients with prostate cancer: a
prospective multicohort study. Lancet Oncol 9:850-856
Hrick H, Choyke, Eberhart SC et al (2007) Imaging prostate can-
cer: a multidisciplinary perspective. Radiology 243:28-53
Hoevels AM, Heesakkers RAM, Adang EM et al (2008) The
diagnostic accuracy of CT and MRI in the staging of pelvic
lymph nodes in patients with prostate cancer: a meta-analysis.
Clin Radiol 63:387-395
Jager GJ, Severens JL, Thornbury JR et al (2000) Prostate cancer
staging: should MR imaging be used? A decision analytic ap-
proach. Radiology 215:445-451
References
1. Jemal A, Siegel R, Ward E et al (2009) Cancer statistics. CA
Cancer J Clin 59:225-249
2. Carter HB, Piantadosi S, Isaacs JT (1990) Clinical evidence
for and implications of the multistep development of prostate
cancer. J Urol 143:742-746
3. Parkin DM, Bray FI, Devesa SS (2001) Cancer burden in the
year 2000. The global picture. Eur J Cancer 37 Suppl 8:S4-S66
4. Konety BR, Bird VY, Deorah S, Dahmoush L (2005) Com-
parison of the incidence of latent prostate cancer detected at
autopsy before and after the prostate specific antigen era. J
Urol 174:1785-1788
5. Crawford ED (2003) Epidemiology of prostate cancer. Urology
62(6 Suppl 1):3-12
6. Stewart SL, King JB, Thompson TD et al (2004) Cancer mor-
tality surveillance – United States, 1990-2000. MMWR Sur-
veill Summ 53:1-108
7. Catalona WJ, Loeb S, Han M (2006) Viewpoint: expanding
prostate cancer screening. Ann Intern Med 144:441-443
8. Hoffman RM (2006) Viewpoint: limiting prostate cancer
screening. Ann Intern Med 144:438-440
9. Graif T, Loeb S, Roehl KA et al (2007) Under Diagnosis and
Over Diagnosis of Prostate Cancer. J Urol 178:88-92
10. Max W, Rice DP, Sung HY et al (2002) The economic burden
of prostate cancer, California, 1998. Cancer 94:2906-2913
11. Mangar SA, Huddart RA, Parker CC et al (2005) Technologi-
cal advances in radiotherapy for the treatment of localised
prostate cancer. European Journal of Cancer 41:908-921
12. Meraney AM, Haese A, Palisaar J et al (2005) Surgical man-
agement of prostate cancer: Advances based on a rational ap-
proach to the data. European Journal of Cancer 41:888-907
13. Bucci MK, Bevan A, Roach M, III (2005) Advances in radia-
tion therapy: Conventional to 3D, to IMRT, to 4D, and Beyond.
CA Cancer J Clin 55:117-134
138 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer
14. Hricak H, Wang L, Wei DC et al (2004) The role of preopera-
tive endorectal magnetic resonance imaging in the decision
regarding whether to preserve or resect neurovascular bundles
during radical retropubic prostatectomy. Cancer 100:2655-
2663
15. Cellini N, Morganti AG, Mattiucci GC et al (2002) Analysis of
intraprostatic failures in patients treated with hormonal thera-
py and radiotherapy: implications for conformal therapy plan-
ning. Int J Radiat Oncol Biol Phys 53:595-599
16. Roach M, III (2004) Reducing the toxicity associated with the
use of radiotherapy in men with localized prostate cancer. Urol
Clin North Am 31:353-366
17. Baxter NN, Tepper JE, Durham SB et al (2005) Increased risk
of rectal cancer after prostate radiation: A population-based
study. Gastroenterology 128:819-824
18. Steyn JH, Smith FW (1982) Nuclear magnetic resonance
imaging of the prostate. Br J Urol 54:726-728
19. Hricak H, Williams RD, Spring DB et al (1983) Anatomy and
pathology of the male pelvis by magnetic resonance imaging.
AJR Am J Roentgenol 141:1101-1110
20. Villeirs GM, Oosterlinck W, Vanherreweghe E, De Meerleer
GO (2008) A qualitative approach to combined magnetic res-
onance imaging and spectroscopy in the diagnosis of prostate
cancer. Eur J Radiol Dec 10
21. Tanimoto A, Nakashima J, Kohno H et al (2007) Prostate can-
cer screening: the clinical value of diffusion-weighted imaging
and dynamic MR imaging in combination with T2-weighted
imaging. J Magn Reson Imaging 25:146-152
22. Scheenen TW, Klomp DW, Roll SA et al (2004) Fast acquisi-
tion-weighted three-dimensional proton MR spectroscopic
imaging of the human prostate. Magn Reson Med 52:80-88
23. Chen ME, Johnston DA, Tang K et al (2000) Detailed mapping
of prostate carcinoma foci: biopsy strategy implications. Can-
cer 89:1800-1809
24. McNeal JE, Redwine EA, Freiha FS, Stamey TA (1988) Zon-
al distribution of prostatic adenocarcinoma. Correlation with
histologic pattern and direction of spread. Am J Surg Pathol
12:897-906
25. Miller GJ, Cygan JM (1994) Morphology of prostate cancer:
the effects of multifocality on histological grade, tumor vol-
ume and capsule penetration. J Urol 152:1709-1713
26. Brossner C, Winterholer A, Roehlich M et al (2003) Distribu-
tion of prostate carcinoma foci within the peripheral zone:
analysis of 8,062 prostate biopsy cores. World J Urol 21:
163-166
27. Horninger W, Reissigl A, Rogatsch H et al (2000) Prostate
cancer screening in the Tyrol, Austria: experience and results.
Eur J Cancer 36:1322-1335
28. Ohori M, Kattan M, Scardino PT, Wheeler TM (2004) Radical
prostatectomy for carcinoma of the prostate. Mod Pathol
17:349-359
29. Tempany CM, Zhou X, Zerhouni EA et al (1994) Staging of
prostate cancer: results of Radiology Diagnostic Oncology
Group project comparison of three MR imaging techniques.
Radiology 192:47-54
30. Cruz M, Tsuda K, Narumi Y et al (2002) Characterization of
low-intensity lesions in the peripheral zone of prostate on pre-
biopsy endorectal coil MR imaging. Eur Radiol 12:357-365
31. Claus FG, Hricak H, Hattery RR (2004) Pretreatment evalua-
tion of prostate cancer: role of MR imaging and 1H MR spec-
troscopy. Radiographics 24:S167-S180
32. Wang L, Mazaheri Y, Zhang J et al (2008) Assessment of bio-
logic aggressiveness of prostate cancer: correlation of MR sig-
nal intensity with Gleason grade after radical prostatectomy.
Radiology 246:168-176
33. Akin O, Sala E, Moskowitz CS et al (2006) Transition zone
prostate cancers: features, detection, localization, and staging
at endorectal MR imaging. Radiology 239:784-792
34. Mullerad M, Hricak H, Kuroiwa K et al (2005) Comparison
of endorectal magnetic resonance imaging, guided prostate
biopsy and digital rectal examination in the preoperative anatom-
ical localization of prostate cancer. J Urol 174:2158-2163
35. Beyersdorff D, Taupitz M, Winkelmann B et al (2002) Patients
with a history of elevated prostate-specific antigen levels and
negative transrectal US-guided quadrant or sextant biopsy re-
sults: value of MR imaging. Radiology 224:701-706
36. Tamada T, Sone T, Jo Y et al (2008) Prostate cancer: relation-
ships between postbiopsy hemorrhage and tumor detectability
at MR diagnosis. Radiology 248:531-539
37. Sato C, Naganawa S, Nakamura T et al (2005) Differentiation
of noncancerous tissue and cancer lesions by apparent diffu-
sion coefficient values in transition and peripheral zones of the
prostate. J Magn Reson Imaging 21:258-262
38. Zelhof B, Pickles M, Liney G et al (2009) Correlation of dif-
fusion-weighted magnetic resonance data with cellularity in
prostate cancer. BJU Int 103:883-888
39. Mazaheri Y, Shukla-Dave A, Hricak H et al (2008) Prostate
cancer: identification with combined diffusion-weighted MR
imaging and 3D 1H MR spectroscopic imaging – correlation
with pathologic findings. Radiology 246:480-488
40. Haider MA, Van Der Kwast TH, Tanguay J et al (2007) Com-
bined T2-weighted and diffusion-weighted MRI for localiza-
tion of prostate cancer. AJR Am J Roentgenol 189:323-328
41. Yoshimitsu K, Kiyoshima K, Irie H et al (2008) Usefulness of
apparent diffusion coefficient map in diagnosing prostate car-
cinoma: correlation with stepwise histopathology. J Magn Re-
son Imaging 27:132-139
42. Langer DL, Van Der Kwast TH, Evans AJ et al (2009) Prostate
cancer detection with multi-parametric MRI: logistic regres-
sion analysis of quantitative T2, diffusion-weighted imaging,
and dynamic contrast-enhanced MRI. J Magn Reson Imaging
30:327-334
43. Gibbs P, Pickles MD, Turnbull LW (2006) Diffusion imaging
of the prostate at 3.0 tesla. Invest Radiol 41:185-188
44. Pickles MD, Gibbs P, Sreenivas M, Turnbull LW (2006) Diffu-
sion-weighted imaging of normal and malignant prostate tis-
sue at 3.0T. J Magn Reson Imaging 23:130-134
45. Miao H, Fukatsu H, Ishigaki T (2007) Prostate cancer detec-
tion with 3-T MRI: comparison of diffusion-weighted and T2-
weighted imaging. Eur J Radiol 61:297-302
46. Barentsz JO, Engelbrecht M, Jager GJ et al (1999) Fast dy-
namic gadolinium-enhanced MR imaging of urinary bladder
and prostate cancer. J Magn Reson Imaging 10:295-304
47. Padhani AR, Gapinski CJ, Macvicar DA et al (2000) Dynam-
ic contrast enhanced MRI of prostate cancer: correlation with
morphology and tumour stage, histological grade and PSA.
Clin Radiol 55:99-109
48. Engelbrecht MR, Huisman HJ, Laheij RJ et al (2003) Dis-
crimination of prostate cancer from normal peripheral zone
and central gland tissue by using dynamic contrast-enhanced
MR imaging. Radiology 229:248-254
49. Van Dorsten FA, Van Der Graaf M, Engelbrecht MR et al
(2004) Combined quantitative dynamic contrast-enhanced MR
imaging and (1)H MR spectroscopic imaging of human
prostate cancer. J Magn Reson Imaging 20:279-287
50. Runge VM (2000) Safety of approved MR contrast media for
intravenous injection. J Magn Reson Imaging 12:205-213
51. Lin SP, Brown JJ (2007) MR contrast agents: physical and
pharmacologic basics. J Magn Reson Imaging 25:884-899
52. Kurhanewicz J, Vigneron DB, Hricak H et al (1996) Three-di-
mensional H-1 MR spectroscopic imaging of the in situ human
prostate with high (0.24-0.7-cm3) spatial resolution. Radiolo-
gy 198:795-805
53. Coakley FV, Qayyum A, Kurhanewicz J (2003) Magnetic res-
onance imaging and spectroscopic imaging of prostate cancer.
J Urol 170:S69-S75
54. Zakian KL, Sircar K, Hricak H et al (2005) Correlation of
proton MR spectroscopic imaging with gleason score based on
step-section pathologic analysis after radical prostatectomy.
Radiology 234:804-814
Magnetic Resonance Imaging of Prostate Cancer
55. Scheenen TW, Klomp DW, Roll SA et al (2004) Fast acquisi-
tion-weighted three-dimensional proton MR spectroscopic
imaging of the human prostate. Magn Reson Med 52:80-88
56. Scheidler J, Hricak H, Vigneron DB et al (1999) Prostate cancer:
localization with three-dimensional proton MR spectroscopic
imaging – clinicopathologic study. Radiology 213:473-480
57. Wefer AE, Hricak H, Vigneron DB et al (2000) Sextant local-
ization of prostate cancer: comparison of sextant biopsy,
magnetic resonance imaging and magnetic resonance spectro-
scopic imaging with step section histology. J Urol 164:
400-404
58. Fütterer JJ, Heijmink SWTPJ, Scheenen TWJ et al (2006)
Prostate cancer localization with dynamic contrast-enhanced
MR imaging and proton MR spectroscopic imaging. Radiology
241:449-458
59. Shukla-Dave A, Hricak H, Kattan MW et al (2007) The utili-
ty of magnetic resonance imaging and spectroscopy for pre-
dicting insignificant prostate cancer: an initial analysis. BJU
Int 99:786-793
60. Weinreb JC, Blume JD, Coakley FV et al (2009) Prostate can-
cer: sextant localization at MR imaging and MR spectroscop-
ic imaging before prostatectomy – results of ACRIN prospec-
tive multi-institutional clinicopathologic study. Radiology
251:122-133
61. Noguchi M, Stamey TA, McNeal JE, Yemoto CM (2001) Re-
lationship between systematic biopsies and histological fea-
tures of 222 radical prostatectomy specimens: lack of predic-
tion of tumor significance for men with nonpalpable prostate
cancer. J Urol 166:104-109
62. Kumar V, Jagannathan NR, Kumar R et al (2007) Transrectal
ultrasound-guided biopsy of prostate voxels identified as sus-
picious of malignancy on three-dimensional (1)H MR spectro-
scopic imaging in patients with abnormal digital rectal exam-
ination or raised prostate specific antigen level of 4-10 ng/ml.
NMR Biomed 20:11-20
63. Prando A, Kurhanewicz J, Borges AP et al (2005) Prostatic
biopsy directed with endorectal MR spectroscopic imaging
findings in patients with elevated prostate specific antigen lev-
els and prior negative biopsy findings: early experience. Radi-
ology 236:903-910
64. Anastasiadis AG, Lichy MP, Nagele U et al (2006) MRI-guid-
ed biopsy of the prostate increases diagnostic performance in
men with elevated or increasing PSA levels after previous neg-
ative TRUS biopsies. Eur Urol Oct 50:738-748
65. Beyersdorff D, Winkel A, Hamm B et al (2005) MR imaging-
guided prostate biopsy with a closed MR unit at 1.5 T: initial
results. Radiology 234:576-581
66. Engelhard K, Hollenbach HP, Kiefer B et al (2006) Prostate
biopsy in the supine position in a standard 1.5-T scanner un-
der real time MR-imaging control using a MR-compatible en-
dorectal biopsy device. Eur Radiol 16:1237-1243
67. Hambrock T, Futterer JJ, Huisman HJ et al (2008) Thirty-two-
channel coil 3T magnetic resonance-guided biopsies of
prostate tumor suspicious regions identified on multimodality
3T magnetic resonance imaging: technique and feasibility. In-
vest Radiol;43:686-694
68. Hambrock T (2009) The Value of 3 Tesla Magnetic Resonance
Imaging Guided Prostate Biopsies in Men with repeptitive
Negative Biopsies and Elevated PSA
69. Pondman KM, Futterer JJ, ten HB et al (2008) MR-guided
biopsy of the prostate: an overview of techniques and a sys-
tematic review. Eur Urol 54:517-527
70. Gleason DF, Mellinger GT (1974) Prediction of prognosis for
prostatic adenocarcinoma by combined histological grading
and clinical staging. J Urol 111(1):58-64
71. Noguchi M, Stamey TA, McNeal JE, Yemoto CM (2001) Re-
lationship between systematic biopsies and histological fea-
tures of 222 radical prostatectomy specimens: lack of predic-
tion of tumor significance for men with nonpalpable prostate
cancer. J Urol 166:104-109
139
72. Ruijter E, van LG, Miller G et al (2000) Errors in histological
grading by prostatic needle biopsy specimens: frequency and
predisposing factors. J Pathol 192:229-233
73. Wang L, Mazaheri Y, Zhang J et al (2008) Assessment of bio-
logic aggressiveness of prostate cancer: correlation of MR sig-
nal intensity with Gleason grade after radical prostatectomy.
Radiology 246:168-176
74. Ikonen S, Karkkainen P, Kivisaari L (2000) Magnetic reso-
nance imaging of prostatic cancer: does detection vary be-
tween high and low gleason score tumors? Prostate 43:43-48
75. Zakian KL, Sircar K, Hricak H et al (2005) Correlation of pro-
ton MR spectroscopic imaging with gleason score based on
step-section pathologic analysis after radical prostatectomy.
Radiology 234:804-814
76. deSouza NM, Riches SF, Vanas NJ et al (2008) Diffusion-
weighted magnetic resonance imaging: a potential non-inva-
sive marker of tumour aggressiveness in localized prostate
cancer. Clin Radiol 63:774-782
77. Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB (2002)
Combined magnetic resonance imaging and spectroscopic
imaging approach to molecular imaging of prostate cancer. J
Magn Reson Imaging 16:451-463
78. Tamada T, Sone T, Jo Y (2008) Apparent diffusion coefficient
values in peripheral and transition zones of the prostate: com-
parison between normal and malignant prostatic tissues and
correlation with histologic grade. J Magn Reson Imaging 28:
720-726
79. Hosseinzadeh K, Schwarz SD (2004) Endorectal diffusion-
weighted imaging in prostate cancer to differentiate malignant
and benign peripheral zone tissue. J Magn Reson Imaging
20:654-661
80. Wang L, Mullerad M, Chen HN et al (2004) Prostate cancer:
incremental value of endorectal MR imaging findings for pre-
diction of extracapsular extension. Radiology 232:133-139
81. Sonnad SS, Langlotz CP, Schwartz JS (2001) Accuracy of
MR imaging for staging prostate cancer: a meta-analysis to
examine the effect of technologic change. Acad Radiol
8:149-157
82. Engelbrecht MR, Jager GJ, Laheij RJ et al (2002) Local stag-
ing of prostate cancer using magnetic resonance imaging: a
meta-analysis. Eur Radiol 12:2294-2302
83. Cornud F, Flam T, Chauveinc L et al (2002) Extraprostatic
spread of clinically localized prostate cancer: factors predic-
tive of pT3 tumor and of positive endorectal MR imaging ex-
amination results. Radiology 224:203-210
84. Langlotz CP, Schnall MD, Malkowicz SB, Schwartz JS (1996)
Cost-effectiveness of endorectal magnetic resonance imaging
for the staging of prostate cancer. Acad Radiol 1:S24-S27
85. Jager GJ, Severens JL, Thornbury JR et al (2000) Prostate can-
cer staging: should MR imaging be used? – A decision analyt-
ic approach. Radiology 215:445-451
86. Wang L, Hricak H, Kattan MW et al (2006) Prediction of or-
gan-confined prostate cancer: incremental value of MR Imag-
ing and MR spectroscopic imaging to staging nomograms. Ra-
diology 238:597-603
87. Wang L, Hricak H, Kattan MW et al (2007) Prediction of sem-
inal vesicle invasion in prostate cancer: incremental value of
adding endorectal MR imaging to the Kattan nomogram. Ra-
diology 242:182-188
88. McKenna DA, Coakley FV, Westphalen AC et al (2008)
Prostate Cancer: Role of Pretreatment MR in Predicting Out-
come after External-Beam Radiation Therapy – Initial Experi-
ence. Radiology 247:141-146
89. Ren J, Huan Y, Li F et al (2009) Combined T2-weighted and
diffusion-weighted MRI for diagnosis of urinary bladder inva-
sion in patients with prostate carcinoma. J Magn Reson Imag-
ing 30:351-356
90. Ren J, Huan Y, Wang H et al (2009) Seminal vesicle invasion
in prostate cancer: prediction with combined T2-weighted and
diffusion-weighted MR imaging. Eur Radiol 19:2481-2486
140 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer
91. Yu KK, Scheidler J, Hricak H et al (1999) Prostate cancer:
prediction of extracapsular extension with endorectal MR
imaging and three-dimensional proton MR spectroscopic
imaging. Radiology 213:481-488
92. Mullerad M, Hricak H, Wang L et al (2004) Prostate cancer:
detection of extracapsular extension by genitourinary and
general body radiologists at MR imaging. Radiology
232:140-146
93. Fütterer JJ, Engelbrecht MR, Huisman HJ et al (2005) Stag-
ing Prostate Cancer with Dynamic Contrast-enhanced En-
dorectal MR Imaging prior to Radical Prostatectomy: Expe-
rienced versus Less Experienced Readers. Radiology
237:541-549
94. Wang L, Zhang J, Schwartz LH et al (2007) Incremental val-
ue of multiplanar cross-referencing for prostate cancer stag-
ing with endorectal MRI. AJR Am J Roentgenol 188:99-104
95. Akin O, Riedl CC, Ishill NM (2009 Interactive dedicated
training curriculum improves accuracy in the interpretation of
MR imaging of prostate cancer. Eur Radiol Nov 17
96. Fütterer JJ, Heijmink SW, Scheenen TW et al (2006) Prostate
cancer: local staging at 3-T endorectal MR imaging – early
experience. Radiology 238:184-191
97. Heijmink SWTPJ, Futterer JJ, Hambrock T et al (2007)
Prostate Cancer: Body-Array versus Endorectal Coil MR
Imaging at 3 T-Comparison of Image Quality, Localization,
and Staging Performance. Radiology 244:184-195
98. Partin AW, Mangold LA, Lamm DM et al (2001) Contempo-
rary update of prostate cancer staging nomograms (Partin Ta-
bles) for the new millennium. Urology 2001 58:843-848
99. Narayan P, Gajendran V, Taylor SP et al (1995) The role of
transrectal ultrasound-guided biopsy-based staging, preoper-
ative serum prostate-specific antigen, and biopsy Gleason
score in prediction of final pathologic diagnosis in prostate
cancer. Urology 46:205-212
100. Salo JO, Kivisaari L, Rannikko S, Lehtonen T (1986) The
value of CT in detecting pelvic lymph node metastases in cas-
es of bladder and prostate carcinoma. Scand J Urol Nephrol
20:261-265
101. Jager GJ, Barentsz JO, Oosterhof GO et al (1996) Pelvic
adenopathy in prostatic and urinary bladder carcinoma: MR
imaging with a three-dimensional TI-weighted magnetiza-
tion-prepared-rapid gradient-echo sequence. AJR Am J
Roentgenol 167:1503-1507
102. Heesakkers RA, Hovels AM, Jager GJ et al (2008) MRI with
a lymph-node-specific contrast agent as an alternative to CT
scan and lymph-node dissection in patients with prostate can-
cer: a prospective multicohort study. Lancet Oncol 9:850-856
103. Wolf JS, Jr., Cher M, Dall’era M et al (1995) The use and ac-
curacy of cross-sectional imaging and fine needle aspiration
cytology for detection of pelvic lymph node metastases be-
fore radical prostatectomy. J Urol Mar 153:993-999
104. Hovels AM, Heesakkers RA, Adang EM et al (2008)The di-
agnostic accuracy of CT and MRI in the staging of pelvic
lymph nodes in patients with prostate cancer: a meta-analy-
sis. Clin Radiol 63:387-395
105. Oyen RH, Van Poppel HP, Ameye FE et al (1994) Lymph
node staging of localized prostatic carcinoma with CT and
CT-guided fine-needle aspiration biopsy: prospective study of
285 patients. Radiology 190:315-322
106. Tiguert R, Gheiler EL, Tefilli MV et al (1999) Lymph node
size does not correlate with the presence of prostate cancer
metastasis. Urology 53:367-371
107. Levran Z, Gonzalez JA, Diokno AC et al (1995) Are pelvic
computed tomography, bone scan and pelvic lymphadenecto-
my necessary in the staging of prostatic cancer? Br J Urol
75:778-781
108. Wolf JS Jr, Cher M, Dall’era M et al (1995) The use and ac-
curacy of cross-sectional imaging and fine needle aspiration
cytology for detection of pelvic lymph node metastases be-
fore radical prostatectomy. J Urol 153:993-999
109. Harisinghani MG, Barentsz J, Hahn PF et al (2003) Nonin-
vasive detection of clinically occult lymph-node metastases in
prostate cancer. N Engl J Med 2003 348:2491-2499
110. Heesakkers RA, Jager GJ, Hövels AM et al (2009) Prostate
cancer: detection of lymph node metastases outside the rou-
tine surgical area with ferumoxtran-10-enhanced MR imag-
ing. Radiology 251:408-414
111. Will O, Purkayastha S, Chan C et al (2005) Diagnostic preci-
sion of nanoparticle-enhanced MRI for lymph-node metas-
tases: a meta-analysis. Lancet Oncol 7:52-60
112. O’Donoghue EP, P Constable AR, Sherwood T et al (1978)
Bone scanning and plasma phosphatases in carcinoma of the
prostate. Br J Urol 50:172-177
113. Rybak LD, Rosenthal DI (2001) Radiological imaging for the
diagnosis of bone metastases. Quarterly Journal of Nuclear
Medicine 45:53-64
114. Lecouvet FE, Geukens D, Stainier A et al (2007) Magnetic
resonance imaging of the axial skeleton for detecting bone
metastases in patients with high-risk prostate cancer: diag-
nostic and cost-effectiveness and comparison with current de-
tection strategies. J Clin Oncol 25:3281-3287
115. Tombal B, Rezazadeh A, Therasse P et al (2005) Magnetic
resonance imaging of the axial skeleton enables objective
measurement of tumor response on prostate cancer bone
metastases. Prostate 65:178-187
116. Cumming J, Hacking N, Fairhurst J et al (1990) Distribution
of bony metastases in prostatic carcinoma. Br J Urol 66:411-
414
117. Pound CR, Brawer MK, Partin AW (2001) Evaluation and
treatment of men with biochemical prostate-specific antigen
recurrence following definitive therapy for clinically local-
ized prostate cancer. Rev Urol 3:72-84
118. Horwitz EM, Vicini FA, Ziaja EL et al (1998) The correlation
between the ASTRO Consensus Panel definition of biochem-
ical failure and clinical outcome for patients with prostate
cancer treated with external beam irradiation. American So-
ciety of Therapeutic Radiology and Oncology. Int J Radiat
Oncol Biol Phys 41:267-272
119. Heidenreich A, Aus G, Bolla M et al (2008) EAU guidelines
on prostate cancer. Eur Urol 53:68-80
120. Roach M, III, Hanks G, Thames H et al (2006) Defining bio-
chemical failure following radiotherapy with or without hor-
monal therapy in men with clinically localized prostate can-
cer: recommendations of the RTOG-ASTRO Phoenix Con-
sensus Conference. Int J Radiat Oncol Biol Phys 65:965-974
121. Crook J, Robertson S, Collin G et al (1993) Clinical relevance
of trans-rectal ultrasound, biopsy, and serum prostate-specif-
ic antigen following external beam radiotherapy for carcino-
ma of the prostate. Int J Radiat Oncol Biol Phys 27:31-37
122. Leventis AK, Shariat SF, Slawin KM (2001) Local recurrence
after radical prostatectomy: correlation of US features with
prostatic fossa biopsy findings. Radiology 219:432-439
123. Chan TW, Kressel HY (1991) Prostate and seminal vesicles
after irradiation: MR appearance. J Magn Reson Imaging
1:503-511
124. Sala E, Eberhardt SC, Akin O et al (2006) Endorectal MR
imaging before salvage prostatectomy: tumor localization and
staging. Radiology 238:176-183
125. Coakley FV, Teh HS, Qayyum A et al (2004) Endorectal MR
imaging and MR spectroscopic imaging for locally recurrent
prostate cancer after external beam radiation therapy: prelim-
inary experience. Radiology 233:441-448
126. Haider MA, Chung P, Sweet J et al (2008) Dynamic contrast-
enhanced magnetic resonance imaging for localization of re-
current prostate cancer after external beam radiotherapy. Int
J Radiat Oncol Biol Phys 70:425-430
127. Rouviere O, Valette O, Grivolat S et al (2004) Recurrent
prostate cancer after external beam radiotherapy: value of con-
trast-enhanced dynamic MRI in localizing intraprostatic tu-
mor – correlation with biopsy findings. Urology 63:922-927
Magnetic Resonance Imaging of Prostate Cancer
128. Kim CK, Park BK, Lee HM (2009) Prediction of locally re-
current prostate cancer after radiation therapy: Incremental
value of 3T diffusion-weighted MRI. J Magn Reson Imaging
29:391-397
129. Sella T, Schwartz LH, Swindle PW et al (2004) Suspected lo-
cal recurrence after radical prostatectomy: endorectal coil
MR imaging. Radiology 231:379-385
130. Casciani E, Polettini E, Carmenini E et al (2008) Endorectal
and dynamic contrast-enhanced MRI for detection of local re-
currence after radical prostatectomy. AJR Am J Roentgenol
190:1187-1192
141
131. Cirillo S, Petracchini M, Scotti L et al (2009) Endorectal
magnetic resonance imaging at 1.5 Tesla to assess local re-
currence following radical prostatectomy using T2-weighted
and contrast-enhanced imaging. Eur Radiol 19:761-769
132. Sciarra A, Panebianco V, Salciccia S et al (2008) Role of dy-
namic contrast-enhanced magnetic resonance (MR) imaging
and proton MR spectroscopic imaging in the detection of lo-
cal recurrence after radical prostatectomy for prostate cancer.
Eur Urol 54:589-600
 IDKD 2010-2013

Imaging of the Male Pelvis: The Scrotum

Brent J. Wagner
Department of Radiology, The Reading Hospital and Medical Center, West Reading, PA, USA

Introduction
Although there are many potential ways one might com-
partmentalize the imaging of scrotal disease (including
pattern recognition, etiologies, and clinical presentation),
this chapter focuses on the requisite entities that, based
on their imaging and clinical features, constitute the core
knowledge needed for the radiologist when faced with an
abnormal scrotal sonogram. Hence, while not compre-
hensive, this review addresses diseases that are of inter-
est because of a distinctive combination of clinical rele-
vance and characteristic imaging features.
With the continued evolution of high-resolution linear
transducers over the past 15 years, clinicians, patients,
and imaging technicians are able to rely on ultrasound
as a tool with near 100% sensitivity for significant
intrascrotal pathology. Not only is the technique non-
invasive and relatively inexpensive, it is also highly
Fig. 1. Seminoma. A palpable nodule in a 33-year-old male. Sono-
effective in both the detection and the characterization of gram demonstrates a small, homogeneous, well-marginated intra-
a wide variety of disorders involving the scrotum. testicular mass

Intratesticular Disorders
Neoplastic
Most neoplasms of the testis are germ cell tumors (GCTs),
and the vast majority of these are malignant. Among the
various histological types, seminoma is the most common
to occur as a pure tumor, accounting for approximately
40% of all testicular neoplasms. Most of the remainder are
mixed tumors, containing two or more histological types.
Lesions typically present as a palpable mass, although
some aggressive tumors may present with metastatic foci
to lung or bone or as nodal masses.
The majority of GCTs are hypoechoic relative to the
homogeneous medium-high echogenicity background of
the testis (Fig. 1). Calcifications are seen in at least one-
third of cases, especially in non-seminomatous tumors,
which also tend to exhibit more heterogeneity.
The management of GCTs nearly always involves or-
chiectomy for the definitive diagnosis and treatment of
the local disease. The well-established radiosensitivity of
seminomas is the most important distinguishing clinical
feature of the two groups of tumors (seminomatous vs.
non-seminomatous). This feature accounts for the previ-
ous, widespread use of prophylactic retroperitoneal ra-
diotherapy for seminomas, even when imaging studies
suggested that the lesion was confined to the testis. Over
the last decade, however, there has been a trend toward
surveillance (using imaging and serum tumor markers) in
the 75% of patients with seminomas who are diagnosed
as having stage I disease [1, 2].
For non-seminomatous tumors, computed tomography
(CT) is often used to determine whether patients require
retroperitoneal lymph node dissection. Patients with
(1) nodal involvement (either enlarged on CT, or confirmed
on pathological assessment) and/or (2) hematogenous
metastasis are generally treated with chemotherapy. The
prognosis for patients with non-seminomatous tumors is
more guarded than that for patients with seminoma, al-
though patients with stage I disease (limited to the scro-
tum) have 5-year survival rates approaching 90%. While
Imaging of the Male Pelvis: The Scrotum
attempts have been made to use imaging to differentiate
GCTs [3], the initial management of both types is surgi-
cal (orchiectomy) and the exercise of pre-operatively dif-
ferentiating seminoma vs. non-seminoma is typically not
relevant [1].
Gonadal stromal tumors are less common (<10% of
lesions) and are often incidental findings. Both Leydig
cell and Sertoli tumors are occasionally the cause of gy-
necomastia. The vast majority of gonadal stromal tumors
are benign but they have gray-scale sonographic features
similar to the more sinister group of GCTs.
Patients with lymphoma of the testis are typically old-
er than those with GCTs. Lymphomas are rarely demon-
strated in patients without a known diagnosis of systemic
lymphoma. Hypoechoic, multifocal (or geographic), bi-
lateral lesions, which often have increased vascularity
compared with GCT’s, are characteristically seen [4].
While not strictly neoplastic, five entities deserve
brief discussion in the context of neoplastic intratesticu-
lar tumors:
1. Simple cysts are typically peripheral, multiple, and con-
tiguous with the tunica albuginea (more central cysts
are usually a manifestation of tubular ectasia). These
have no malignant potential and are rarely of clinical
significance, although they may be palpable and clini-
cally mimic a GCT. Rarely, they may be complicated by
hemorrhage; however, most cases are easily diagnosed
as simple cysts. Features that raise concern for malig-
nancy (and indicate the need for prompt surgical re-
moval) include solitary lesions complicated by internal
soft tissue, calcification, or a thick irregular wall.
2. Tubular ectasia represents a dilatation of the rete testis
as it converges along the mediastinum testis. Patients
may have a history of prior epididymitis or vasectomy,
but most have no specific symptoms. An ovoid or linear
area of decreased echogenicity, contiguous with the epi-
didymis, can usually be differentiated from a neoplasm
based on the pattern of branching, anechoic channels
(Fig. 2). In occasional cases that may mimic neoplasm,
T2-weighted magnetic resonance imaging will show a
hyperintense focus (in contrast to the hypointensity that
is characteristic of most testicular neoplasms) [5].
3. Epidermoid cyst is not considered to be a neoplasm by
most authorities. Instead, it is an inclusion cyst, lined by
squamous epithelium and filled with keratinized debris.
Although it accounts for <10% of intratesticular mass-
es, the differentiation of epidermoid cyst from germ cell
neoplasm is important in order to avoid radical surgery
for this lesion, which can, instead, be effectively diag-
nosed and treated with a more limited procedure (enu-
cleation, sparing the testis) [5, 6]. Epidermoid cyst can
be recognized in many instances by its “onion-skin”
appearance, i.e., alternating concentric rings of hypo-
and hyperechogenicity [5, 7]. A minority of cysts will
show wall calcification. The absence of Doppler flow
may be of some help (the presence of central flow
excludes the diagnosis), but one should remember that
some malignant neoplasms, especially when small, will
143
Fig. 2. Tubular ectasia. The somewhat focal decreased echogenicity
resembles a mass but the non-spherical shape and the elongate
anechoic channels are characteristic and do not suggest the need
for tissue diagnosis
not have flow that can be detected sonographically.
Additionally, the onion-skin appearance is neither
sensitive (it is seen in approximately half the cases) nor
entirely specific (some neoplasms, including teratoma,
may rarely have the same feature) [7].
4. Congenital adrenal rest tumors are seen in patients
with poorly controlled congenital adrenal hyperplasia
(CAH). Aberrant adrenal cortical cells migrate with
gonadal tissue during fetal development and may
hypertrophy at some point during the disease. The
sonographic appearance is extremely variable but is
typically distinguished from that of malignant GCT by
the multiplicity of the lesions, their eccentricity, and
their contiguity with the mediastinum testis. Treatment
is glucocorticoid replacement or, rarely, surgery
(partial orchiectomy) [8].
5. Testicular microlithiasis (TML) describes the presence
of numerous small (1-2 mm) calcifications scattered
throughout the testicular parenchyma. This idiopathic
condition is associated with oligospermia in a minori-
ty of patients but in most it is merely an incidental
finding [5]. Several published reports concluded that
there is a small but real increased risk of testicular
GCT in patients with TML (Fig. 3) [9]. This prompted
multiple authors to suggest that patients should be
carefully screened for coexistent tumors and followed
sonographically at annual intervals for the develop-
ment of malignancy. Recently, there has been a shift
away from this recommendation, based on the high
cost of serial sonographic exams in a large number of
men who have an increased but still very low risk of
developing a germ cell tumor [9, 10]. Nonetheless, the
need for sonographic follow-up remains contentious
within the field.
144
Fig. 3. Mixed germ cell tumor arising in a background of testicular
microlithiasis
Non-neoplastic
Infection typically begins in the epididymis. Epididymo-
orchitis is usually a clinical diagnosis, with sonography
sometimes used to rule out torsion or abscess. Sono-
graphic findings are often absent, although some patients
will show a hydrocele (with or without complicating ele-
ments indicating pyocele). There may also be increased
Doppler flow to the epididymis. Progression to involve-
ment of the testis; which occurs in a minority of cases,
may result in abscess formation or infarction.
Ischemia/infarction may result from torsion or, less com-
monly, from a variety of other causes, including vasculitis,
diabetes, or orchitis, which may be segmental [11]. Patients
with torsion typically have an acute clinical presentation
that includes severe unilateral scrotal pain, often follow-
ing minor trauma or physical exertion. The typical find-
ing of ischemia/infarction is an asymmetrical decrease in
the color or amplitude (power) of the Doppler signal on
the symptomatic side. However, subtle variations of arte-
rial spectral Doppler waveforms may be seen early, in-
cluding the absence of the dicrotic notch and/or increased
resistance (decreased or absent diastolic flow) [12]. The
latter finding may also be seen in the early stages of tor-
sion, when venous flow is altered but arterial flow is still
observed using color Doppler. Emergency surgery is
indicated to detorse and save the testis; however, when
gray-scale findings are present, including heterogeneity
and decreased echogenicity, the ischemia has almost
always progressed to infarction. At this point, testicular
salvage is not possible [13].
Important pitfalls in Doppler evaluation must be rec-
ognized in order to avoid misdiagnosis. The examiner
must be careful not to alter the Doppler settings (gain,
scale, etc.) when comparing normal (asymptomatic)
testis to the painful side. One must also remember that
Brent J. Wagner
paratesticular tissues may show reactive hyperemia, de-
spite the absence (or significant decrease) in Doppler
flow to the testis itself.
Traumatic injury to the testis is frequently an indica-
tion for sonography. The differentiation of testicular rup-
ture from a less serious injury (for example, merely a
hematoma with an intact tunica albuginea) generally rep-
resents the most important role of the imager. Findings of
a heterogeneous echo-texture within the testis, testicular
contour abnormality, and disruption of the tunica albug-
inea are considered very sensitive and specific for the di-
agnosis of testicular rupture [14].
Sarcoidosis is an unusual cause of a scrotal mass and
the scrotum may rarely be the presenting site of this dis-
ease in some patients. The pattern of multiple, hypo-
echoic, intratesticular masses associated with epididymal
enlargement and heterogeneity is characteristic, although
it may also be seen in lymphoma. A useful differentiating
feature is the relative prominence of testicular disease in
lymphoma (sarcoidosis tends to affect the epididymis
more than the testis, as discussed below) [5].
Extratesticular Disorders
Neoplastic
In adults, malignant extratesticular neoplasms are rare and
have a non-specific appearance [15]. Mesothelioma is an
uncommon neoplasm that usually presents as a hydrocele,
with soft-tissue nodules of the tunica vaginalis. Alterna-
tively, it may present as a large heterogeneous mass that
may be difficult to separate from the testis. Mesothe-
liomas tend to occur in individuals who are decades older
than those typically diagnosed with testicular GCTs.
Lymphoma may occasionally involve the epididymis,
although in the majority of patients this does not lead to
a diagnostic dilemma as: (1) the patient will be known to
have lymphoma and (2) there will be coexistent involve-
ment of the testis itself. Very rarely, solid tumors may
metastasize to the epididymis. Affected patients almost
always have advanced metastatic disease elsewhere
throughout the body.
The most common extratesticular intrascrotal neo-
plasm is lipoma, which arises from the spermatic cord
and can often be diagnosed clinically based on palpation.
Adenomatoid tumors are nearly as common as lipomas,
accounting for about one-third of extratesticular masses.
These are benign, but may be surgically removed either
to establish the diagnosis or because of local pain or ten-
derness. They are solid, well-marginated lesions that are
typically <20 mm in size. They most frequently arise
from the epididymis.
Papillary cystadenomas of the epididymis are seen in
about one-quarter of patients with von Hippel-Lindau
disease (the lesions are otherwise extremely rare). They
are typically solid, measure between 1 and 5 cm, and may
be indistinguishable from adenomatoid tumors [15].
Imaging of the Male Pelvis: The Scrotum
Sarcoidosis is more likely to affect the epididymis
than the testis. More than one-third of patients will
have bilateral disease. Although discrete nodules are
occasionally seen, the appearance is more commonly
one of heterogeneous enlargement. A diagnostic pattern
that may be of use in a previously undiagnosed patient
with hilar adenopathy – which could be either lym-
phoma or sarcoid – is to compare the testicular and the
epididymal involvement: in sarcoidosis, the degree of
epididymal disease typically exceeds that of testis
involvement, whereas in lymphoma the converse is
expected.
References
1. Sohaib SA, Koh DM, Husband JE (2008) The role of imaging
in the diagnosis, staging, and management of testicular cancer.
AJR Am J Roentgen 191:387-395
2. Krohmer SJ, McNulty NJ, Schned AR (2009) Best cases from
the AFIP: testicular seminoma with lymph node metastases.
Radiographics 29:2177-2183
3. Tsili AC, Tsampoulas C, Giannakopoulos X et al (2007) MRI
in the histologic characterization of testicular neoplasms. AJR
Am J Roentgen 189:W331-W337
4. Mazzu D, Jeffrey RB, Ralls PW (1995) Lymphoma and
leukemia involving the testicles: findings on gray-scale
and color Doppler sonography. AJR Am J Roentgen 164:
645-647
145
5. Woodward PJ, Sohaey R, O’Donoghue MJ, Green DE (2002)
Tumors and tumorlike lesions of the testis: radiologic-
pathologic correlation. Radiographics 22:189-216
6. Cho J-H, Chang J-C, Park B-H et al (2002) Sonographic and
MR imaging findings of testicular epidermoid cysts. AJR Am
J Roentgen 178:743-748
7. Maizlin ZV, Belenky A, Baniel J et al (2005) Epidermoid cyst
and teratoma of the testis: sonographic and histologic similar-
ities. J Ultrasound Med 24:1403-1409
8. Nagamine WH, Mehta SV, Vade A (2005) Testicular adrenal rest
tumors in a patient with congenital adrenal hyperplasia: sono-
graphic and magnetic resonance imaging findings. J Ultrasound
Med 24:1717-1720
9. Lam DL, Gerscovich EO, Kuo MC, McGahan JP (2007) Tes-
ticular microlithiasis: our experience of 10 years. J Ultrasound
Med 26:867-873
10. Costabile RA (2007) How worrisome is testicular microlithia-
sis? Curr Opin Urol 17:419-423
11. Fernández-Pérez GC, Tardáguila FM, Velasco M et al (2005)
Radiologic findings of segmental testicular infarction. AJR
Am J Roentgen 184:1587-1593
12. Dogra VS, Rubens DJ, Gottlieb RH, Bhatt S (2004) Torsion
and beyond: new twists in spectral Doppler evaluation of the
scrotum. J Ultrasound Med 23:1077-1085
13. Middleton WD, Middleton MA, Dierks M et al (1997) Sono-
graphic prediction of viability in testicular torsion: preliminary
observations. J Ultrasound Med 16:23-27
14. Bhatt S, Dogra VS (2008) Role of US in testicular and scrotal
trauma. Radiographics 28:1617-1629
15. Woodward PJ, Schwab CM, Sesterhenn IA (2003) Extratestic-
ular scrotal masses: radiologic-pathologic correlation. Radio-
graphics 23:215-240
 IDKD 2010-2013

Spread of Metastatic Disease in the Abdomen and Pelvis

James A. Brink1, Ali Shirkhoda2
1 Department of Diagnostic Radiology, Yale University School of Medicine, New Haven, CT, USA
2 Diagnostic Radiology, William Beaumont Hospital, Royal Oak, MI, USA

Introduction
Basic knowledge of the normal intra-abdominal anatomy
and of the anatomical variants is essential to understand-
ing the spread of pathology within the peritoneum. Of
special importance are constant landmarks, i.e., the
anatomical relationships maintained and bounded by
peritoneal and fascial attachments as well as by the ab-
dominal adipose tissue. The peritoneal and extraperi-
toneal spaces and their fascial planes create complex
three-dimensional structures with unique radiological
characteristics. Intraperitoneal and extraperitoneal adi-
pose tissue provides contrast interfaces between the or-
gans and visceral structures. The intra-abdominal adipose
also yields clues as to the spread and localization of many Fig. 1. The superior recess of the lesser sac is filled with ascetic flu-
pathological conditions. id and seen here as it extends superiorly (long arrow) to the lower
The four different pathways for the spread of neoplas- mediastinum, displacing the esophagus (short arrow) to the left side
tic diseases within the abdomen and pelvis are blood-
borne metastasis, lymphatic extension, direct invasion,
and intraperitoneal spread and seeding. Direct invasion
may occur from contiguous primary tumors and usually
implies that a locally aggressive tumor has broken
through fascial planes. Direct invasion from non-contigu-
ous primary tumors typically occurs via spread along the
peritoneal ligaments and mesenteries. Intraperitoneal
spread of malignancy occurs first by seeding of the peri-
toneal cavity with metastatic cells. Tumor spread occurs
via the natural flow of ascitic fluid within the peritoneal
spaces, which are defined by the peritoneal ligaments and
mesenteries [1].
The peritoneum is the largest and the most complexly
arranged serous membrane in the body. The potential
space between the parietal peritoneum lining the abdom-
inal wall and the visceral peritoneum enveloping the ab-
dominal organs is called the peritoneal cavity. It consists
of a main region, termed the greater sac, and a diverticu-
lum, called the omental bursa or lesser sac, situated be- Fig. 2. In this patient with a gastrostomy (PEG) tube, diluted water-
hind the stomach [2-4]. These two areas communicate via soluble contrast material was injected through the tube before an
the epiploic foramen (foramen of Winslow) and normal- abdominal CT scan. However, unbeknownst to the nurse, the tip of
ly contain only a small amount of fluid; therefore, they the tube was out of the stomach such that contrast was injected in-
to the peritoneal cavity. The CT scan reveals opacification of the
are normally not visible on cross-sectional imaging, ex- peritoneal cavity, with the exception of the bare area of the liver
cept when there is ascites (Fig. 1) or they are filled inad- (double arrows). The falciform ligament (single arrow) is outlined
vertently by contrast (Fig. 2). Fluid dynamics, respiratory by contrast material
Spread of Metastatic Disease in the Abdomen and Pelvis 147

motion, gravity, and anatomical barriers dictate the

spread of disease processes within the peritoneal cavity
and their appearance on cross-sectional imaging.
Tumors that arise within the abdominal cavity have the
propensity to spread (1) via the peritoneal ligaments and
mesenteries that suspend their organs of origin, either by
lymphatic extension or direct invasion, or (2) via seeding
of the peritoneal spaces in which they reside [5]. A work-
ing knowledge of the anatomy of the peritoneal liga-
ments, mesenteries, and spaces, although complex, per-
mits one to narrow the diagnostic possibilities when
metastatic disease is recognized.
The peritoneal reflections in the abdomen comprise
four mesenteries, eight ligaments, and two omenta orga-
nized as follows:
1. four mesenteries: the small bowel mesentery, the trans-
verse mesocolon, the sigmoid mesocolon, and the
mesoappendix;
2. eight ligaments: right coronary, left coronary, falci-
form, hepatoduodenal, duodenocolic, gastrosplenic,
splenorenal, and phrenicocolic ligaments;
3. two omenta: lesser omentum and greater omentum.
The peritoneal ligaments suspend and support the in-
traperitoneal organs and subdivide the peritoneal cavity
into interconnected compartments that dictate the flow
of fluid and location of disease (Figs. 3, 4) [2, 6]. These
reflections are generally recognizable on computed
tomography (CT) scans as fat-containing structures,
either by their typical location and organ relationships
or by the landmarks provided by their major constituent
vessels. The ligaments can serve as conduits or as barriers
to the spread of disease [7].
Metastatic Spread via the Peritoneal Ligaments
Coronary
Ligament
Transverse
Mesocolon
Small Bowel
Mesentery
Sigmoid
Mesocolon
Fig. 3. Posterior peritoneal reflections and recesses. Intraperitoneal
fluid flows naturally from the pelvis to the upper abdomen, prefer-
entially through the right rather than the left paracolic gutter ow-
ing to the broader diameter of the former. In addition, flow in the
left paracolic gutter is cut off from reaching the left subphrenic
space by the phrenicocolic ligament. The transverse mesocolon di-
vides the abdomen into supra- and inframesocolic spaces. In the
right inframesocolic space, fluid is impeded from draining into the
pelvis via the small bowel mesentery. Owing to natural holdup of
fluid at the root of the small bowel mesentery and sigmoid meso-
colon, these structures are naturally predisposed to involvement
with serosal-based metastases in the setting of peritoneal carcino-
matosis (Reprinted with permission from [6])
Left lobe
of liver

The major attachments of the upper abdomen include the Stomach

lesser omentum, the gastrosplenic ligament, and the
splenorenal ligament. The lesser omentum is subdivided
into the gastrohepatic ligament and the hepatoduodenal
GSL
ligament. In the embryo, the gastrosplenic ligament gives
rise to the gastrocolic ligament (also known as the greater
omentum) and the transverse mesocolon.

Gastrohepatic Ligament Lesser

P SAC
The gastrohepatic ligament can be recognized on CT
scans as a fatty plane that joins the lesser curvature of
the stomach to the liver. It extends from the fissure for
the ligamentum venosum to the porta hepatis and con- LK
tains the left gastric artery, coronary vein, and associat- Spleen
ed lymphatics. The size defining nodal enlargement in Splenic
this region is somewhat smaller than elsewhere in the artery
abdomen; nodes in the gastrohepatic ligament are gen-
erally considered abnormal when they exceed 8 mm in SRL
diameter [3]. On occasion, pathology in the gastro- Fig. 4. Peritoneal components in the left upper quadrant. The gas-
hepatic ligament may be mimicked by the absence of trosplenic ligament (GSL) and splenorenal ligament (SRL) are
opacification of the bowel loops, the pancreatic neck, or clearly seen (From [2])
148
Fig. 5. Intraperitoneal seeding of numerous hydatid cysts. This
patient had a hydatid cyst in the liver that ruptured during resec-
tion, resulting in extensive peritoneal implants, including in the
lesser omentum, which is seen here between the stomach, liver,
and spleen
the papillary process of the caudate lobe of the liver
projecting into the expected plane of the gastrohepatic
ligament [8, 9].
The gastrohepatic ligament provides an important con-
duit of disease from the stomach to the liver in that the
subperitoneal areolar tissue within the ligament is con-
tinuous with the Glisson capsule (the perivascular fibrous
capsule within the liver). Thus, gastric malignancy can
spread directly into the left lobe of the liver and vice ver-
sa via this pathway. Neoplastic or infectious conditions
that may spread throughout the entire peritoneum also in-
volve this ligament. On CT, the abnormality is seen be-
tween the stomach and the liver (Fig. 5). Common neo-
plasms spreading via the gastrohepatic ligament include
nodal metastases from gastric, esophageal, breast, pan-
creatic, and lung cancer as well as nodal involvement of
lymphoma. Gastric and esophageal cancer can directly
invade the gastrohepatic ligament and spread into the left
hepatic lobe [3].
Hepatoduodenal Ligament
The hepatoduodenal ligament is the free edge of the gas-
trohepatic ligament along its rightward aspect. It contains
important structures of the porta hepatis, including the
common bile duct, hepatic artery, and portal vein. The
hepatoduodenal ligament extends from the flexure be-
tween the first and second duodenum to the porta hepatis;
the foramen of Winslow is immediately posterior to this
ligament, permitting communication between the greater
and lesser sacs [10]. The nodes of the foramen of
Winslow, or portocaval space, have an unusual morphol-
ogy such that their transverse dimension is greater than
their anteroposterior dimension. Generally, the upper lim-
it of normal for the latter is 1.0-1.3 cm, whereas the for-
mer can be up to 2.0 cm in width. Size criteria are some-
what less helpful than in other lymph nodes. In the ab-
sence of frank enlargement, a more spherical shape or
James A. Brink, Ali Shirkhoda
Fig. 6. Invasive pancreatic carcinoma arising from the pancreatic
tail, with numerous hematogenous metastases to the liver. A porto-
caval lymph node at the base of the hepatoduodenal ligament (ar-
row) is not enlarged according to size criteria but contains metasta-
tic disease, as evidenced by its central necrosis
central necrosis suggests the presence of tumor within
these nodes (Fig. 6) [11, 12].
A broad range of tumors may spread via the hepato-
duodenal ligament. Liver or biliary cancer, whether pri-
mary or metastatic, may spread in an antegrade fashion
through lymphatics in the hepatoduodenal ligament to
deposit in periduodenal or peripancreatic lymph nodes.
Similarly, malignant disease in the nodes about the su-
perior mesenteric artery (commonly involved in pan-
creatic and colon cancer) can spread in a retrograde
fashion up the lymphatics in the hepatoduodenal liga-
ment. Lymphoma can involve these nodes as well. Pri-
mary gastric cancer arising in the lesser curvature of
the stomach can directly spread through the gastro-
hepatic ligament to the hepatoduodenal ligament and
then to peripancreatic and periduodenal nodes. Vascular
complications related to the portal vein and hepatic
artery can result; portal venous thrombosis and hepatic
arterial pseudoaneurysms may occur in advanced cases
owing to their coexistence in the hepatoduodenal liga-
ment [2, 10].
Gastrosplenic and Splenorenal Ligaments
In the embryo, the gastrosplenic ligament is a long liga-
mentous attachment between the stomach and the
retroperitoneum. It gives rise to the gastrocolic ligament
(greater omentum) and the transverse mesocolon. In the
adult, the gastrosplenic ligament is a thin ligamentous at-
tachment between the greater curvature of the stomach and
the splenic hilus (Fig. 4). It contains the left gastroepiploic
and short gastric vessels as well as associated lymphatics.
The gastrosplenic ligament is continuous with the gastro-
colic ligament inferiorly and medially and with the
splenorenal ligament posteriorly and medially [13, 14]. As
such, it provides an important pathway of communication
between the stomach, spleen, and retroperitoneum. Gastric
malignancies commonly spread through the gastrosplenic
Spread of Metastatic Disease in the Abdomen and Pelvis
a b
Fig. 7 a, b. Gastric adenocarcinoma invading
the spleen via the gastrosplenic ligament.
a Initial contrast-enhanced CT scan reveals
circumferential tumor involving the gastric
fundus. b A repeat CT scan 6 months later
shows invasion and dissection of the spleen
secondary to tumor spread via the gastro-
splenic ligament
ligament (Fig. 7), thereby involving the spleen and ulti-
mately resulting in disease about the tail of the pancreas.
Conversely, pancreatic neoplasms may spread via the
splenorenal ligament to the gastrosplenic ligament and in-
volve the greater curvature of the stomach [2].
Gastrocolic Ligament
The gastrocolic ligament (or greater omentum) joins the
greater curvature of the stomach to the transverse colon.
On the left, it is continuous with the gastrosplenic liga-
ment; on the right, it ends at the gastroduodenal junction,
near the hepatoduodenal ligament. Since, developmental-
ly, it results from fusion of the anterior and posterior
leaves of the gastrosplenic ligament, it contains the four
layers of peritoneum that invest the stomach and has a po-
tential space within it (Fig. 8) [15].
RPS
G On the right, the transverse mesocolon is continuous
P with the duodenocolic ligament; in the middle, it is
D
TC
GCL
Fig. 8. The gastrocolic ligament (GCL) joins the greater curvature
of the stomach (G) to the transverse colon (TC). In concert with the
transverse mesocolon, a pathway of disease is formed between
retroperitoneal structures, such as the pancreas (P) and the duode-
num (D), and the anterior aspect of the intraperitoneal cavity
(Modified from [15])
149
The gastrocolic ligament contains the gastroepiploic
vessels and associated lymphatics. It provides an impor-
tant conduit of malignant disease from the greater curva-
ture of the stomach to the transverse colon and vice ver-
sa. When viewed in concert with the transverse meso-
colon, a conduit exists between the greater curvature of
the stomach and the retroperitoneum. In addition to al-
lowing direct spread of disease between the stomach,
transverse colon, and pancreas, the gastrocolic ligament
serves as an important nidus for peritoneal metastases –
as commonly occur with ovarian, gastric, colon, and pan-
creatic cancers [16, 17]. Finally, dilated veins within this
ligament may represent gastroepiploic collaterals result-
ing from splenic venous compromise, such as might oc-
cur in the setting of invasive pancreatic tumors or in the
presence of intraperitoneal tumors, which spread to the
retroperitoneum via the transverse mesocolon (Fig. 9).
Transverse Mesocolon
The transverse mesocolon serves as a broad conduit of
disease across the mid-abdomen; bare areas link the pan-
creas to the transverse colon, spleen, and small bowel.
Fig. 9. A pancreatic islet-cell tumor arising in the pancreatic tail has
resulted in splenic vein thrombosis, with secondary short gastric
venous collaterals in the gastrosplenic and splenorenal ligaments as
well as gastroepiploic venous collaterals (arrow) in the gastrocolic
ligament
150
continuous with the small bowel mesentery; and on the
left, it is continuous with the phrenicocolic and splenore-
nal ligaments (Fig. 10). It contains the middle colic ves-
sels and associated lymphatics. On CT, the transverse
mesocolon may be recognized as a fatty plane at the lev-
el of the uncinate process. Pancreatic tumors often
spread ventrally into the transverse mesocolon to involve
SRL
TM
PCL
SBM
Fig. 10. The transverse mesocolon (TM) provides an important con-
duit for the spread of disease across the mid-abdomen. It is con-
tinuous with the splenorenal ligament (SRL) and phrenicocolic lig-
ament (PCL) on the left and with the duodenocolic ligament on the
right. In its mid-portion, it is continuous with the small bowel
mesentery (SBM) (Reprinted with permission from [6])
a b
a b
James A. Brink, Ali Shirkhoda
the transverse colon and have the propensity to continue
through the gastrocolic ligament to involve the stomach
(Fig. 11). Alternatively, they may spread through the
transverse mesocolon to involve the proximal jejunum,
just beyond the ligament of Treitz (Fig. 12). Like the
gastrocolic ligament, a potential space exists within the
transverse mesocolon due to embryological fusion of the
gastrosplenic ligament with the embryological trans-
verse mesocolon [13]. A less common but important
route of spread also exists between the right colon and
the periduodenal and peripancreatic nodes via the right-
ward aspect of the transverse mesocolon (duodenocolic
ligament). This is important because lymphadenopathy
in the periduodenal and peripancreatic regions may her-
ald a right colon cancer when other, more common caus-
es of lymphadenopathy in this region are excluded [2].
Thus, three routes of spread between the intraperi-
toneal viscera and retroperitoneum are provided by three
pairs of ligaments. The gastrohepatic and hepatoduodenal
ligaments link the liver and lesser curvature of the stom-
ach to the retroperitoneum; the gastrosplenic and spleno-
renal ligaments link the superior greater curvature of the
stomach and spleen to the retroperitoneum; and the
gastrocolic and transverse mesocolon link the inferior
greater curvature of the stomach and transverse colon to
the retroperitoneum. The ligamentous pair in which
metastatic disease is recognized can therefore suggest the
organ of origin and, in the case of gastric cancer, the
location of the primary tumor within the stomach.
Fig. 11 a, b. Invasive pancreatic carcinoma in-
vading the retroperitoneum, transverse meso-
colon, and greater omentum. Transaxial im-
age (a) and parasagittal reformation (b). M
mass, P normal pancreas, S stomach, TC
transverse colon. The transverse mesocolon
(black arrows) provides a pathway for the
spread of metastatic disease to the greater
omentum (white arrows). Also noted is en-
casement of the left renal artery and ovarian
vein (double arrows)
Fig. 12 a, b. An incidental pancreatic adeno-
carcinoma involving the splenic vein (a) was
thought to be resectable owing to the lack of
extrapancreatic involvement. At surgery, up-
on elevation of the transverse colon, the tu-
mor was found to have penetrated the base of
the transverse mesocolon and involved the
proximal jejunum just beyond the ligament
of Treitz (seen in retrospect in b, arrows)
Spread of Metastatic Disease in the Abdomen and Pelvis 151

Metastatic Spread via the Peritoneal Spaces
Diseases such as ovarian carcinoma (Fig. 13 a) and lym-
phoma (Fig. 13 b) can diffusely extend through the peri-
toneum, with propensity for involvement of the greater
omentum. However, the initiation and growth of seeded
metastases on the peritoneal surfaces usually depend on
the natural flow of ascites through the peritoneal spaces.
Primary abdominal malignancies and secondary nodal
metastases can break through the visceral peritoneum and
shed cells into the peritoneal cavity (Fig. 13 c). Once in-
traperitoneal, such cells propagate through the peritoneal
spaces along predicable routes. A thorough understand-
ing of the anatomy of the peritoneal spaces may help re-
fine differential diagnoses for the source of intra-abdom-
inal metastases [18, 19]. By recognizing that a process is
intraperitoneal, one may better predict its organ of origin
and likely routes of spread (Fig. 13).
The most dependent portion of the peritoneal cavity is
in the pelvis. The cavity is anatomically continuous with
the paracolic gutters and is subdivided into the midline
pouch of Douglas (rectovaginal pouch in women, rec-
tovesical pouch in men), the lateral paravesical recesses,
and the medial and lateral inguinal fossae. Due to the
deep and dependent nature of the pelvic peritoneal cavi-
ty, many infections and half of all seeded metastases will
involve the pouch of Douglas. Fluid flows preferentially
to the right lower quadrant and from there to the inferior
portion of the small bowel mesentery and right paracolic
gutter.
Left Peritoneal Space
The left peritoneal space can be subdivided into four
compartments. Although these freely communicate with
each other, the inflammatory nature of exudative fluid
collections within them favors the development of fi-
brous adhesions, which may seal off one or more portions
of the left peritoneal space from the others.
The left anterior perihepatic space is limited on the
right by the falciform ligament and on the left by the an-
terior wall of the stomach. It follows the posterior curve
of the diaphragm and is limited posteriorly by the left
coronary ligament (Fig. 14).

a b c

Fig. 13 a-c. Peritoneal malignancies. a Diffuse omental metastasis from a primary ovarian carcinoma associated with ascites. b Extensive
involvement of the greater omentum by Burkitt’s lymphoma. c Seeding of the inferior aspect of the peritoneum (cul de sac) by gastric
carcinoma (arrows)

a Left Left b
Subphrenic Coronary Rt. Coronary Lig.
Space Ligament Superior reflection
Fig. 14 a, b. The left (a) and
right (b) perihepatic spaces
are bounded posteriorly by Inferior reflection
the coronary ligaments.
The reflections of the coro- L A
LL
nary ligaments mark the Rt. Posterior
site of the non-peritoneal- Perihepatic Space
ized “bare area” of the Lesser (Morison’s Pouch)
liver. LL left lobe of the Omentum D K
liver, LK left kidney,
S stomach, TC transverse S Lesser
colon, P pancreas, D duo- LK Rt. Anterior
Sac C
denum, Lu lung, L liver P Perihepatic Space
(right lobe), A adrenal,
K kidney, C colon (Reprint- TC D
ed with permission from [6])
152
The left posterior perihepatic space (gastrohepatic
recess) is limited on the left by the lateral wall of the
stomach. This space follows the posterior margin of the
left hepatic lobe deep into the fissure for the ligamentum
venosum to form the posterior margin of the left hepatic
lobe. Thus, it is in close proximity to the lesser curve of
the stomach, the anterior wall of the duodenal bulb, and
the anterior wall of the gallbladder [13].
Although the gastrohepatic recess is close to the less-
er sac (divided from it by the lesser omentum), it is a por-
tion of the left peritoneal space, while the lesser sac is a
portion of the right peritoneal space (Fig. 15). This dis-
tinction is important in that lesser sac collections are very
difficult to approach percutaneously whereas gastro-
hepatic recess collections are usually accessible by guiding
a catheter along the inferior margin of the left hepatic lobe.
The anterior left subphrenic space is in direct conti-
nuity with the left anterior perihepatic space, which
forms its right boundary. Far to the left, on the antero-
lateral surface of the stomach, this space is limited by the
greater omentum. This is a common site for fluid locula-
tion in the setting of malignant ascites (Fig. 16).
Caudate Lesser Fissure for
lobe omentum ligamentum venosum
Liver
IVC
Ao
Diaphragm Upper recess Esophagus
Fig. 15. The boundaries of the superior recess of the lesser sac may
be recognized when fluid engulfs the caudate lobe. The lesser
omentum separates this fluid from fluid in the fissure for the liga-
mentum venosum, which is in continuity with the left posterior
perihepatic space (gastrohepatic recess). IVC inferior vena cava,
Ao aorta (Reprinted with permission from [6])
Fig. 16. Cystic mesothelioma. Cystic and solid lesions are seen fill-
ing the left anterior and posterior subphrenic spaces
James A. Brink, Ali Shirkhoda
The posterior left subphrenic (perisplenic) space is the
posterior continuation of the anterior subphrenic space
and generally surrounds the lateral and superior margins
of the spleen. The “bare areas” of the spleen are reliably
observed in perisplenic fluid collections [20-22]. Superi-
orly, the perisplenic space is entirely subphrenic and sur-
rounds the top of the spleen (Fig. 16) [23].
Right Peritoneal Space
There are three major subdivisions of the right peritoneal
space: the right subphrenic space, the hepatorenal recess,
and the lesser sac (Figs. 14, 15). The right subphrenic
space occupies the smoothly contoured area between the
superolateral margin of the liver and the right hemi-
diaphragm. The medial extension of this compartment is
limited by the right coronary ligament, which is simply
the right lateral margin of the liver’s bare area [24]. The
hepatorenal recess (or Morison’s pouch) is the postero-
medial extension of the subphrenic space, inferior to the
coronary ligament. As its name implies, it extends be-
tween the right hepatic lobe and the anterior border of the
right kidney.
The lesser sac has two major components [14, 25]: a
small superior recess is immediately posterior to the
hepatoduodenal ligament. The caudate lobe of the liver is
enveloped by this peritoneal reflection (Fig. 15). The
larger inferior recess occupies the space behind the stom-
ach, anterior to the transverse mesocolon and medial to
the gastrosplenic ligament. As both portions of the lesser
sac are surrounded by abdominal viscera, percutaneous
drainage of collections within this space is difficult. In-
feriorly, the superior recess communicates with the right
perihepatic space through the foramen of Winslow. Cau-
dally, behind the duodenum and pancreatic head, there
may be an extension that is responsible for peritoneal
fluid collections behind the pancreatic head [7].
References
1. Levy AD, Shaw JC, Sobin LH (2009) Secondary tumors and
tumorlike lesions of the peritoneal cavity: imaging features
with pathologic correlation. RadioGraphics 29:347-373
2. Meyers MA, Oliphant M, Berne AS, Feldberg MAM (1987)
The peritoneal ligaments and mesenteries: pathways of in-
traabdominal spread of disease. Radiology 163:593-604
3. Balfe DM, Mauro MA, Koehler RE et al (1984) Gastrohepat-
ic ligament: normal and pathologic CT anatomy. Radiology
150:485-490
4. Dodds WJ, Foley WD, Lawson TL et al (1985) Anatomy and
imaging of the lesser peritoneal sac. Am J Roentgenol
144:567-575
5. Low RN (2007) MR imaging of the peritoneal spread of ma-
lignancy. Abdom Imaging 32:267-283
6. Meyers MA (1994) Dynamic radiology of the abdomen: normal
and pathologic anatomy. 4th edn. Springer-Verlag, New York
7. Sivit CJ (1996) CT of mesentery-omentum peritoneum. Radi-
ol Clin North Am 34:863-884
8. Auh YH, Rosen A, Rubenstein WA et al (1984) CT of the pap-
illary process of the caudate lobe of the liver. AJR Am J
Roentgenol 142:535-538
Spread of Metastatic Disease in the Abdomen and Pelvis
9. Donoso L, Martinez-Noguera A, Zidan A, Lora F (1989) Pap-
illary process of the caudate lobe of the liver: sonographic ap-
pearance. Radiology 173:631-633
10. Weinstein JB, Heiken JP, Lee JKT et al (1986) High resolution
CT of the porta hepatis and hepatoduodenal ligament. Radio-
Graphics 6:55-74
11. Zirinsky K, Auh YH, Rubenstein WA et al (1985) The porta-
caval space: CT with MR correlation. Radiology 156:453-460
12. Ito K, Choji T, Fujita T et al (1993) Imaging of the portacaval
space. AJR Am J Roentgenol 161:329-334
13. Vincent LM, Mauro MA, Mittelstaedt CA (1984) The lesser
sac and gastrohepatic recess: sonographic appearance and dif-
ferentiation of fluid collections. Radiology 150:515-519
14. Dodds WJ, Foley WD, Lawson TL et al (1985) Anatomy and
imaging of the lesser peritoneal sac. AJR Am J Roentgenol
144:567-575
15. Langman J (1971) Medical Embryology. Part two: Special em-
bryology. Chap. 13 Digestive system. WB Saunders, Philadel-
phia, PA, p 29
16. Cooper C, Jeffrey RB, Silverman PM et al (1986) Computed
tomography of omental pathology. J Comput Assist Tomogr
10:62-66
17. Rubesin SE, Levine MS, Glick SN (1986) Gastric involvement
by omental cakes: radiographic findings. Gastrointest Radiol
11:223-228
153
18. DeMeo JH, Fulcher AS, Austin RF (1995) Anatomic CT
demonstration of the peritoneal spaces, ligaments, and mesen-
teries: normal and pathologic processes. RadioGraphics
15:755-770
19. Pai RK, Longacre TA (2007) Pseudomyxoma peritonei syn-
drome: classification of appendiceal mucinous tumours. In:
Ceelen WP (ed) Peritoneal carcinomatosis: a multidisciplinary
approach. Springer-Verlag, New York, pp 71-107
20. Vibhakar SD, Bellon EM (1984) The bare area of the spleen:
a constant CT feature of the ascitic abdomen. AJR Am J
Roentgenol 141:953-955
21. Rubenstein WA, Auh YH, Zirinsky K et al (1985) Posterior
peritoneal recesses: assessment using CT. Radiology 156:
461-468
22. Love L, Demos TC, Posniak H (1985) CT of retrorenal fluid
collections. AJR Am J Roentgenol 145:87-91
23. Crass JR, Maile CW, Frick MP (1985) Catheter drainage of the
left posterior subphrenic space: a reliable percutaneous ap-
proach. Gastrointest Radiol 10:397-398
24. Rubenstein WA, Auh TH, Whalen JP, Kazem E (1983) The
perihepatic spaces: computed tomographic and ultrasound
imaging. Radiology 149:231-239
25. Jeffrey RB, Federle MP, Goodman PC (1981) Computed
tomography of the lesser peritoneal sac. Radiology 141:
117-122

Abdominal Vascular Disease: Diagnosis and Therapy
Johannes Lammer
Cardiovascular and Interventional Radiology, AKH-Universitätsklinik, Wien, Austria
Acute Arterial Occlusion
If acute occlusion of the abdominal aorta or branch ves-
sels occurs, two causes have to be considered: (1) em-
bolization and (2) dissection.
Embolization
Embolization is most common in elderly patients with
atrial fibrilation, a history of myocardial infarction, and
aortic aneurysm. Embolization from the heart or thoracic
aorta may cause acute subtotal or total occlusion of the
celiac artery, superior mesenteric artery (SMA), inferior
mesenteric artery (IMA), or the renal arteries. The most
common causes of embolization are:
• atrial fibrillation;
• thoracic aortic aneurysm;
• myocardial infarction with mural thrombus formation;
• advanced aortic arteriosclerosis;
• perforating arteriosclerotic ulcer;
• hypercoagulability syndrome.
The symptoms of acute ischemia are dependent on the
involved vascular territory. Embolization to the liver or
spleen may cause acute right or left upper abdominal
pain. Acute mesenteric ischemia typically causes severe
abdominal pain and bowel paralysis. Embolization to the
kidney results in flank pain, hematuria, and hypertension.
In any case, severe elevation of serum lactate dehydroge-
nase levels points to the ischemic nature of the acute
pain. In case of complete ischemia without a sufficient
collateral circulation, the warm ischemic time tolerated
by the abdominal organs is <6 h. Therefore acute diag-
nosis and therapy are mandatory.
The primary diagnosis is made by CT with contrast en-
hancement injected at a concentration of 300-400 mg io-
dine/mL and at an injection rate of 4 mL/s, with a total
bolus volume of 80-120 mL. A bolus care technique with
a delay of 20-40 s is used. The CT settings are 2-mm col-
limation, pitch = 2, with a reconstruction interval of 1 mm.
In the arterial phase, the obstructing embolus and the is-
chemic territory can be visualized. In the delayed phase,
residual perfusion through the collateral arteries may be
demonstrated.
IDKD 2010-2013
Interventional treatment with a thrombectomy device
and/or local intra-arterial fibrinolysis with recombinant
tissue plasminogen activator (loading dose 10 mg, infu-
sion dose 5 mg/h) or urokinase (loading dose 250,000 IU,
infusion dose 100,000 IU/h) together with a glycoprotein
IIb/IIIa antagonist (Aciximab: loading dose 0.25 mg/kg,
infusion dose 0.125 mg/kg/h) is one option. The other op-
tion is surgery, which may be faster, enables inspection,
and, if necessary, resection of the ischemic organs.
Dissection
Patients with chest trauma, chronic severe hypertension,
or a connective tissue disease such as Marfan syndrome
or Ehlers-Danlos syndrome are vulnerable to aortic dis-
section. Acute type A and B aortic dissection may cause
dynamic compression of the true lumen of the aorta by
the pressurized false lumen. This can result in acute is-
chemia of the liver and spleen, the bowel, and one or both
kidneys. The dissection plane may also run into one of
the arteries perfusing the organ, thereby causing obstruc-
tion of the true lumen.
There are several options in the interventional treat-
ment of dissection. First, in case of dynamic compression
of the true aortic lumen, occlusion of the proximal entry
into the false lumen with an aortic stent graft will de-
compress the false lumen and result in re-opening of both
the true lumen of the aorta and the side branches (Fig. 1).
Second, in case of static compression due to a side-
branch dissection, stent placement in the true lumen of
the organ artery will reconstitute organ perfusion. Third,
in case of organ perfusion through the false lumen, bal-
loon fenestration of the intimal flap will re-establish flow
into the malperfused territory.
Chronic Arterial Occlusive Disease
In young patients, potential causes of chronic arterial oc-
clusion are fibromuscular disease, Takayasu arteritis, and
Recklinghausen neurofibromatosis. In elderly patients,
the primary cause of chronic arterial occlusive disease is
arteriosclerosis.
Abdominal Vascular Disease: Diagnosis and Therapy 155

a b c

d e f

Fig. 1 a-f. Complicated type B aortic dissection. CT demonstrating (a) primary entry tear of type B aortic dissection and (b) compression of
true lumen of the aorta. c Aortic angiogram showing true and false lumen of type B dissection. d Aortography showing compression of the
true lumen, with malperfusion of the superior mesenteric and renal arteries (“floating visceral sign”). Aortography (e) after stentgraft im-
plantation and closure of primary entry tear, and (f) following stent graft implantation, which demonstrates spontaneous revascularization
of the visceral arteries

Mesenteric Artery Stenosis post-prandial cramps in 86%, weight loss in 74%,

Between the three large mesenteric arteries (celiac artery,
SMA, IMA) there are two main collateral pathways:
(1) the pancreatico-duodenal arteries and the arc of
Buehler, between the celiac artery and SMA; (2) the arc
of Riolan and the marginal artery of Drummond, between
the SMA and IMA. Therefore, an obstruction of at least
two mesenteric arteries is necessary to cause ischemic
symptoms. The typical clinical symptom is abdominal
angina, with abdominal pain in 94% of patients,
abdominal bruit in 70%, and diarrhea.
The primary diagnosis is made by computed tomogra-
phy angiography (CTA), magnetic resonance angiogra-
phy (MRA), or by intra-arterial catheter angiography of
the abdominal aorta in a lateral projection. Intervention-
al treatment consists of percutaneous transluminal angio-
plasty (PTA) with or without secondary stent placement
in at least one of the obstructed arteries.
The causes and symptoms of chronic arterial occlusion
depend on the obstructed artery and its location. In
156
celiac trunk stenosis, chronic obstruction may remain
asymptomatic because of the collateral pathways through
the gastroduodenal and pancreatic arteries from the
SMA. The causes are an arteriosclerotic plaque, com-
pression by the arcuate ligament, or carcinoma of the
pancreas. Superior mesenteric artery stenosis will result
in post-prandial abdominal pain (abdominal angina) only
if two or all three gastrointestinal arteries are obstructed.
The causes for SMA obstruction are arteriosclerosis, fi-
bromuscular disease, Takayasu arteritis, pancreatic carci-
noma, or chronic pancreatitis. Inferior mesenteric artery
stenosis, with obstruction of the IMA, is most common-
ly observed in patients with advanced atheromatosis or a
partially thrombosed abdominal aortic aneurysm. Due to
the collateral circulation through the arc of Riolan and
the marginal artery, IMA obstruction normally remains
asymptomatic.
Renal Artery Stenosis
Hypertension and/or renal insufficiency are the most fre-
quent consequences of renal artery stenosis. Acute onset
of the clinical symptoms and repeated flash pulmonary
edema are suggestive. The most common etiology in pa-
tients over age 50 years is arteriosclerosis (65-75% of all
patients), with males more often affected than females. In
the majority of cases, the proximal 2 cm of the renal
artery are involved, accompanied by atherosclerotic
changes in the aorta. In 30% of patients, the stenosis is
bilateral.
In patients under the age of 50, renal artery stenosis is
most often due to fibromuscular disease (20-30% of all
patients). In this case, females are five times more likely
than males ratio to be affected. Most commonly, the mid-
dle to distal renal artery, including its branches, is ob-
structed, with bilateral involvement in 50-70% of pa-
tients. Imaging shows the typical “string of pearls” ap-
pearance of the stenosis and, possibly, aneurysms and dis-
sections. However, importantly, there is no aortic disease.
Other causes of renal artery stenosis include Takayasu
arteritis, mid-aortic syndrome, Recklinghausen neuro-
fibromatosis, and as a consequence of radiation therapy.
However, an algorithm for the diagnosis of a renal artery
stenosis has yet to be established.
Color duplex ultrasound is a non-invasive but complex
examination that requires operator experience. The diag-
nostic criteria for renal artery stenosis are: increased peak
systolic velocity >250 cm/s, a renal-to-aortic ratio of
peak systolic velocity >3.5, intrastenotic turbulence, and
a flattened pulse wave in the periphery (pulsus tardus).
The sensitivity of color duplex sonography for detection
of a renal artery stenosis >70% is 72-92%. Color duplex
ultrasound with an angiotensin-converting enzyme
(ACE) inhibitor provides a positive predictive value of
67-95% for cure or improvement after revascularization.
A nuclear scan, specifically, renal scintigraphy with
technetium-99m mercaptoacetyltriglycine (MAG3) or
Tc-99m diethylenetriaminepentaacetic acid (DTPA) in
Johannes Lammer
patients previously administered an ACE inhibitor (cap-
topril 25 mg) shows a delayed wash-out of the tracer
within the post-stenotic kidney. In bilateral disease and in
chronic ischemic nephropathy, however, lateralization of
the tracer is less evident. In a selected population at clin-
ically high risk for renal artery stenosis, the sensitivity for
detection of a unilateral stenosis >70% is 51-96% (mean
82%). Its positive predictive value for a renal artery
stenosis with improvement of hypertension after revascu-
larization is 51-100% (mean 85%). However, scintigra-
phy is much less sensitive in unselected patients and in
those with bilateral disease, impaired renal function, uri-
nary obstruction, and chronic ACE inhibitor intake.
Newer tests are gadolinium-enhanced MRA and spiral
CTA. For state-of-the-art MRA, high-field-strength sys-
tems with high performance gradients are necessary to
obtain breath-hold 3D T1-weighted spoiled gradient-echo
imaging with short TR and TE. Intravenous administra-
tion of gadolinium contrast material (0.1 mmol/kg; flow
rate 2 mL/s), a central k-space readout, and background
subtraction are additional techniques to improve signal-
to-noise ratio and spatial resolution. The sensitivity of
MRA to detect a renal artery stenosis >50% is over 95%
(Fig. 2). The main limitations of renal MRA are its lack
of accuracy in the evaluation of small accessory renal ar-
teries and branch vessels, artifacts due to the presence of
stents, and a tendency of the techniques to overestimate
moderate stenoses. In a double-blind randomized study,
contrast-enhanced MRA with the blood-pool contrast
agent gadofosveset was not superior to gadobenate
dimeglumine.
The sensitivity of CTA to detect stenosis of the renal
artery and its accessory arteries is >95%. For high-
quality opacification of the renal arteries and to avoid
renal vein overlap, correct bolus planning is mandatory:
density measurement during bolus rise, flow 4 mL/s, total
volume 80-120 mL (multidetector scanners need less
contrast). A short breath-hold acquisition, 1- to 2-mm
collimation, pitch =1.5-6 (depending on single or multi-
detector technology), and an overlap of reconstruction of
0.5-0.75 are important parameters to obtain good spatial
resolution of the study. Curved planar reconstruction
(most useful for stents), volume rendering, and maximum
intensity projection (MIP) are used for 3D imaging (Fig. 2).
Nonetheless, intra-arterial catheter arteriography together
with pressure gradient measurement is still the gold stan-
dard for the evaluation of a renal artery stenosis.
The revascularization technique of choice is renal PTA,
without or with stent placement (Fig. 2). Aorto-renal by-
pass surgery is indicated only if PTA fails. In a recently
published meta-analysis, renal arterial stent placement
proved to be technically superior and clinically compa-
rable to renal PTA alone. The technical success rate of
stent vs. PTA was 98 vs. 77%, and the re-stenosis rate
17 vs. 26% (p <0.001). In hypertension, the cure rate of
PTA vs. stent was 10 vs. 20%, the rate of improvement
53 vs. 49 %. In renal insufficiency, the rate of improve-
ment was 38 vs. 30%, that of stabilization 41 vs. 38%.
Abdominal Vascular Disease: Diagnosis and Therapy
a b
c d
Fig. 2 a-d. Imaging and intervention in a pa-
tient with hypertension due to renal artery
stenosis. a MRA, b CTA, and c arteriography
show the stenosis of the right renal artery.
d Arteriography of right renal artery after
stent placement
The complication rate was 11-13% (95% CI 6-19%), the
in-hospital mortality rate 1%. In a randomized study
comparing stents vs. PTA in ostial stenoses, the techni-
cal success rate was 88 vs. 57%, and the 6-month pri-
mary patency rate 75 vs. 29%. Surprisingly, randomized
trials comparing the effect of PTA and drug therapy on
renal hypertension did not reveal a significant benefit of
PTA and stenting over continuous drug therapy. Howev-
er, in a Dutch study, PTA patients required only 2.1
vs. 3.2 daily drug doses (p <0.001), and 22 of 53 patients
in the drug group had to be switched to the PTA group
because of persistent hypertension or deterioration of
renal function.
Aneurysms
Abdominal Aortic Aneurysm
The incidence of abdominal aortic aneurysm in European
adults 60 years and older is 2.5%. Up to 10% of patients
with symptomatic peripheral arterial disease die from
rupture of the aneurysm.
Currently, the standard treatment is open surgery.
However, endovascular implantation of stent grafts is a
157
new, emerging technique that may replace open surgery
in the future. Since the first clinical implant of a tube
stent graft, in 1990, many different stent graft designs
have been developed and tested in feasibility studies.
Most recently, randomized studies (EVAR 1, Dream)
compared the results of open vs. endovascular repair. In
the EVAR trial, the 30-day mortality in the EVAR group
was 1.7% (9/531) vs. 4.7% (24/516) in the open repair
group (p = 0.009). Four years after randomization, all-
cause mortality was similar in the two groups (about
28%; p = 0.46), although there was a persistent reduc-
tion in aneurysm-related deaths in the EVAR group
(4 vs. 7%; p = 0.04).
Indications
The indications for endovascular treatment of abdominal
aortic aneurysm are currently the same as for open
surgery: (1) diameter of the aneurysm >5 cm (Fig. 3), (2)
documented growth >0.5 mm/year, (3) symptomatic
aneurysm (i.e., embolization, pain, ureteral compres-
sion), and (4) rupture. The specific clinical indications
for the endovascular approach are typically: patients >75
years old, ASA class 3 and 4, “hostile abdomen”, and in-
flammatory aneurysm or horse-shoe kidney.
158
a b
Fig. 3 a-c. Patient with abdominal aortic aneurysm and renal artery stenosis. a CTA with MIP reconstruction; aortography (b) before and
(c) after stent graft placement
The anatomic indications for stent graft treatment are:
• infrarenal neck >15 mm in length;
• infrarenal neck without thrombus or severe calcifica-
tion;
• angulation of the infrarenal neck <65°;
• patent celiac trunk and SMA;
• stent graft diameter 10% more than neck diameter;
• iliac artery angulation <90°;
• iliac artery without thrombus or severe calcification;
• overlap of >15 mm within the iliac arteries.
Endovascular implantation of stent grafts can be per-
formed under general, epidural, or local anesthesia. The
use of epidural anesthesia is a major advantage in elder-
ly and high-risk patients.
Stent Graft Designs
Stent grafts have a self-expandable stent structure cov-
ered by an ultrathin polyester or ePTFE fabric. Current-
ly, only bifurcated stentgrafts are used for the treatment
of abdominal aortic aneurysm.
Imaging before Stent Graft Implantation
Contrast-enhanced spiral CT with multiplanar recon-
struction (MPR) or MIP reconstruction is the most
important examination before stent graft implantation
(Fig. 3). The diameter of the landing zones (infrarenal
neck, iliac arteries), the maximum diameter of the
aneurysm, the extent of the thrombus, and calcifications
are well depicted on CT.
Complications
The most frequent complication is incomplete exclusion
of the aneurysm, with remaining pressurization of the
Johannes Lammer
c
aneurysm sack through an endoleak. White and May pro-
posed a classification of primary (<30 days) and sec-
ondary (>30 days) endoleaks. Type 1 endoleaks are char-
acterized by direct perfusion through the proximal (in-
frarenal) or distal (iliac) anastomosis. In type 2, there is
retrograde perfusion through branch vessels (lumbar ar-
teries, IMA, accessory renal artery). Type 3 consists of
mid-graft leak due to disintegration of the stent graft (dis-
connection of the second iliac limb, fabric erosion). In
type 4, there is fabric porosity, while type 5 is character-
ized by endotension.
Visceral Artery Aneurysm
Aneurysms of the celiac trunk, splenic artery, hepatic
artery, gastroduodenal artery, and SMA are caused by ar-
teriosclerosis, arteritis, periarterial inflammation (such as
pancreatitis), trauma, and soft-tissue diseases (such as
Marfan and Ehlers-Danlos syndromes). An aneurysm
>2.5 cm in diameter should be considered for treatment
to prevent rupture. Meticulous imaging, including selec-
tive catheter angiography and 3D imaging with CTA or
MRA, is necessary before surgery or endovascular treat-
ment. The endovascular options are embolization and ex-
clusion with a stent graft.
Renal Artery Aneurysm
The causes are arteriosclerosis, systemic vasculitis
(such as polyarteritis nodosa or lupus erythematosus),
fibromuscular disease, soft-tissue disorders, and trau-
ma. Arteriosclerotic and large aneurysms are usually
calcified. The risk of rupture and chronic embolization
are indications for treatment. Bypass surgery, coil em-
bolization, and stent graft implantation are the thera-
peutic options.
Abdominal Vascular Disease: Diagnosis and Therapy
Therapeutic Embolization of Gastrointestinal Bleeding
Bleedings in the upper gastrointestinal (GI) tract are
usually treated by endoscopic coagulation therapy.
However, if endoscopic therapy fails or the bleeding
source is inaccessible to endoscopy, angiographic em-
bolization is indicated. Patients after gastric resection
involving bilioenteric anastomoses and patients bleed-
ing from hepatobiliary and pancreatic pathologies
(pseudoaneurysms) should be treated primarily by an-
giographic techniques. Due to the extensive collateral
circulation in the upper abdomen, a detailed angio-
graphic evaluation followed by embolization of all feed-
ing arteries is required.
The most common causes of upper GI bleeding are:
• gastroduodenal ulcer;
• proximal jejunal tumors (endocrine carcinomas, angio-
fibroma, melanoma metastasis, arteriovenous malfor-
mation etc.);
• pseudoaneurysm of the gastroduodenal, pancreatico-
duodenal, intrapancreatic, and splenic arteries after
pancreatitis or trauma, bleeding through the Wirsung
pancreatic duct;
• pseudoaneurysm of the hepatic artery, bleeding
through the common bile duct (triad of hemobilia: ab-
dominal colic followed by jaundice and hematemesis
or melena).
In lower GI bleeding, diagnosis and treatment are
preferentially done by colonoscopy. However, in select-
ed cases embolization is required. For the primary diag-
nosis, a contrast-enhanced CT of the abdomen (400 mg
iodine/mL, flow 4 mL, volume 100 mL) in the arterial
and delayed phases usually demonstrates the area of
bleeding nicely. This helps to find the bleeding source
a b
Fig. 4 a-c. Hematuria and shock due to renal bleeding after nephrostomy. a CT showing large
perirenal hematoma and active bleeding. b Selective angiography demonstrates the bleeding
site. c Control angiography after selective coil embolization
159
during catheter angiography. Superselective catheteriza-
tion and bowel paralysis with Buscopan (Boehringer Ingel-
heim, Germany) are mandatory for a bleeding angiogram.
The causes of lower GI bleeding are:
• small bowel tumours (endocrine carcinomas, angio-
fibroma, melanoma metastasis, arteriovenous malfor-
mation, etc.);
• Meckels diverticulum;
• large bowel diverticulum bleeding (most common
cause in elderly patients);
• hemangiomatosis and arteriovenous malformation of
the colon;
• colon cancer (rectal bleeding due to hemorrhoidal dis-
ease has to be ruled out primarily).
GI bleedings are preferentially embolized by coils.
Treatment for Bleeding Complications after Surgery,
Trauma, and Post-partum
Bleeding in the abdomen may occur from iatrogenic
causes, particularly in the kidneys and the liver after per-
cutaneous interventions. It may also be due to trauma or
tumor. Frequent and typical locations and causes are re-
nal arteriovenous fistulas due to nephrostomy (Fig. 4) or
biopsy, laceration of the hepatic arteries by percutaneous
manipulations, psoas and pelvic bleeding due to traumat-
ic arterial injury, and uterine artery bleeding post-partum.
Temporary occlusion of the uterine artery can be a valid
alternative to emergency hysterectomy in patients with
intractable bleeding resulting from an atonic uterus. In
other anatomical locations, the type, source, and location
of the bleeding determine the method used to safely in-
terrupt the extravasations.
c
160
Therapeutic Embolization of Tumors
Tumor embolization techniques range from palliative
embolization in bleeding genitoureteral tumors, to
chemoembolization techniques in hypervascularized he-
patic neoplasms, in particular hepatocellular carcinoma
(HCC), to the definitive treatment of benign lesions such
as uterine fibroids. In the liver, tumor embolization is
mainly used in patients with inoperable HCC but pre-
served liver function (Child-Pugh A and B) and in those
with neuroendocrine tumor metastases. Classical treat-
ment is chemoembolization with doxorubicin mixed
with Lipiodol (Guerbet, France), sometimes in combina-
tion with a temporary blockade of the hepatic artery by
Gelfoam or other embolization particles. In randomized
trials, chemoembolization of unresectable HCC has
shown to be superior to supportive treatment only. New
techniques include doxorubicin-loaded particles as an
alternative embolization agent for HCC (Fig. 5). In the
Precision V trial, more than 200 patients were randomly
assigned to either transarterial chemoembolization
(TACE) with drug-eluting beads or to conventional
chemoembolization with Lipiodol and doxorubicin. The
former group had a higher rate of objective response and
disease control within 6 months of follow-up. In addition,
the use of drug-eluting beads resulted in a significantly
a b
Johannes Lammer
higher efficacy in high-risk patients (defined by Child-
Pugh B cirrhosis, ECOG 1 tumor symptoms, bilobar and
recurrent disease), a lower liver toxicity, but also systemic
doxorubicin-related toxicity. Regional chemotherapy
with irinorecan drug-eluting beads in patients with col-
orectal liver metastases is under evaluation. Intrahepatic
radiation by radioactive β-emitting particles directly in-
jected into the hepatic arteries is another treatment
modality for primary and metastatic liver tumors.
Venous Interventions
TIPS
Transjugular intrahepatic portocaval shunt (TIPS) to de-
pressurize the portal venous system was introduced into
clinical medicine at the end of the 1980s. Since then, a
standardized technique has been developed that allows
safe implantation of this artificial connection between the
portal and hepatic veins. The introduction of stent grafts
instead of bare stents has also led to improved patency of
the shunt tract.
TIPS is indicated in patients symptomatic from portal
hypertension with acute or chronic bleeding of esophageal
or gastric varices and in those with intractable ascites.
Fig. 5 a-d. Hepatocellular carcinoma before
and after transarterial chemoembolization
(TACE) with drug-eluting beads. a Contrast-
enhanced CT shows HCC in right hepatic
lobe. b Arteriography shows hypervascular
HCC. c Control arteriography after TACE of
HCC. d Contrast-enhanced CT shows com-
plete necrosis of the HCC
Abdominal Vascular Disease: Diagnosis and Therapy
Endoscopic techniques to treat varices are competitive
in bleeders whereas in some patients with ascites there
are few alternatives. Randomized trials, meta-analyses,
and Cochrane data review analyses have shown that TIPS
is superior to endoscopic therapy in the prevention of re-
bleeding and is superior to paracenteses to remove as-
cites. In patients with acute or subacute Budd-Chiari syn-
drome, TIPS can be a life-saving procedure and help to
overcome the acute phase, but the approach is burdened
by a relatively high re-thrombosis rate. The risks after
TIPS procedure are liver failure from shunted blood vol-
ume and encephalopathy.
Embolization of the Portal Veins
An intervention of increasing importance is pre-operative
embolization of the right or left portal vein in order to in-
duce hypertrophy of the contralateral hepatic lobe prior to
extended hemi-hepatectomy.
Suggested Reading
ASTRAL Investigators, Wheatley K, Ives N, Gray R et al (2001)
Revascularization versus medical therapy for renal artery
stenosis. N Engl J Med 361:1953-1962
Bax L, Woittiez AJ, Kouwenberg HJ et al (2009) Stent placement
in patients with atherosclerotic renal artery stenosis and im-
paired renal function: a randomized trial. Ann Intern Med
150:840-848
Blankensteijn JD, de Jong SE, Prinssen M et al (2005) Dutch Ran-
domized Endovascular Aneurysm Management (DREAM) Tri-
al Group. Two-year outcomes after conventional or endovas-
cular repair of abdominal aortic aneurysms. N Engl J Med
352:2398-2405
Blum U, Voshage G, Lammer J et al (1997) Endoluminal stent-
grafts for infrarenal abdominal aortic aneurysms. N Engl J
Med 336:13-20
Covey AM, Tuorto S, Brody LA et al (2005) Safety and efficacy
of preoperative portal vein embolization with polyvinyl alco-
hol in 58 patients with liver metastases. AJR Am J Roentgenol
185:1620-1626
D’Amico G, Luca A, Morabito A et al (2005) Uncovered trans-
jugular intrahepatic portosystemic shunt for refractory ascites:
a metaanalysis. Gastroenterology 129:1282-1293
EVAR trial participants (2005) Endovascular aneurysm repair versus
open repair in patients with abdominal aortic aneurysm (EVAR
trial 1): randomised controlled trial. Lancet 365:2179-2186
Greenhalgh RM, Brown LC, Kwong GP et al, EVAR trial partici-
pants (2004) Comparison of endovascular aneurysm repair
with open repair in patients with abdominal aortic aneurysm
(EVAR trial 1), 30-day operative mortality results: randomised
controlled trial. Lancet 364:843-848
Kaatee R, Beek FJ, de Lange EE et al (1997) Renal artery steno-
sis: detection and quantification with spiral CT angiography
versus optimized digital subtraction angiography. Radiology
205:121-127
161
Kasirajan K, O’Hara PJ, Gray BH et al (2001) Chronic mesenteric
ischemia: open surgery versus percutaneous angioplasty and
stenting. J Vasc Surg 33:63-71
Khan S, Tudur Smith C, Williamson P, Sutton R (2005) Portosys-
temic shunts versus endoscopic therapy for variceal rebleeding
in patients with cirrhosis. Cochrane Database Syst Rev
18:CD000553
Khuroo MS, Al-Suhabani H, Al-Sebayel M et al (2005) Budd-
Chiari syndrome: long-term effect on outcome with transjugu-
lar intrahepatic portosystemic shunt. J Gastroenterol Hepatol
20:1494-1502
Lammer J, Malagari K, Vogl T et al, on behalf of the PRECISION
V investigators (2009) Prospective randomized study of doxo-
rubicin-eluting-bead embolization in the treatment of HCC:
results of the PRECISION V study. Cardiovasc Intervent
Radiol 12 [Epub ahead of print]
Leertouwer TC, Gussenhoven EJ, Bosch JL et al (2000) Stent
placement for renal artery stenosis: where do we stand? A
metaanalysis. Radiology 216:78-85
Mann SJ, Pickering TG (1992) Detection of renovascular hyper-
tension: state of the art 1992. Ann Intern Med 117:845-853
Perler AB, Becker GJ (1998) Vascular intervention – a clinical ap-
proach. Visceral vascular disease. Thieme, New York, Stuttgart,
pp 517-637
Prinssen M, Verhoeven EL, Buth J et al (2004) Dutch Randomized
Endovascular Aneurysm Management (DREAM) Trial Group.
A randomized trial comparing conventional and endovascular
repair of abdominal aortic aneurysms. N Engl J Med
351:1607-1618
Rose SC, Quigley TM, Raker EJ (1995) Revascularization for
chronic mesenteric ischemia: comparison of operative arterial
bypass grafting and percutaneous transluminal angioplasty. J
Vasc Interv Radiol 6:339-349
Saab S, Nieto JM, Lewis SK, Runyon BA (2006) TIPS versus para-
centesis for cirrhotic patients with refractory ascites. Cochrane
Database Syst Rev 18:CD004889
Salerno F, Cammà C, Enea M et al (2007) TIPS for refractory as-
cites: a meta-analysis of individual patient data. Gastroen-
terology 133:825-834
Soulez G, Oliva VL, Turpin S et al (2000) Imaging of renovascu-
lar hypertension: respective values of renal scintigraphy, renal
Doppler US, and MR angiography. Radiographics 20:1355-
1368
van de Ven PJ, Kaatee R, Beutler JJ et al (1999) Arterial stenting
and balloon angioplasty in ostial arteriosclerotic renovascular
disease: a randomized trial. Lancet 353:282-286
Van Jaarsveld BC, Krijen P, Pieterman H et al (2000) The effect of
balloon angioplasty on hypertension in atherosclerotic renal-
artery stenosis. Dutch Renal Artery Stenosis Intervention
Cooperative Study Group. N Engl J Med 342:1007-1014
Webster J, Marshall F, Abdalla M et al (1998) Randomized com-
parison of percutaneous angioplasty vs continued medical
therapy for hypertensive patients with atheromatous renal
artery stenosis. Scottish and Newcastle Renal Artery Stenosis
Collaborative Group. J Hum Hypertens 12:329-335
Williams DM, Lee DY (1997) Dissected aorta, parts I-III. Radiol-
ogy 203:23-44
Zheng M, Chen Y, Bai J et al (2008) TIPS vs. endoscopic thera-
py in the secondary prophylaxis of variceal rebleeding in cir-
rhotic patients: meta-analysis update. J Clin Gastroenterol
42:507-516

Non-vascular Abdominal Disease: Diagnosis and Therapy
Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi
Division of Diagnostic and Interventional Radiology, University of Pisa, Pisa, Italy
Introduction
Recent technological advances have given rise to a wide
range of diagnostic and interventional imaging-guided pro-
cedures for an increasing number of abdominal parenchy-
mal diseases, such as those involving the kidneys, adren-
als, or pancreas. However, the liver, and the cirrhotic liver
in particular, still represents the main field of application
of abdominal imaging modalities. Imaging plays a key role
in the diagnostic work-up of the nodular lesion, especially
in its initial detection and characterization. In addition, the
results of imaging assessment guide the therapeutic op-
tions (surgical, interventional, or palliative) and provide the
standard of reference in the evaluation of tumor response.
Hepatocellular Carcinoma: Pathology
The diagnosis of hepatocellular carcinoma (HCC) is
based on imaging examinations in combination with clin-
ical and laboratory findings. However, imaging of the cir-
rhotic liver remains challenging since regenerative nod-
ules and pre-neoplastic hepatocellular lesions, such as
dysplastic nodules, can mimic small HCCs. New imaging
technologies have improved investigations into the multi-
step process that takes place during carcinogenesis, from
regeneration towards dysplasia and full malignancy.
The most important pathological alteration in the de-
velopment of HCC is the derangement in the liver’s vascu-
lar supply due to progressive capillarization of the sinu-
soids, associated with the formation of an increasing
number of unpaired arterioles. In addition, there are pro-
gressive, intracellular histological changes, including loss
of the biliary polarization of hepatocytes, a derangement
of their microscopic secretory structure, and the progres-
sive depletion of Kupffer cells within nodular lesions [1].
Due to the wide spectrum of morphological changes
seen in cirrhotic parenchyma, all cross-sectional imaging
modalities should be applied in the detection and charac-
terization of nodular lesions. Ultrasound (US) is the
modality most commonly applied in the surveillance of
cirrhotic patients and is thus considered as the first-line
approach in their diagnostic workup. The high spatial
IDKD 2010-2013
resolution of US enables the demonstration even of very
small lesions (<1 cm in maximum diameter). In nodules
>1 cm in maximum diameter, the vascular supply should
be assessed as well. Moreover, the introduction of mi-
crobubble contrast agents and the development of con-
trast-specific scanning techniques have expanded the
capabilities of US in examinations of the cirrhotic liver.
The advent of second-generation contrast agents and low-
mechanical-index real-time scanning techniques has been
fundamental in improving the ease and reproducibility of
US examinations [2]. Consequently, contrast-enhanced
US has been introduced into the diagnostic flow chart as
one of the imaging modalities able to demonstrate the
vascular pattern of nodules >1 cm in diameter [3].
Dynamic multidetector computed tomography (MDCT)
and magnetic resonance imaging (MRI) are further imag-
ing modalities used to detect and characterize nodular
vascular changes. They provide a detailed view of the
hepatic parenchyma and allow a confident diagnosis of
neoplasm. Importantly, these imaging evaluations are a
fundamental prerequisite for guiding the therapeutic
approach. The added value of MRI over other cross-
sectional imaging modalities is its ability to assess the
components of a hepatic lesion in baseline image acqui-
sitions. The use of tissue-specific MRI contrast media
(hepatobiliary and reticuloendothelial agents) can reveal
metabolic disorders that occur as a result of cellular
dedifferentiation. Current cross-sectional techniques
demonstrate the pathological changes associated with
carcinogenesis, such that needle biopsy is no longer
considered as the standard reference procedure for the
definitive diagnosis of HCC [3].
Pre-neoplastic Lesions: Regeneration
Regeneration is the result of the architectural rearrangement
of the hepatic parenchyma in response to chronic injury.
Generally, regenerative nodules are small (around 1 cm) but
in some cases may have a maximum diameter v3 cm.
At US examination, a typical regenerative nodule may
appear as a hypoechoic or hyperechoic small nodule that
enhances relative to background because of its thin
hyperechoic rim, which represents perinodular fibrous
Non-vascular Abdominal Disease: Diagnosis and Therapy 163

Arterial Late

a b

Fig. 1 a-c. a Baseline US examination shows the presence of a large (up to 3 cm) hyperechoic
nodule of the VI segment, partially exophytic, in a background of cirrhosis (arrow). b At
dynamic MRI study, the nodule does not show signs of hypervascularization (appearing as
hypointense on arterial phase); instead, the portal supply is still well evident, while the
nodule appears slightly isointense during the late phase. c On this baseline T2-weighted
image, the nodule appears as hypointense, due to its intranodular iron content, which
increases signal intensity compared to the surrounding parenchyma (arrow) c

tissue and which may mimic a pseudocapsule. Generally,
no further examination is required. In cases of larger nod-
ules, dynamic evaluation is required to exclude neoplas-
tic dedifferentiation (Fig. 1 a).
At dynamic contrast studies (contrast-enahnced US,
MDCT, or MRI), there is no evidence of a vascular sup-
ply different than that of the surrounding parenchyma,
due to the conspicuous and predominant feeding of these
nodules by portal vessels (Fig. 1 b) [4]. MRI can depict
some peculiarities that may differentiate regenerative
nodules from other focal lesions. On baseline examina-
tion, regenerative nodules are isointense on T1- and T2-
weighted images due to the presence of normal liver
cells. The frequent intranodular content of iron may de-
crease the relaxation time on T2, thus reducing the signal
intensity of the nodules on these sequences (Fig. 1 c).
Nodular signal intensity after the administration of
tissue-specific contrast agent (hepatobiliary or reticulo-
endothelial) is not modified due to the preserved meta-
bolic activity of hepatocytes and Kupffer cells.
Pre-neoplastic Lesions: Dysplasia
Dysplastic nodules are classified as low-grade (LGDN)
and high grade (HGDN), depending on the degree of
cellular atypia and on changes in architectural structure.
HGDN are considered to be pre-malignant lesions, and
the subsequent development of HCC from a HGDN
within a period of years or even a few months has been
documented [5]. Plain US and CT examinations are gen-
erally useless in the characterization of dysplastic nod-
ules, due to the lack of specific imaging finding (Fig. 2 a).
At dynamic study, the post-contrast enhancement pat-
terns of HGDN may be highly variable. In fact, even if
the main blood supply is still provided by branches of the
portal vein, the presence of a vascular supply on arterial
phase may be seen due to the development of sporadic
unpaired arteries. However, this finding is not associat-
ed with wash-out, which instead represents the diagnos-
tic finding for HCC (Fig. 2 b). Thus, in addition to dis-
playing the degree of vascular supply, MRI demonstrates
parenchymal alterations, thereby playing an important
role in the differential diagnosis between pre-neoplastic
and neoplastic lesions. In fact, at baseline MRI exami-
nation, dysplastic nodules typically show a characteristic
hyperintensity on T1-weighted sequences due to the in-
tranodular presence of glycogen or lipids, while on T2-
weighted sequences they may be slightly hyperintense,
isointense, or even hypointense.
In the hepatobiliary phase, after the administration of
hepatospecific contrast agents, DNs are generally iso-
intense or hyperintense compared to the surrounding liver
164 Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi

Arterial Late

a b

Fig. 2 a-c. a At baseline US, it is possible to appreciate the presence of a suspected isoechoic nodule
surrounded by a thin hypoechoic rim. b The US post-contrast examination does not show any arte-
rial supply within the nodule, which remains hypoechoic in all post-contrast acquisitions. c On he-
patobiliary phase at MRI examination, the nodule is isointense with the surrounding parenchyma,
c due to the preserved biliary function

parenchyma, reflecting the maintenance of biliary func-
tion but also cholestasis, which occurs in so-called green
nodules (Fig. 2 c).
Sometimes, the progression from dysplasia to HCC is
detected in a very early phase, when it is possible to iden-
tify a focus of HCC within a pre-malignant lesion, known
as “a nodule within a nodule”. On T2-weighted images,
the typical appearance is a focus of high signal intensity
located within a low-signal-intensity nodule and also
showing the post-contrast signal behavior of HCC.
Hepatocellular Carcinoma: Imaging Findings
As noted above, one of the key pathological features di-
agnostic of HCC is the vascular supply of the nodule. The
progression from regeneration to overt HCC is character-
ized by neoangiogenesis, that is, the concomitant devel-
opment of feeding arteries and efficient arteriovenous
shunts. This pathological blood supply is well demon-
strated on contrast-enhanced dynamic studies by the typ-
ical findings of wash-in during the arterial phase and sub-
sequent wash-out (Fig. 3 a). However, a typical vascular
behavior may not be present at dynamic imaging. In these
cases, MRI both at baseline and following the adminis-
tration of hepatospecific contrast media may lead to a de-
finitive diagnosis of HCC.
In addition, early or moderately differentiated HCC
may show peculiar signal intensity on T1- and T2-
weighted baseline acquisitions (Fig. 3 b). On baseline
T1-weighted images, HCC usually appears as a hypo-
intense nodule because of its increased cellularity, and
thus its higher amount of intracellular water; however,
small, well-differentiated HCCs may show different
signal intensities, appearing as either hyperintense or
isointense. Hyperintensity, as in the case of HGDN, may
be related to the intracellular presence of glycogen and
fat, which accumulate because of the loss of normal
cellular metabolic activity. On baseline T2-weighted
images, HCC usually shows mild signal hyperintensity,
while small and well-differentiated tumors may be
isointense to the surrounding parenchyma. In a compar-
ison of histological data obtained from explanted
cirrhotic livers with the MRI signal intensity of corre-
sponding lesions, a relationship was found between
lesion malignancy and nodular intensity on T2-weighted
images [6]. It also has been shown that nodular signal
intensity on T2-weighted images is significantly associ-
ated with the intranodular blood supply; in fact, signal
intensity increases as the intranodular portal venous
blood supply decreases [7].
Although their application has not yet been introduced
into diagnostic guidelines, the use of tissue-specific con-
trast medium may give additional information, such
as the atypical baseline or vascular pattern at dynamic
study, as in the case of borderline lesions (dysplastic
nodules) or well-differentiated HCCs. Regarding hepato-
biliary contrast agents, the lack of contrast uptake is
strongly related to overt HCC, due to the loss of normal
metabolic function whereas the uptake is preserved in
early HCCs, resembling that of HGDNs [8]. In daily
practice, the advantages of the most recent generation of
MRI contrast media can be exploited, as they illustrate
the nodule’s characteristic vascular and hepatospecific
phases, i.e., neoangiogenesis and lack of hepatobiliary
function, respectively, which allow a highly confident
diagnosis of HCC (Fig. 3 c). Specifically, reticulo-
endothelial system (RES) agents allow the carcinogenetic
Non-vascular Abdominal Disease: Diagnosis and Therapy
Arterial Late
T1 w.i
b c
pathway to be followed; for example, the progressive in-
crease in sinusoid capillarization provides a hostile envi-
ronment for reticuloendothelial cells, such that their pro-
gressive loss explains the very high signal intensity of
HCC [9]. Nonetheless, due to the less consistent vascu-
lar phases, the application of RES agents is strongly lim-
ited and has largely been discontinued.
Therapy and Follow-up
Nowadays, the therapeutic approach to HCC is based on
surgical (transplantation and resection) and non-surgical
approaches, mini-invasive modalities (percutaneous and
intra-arterial therapies), and palliative approaches. The
decision is based upon clinical and functional data as well
as on the imaging findings, i.e., number of lesions and
their size, location, degree of vascularization, and rela-
tionships with vascular and biliary structures. For exam-
ple, a candidate for liver transplantation should fulfill
imaging criteria, which include the presence of a single
lesion measuring <5 cm or of up to three lesions, each
with a greatest dimension f3 cm. Moreover, post-
processing of native images is fundamental in candidates
for surgical or intra-arterial therapies, in order to provide
an overall representation of the vascular anatomy or of
the feeding vessels of the lesions.
Imaging modalities are also very important as guid-
ance for percutaneous ablation. In these cases, US repre-
sents the most appropriate technique, as its real-time
165
Fig. 3 a-c. a MR dynamic examination reveals
a nodule of the VIII segment that is hyper-
vascular on arterial phase and shows clear
cut washout in the late phase. b At baseline
MRI examination, the nodules has a typical
signal intensity, i.e., hypointense on T1-
weighted sequences and hyperintense on T2-
weighted sequences. c On hepatobiliary
phase, the nodule is hypointense to the sur-
rounding parenchyma, due to the complete
loss of biliary function
T2 w.i
visualization of the target allows continuous assessment
of the ongoing changes that occur during the ablation
procedure. In addition, the inclusion of periprocedural
contrast enhancement provides immediate evaluation of
the presence of residual viable tumor, which may be re-
treated during the same ablation session.
Contrast-enhanced CT and MRI play a major role in
follow-up, permitting assessment of the tumor’s response
in terms of necrosis or relapse as well as the detection
of new lesions in the surrounding parenchyma [3].
MRI, when performed with hepatospecific contrast
agents, can provide additional information about post-
ablation tissue components. This may be useful in ques-
tionable cases, in which periablation hyperemia (espe-
cially after radiofrequency ablation) or arteriovenous
shunts/thrombosis (especially after ethanol ablation)
must be differentiated from tumoral persistence or re-
currence [10].
Conclusions
In conclusion, HCC in a cirrhotic liver represents one of
the most important fields of application of imaging
modalities, based on their key role in the detection and
characterization of nodular lesions. Moreover, any thera-
peutic decision is strongly related to the imaging results,
as they also guide the use of mini-invasive procedures
and allow evaluation of therapeutic success.
166
References
1. International Consensus Group for Hepatocellular Neoplasia.
The International Consensus Group for Hepatocellular Neo-
plasia (2009) Pathologic diagnosis of early hepatocellular car-
cinoma: a report of the international consensus group for he-
patocellular neoplasia. Hepatology 49:658-664
2. Cosgrove D (2006) Ultrasound contrast agents. An overview.
Eur J Radiol 60:324-330
3. Bruix J, Sherman M (2005) Practice Guidelines Committee,
American Association for the Study of Liver Diseases. Manage-
ment of hepatocellular carcinoma. Hepatology 42:1208-1236
4. Roncalli M, Roz E, Coggi G et al (1999) The vascular profile
of regenerative and dysplastic nodules of the cirrhotic liver:
implications for diagnosis and classification. Hepatology
30:1174-1178
5. Theise, Park YN, Kojiro M (2002) Dysplastic nodules and he-
patocarcinogenesis. Clin Liver Dis 6:497-512
Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi
6. Bartolozzi C, Cioni D, Donati F et al (2001) Focal liver
lesions: MR imaging-pathologic correlation. Eur Radiol 11:
1374-1388
7. Shinmura R, Matsui O, Kobayashi S et al (2006) Cirrhotic
nodules: association between MR imaging signal intensity and
intranodular blood supply. Radiology 237:512-519
8. Bartolozzi C, Crocetti L, Lencioni R et al (2007) Biliary and
reticuloendhotelial impairment in hepatocarcinogenesis: the
diagnostic role of tissue specific MR contrast media. Eur
Radiol 17:2519-2530
9. Chen RC, Lii JM, Chou CT et al (2008) T2-weighted and T1-
weighted dynamic superparamagnetic iron oxide (ferucarbo-
tran) enhanced MRI of hepatocellular carcinoma and hyper-
plastic nodules. J Formos Med Assoc 107:798-805
10. Cioni D, Lencioni R, Bartolozzi C. (2001) Percutaneous abla-
tion of liver malignancies: imaging evaluation of treatment re-
sponse. Eur J Ultrasound 13:73-93
 IDKD 2010-2013

An Approach to Imaging the Acute Abdomen in the Pediatric Population

Alan Daneman1, Simon G. Robben2
1 Department of Radiology, University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada
2 Department of Radiology, Maastricht University Medical Centre, Maastricht, The Netherlands

Introduction
The acute abdomen is a common and often challenging
emergency in the pediatric population. This chapter pro-
vides an approach to the imaging evaluation of children,
highlighting briefly the more common causes of abdom-
inal pain that may require surgery.

Causes of Acute Abdomen

There is a wide spectrum of pathologies that may give rise
to acute abdominal pain. These include congenital and ac-
quired lesions that may present in the immediate neonatal
period or in older infants and children. In neonates, the
acute abdomen is not an uncommon event in infants in
neonatal intensive care units and may be the result of sev-
eral causes, with congenital bowel obstruction, complica-
tions of mid-gut malrotation, and necrotizing enterocolitis
being the three most significant conditions. In older in-
fants and children, the common causes of acute abdomi- Fig. 1. A boy with right lower quadrant pain and fever. The abdom-
nal pain that may require surgery include acute appen- inal radiograph shows a triangular consolidation (arrows) in a basal
dicitis, complications of malrotation, intussusception, and segment of the right lower lobe: basal pneumonia. Dilated air-filled
Meckel diverticulum. Inflammatory bowel diseases may ascending colon was considered secondary to referred pain. Un-
eventful recovery after antibiotic therapy
also present with acute abdominal symptomatology.
Although the common, above-mentioned causes of
acute abdominal pain are due to lesions involving the
gastrointestinal tract, the acute abdomen may also be due cause abdominal pain out of proportion to the degree of
to abnormalities of other viscera, e.g., gynecological ab- trauma. Imaging in these children may reveal an underly-
normalities such as ovarian torsion, omental lesions such ing abnormality such as a neoplasm or an anomaly such
as omental infarcts, obstructions of the biliary and uri- as uretero-pelvic obstruction. Furthermore, one should
nary tracts due to stones and inflammatory processes consider non-accidental injury or abuse when certain
such as pancreatitis or hepatitis. traumatic lesions are found, especially when they appear
It has to be emphasized that abdominal pain may also in combination with hematomas of the left lobe of the liv-
be due to referred pain from extra-abdominal patholo- er and duodenum and pancreatitis.
gies, such as pneumonia or pleural effusion (Fig. 1).
Medical diseases, e.g., sickle cell disease, Henoch-
Schonlein purpura, and hemolytic uremic syndrome, may Modalities
also be the cause of significant acute abdominal pain.
Imaging of the child with acute abdominal pain should Ultrasonography (US) has come to play an ever-increasing
include a search for evidence of these other pathologies. role in the management of children with acute abdominal
Abdominal trauma may also be the cause of an acute pain and has replaced the plain abdominal radiograph as
abdomen. Occasionally, mild abdominal trauma may the modality of initial choice in many clinical situations.
168
The major advantages of US are that it does not use ion-
izing radiation, it is relatively inexpensive, and the ab-
dominal viscera, including the bowel, are well delineated
in children. Therefore, many pathological entities can be
easily confirmed or excluded.
Plain abdominal radiograph (AXR) remains a standard
method for evaluation of the acute abdomen in some clin-
ical situations. It is essential when peritonitis is present
and perforation is suspected. All neonates with an acute
abdomen are evaluated with AXR. This modality is es-
sential for the detection of conditions such as necrotizing
enterocolitis and congenital bowel obstruction. In the for-
mer, AXR may be diagnostic; in the latter, the findings
guide the choice of subsequent contrast examinations of
the gastrointestinal tract. Views with a horizontal beam
are essential to exclude the presence of free air due to
bowel perforation and can be performed with the neonate
in the dorsal or lateral decubitus position. In older chil-
dren, the diagnosis of intestinal obstruction can often be
made based on the supine film alone. A search for air-
fluid levels on the upright view does not always add extra
information and a search for free air in the abdomen is
often more easily achieved with a lower radiation dose
and a single upright view of the chest, which will also
serve to exclude lung pathology. However, in both the
neonate and the older child, AXR findings are often non-
specific, which limits the role of this modality.
Contrast studies of the gastrointestinal (GI) tract are
essential in certain conditions, such as suspected mid-gut
malrotation and congenital bowel obstruction. In the lat-
ter situation, contrast enema may be important in some
patients for diagnosis and in others for therapy as well.
Computed tomography (CT) may be reserved for more
complicated imaging situations, when US may not provide
a b
a b
Alan Daneman, Simon G. Robben
all the information required, e.g., appendicitis when
gas obscures the right lower quadrant, or in older, obese
children and those children with abscesses. CT without
contrast injection is also extremely helpful in delineating
urinary stones when these are not well shown by US.
Magnetic resonance imaging (MRI) has a very limited
role in children with acute abdominal pain. However,
it can depict the anatomy exceptionally well in certain
conditions, in which case it may be used to complement
findings on US. These include biliary and pancreatic duct
anomalies and gynecological disorders, such as complex
anomalies associated with hydrocolpos.
Acute Appendicitis
Acute appendicitis is a common clinical entity in pediatrics.
In many patients, it is easy to make the diagnosis clinically
with certainty and no imaging is required prior to appen-
dicectomy. However, imaging is extremely important in
those children with non-specific symptoms or signs of
acute appendicitis. In such cases, we have used US as the
modality of initial choice, reserving CT for those patients
in whom the US examination is inconclusive or when
abscesses are present in order to better define their extent
prior to drainage by the interventional radiology team.
The diagnosis of appendicitis is made on US when the
appendix is >6 mm in diameter and is non-compressible.
These features should not be considered absolute and
others should be taken into account, including edema of
the mesentery, hyperemia of the wall of the appendix on
color or power Doppler examination, the presence of an
appendicolith and local fluid collections, or abscess for-
mation (Figs. 2, 3). There are other conditions that may
Fig. 2 a, b. Abortive appendicitis in a 9-year-
old boy who had local peritoneal tenderness
for one day, exactly at the site of the appen-
dix, and normal temperature. a US shows
borderline appendix diameter (6 mm) and
slight edema of the apendiceal mesentery
(echogenic triangle, arrow) but no surround-
ing edema. b Color Doppler US demonstrat-
ed marked hyperemia. Complaints subsided
within 2 days without therapy
Fig. 3 a, b. Rapidly progressing appendicitis in
an 8-year-old boy who had local peritoneal
tenderness exactly at the site of the appendix
for 1 day and low grade fever. a US shows
thickened appendix (8 mm) with edema of
the appendicular mesentery (echogenic trian-
gle, arrow) and extensive surrounding ede-
ma. b Color Doppler US demonstrated mod-
erate hyperemia. At surgery, an inflamed ap-
pendix was removed
An Approach to Imaging the Acute Abdomen in the Pediatric Population 169

cause the appendix to become thick-walled and dilated; intussusception promptly and accurately. The diagnosis
these include cystic fibrosis, Henoch-Schonlein purpura, can be made by US, AXR, or contrast studies of the colon.
and inflammatory bowel diseases. Ultrasonography has been shown in many series to be
100% accurate in depicting the presence or absence of
the common types of ileocolic or ileo-ileocolic intussus-
Malrotation ceptions in children. These lesions have a characteristic
sonographic appearance and are usually found just under
See the chapter “Malrotation: Techniques, Spectrum of the abdominal wall, most commonly on the right side of
Appearance, Pitfalls, and Management”, in the “Kangaroo” the abdomen (Fig. 4). Since it is a non-invasive procedure
section of this syllabus for a review of this important and because of its accuracy, sonography is the modality
cause of acute abdomen in the pediatric population. of choice for the evaluation of patients suspected of hav-
ing an intussusception. The sonographic appearance of
intussusception was excellently reviewed in a 1996 arti-
Intussusception cle by del-Pozo et al. (see “Suggested Reading”).
Some of the characteristic signs of an intussusception
Intussusception is not an uncommon phenomenon in can be seen on AXR, including the meniscus sign, target
children and is one of the commoner causes of the acute sign, and, less commonly, a soft-tissue mass (Fig. 5).
abdomen in those between 6 months and 5 years of age.
The vast majority of intussusceptions arise in the ileum
and are either ileocolic or ileo-ileocolic. They are thought
to occur because of hyperplasia of the lymphoid tissue
in the ileum, possibly as a result of a viral infection.
There are other types of intussusceptions that may re-
late to pathological lead points or gastro-jejunostomy
tubes, or that are seen in the post-operative period. These
are discussed separately at the end of this chapter. An-
other type of intussusception is the benign small bowel
intussusception. This is often an incidental finding and
does not present as an acute abdominal emergency. This
entity is discussed separately in the chapter “Pediatric
Intestinal Ultrasonography”, in the “Kangaroo” section of
this syllabus.

Ileocolic and Ileo-ileocolic Intussusceptions

Diagnosis

Children presenting with ileocolic or ileo-ileocolic intus-

susception commonly do not present with the classical
clinical triad of abdominal pain, red currant jelly stool,
and a palpable abdominal mass. Instead, the presentation
may be non-specific. For this reason, the clinician often
has to rely on imaging procedures to diagnose or exclude
Fig. 5. Child with ileocolic intussusception that extended into the
transverse colon. Abdominal radiograph shows a gasless right up-
per abdomen due to the presence of the intussusception on the
right. The arrows indicate a soft-tissue mass that represents the in-
tussusception, which is in the transverse colon

a b

Fig. 4 a, b. Sonograms of children with ileo-

colic intussusception. a Transverse scan
through an intussusception shows the typical
target sign, which can be easily detected just
deep to the abdominal wall anteriorly. The
characteristic of this target sign is multiple
concentric rings representing alternating mu-
cosal and muscular layers. b Transverse scan
through an intussusception shows the pres-
ence of a lymph node (arrow) within the
mesentery which has been drawn into the
intussusception between the layers of the
intussusceptum
170
However, in our institutions, we have relied on the plain
radiograph only in those instances in which there is a
clinical consideration of peritonitis. In this clinical set-
ting, AXR is essential to exclude perforation, which is the
major contraindication for attempted enema reduction.
Contrast enema represents a more invasive diagnostic
procedure, requiring radiation. Moreover, in a study by
Daneman and Navarro, in 2003, in which patients were
administered contrast enema for diagnosis, 50% of the
enemas were negative for intussusceptions (see “Suggest-
ed Reading”). Using sonography as the diagnostic modal-
ity enables one to avoid performing unnecessary contrast
enemas in those patients without intussusception.
Reduction
According to the many series in the recent literature, the
reduction rate of intussusception ranges from 80% to as
high as 95%. These series have used either fluoroscopic
or sonographic guidance for reducing the intussusception
and either hydrostatic (barium, water-soluble contrast,
saline) or pneumatic reduction. The fact that different
techniques have been used with similar success rates sug-
gests that it is not important which technique is used.
Non-operative reduction of an intussusception should on-
ly be attempted after the surgical team has evaluated the
patient and the patient is clinically stable, well-hydrated,
has no evidence of peritonitis, and has an intravenous line
in place. The major contraindications to administering the
enema are the clinical findings of peritonitis or shock or
signs of perforation on an abdominal radiograph.
In order to improve the reduction rate, delayed, repeat-
ed reduction attempts can be used as long as the intussus-
ception moves in response to the initial attempted reduc-
tion and the child becomes asymptomatic and maintains
stable vital signs. It has been shown that this approach is
safe and effective, with a good success rate. Navarro et al.,
in a study published in 2004 (see “Suggested Reading”),
used this approach in approximately 15% of patients with
intussusceptions, achieving successful reduction in 50%
of those intussusceptions not reduced on the first attempt.
There does not appear to be a fixed optimal timing be-
tween attempts, and delayed second or third attempts can
be made several hours after the first.
Pathological Lead Points
Pathological lead points are found in about 5-7% of all
intussusceptions. The commonest are Meckel diverticu-
lum, polyps, Henoch-Schonlein purpura, and cystic fi-
brosis. Less common causes are lymphoma, duplication
cyst, and various inflammatory lesions of the bowel.
Management of these patients remains a challenge.
Contrast or air enema techniques are not always diag-
nostic in documenting the presence of a pathological lead
point. Sonography is extremely useful in this regard as it
may depict two-thirds of pathological lead points, providing
a specific diagnosis in one-third of these cases (Fig. 6).
Alan Daneman, Simon G. Robben
Fig. 6. A 3-month-old child with an intussusception due to a dupli-
cation cyst. In the transverse scan, the duplication cyst (C) is eas-
ily identified as a pathological lead point
However, it remains a diagnostic challenge as to how to
search for pathological lead points in those patients in
whom there is a high index of suspicion for this lesion and
in whom the sonogram is negative. In such cases, the choice
of which other imaging modalities to use will depend on the
clinical situation in each particular patient. We recommend
attempted enema reduction in all patients with a lead point
if there is no contraindication to non-operative reduction.
Postoperative Intussusception and Intussusception with
Gastroenterostomy Tubes
Intussusception may occur as a complication in <1% of
laparotomies. These are usually more difficult to diag-
nose on sonography than the usual ileocolic intussuscep-
tions, as they are small bowel intussusceptions that are
often surrounded by large, dilated loops of obstructed
bowel. They frequently require surgical reduction.
Intussusceptions may complicate the presence of a
gastrojejunostomy (GJ) tube. Most of the patients pre-
senting with this complication are usually clinically sta-
ble and do not require urgent reduction of the intussus-
ception. The majority can be managed by replacing the
tube with a standard or a shortened GJ tube or with a gas-
trostomy tube. However, manipulation of the GJ tube and
flushing with air or saline may also be helpful. Surgery is
rarely required for reduction.
Necrotizing Enterocolitis
Necrotizing enterocolitis (NEC) usually presents in
infants in the neonatal intensive care unit, and more
commonly in premature neonates. The classical presenta-
tion includes abdominal distention and blood in the stool.
The radiologist plays an important role at the time of
An Approach to Imaging the Acute Abdomen in the Pediatric Population
a b
Fig. 7 a, b. a Necrotizing enterocolitis in a pre-
mature infant. Bowel loops show distention
with gas and thickened bowel walls. Large
bubbles of intramural air are seen at the de-
scending colon; a bubbly appearance in the
right flank also suggests the presence of in-
tramural air in the ascending colon and distal
ileal loops. b Detail shows the intramural gas
(arrow) to better advantage. Intramural gas is
most commonly seen in newborns with
necrotizing enterocolitis
diagnosis of NEC, during follow-up, and in the detection
of later complications such as strictures.
At the time of diagnosis, there are three abnormalities
that may be seen on AXR: bowel dilatation, intramural
gas, and portal venous gas. Bowel dilatation is present in
almost 100% of the patients with NEC and the degree of
distention of the bowel usually correlates well with the
clinical severity. Follow-up AXR may show asymmetrical
dilatation and fixed loops in those infants who deterio-
rate. Intramural gas is not present in 100% of patients and
the amount of intramural gas does not always correlate
well with the degree of clinical severity (Fig. 7). Portal
venous gas usually is present in those with severe NEC.
Disappearance of the intramural gas and portal venous
gas does not necessarily correlate with clinical improve-
ment, as in either case the gas will eventually disappear
even in those children who deteriorate clinically.
Ultrasound is an extremely useful modality in the in-
vestigation of patients with NEC as it can provide infor-
mation regarding the presence of intraperitoneal fluid,
bowel wall thickness, and bowel perfusion (using color or
power Doppler sonography). US is much more accurate
than AXR in documenting the presence of free and focal
intra-peritoneal fluid and can also define the character of
this fluid. It is well known that not all patients with NEC
will show free air on AXR following perforation; instead,
they may only present with free fluid.
In the early phases of NEC, we have found that the bow-
el wall will be quite thickened; in patients who are more
severely affected, however, the mucosa and submucosa of
the bowel sloughs into the lumen of the bowel, leaving a
markedly thinned bowel wall, which is much more prone
to perforation. Thinning of the bowel wall can be docu-
mented with sonography. In NEC, the bowel (particularly,
the thickened bowel) becomes markedly hyperemic, which
indicates the presence of viable bowel. However, the ab-
sence of bowel perfusion in single or multiple loops of
171
bowel (especially when the bowel wall is thinned) indicates
the presence of necrosis – a condition that may warrant
surgical intervention even if free air is not seen on AXR.
Sonography comes to play a more major role in the
follow-up of both those patients who are not responding
to medical management and those who deteriorate clini-
cally. In such cases, US may provide information that is
not depicted with AXR.
Meckel Diverticulum
Meckel diverticulum most commonly presents as pain-
less rectal bleeding due to ulceration because of the pres-
ence of ectopic gastric mucosa. These patients are usual-
ly adequately diagnosed and managed following a ra-
dionuclide scan.
In <50% of the children presenting with Meckel di-
verticulum, the clinical findings will be more complex,
with a combination of abdominal pain, vomiting, and, oc-
casionally, rectal bleeding. In these children with acute
pain, the diagnosis is often difficult and non-specific. US
can be used successfully to document the presence of an
inflamed or hemorrhagic Meckel diverticulum, which in
this situation has a variable appearance and may simulate
the presence of an inflamed duplication cyst, appendici-
tis, or sometimes a small intussusception. The finding of
this somewhat atypical appearance on US is diagnosti-
cally suggestive of a complicated Meckel diverticulum
rather than the other pathologies it simulates.
Congenital Bowel Obstruction in the Neonate
Obstruction due to congenital lesions may occur at all
levels of the GI tract and are, from a practical point of
view, usefully divided into those lesions that are termed
172
Fig. 8. High gastrointestinal obstruction in a newborn with trisomy
21 (Down syndrome). A double-bubble appearance is present due to
distention of the stomach and duodenum by gas. This finding sug-
gests duodenal atresia. However, small amounts of gas are noted
distally, indicating a partial obstruction, such as a duodenal web,
duodenal stenosis, or malrotations with Ladd’s bands. In this pa-
tient, an annular pancreas was found at surgery. Patients with Down
syndrome can have a variety of congenital duodenal abnormalities
“high” and those that are termed “low”. High obstruc-
tions denote lesions of the esophagus, stomach, duode-
num, and upper small bowel (Fig. 8). Low obstructions
include lesions of the lower small bowel and large bow-
el, and anorectal malformations.
The distribution of dilated bowel loops on plain radio-
graphs usually enables one to differentiate high from
low obstruction relatively easily. This is simply done by
evaluating the number of gas-filled loops that are visible.
Fluid-filled loops may be difficult to visualize on AXR,
possibly masquerading as free fluid or masses; thus, they
can occasionally confuse the picture. It should be em-
phasized that the differentiation of dilated gas-filled
small from large bowel loops may be impossible in
neonates. Free air is not usually evident in these patients
unless the diagnosis is delayed. Intramural air (and even
portal venous gas) may be seen proximal to a high grade
obstruction but it is much more commonly seen in NEC.
Calcification may be present in the peritoneum (meconi-
um peritonitis) due to prenatal perforation, in the wall of
the bowel proximal due to an atresia, or in the bowel con-
tent within the lumen, occasionally proximal to a low ob-
struction. Most complete high obstructions are easily di-
agnosed on AXR. If the diagnosis is in doubt, air can be
injected slowly via a feeding tube into the lumen of the
GI tract to confirm or exclude an obstruction. In incom-
plete obstructions (e.g., malrotation and stenoses), posi-
tive water-soluble contrast agents are required to confirm
the level and nature of the obstruction.
The low obstructions include ileal and colonic atresias,
meconium ileus, functional immaturity of the large
bowel, Hirschsprung’s disease, and anorectal malforma-
tions. Although there are some features on AXR that
Alan Daneman, Simon G. Robben
might suggest each of the above conditions, the radio-
graphic findings in all of them are often non-specific.
Differentiation of these conditions will therefore depend
on other factors, such as the clinical history and the re-
sults of the physical examination. Invariably, to increase
diagnostic confidence, these conditions require the ad-
ministration of a water-soluble contrast enema to define
the distal colon and ileum. The contrast enema may also
prove therapeutic in patients with meconium ileus or
meconium plugging the colon.
Suggested Reading
Ang A, Chong NK, Daneman A (2001) Pediatric Emergency Care
17:334-340
Baldisserotto M, Marchiori E (2000) Accuracy of Noncompressive
Sonography of Children with Appendicitis According to the
Potential Positions of the Appendix. AJR Am J Roentgenol
175:1387-1392
Bombelburg T, Von Lengerke HJ (1992) Sonographic findings in
infants with suspected necrotizing enterocolitis. European J
Radiol 15:149-153
Buonomo C (1999) The radiology of necrotizing enterocolitis.
RCNA 37:1187-1198
Couture A, Baud C, Ferran JL et al (2008) Gastrointestinal tract
sonography in fetuses and children. 1st edn. Springer-Verlag,
Heidelberg Berlin New York
Daneman A, Alton DJ, Ein S et al (1995) Perforation during at-
tempted intussusception reduction in children – a comparison
of perforation with barium and air. Pediatr Radiol 25:81-88
Daneman A, Alton DJ, Lobo E et al (1998) Patterns of recurrence
of intussusception in children: a 17 year review. Pediatr Radi-
ol 28:913-919
Daneman A, Lobo E, Alton DJ, Shuckett B (1998) The value of
sonography, CT and air enema for detection of complicated
Meckel diverticulum in children with nonspecific clinical pre-
sentation. Pediatr Radiol 28:928-932
Daneman A, Myers M, Shuckett B, Alton DJ (1997) Sonographic
appearances of inverted Meckel diverticulum with intussus-
ception. Pediatr Radiol 27:295-298
Daneman A, Navarro O (2003) Intussusception Part 1: A review of
diagnostic approaches. Pediatr Radiol 33:79-85
Daneman A, Navarro O (2004) Intussusception Part 2: An update
on the evolution of management. Pediatr Radiol 34:97-108
Daneman A, Navarro O (2005) Intussusception: the debate en-
dures. Pediatr Radiol 35:95-96
del-Pozo G, Abillos JC, Tejedor D (1996) Intussusception: US
findings with pathologic correlation – the crescent-in-dough-
nut sign. Radiology 199:688-692
del-Pozo G, Albillos JC, Tejedor D et al (1999) Intussusception in
children: current concepts in diagnosis and enema reduction.
Radiographics 19:299-319
Doria AS, Amernic H, Dick P et al (2005) Cost effectiveness analy-
sis of weekday and weeknight shifts for assessment of appen-
dicitis. Pediatr Radiol 35:1186-1195
Elsayes KM, Menias CO, Harvin HJ, Francis IR (2007) Imaging
manifestations of Meckel’s diverticulum. AJR Am J Roentgenol
189:81-88
Faingold R, Daneman A, Tomlinson G et al (2005) Bowel viabili-
ty assessment by colour doppler sonography in necrotizing en-
terocolitis. Radiology 235:587-594
Fotter R, Sorantin E (1994) Diagnostic imaging in necrotizing en-
terocolitis. Acta Paediatr Supp 396:41-44
Frush DP, Frush KS, Oldham KT (2009) Imaging of acute appen-
dicitis in children: EU versus U.S… or US versus CT? A North
American perspective. Pediatr Radiol 39:500-505
An Approach to Imaging the Acute Abdomen in the Pediatric Population
Fujii Y, Hata J, Futagami K et al (2000) Ultrasonography improves
diagnostic accuracy of acute appendicitis and provides cost
savings to hospitals in Japan. J Ultrasound Med 19:409-414
Gu L, Alton DJ, Daneman A et al (1988) Intussusception reduction
in children by rectal insufflation of air. AJR Am J Roentgenol
150:1345-1348
Holscher HC, Heij HA (2009) Imaging of acute appendicitis in
children: EU versus U.S… or US versus CT? A European per-
spective. Pediatr Radiol 39:497-499
Hughes UM, Connolly BL, Chait PG, Muraca S (2000) Further re-
port of small-bowel intussusceptions related to gastrojejunos-
tomy tubes. Pediatr Radiol 30:614-617
Kim G, Daneman A, Alton DJ et al (1997) The appearance of in-
verted Meckel diverticulum with intussusception on air enema.
Pediatr Radiol 27:647-650
Kornecki A, Daneman A, Navarro O et al (2000) Spontaneous re-
duction of intussusception: clinical spectrum, management
and outcome. Pediatr Radiol 30:58-63
Kosloske AM, Love CL, Rohrer JE et al (2004) The diagnosis of
appendicitis in children: outcomes of a strategy based on pe-
diatric surgical evaluation. Pediatrics 113:29-34
Levy AD, Hobbs CM (2004) From the Archives of the AFIP.
Meckel Diverticulum: Radiologic Features with Pathologic
Correlation. Radiographics 24:565-587
McDermott VGM (1994) Childhood intussusception and ap-
proaches to treatment: A historical review. Pediatr Radiol
24:153-155
Meradji M, Hussain SM, Robben SG, Hop WC (1994) Plain film
diagnosis in intussusception. Br J Radiol 67:147-149
Miller SF, Seibert JJ, Kinder DL, Wilson AR (1993) Use of ultra-
sound in detection of occult bowel perforation in neonates. JU
Med 12:531-535
Navarro O, Daneman A (2004) Intussusception. Part 3: Diagnosis
and management of those with an identifiable or predisposing
cause and those that reduce spontaneously. Pediatr Radiol
34:305-312
Navarro OM, Daneman A, Chae A (2004) Intussusception: The use
of delayed, repeated reduction attempts and the management of
intussusceptions due to pathologic lead points in pediatric pa-
tients. AJR Am J Roentgenol 182:1169-1176
Navarro O, Dugougeat F, Kornecki A et al (2000) The impact of
imaging in the management of intussusception owing to patho-
logic lead points in children. A review of 43 cases. Pediatr Ra-
diol 30:594-603
O’Hara SM (1996) Pediatric gastrointestinal nuclear medicine. Ra-
diol Clin North Am 34:845-862
173
Park NH, Park SI, Park CS et al (2007) Ultrasonographic findings
of small bowel intussusception, focusing on differentiation
from ileocolic intussusception. Br J Radiol 80:798-802
Pena BMG, Taylor GA, Fishman SJ, Mandl KD (2002) Effect of an
imaging protocol on clinical outcomes among pediatric pa-
tients with appendicitis. Pediatrics 110:1088-1093
Poortman P, Lohle PNM, Schoemaker CMC et al (2003) Compar-
ison of CT and sonography in the diagnosis of acute appen-
dicitis: a blinded prospective study. AJR Am J Roentgenol 181:
1355-1359
Puylaert JB (1986) Acute appendicitis: US evaluation using grad-
ed compression. Radiology 158:355-360
Rohrschneider W, Troger J, Betsch B (1994) The post reduction
donut sign. Pediatr Radiol 24:166-170
Rohrschneider WK, Troger J (1995) Hydrostatic reduction of in-
tussusception under US guidance. Pediatr Radiol 25:530-534
Sargent MA, Babyn P, Alton DJ (1994) Plain abdominal radiogra-
phy in suspected intussusception: A reassessment. Pediatr Ra-
diol 24:17-20
Shanbhogue RL, Hussain SM, Meradji M et al (1994) Ultrasonog-
raphy is accurate enough for the diagnosis of intussusception.
J Pediatr Surg 29:324-327
Silva CT, Daneman A, Navarro OM et al (2007). Correlation of
sonographic findings and outcome in necrotizing enterocolitis.
Pediatr Radiol 37:274-282
Sivit CJ (2004) Imaging the child with right lower quadrant pain
and suspected appendicitis: current concepts. Pediatr Radiol
34:447-453
Taylor GA (2004) Suspected appendicitis in children: In search of
the single best diagnostic test. Radiology 231:293-295
Thurley PD, Halliday KE, Somers JM et al (2009) Radiological
features of Meckel’s diverticulum and its complications. Clin
Radiol 64:109-118
Todani T, Sato Y, Watanabe Y et al (1990) Air reduction for intus-
susception in infancy and childhood: Ultrasonographic diag-
nosis and management without X-ray exposure. Z Kinderchir
45:222-226
Verschelden P, Filiatraut D, Garel L et al (1992) Intussusception in
children: reliability of US in diagnosis – a prospective study.
Radiology 184:741-744
Wang G, Liu S (1988) Enema reduction of intussusception by hy-
drostatic pressure under ultrasound guidance. A report of 377
cases. J Pediatr Surg 23:814-818
Wiersma F, Allema JH, Holscher HC (2006) Ileoileal intussuscep-
tion in children: ultrasonographic differentiation from ileocolic
intussusception. Pediatr Radiol 36:1177-1181

Imaging Uronephropathies in Children
Jeanne S. Chow1, Fred E. Avni2
1 Departments of Urology and Radiology, Children’s Hospital, Boston, MA, USA
2 Department of Radiology, University Clinics of Brussels, Erasme Hospital, Brussels, Belgium
Introduction and Techniques
The principle that guides imaging of the pediatric genito-
urinary tract is the same that guides imaging elsewhere in
the body of the child: judicious use of imaging while
minimizing ionizing radiation exposure [1] and unneces-
sary sedation. Ultrasound (US) is the main imaging
modality of the genitourinary tract, providing excellent
images of the kidneys and bladder. Also, at the level of
these organs, an abnormally dilated ureter is easily im-
aged. Doppler provides additional information regarding
vascular flow and is especially helpful in evaluating the
main renal artery and veins as well as the arcuate vessels.
US contrast agents are used to further evaluate the vas-
cular flow of the kidneys, intrarenal masses, and vesico-
ureteral reflux [2]. The main limitation of US is that it
provides little functional imaging, instead mainly reveal-
ing the physical appearance of the urinary tract.
If further imaging is necessary, contrast-enhanced com-
puted tomography (CT) and magnetic resonance (MR)
provide exquisite, detailed images of the urinary tract in
addition to more functional information. MR urography
has become very popular because it emits no ionizing ra-
diation and yields both anatomical and functional data.
The latter includes uptake and excretion rates, which to-
gether with relative function and indirect measurements of
glomerular filtration rates provide more information than
obtained from a MAG-3/Lasix renogram [3]. However,
because MR is a long procedure that often requires seda-
tion, Tc-99m(V) dimercaptosuccinic acid (DMSA) and
MAG-3 are more commonly used to provide functional
imaging of the genitourinary tract. In addition, functional
information cannot be obtained without gadolinium, but
this approach places children with chronic renal failure
and end-stage renal disease at risk for developing nephro-
genic systemic fibrosis [4]. Although nephroureterolithia-
sis may be commonly studied by CT in adults, US is the
first imaging modality in children [4, 5].
Key Point
– Ultrasound is the main imaging modality of the geni-
tourinary tract in children.
IDKD 2010-2013
Management of In Utero Renal Pelvis Dilatation
Dilatation of the renal pelvis is a common finding on ob-
stetric US, with a frequency of around 1-4% of all preg-
nancies. Yet, every dilatation does not have the same clini-
cal relevance; furthermore, the antenatal and postnatal evo-
lution is variable. This has led to controversy in the litera-
ture about the best work-up and follow-up after birth [6-8].
Definition
The best criterion for objectivating urinary tract (UT) di-
latation is the anteroposterior diameter of the renal pelvis,
which is measured on a transverse scan of the fetal ab-
domen. The upper limit for normal should be 4 mm in the
second trimester and 7 mm in the third trimester.
Other US evidence of a UT abnormality includes the
visibility of the fetal ureter, dilatation of the renal ca-
lyces, abnormal echogenicity of the renal parenchyma,
and the demonstration of an enlarged bladder (Table 1).
Findings on Obstetric US
Abnormalities of the UT can be found at any time during
pregnancy. Once a dilatation has been detected in utero,
subsequent evaluations should address three issues: the
origin of the dilatation, the co-existence of associated
anomalies, and the prognosis.
The most common causes of dilatation in the fetus are
related to obstructions of the ureteropelvic and uretero-
vesical junctions and to complicated duplex kidneys. UT
Table 1. Characteristics suggesting an abnormality of the urinary
tract on neonatal ultrasound
Pelvic dilatation >7 mm
Calyceal dilatation
Parenchymal thinning
Lack of cortico-medullary differentiation
Small kidney
Thickening of the pelvic wall
Thickening of the ureteral wall
Ureteral dilatation >3 mm
Enlarged bladder
Imaging Uronephropathies in Children
dilatation can also result from vesicoureteric reflux
(VUR) and intravesical obstruction, especially in male fe-
tuses. In most cases, the US evaluation will be able to dif-
ferentiate between etiologies. In some patients, especial-
ly those with bilateral and complex uropathies, fetal MR
imaging will provide additional information.
Other organ malformations also can be associated with
UT dilatation; therefore, the US survey should be as
meticulous and complete as possible. Chromosomal
analysis may be indicated in selected patients.
The prognosis of a uropathy will depend upon the type
and extent of the anomalies. Amniotic fluid volume is im-
portant to the prognosis as well; oligohydramnios,
thought to be related to decreased urine production, is a
poor prognostic indicator.
It is of utmost importance that any relevant informa-
tion is correctly transmitted to the postnatal team that will
be in charge of caring for the newborn.
Postnatal Management of Fetal Pelvis Dilatation
Certain conditions require immediate postnatal confir-
mation and therapeutic maneuvers, such as obstructive
posterior urethral valves and prolapsed ectopic uretero-
cele into the urethra. In those cases, US and micturating
cystourethrography (MCU) should be performed directly
after birth. In all other cases, the work-up can be planned
without urgency. Patients with ureterovesical junction ob-
struction (UVJO) and complex uropathies should be put
on prophylactic chemotherapy at least until the final di-
agnosis is made.
An algorithm based on US examination is presently
applied by most teams (Fig. 1). Micturating voiding
urethrography is only applied if US displays a significant
anomaly (Table 1). MCU is used to detect high-grade
VUR and urethral anomalies. If VUR is not present, com-
plementary imaging is necessary to determine the precise
US : 1st US around day 5
abnormal: pelvis ≥7 mm + dilated calices, or other anomalies
VCUG
normal abnormal abnormal
pelvis ≥ 10 mm
other malformation,
US at 3 mo “extended criteria”
pelvis ≥10 mm pelvis >10 (15) mm
further morphological & functional
Stop follow-up evaluation: scintigraphy, IVU, MRU …
175
origin of the dilatation. Renal function is assessed
through isotopic studies, while complicated UT malfor-
mations are best evaluated by MR imaging. The latter is
particularly helpful for the assessment of a very dilated
UT and complicated duplex-kidney systems.
The type of treatment (conservative or surgical) will
depend upon the diagnosis, renal function on follow-up,
and complications. Today, the trend is increasingly to-
ward a conservative approach. The length of follow-up
must be adapted to the type of anomaly as well as to clin-
ical and imaging follow-up.
Key Points
– Fetal renal dilatation is a common finding during ob-
stetric US.
– Thresholds of 4 and 7 mm during the 2nd and 3rd
trimester, respectively, are widely accepted.
– Postnatally, these patients must be further evaluated by
US; voiding cystourethrography is performed only if an
anomaly is found at birth or at the age of one month.
– The trend is toward a more conservative approach to
treatment, based on clinical and imaging follow-up.
Imaging Cystic Kidneys in Children
Introduction
Renal cystic diseases may be discovered or suspected at
any stage during fetal life or at any age in childhood.
They encompass a large number of conditions that can be
separated into those with or without hereditary transmis-
sion. Imaging, mainly US, plays an important role in dif-
ferentiating between the various types of cystic diseases
as it shows the features of renal involvement as well as
associated anomalies.
normal
US at 1 mo
normal
Fig. 1. Algorithm in fetal hydronephrosis
(HN). Antenatal diagnosis of mild to moder-
ate renal pelvis dilatation [6]. VCUG, Void-
Stop follow-up ing Cystourethrography
176
Cystic Kidneys in the Fetus
In the fetus (and during the perinatal period), cystic renal
disease should be suspected whenever bilateral hyperechoic
kidneys or cysts (uni- or bilateral) are discovered during an
obstetric US examination. The imaging approach to the di-
agnosis should be based on a detailed sonographic analysis
that includes measurement of renal length, the presence or
absence of cortico-medullary differentiation (CMD), and
the presence, number, size, and location of the cysts. This
evaluation should be completed through an analysis of the
entire fetus, looking for associated anomalies.
The timing of detection and the amount of amniotic
fluid are the most important prognostic factors. Further-
more, a detailed clinical and familial inquiry is essential
in disease evaluation.
Autosomal recessive polycystic kidney disease (ARPKD)
is the main diagnosis to consider in case of bilateral,
markedly enlarged, hyperechoic kidneys without CMD.
The prognosis is usually poor if the amniotic fluid vol-
ume is markedly decreased.
In case of moderately enlarged hyperechoic kidneys,
three diagnoses have to be considered:
1. nephropathy due to a mutation in the TCF2 gene,
2. a milder form of ARPKD, and
3. autosomal dominant polycystic kidney disease (ADPKD).
In TCF2-mediated nephropathy, CMD is absent or pre-
sent and whenever cysts are visible they are typically sub-
cortical. In the milder form of ARPKD, cysts may be ob-
served in the medullary area. The main sonographic fea-
ture of ADPKD is a striking cortical hyperechogenicity
associated with increased CMD.
Whenever cysts are the main US finding in the fetal kid-
neys, their number and location are the main criteria for the
differential diagnosis. The diagnosis of unilateral multiple
cysts suggests a multicystic dysplastic kidney. Bilateral
multiple cysts can be visualized in a large number of renal
or syndromic diseases, the most common diagnoses being
bilateral multicystic dysplastic kidney, ADPKD, bilateral
obstructive dysplasia, and glomerulocystic kidneys.
Renal Cystic Diseases in Children
In children, renal cystic diseases are usually discovered
during US examination performed in the follow-up of a
known perinatally diagnosed disease, during the work-up
of syndromes diagnosed after birth, during screening in
an at-risk family, or as an incidental finding. The US ap-
proach is the same as that described for the fetus. The role
of imaging in the diagnosis and follow-up of renal in-
volvement is to search for complications such as hemor-
rhage or urolithiasis. Specifically, an important role for
US is the detection of hepatic-biliary complications.
Key Points
– Renal cystic diseases can be diagnosed in the perinatal
period or in later childhood.
Jeanne S. Chow, Fred E. Avni
– Hyperechogenicity or cysts are cardinal findings.
– A familial history, detailed clinical inquiry, and asso-
ciated findings help in establishing the diagnosis.
Renal Ectopia and Duplications
One of the most interesting areas of pediatric uroradiolo-
gy is studying and understanding the multitude of con-
genital abnormalities of the urinary tract. During normal
renal development, the kidneys ascend from the renal
pelvis while rotating medially. If the kidneys do not as-
cend or ascend past their normal location in the renal fos-
sae, they are ectopic. In some cases they are as low as the
pelvis and in others as high as the thoracic cavity. If the
kidneys fuse during ascent, pelvic cake kidneys, midline
horseshoe kidneys, or left- or right-sided cross-fused ec-
topic kidneys form. Since the embryological origin of the
kidneys (metanephros) is separate from that of the ureters
(ureteric buds), the site of ureteral insertion is normal
even if the kidney is ectopic. However, the renal blood
supply from the aorta will vary depending on the level of
ectopia.
The ureteric bud must meet the metanephros in order
for the kidney to form. Without this interaction, kidney
formation is not induced. If two ureteric buds meet at the
metanephric blastema, then the kidney becomes “duplex”.
Ureteral duplication may be complete or, more commonly,
incomplete.
In incomplete ureteral duplication, a single ureteric
bud bifurcates and meets the metanephros during ap-
proximately the 5th to 6th week of gestation. The two
branches of the ureter may join at the level of the renal
pelvis (bifid pelvis) or at the proximal, mid-, or distal
ureter (bifid ureter) and terminate in a single distal
ureter that inserts orthotopically into the bladder. Since
the two moieties of the kidney share a common distal
ureter, they behave similarly and usually appear normal.
Rarely, one of the ureteral buds may be blind-ending and
never appear to “reach” the kidney (blind-ending ureter-
al duplication). The associated kidney has a single col-
lecting system.
In complete ureteral duplication, two separate ureteric
buds arise from the Wolffian duct. The lower-pole ureter
is considered the analogue to the normal single-system
ureter. Thus, the lower pole of the kidney has all of the
same abnormalities that can affect a single-system kid-
ney, including VUR, ureteropelvic junction obstruction
(UPJO), and UVJO. The upper-pole ureter is “abnormal”
and ectopic (Weigert-Meyer rule). The ectopic ureter in-
serts medially and inferiorly to the normal ureteral ori-
fice, usually in the bladder. In girls, the ectopic ureter
may insert below the bladder base, into the urethra or the
vagina. A vaginal ectopic ureter can cause constant uri-
nary dribbling in girls and incontinence [9]. In boys, ec-
topic ureters never terminate below the urinary sphincter
and thus never result in incontinence; however, the ureter
can terminate in Wolffian duct derivatives, including the
Imaging Uronephropathies in Children
seminal vesicals and vas deferens. Very rarely, three com-
pletely or incompletely separated ureters form, resulting
in ureteral triplication [10].
Ectopic ureters are often obstructed but rarely reflux. If
the ectopic ureter inserts into the urethra at the level of the
urinary sphincter, urinary flow is obstructed or refluxes
depending whether the sphincter is closed or open [11]. The
more distal the ureteral insertion, the more dysplastic and
dysfunctional the associated renal parenchyma. Ectopic
ureters, and all the associated abnormalities, can also occur
in single-system kidneys (single ectopic ureter) [12].
A ureterocele is the dilated submucosal terminal
segment of the ureter. It is associated with varying de-
grees of ureteral obstruction and subsequent dilatation
of the renal pelvis and calyces. In girls, ureteroceles are
most commonly seen in association with ectopic upper-
pole ureters. In boys, they are most commonly associat-
ed with single-system kidneys and are orthotopic.
Although ureteroceles protrude into the bladder, when
the intravesical pressure equals that of the ureterocele,
the ureterocele can flatten and become imperceptible
(efface). When the intravesical pressure exceeds that of
the ureterocele, the latter everts or intussuscepts into its
ureter. Ectopic bladder-neck ureteroceles or large simple
ureteroceles can prolapse into the urethra and cause
bladder outlet obstruction.
Key Points
– Renal ectopia is due to abnormalities in the normal as-
cent of the kidney.
– Ureteral duplication may be incomplete (more com-
mon) or complete.
– The Weigert-Meyer rule states that the upper-pole
ureter of a duplex kidney inserts ectopically, medially,
and inferiorly to the orthotopic location.
– The lower-pole ureter is the analogue of the single-sys-
tem kidney.
Urinary Tract Obstruction
Urinary tract obstructions occur at three main areas: the
ureteropelvic junction, the ureterovesical junction, and
the bladder outlet (i.e., the urethra). Rarely, the mid-
ureter can be obstructed by webs, fibrosis, or compres-
sion from the inferior vena cava, or there may be ob-
struction at the level of the infundibula in the kidney. On
US, the normal hypoechoic medullary pyramids seen
routinely in infancy and childhood should not be con-
fused with dilated calyces or a sign of obstruction.
The most common congenital obstruction of the kid-
ney is UPJO, which is due to a stenosis at the junction of
the renal pelvis and proximal ureter. Since most children
are now diagnosed prenatally and followed postnatally,
they rarely present with symptoms and signs of obstruc-
tion, such as infection, pain, or renal stones. Some chil-
dren have UPJO due to a crossing renal artery and pre-
177
sent with intermittent pain from intermittent obstruction,
with hydronephrosis only evident during obstruction. To
be correctly diagnosed, these children must be imaged at
the time of their painful episodes [13].
The conundrum of UPJO is that we are still unable to
predict whether the degree of obstruction and thus its
eventual effects on renal function will improve or worsen
over time. US is routinely used to describe the degree of
obstruction and the appearance of the renal parenchyma.
However, functional imaging studies, primarily MAG-3
studies with Lasix (MAG-3/Lasix renogram) and MR
urography, are used to help quantify the degree of ob-
struction and the contributing function of each kidney.
An obstruction of the distal ureter as it enters into the
bladder results in UVJO. Most such cases are primary
and due to a ureteral obstruction, although secondary
UVJO can occur with an abnormally thickened bladder.
The insertion of the obstructed ureter may be orthotopic
(primary mega-ureter) or ectopic. An orthotopic or ec-
topic ureterocele may also be associated with obstruc-
tion. Primary mega-ureter accounts for the majority of
the cases of UVJO. In most patients with this condition,
the degree of dilatation improves over time [14] such
that surgical repair is required only for a minority of af-
fected patients. Surgery is indicated if the degree of di-
latation worsens, renal function is impaired, or the ob-
struction is thought to be contributing to stasis and UT
infections.
Key Points
– Ureteropelvic junction obstruction is the most com-
mon cause of urinary tract obstruction.
– Most cases of ureterovesical junction obstruction im-
prove with time.
Voiding Abnormalities and Secondary Vesicoureteric
Reflux
Children with lower-spine abnormalities may have detrusor-
sphincter dysynergia, in which the bladder contracts but
the urinary sphincter does not relax during voiding, re-
sulting in chronic obstruction of the bladder outlet. The
uncoordinated voiding causes urinary retention, in-
creased bladder pressures, secondary VUR, and sec-
ondary UVJO. Similarly, children with voiding dysfunc-
tion without a neurogenic cause can also develop sec-
ondary reflux.
Urethral obstruction can occur in the posterior or an-
terior urethra in boys whereas the urethra is rarely ob-
structed in girls. Posterior urethral valve obstruction is
the most common congenital urethral obstruction and is
caused by an obstructing membrane just below the level
of the verumontanum. Anterior urethral obstruction is
most commonly due to traumatic strictures and mostly lo-
cated in the bulbar urethra. Anterior urethral valves or di-
verticula are rare. Depending on the severity of the ob-
178
struction, the portion of the urethra proximal to the ob-
struction may be dilated, the bladder wall may be hyper-
trophied, and secondary reflux and UVJO may occur. In
circumcised boys, meatal stenosis is another cause of ure-
thral obstruction. Retrograde urethrogram under fluo-
roscopy or ultrasound is the best way of studying the an-
terior urethra whereas the posterior urethra can only be
studied during voiding.
Key Points
– Voiding dysfunction can lead to vesicoureteral reflux.
– Most obstructions of the bladder outlet in boys are
congenital (e.g., posterior and anterior urethral valves)
or post-traumatic (bulbar urethral stricture).
Renal Masses in Children
Once a mass is established to be intrarenal, its histology
can be predicted based on its appearance and on the pa-
tient’s age. Most intrarenal masses occurring in the new-
born period are benign. Although rare, the most common
solid intrarenal mass seen in newborns is a mesoblastic
nephroma [15]. These large, solid, enhancing masses are
benign, although the cellular subtype is the most aggres-
sive and can cause paraneoplastic syndromes. These must
be removed but the prognosis is excellent.
In the newborn, the most frequently occurring cystic
abnormality of the kidney is multicystic dysplasic kid-
ney, which can involve the entire kidney or be segmen-
tal. The condition is due to a congenital abnormality of
the kidney in which the collecting system forms as cysts,
and the renal parenchyma is dysplastic and non-
functional. Multicystic dysplasic kidney is now com-
monly diagnosed in utero. Over time, the cyst fluid
resorbs and a tiny nub of tissue remains. These are typi-
cally treated non-surgically.
The most common renal mass in toddlers is Wilms’ tu-
mor. Children with aniridia, WAGR (Wilms’ tumor,
aniridia, genitourinary abnormalities, and mental retarda-
tion), Deny-Drash syndrome, Beckwith-Weidemann syn-
drome, hemihypertrophy, or nephroblastomatosis are pre-
disposed to developing this tumor. Wilms’ tumor is a sol-
id, cystic, and often hemorrhagic mass and is far more
common but radiographically indistinguishable from ei-
ther clear cell sarcoma or malignant rhabdoid tumor.
However, if a tumor has a large subcapsular hematoma,
and if there are brain metastases, malignant rhabdoid tu-
mor should be considered [16]. Centrally located multi-
locular masses of the kidney may be a multilocular cys-
tic nephroma, which is more common in boys in child-
hood and in women in adult life [17].
In children over 11 years of age, renal cell carcinoma
becomes more common than Wilms’ tumor, although the
likelihood of either tumor is extremely rare [18]. It is cru-
cial to confirm that the child has no clinical indicators of
UT infection, because focal pyelonephritis or lobar
Jeanne S. Chow, Fred E. Avni
nephronia mimics a tumor in appearance and thus must
always be considered in the differential diagnosis of a re-
nal mass.
If the renal mass is bilateral, the appearance and clin-
ical presentation are extremely helpful in predicting the
histology. If there are multiple large masses and the kid-
neys are also enlarged, bilateral nephrogenic rests due to
nephroblastomatosis are most likely. Nephrogenic rests
are remnant fetal renal tissue that never fully matured. As
they have a high propensity to develop into Wilms’ tu-
mors, these masses need frequent surveillance. If the
masses are partially echogenic, angiomyolipoma should
be considered, especially if the patient has tuberous scle-
rosis. Wilms’ tumors, lymphoma, and infections may al-
so be bilateral.
Solitary simple cysts are much less commonly seen in
children than in adults. Calyceal diverticula may appear
as simple cysts but they actually communicate with the
adjacent calyx and can become superinfected. Delayed
intravenous pyelogram, CT, or MR imaging, which show
contrast within the cyst, is able to distinguish calyceal di-
verticula from simple cysts. If there are multiple simple
cysts, especially in enlarged kidneys, ADPKD should be
considered.
Key Points
– Most newborn renal masses are benign.
– The most common renal malignancy in toddlers is
Wilms’ tumor.
– Focal pyelonephritis mimics renal tumors.
Imaging Renal Failure in Children
Introduction
Ultrasound plays a central role in pediatric imaging, par-
ticularly in pediatric nephrology, in which it helps to dif-
ferentiate between the etiologies of renal failure. For
some diseases, the US pattern will be specific, while for
others there will be little or no parenchymal changes. The
US evaluation should therefore be very meticulous and
correlated to the biological and clinical data [18-20].
Sonographic Technique
Renal US has to be carried out with the highest-resolu-
tion transducers, depending on the patient’s size. The use
of both curved and linear transducers is essential. US
studies include measurements of the kidneys and of any
dilatation as well as the evaluation of renal echogenicity
(cysts? calcifications?) and CMD. Doppler analysis also
must be performed. In case of UT dilatation, the cause
and level of obstruction must be determined, including
the bladder, within the field of investigation. It might be
of interest in some patients to evaluate the liver, spleen,
and biliary tract, too.
Imaging Uronephropathies in Children
Acute Renal Failure
Acute renal failure (ARF) is defined as urine production
<1 mg/kg/day. Its causes can be pre-renal, renal, or post-
renal in origin. In the case of ARF of renal origin, US is
often diagnostic.
Hemolytic Uremic Syndrome
Hemolytic microangiopathic anemia, thrombocytopenia,
and ARF occurring together constitute hemolytic uremic
syndrome, which is the commonest cause of ARF in the
United States and in several European countries, espe-
cially in young infants.
During the acute phase of the disease, the US appear-
ance of the renal cortex is markedly hyperechoic, with in-
creased CMD. On Doppler analysis, there is no diastolic
flow, which correlates well with the lack of urine pro-
duction. The return of the diastolic wave indicates a re-
turn to normal diuresis. Other organs may be involved as
well, including the gallbladder and digestive tract.
Medullary or Cortical Necrosis, Shock Kidneys
Medullary and cortical necrosis in the neonate results
from a lack of renal perfusion. On US, the cortex in cor-
tical necrosis first appears hyperechoic, then shrinks, and
finally calcifies. In medullary necrosis, calcifications de-
velop within the medulla.
Renal Vein Thrombosis
While largely a neonatal disease, renal vein thrombosis
may occur already in utero. When both renal veins are in-
volved, the condition is associated with ARF. On US, the
kidneys appear enlarged, CMD is absent, and hyper-
echoic streaks are demonstrated in the interlobar areas.
Obstructive Uropathies
Anuria and ARF may follow ureterocele prolapse within the
urethra, obstruction of a single renal system, tumoral en-
trapment of the ureters, or bilateral obstructive urolithiasis.
Chronic Renal Failure
Chronic renal failure (CRF) is defined as a glomerular
filtration rate <50 mL/min/1.73 m2/kidney. One of the
most common causes of CRF is renal hypodysplasia. On
US, the kidneys are small, CMD absent, and small cysts
can be visualized.
Renal Cystic Diseases (see above)
Congenital Nephrotic Syndromes
Congenital nephrotic syndromes (CNSs) encompass dis-
eases in which there is massive proteinuria occurring
179
after birth. The most common form of CNS is the Finnish
type. Proteinuria starts in utero and the placenta is thick-
ened. On US, at birth, the kidneys are swollen and
hyperechoic; CMD is present but the pyramids are
irregular and within weeks they will “disappear”.
Other causes of CNS include diffuse mesangial scle-
rosis (DMS), which can be part of the Drash syndrome
(DMS, genital anomalies, and a risk for Wilms’ tumor).
In some patients, CNSs evolve toward end-stage renal
disease and necessitate renal transplantation.
Syndromes Affecting the Tubules and Metabolic Diseases
Renal diseases that include a primary and secondary
tubulopathy are numerous. Hypercalciuria is a constant
finding and may lead to nephrocalcinosis, which is easi-
ly detected by US. For instance, type 1 primitive hyper-
oxaluria is seen as strikingly hyperechoic kidneys already
at birth and results in urolithiasis.
Key Points
– US is the key imaging examination in children with
acute or chronic renal failure.
– The etiologies of renal failure are numerous.
– The US patterns will orient the diagnosis.
Urinary Tract Infection
Introduction
Urinary tract infection is one of the commonest bacterial
diseases in children: 5% of girls and 0.5% of boys will
suffer at least one episode. Despite numerous studies and
publications, the role of imaging is controversial. The
main challenges of imaging are to identify patients with
complicated UT infection, those with an underlying
cause, and those at risk for recurrence [21-23].
Imaging may be done at the time of diagnosis of an
acute episode, or during treatment, or in a late assess-
ment. No single imaging technique allows complete eval-
uation of the UT; instead, the use of each one must be op-
timized to obtain the maximum amount of information in
association with the lowest morbidity.
Imaging Acute Pyelonephritis
The main role of conventional US is to diagnose any as-
sociated malformative uropathy and/or the complications
(abscess) of a UT infection. Color Doppler US may pro-
vide significant additional information regarding the de-
gree of renal insult. The gold standard examination for
the diagnosis of renal lesions during an acute episode (as
well as for the detection of scars) is DMSA scintigraphy.
CT scan with contrast enhancement is as efficient as a
DMSA scan for the demonstration of acute renal lesions.
Its drawbacks are the radiation hazard and the need for
180
contrast injection. MR imaging has great potential to de-
tect renal inflammatory lesions and provides information
on renal parenchyma status as well as on UT morpholo-
gy. In addition, it may clarify ultrasound findings. MCU
is an essential complementary examination for patients
with UT infections, as there may be associated reflux in
as many as 40% of cases. Therefore, the technique is gen-
erally advocated in case of UT infections to look for VUR
and voiding dysfunction. However, MCU is an invasive
technique and irradiation is a drawback.
Proposed Work-Up
Based on these considerations, it is important to identify
among patients with clinical suspicion of acute
pyelonephritis those who need an imaging work-up.
Clearly, newborn boys and school-aged girls are groups
at risk. Still, it is impossible to define why certain pa-
tients will develop renal involvement and in which of
these patients the disease will recur. Therefore, every pa-
tient with a UT infection should undergo imaging evalu-
ation albeit tailored and optimized to his/her condition. A
work-up protocol is summarized in Fig. 2.
Complications of Acute Pyelonephritis
Renal abscess is the typical complication of delayed or
non-adapted treatment. Lesions are demonstrated by US,
although further, cross-sectional imaging (CT scan or
MR) is occasionally useful. Pyonephrosis corresponds to
US + Doppler
Normal Abnormal
DMSA APN
Normal Abnormal VCUG
Stop
Fig. 2. Proposed imaging workflow for a child with a urinary tract
infection. US, Ultrasound; APN, Acute Pyelonephritis; DMSA,
Tc-99m(V) Dimercaptosuccinic acid; VCUG, Voiding Cysto-
urethrography
Jeanne S. Chow, Fred E. Avni
infection of the urine in a dilated urinary tract. Suspicion
of this diagnosis should prompt a diagnostic and thera-
peutic nephrostomy.
Late Complications of UT Infection
The development of renal scars is the long-term risk of
untreated acute pyelonephritis. Patients with renal scars
are at risk for developing renal hypertension, complica-
tions during pregnancy, and renal failure. Presently,
DMSA scintigraphy is the gold standard method for the
diagnosis of scars, if performed at a time sufficiently
removed from the acute episode.
Xanthogranulomatous pyelonephritis is an atypical
chronic infectious renal lesion. It may appear as a tumor
or diffusely. CT is the best imaging modality.
Key Points
– Imaging of UT infections is controversial.
– The aim is to detect groups at risk, renal involvement,
and risks for recurrence.
– US Doppler, DMSA scintigraphy, and MCU form the
basis of the imaging work-up.
Imaging Urolithiasis in Children
Introduction
Urolithiasis is less common in children than in adults
(1 child for every 20,000 adults). It is more frequent in
certain geographic areas. Hematuria and pain are the pre-
senting symptoms in about half of the patients whereas
the disease is asymptomatic and an incidental finding in
20% of patients. A familial history is frequent and ac-
counts for over 50% of patients. Metabolic disorders and
UT infections are common etiologies [24, 25].
The work-up of a patient with urolithiasis includes uri-
nalysis, blood tests, and genetic analysis, looking for fa-
voring diseases or infections. Imaging must be applied sys-
tematically, starting with US and using a tailored decision-
making process for additional imaging modalities [25].
Imaging
Ultrasound, with its entire spectrum of optimized tech-
niques, is the ideal primary imaging modality in children
(Fig. 3). It allows reliable demonstration of the kidneys
and urinary tract and is useful in diagnosis as well as in
post-therapeutic follow-up. On US, urolithiasis manifests
as either nephrocalcinosis or lithiasis. In nephrocalci-
nosis, echogenic deposits in the pyramids lead to a re-
versed CMD. Nephrocalcinosis should be differentiated
from other causes of echogenic pyramids. Lithiasis is eas-
ily demonstrated within the dilated renal upper cavities
and at the ureterovesical junction. It has the same ap-
pearance as in adults.
Imaging Uronephropathies in Children 181

US (gray scale + DDS/CDS)

US inconclusive
negative + low negative + high
clinical suspicion or non-diagnostic
clinical suspicion positive mismatch US & clinics

only secondary signs

Fig. 3. Imaging algorithm for infants and chil-
dren with suspected urolithiasis [7]. CDS,
Color Doppler Sonography (power Doppler);
ce-VUS, contrast-enhanced Voiding Uroso-
nography; (uro-)CT, (urinary tract) CT; DD,
Differential Diagnosis; DDS, Duplex
Doppler Sonography; KUB, Kidney-Ureter-
Bladder film; IVU, Intravenous Urography;
MRU, Magnetic Resonance Urography; US,
Ultrasound
+ no stone visible,
or before intervention, if
therapeutically necessary
Stop
follow-up after/during treatment uro-CT x1
(ultra-)low dose + unenhanced x2
potentially + KUB for
confirmation, for treatment needs … + KUB (± IVU) before intervention / lithotripsy
x1
or KUB (+ adapted IVU, particularly if low-dose CT unavailable)
x2
potentially contrast-enhanced CT-urography, if other DD or complication; MRU in selected cases

Some kidney diseases have a specific, characteristic
pattern, such as primary hyperoxaluria type 1. A plain
film of the abdomen may be necessary for the proper
demonstration of the stone prior to treatment.
Tailored intravenous urography is still used by some
clinicians to demonstrate the morphology of the UT, es-
pecially if CT is not available. Furthermore, CT is much
less frequently used than in adults. It serves as a comple-
mentary tool in case of a non-diagnostic US examination
or prior to treatment. The radiation dose must be mini-
mized and optimally adapted to the child’s size [5, 26].
Treatment
The treatment of urolithiasis should aim to avoid or cor-
rect conditions that have led to the disease. If the lithia-
sis cannot be medically eliminated, more interventional
treatments will be needed. For example, extracorporeal
lithotripsy, the primary treatment approach in adults, has
been adapted to children as well. However, in selected
cases surgery will be unavoidable.
Key Points
– Urolithiasis is less frequent in children.
– An etiology is more often detected than in adults.
– US is the main imaging modality.
References
1. Brenner DJ, Elliston CD, Hall EJ, Berdon WE (2001) Estimat-
ed risks of radiation-induced fatal cancer from pediatric CT.
AJR Am J Roentgenol 176:289-296
2. Ripolles T, Puig J (2009) Actualización del uso de contrastes
en ecografía. Revisión de las guias clinicas de la Federación
Europea de Ecografía, Radiología 51:362-375
3. Grattan-Smith JD, Jones RA (2008) MR urography: technique
and results for the evaluation of urinary obstruction in the
pediatric population, Magn Reson Imaging Clin N Am 16:
643-660
4. Thomsen HS (2007) ESUR guideline: gadolinium-based con-
trast media and nephrogenic systemic fibrosis Eur Radiol
17:692-2696
5. Passerotti C, Chow JS, Silva A et al (2009) Ultrasound versus
computed tomography for evaluating urolithaisis. J Urol
182:1829-1841
6. Riccabona M, Avni FE, Blickman JG et al (2008) Imaging
recommandations in pediatric uroradiology (Part 1). Pediatr
Radiol 38:138-145
7. Riccabona M, Avni FE, Blickman JG et al (2009) Imaging rec-
ommendations in pediatric uroradiology (Part 2). Pediatr Ra-
diol 39:891-898
8. De Bruyn R, Marks SD (2008) Post-natal investigations of fe-
tal renal disease. Semin Fetal Neonatal Med 13:133-141
9. Carrico C, Lebowitz RL (1998) Incontinence due to an infra-
sphincteric ectopic ureter: why the delay in diagnosis and what
the radiologist can do about it, Pediatric Radiology 28:942-949
10. Gill RD (1952) Triplication of the ureter and renal pelvis. J
Urol 140:147
11. Wyly JB, Lebowitz RL (1984) Refluxing urethral ectopic
ureters: recognition by the cyclic voidng cystourethrogram.
AJR 142:1263-1267
12. Prewitt LH, Lebowitz RL (1976) The single ectopic ureter.
AJR 127:941-948
13. Rooks VJ, Lebowitz RL (2001) Extrinsic ureteropelvic junc-
tion obstruction from a crossing renal vessel: demography and
imaging. Pediatr Radiol 31:120-124
14. Shukla AR, Cooper J, Patel RP et al (2005) Prenatally detect-
ed primary meagureter: a role for extended follow-up. J Urol
173:1353-1356
15. Chaudry G, Perez-Ataude AR, Ngan BY et al (2009) Imaging
of congenital mesoblastic nephroma wath pathological corre-
lation. Pediatr Radiol 39:1080-1086
16. Eftekhari F, Erly WK, Jaffe N (1990) Malignant rhabdoid
tumor of the kidney: imaging features in two cases. Pediatric
Radiol 21:39-42
17. Boggs LK, Kimmelstiel P (1956) Benign multilocular cystic
nephroma: report of two cases of so-called multilocular cyst of
the kidney. J Urol 76:530-541
182
18. Estrada CR, Sthar AM, Eaton SH et al (2005) Renal cell car-
cinoma: Childrens Hospital Boston experience. Urology 66:
1296-1300
19. Wedekin M, Ehrich JH, Offner G, Pape L (2008) Aetiology
and outcomes of acute and chronic renal failure in infants.
Nephrol Dial Transplant 23:1575-1580
20. Ardissimo G, Dacco V, Testa S et al (2003) Epidemiology of CRF
in children: data of the Ital Kid Project. Pediatrics 111:382-387
21. Mercado-Duane MG, Benson JE, John SD (2002) US of renal
insufficiency in neonates. RadioGraphics 22:1429-1438
22. Lim R (2009) Vesico-ureteral reflux and UTI: evolving prac-
tices and current controversies in pediatric imaging. AJR Am
J Roentgen 192:1197-1208
Jeanne S. Chow, Fred E. Avni
23. Lee M, Lin C, Huang F et al (2009) Screening young children
with a first febrile UTI for high grade VUR with renal US
scanning and Technecium-99m-labelled DMSA. J Pediatr
154:797-802
24. Ajdinovic B, Jaukovic L, Krstic Z, Dopuda M (2008) Impact
of MCU and DMSA renal scintigraphy on the investigation
schema in children with UTI. Ann Nucl Med 22:661-665
25. Hoppe B, Kemper MJ (2008) Diagnostic examination of the
child with urolithiasis or nephrocalcinosis. Pediatr Nephrol
DOI:10.1007/s00467-008-1073-x
26. Karmazyn B, Frush DP, Applegate KE et al (2009) CT with
computed simulated dose reduction technique for detection of
pediatric nephrourolithiasis. AJR 192:143-149

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging
Gerald Antoch
Department of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen,
University at Duisburg-Essen, Essen, Germany
Introduction
Malignant tumors are the second most common cause of
death in the western world [1]. Based on the assumption
that patients’ prognoses can be improved by the adoption
of stage-adapted therapy, accurate clinical and radiologi-
cal tumor staging must be considered as essential when
assessing primary tumors and recurrent disease. In addi-
tion, different imaging procedures are used for therapy
response assessment in patients undergoing treatment for
malignant disease. For both, tumor staging and therapy
response assessment, morphological and functional
imaging procedures are available; however, these have
well-known limitations in their diagnostic accuracy.
Computed tomography (CT), magnetic resonance imag-
ing (MRI), and ultrasound (US) provide mainly morpho-
logical information on the tumor and its potential metas-
tases, but the lack of functional data has been shown to
hamper accurate assessment of local lymph node in-
volvement [2, 3]. Therapy response assessment is mainly
based on lesion size, which nonetheless has been shown
to be an insensitive indicator of response, at least in the
initial phase of tumor treatment. The functional data pro-
vided by positron emission tomography (PET) are known
to be more sensitive and specific than morphological da-
ta and to complement the latter in the assessment of lo-
cal lymph node involvement by malignant tumors [4-6].
Similarly, when PET and PET/CT are used to assess ther-
apy response, they have been found to offer earlier and
more reliable response assessment than achieved based
on morphology alone.
This review summarizes the applications of PET and
PET/CT imaging in different genitourinary tumors. The
indications and diagnostic accuracies of integrated
PET/CT are discussed and the different radionuclides
used in this type of imaging are addressed.
Prostate Cancer
Prostate cancer is the most common malignancy in men
[7]. The disease is characterized by a widely variable bi-
ological behavior which ranges from clinically silent tu-
IDKD 2010-2013
mors to highly aggressive tumors with rapid growth and
distant metastases. Although different radionuclides have
been proposed, none has evolved as “the” radionuclide
for the imaging of prostate cancer.
The uptake of one such agent, fluorodeoxyglucose
(FDG), was shown in different studies to be increased in
poorly differentiated prostate cancers associated with
high Gleason scores or high prostate-specific antigen
(PSA) levels. However, well-differentiated prostate can-
cers frequently do not demonstrate an increase in gluco-
lysis, with negative FDG-PET scans as a consequence.
Therefore, a positive FDG-PET scan may add relevant in-
formation to morphology alone, when assessing the
pelvis for potential metastases, but a negative FDG-PET
scan has to be interpreted with caution. In addition, even
in FDG-PET-positive tumors, urinary excretion of the ra-
dionuclide may limit the diagnostic accuracy with regard
to assessment of the primary tumor and its potential
infiltration into adjacent organs, such as the seminal
vesicles. Based on its high soft-tissue contrast MRI has
become the modality of choice for imaging the primary
tumor. Its use is further supported by the fact that poorly
differentiated prostate cancers, benign hyperplasia, and
prostatitis show substantial overlap in FDG uptake.
The limitations of FDG in imaging prostate cancer
have led to the introduction of alternative radionuclides
for prostate imaging, shifting the focus away from glu-
colysis to cell-membrane turnover. Accordingly, choline
has become the most commonly used radionuclide in
prostate imaging. Choline is involved in transmembrane
signaling and phospholipid synthesis in cell membranes.
Different malignant tumors show substantial increases in
choline kinase activity, resulting in an increase in mem-
brane phospholipids. Thus, tracers such as 11C-choline or
18F-choline can estimate the proliferative potential of a
tumor by detecting membrane lipid synthesis [8]. An
advantage of 18F-choline over 11C-choline is its longer
half-life and better image quality due to an increase in
spatial resolution, which is a consequence of the shorter
positron range of 18F. However, the urinary excretion of
18F-choline is higher than that of 11C-choline, which has
been attributed to incomplete tubular reabsorption of the
former. Regardless, compared with FDG, both tracers
184 Gerald Antoch

have the advantage of reduced urinary excretion with less
interpretive issues caused by radioactive urine. There
have been reports on the use of 11C-choline in primary tu-
mor assessment and in localization of the primary tumor
within the prostate in patients with negative or unclear
endorectal coil [9]. However, its high soft-tissue contrast
and the availability of MR-spectroscopy favor MRI as the
imaging method of choice for lesion localization within the
prostate gland. The main indication for 11C-choline or
18F-choline is detection of tumor recurrence in patients

findings on biopsy. Some studies have even reported a with rising PSA levels and negative findings on CT
higher accuracy for lesion localization than MRI using an or MRI after treatment of prostate cancer (Figs. 1, 2).

a b

Fig. 1 a-c. CT, 18F-choline-PET, and 18F-choline-PET/CT of a retroperitoneal lymph

node in a patient with an increase in the PSA after radical prostatectomy. Un-
remarkable lymph node on CT (a) diagnosed as metastasis on PET (b) and PET/CT
(c) based on increased choline uptake

a b c

Fig. 2 a-c. a CT, b 18F-choline-PET, and c 18F-

choline-PET/CT of a patient with prostate
cancer and a bone metastasis in the sacrum
Integrated Imaging in Genitourinary Oncology: PET/CT Imaging
In selected patients undergoing diagnosis of prostate can-
cer, both radionuclides may be used for initial tumor stag-
ing. For tumor recurrence, the detection rate of 11C-choline
is reportedly as high as 50% in patients with negative find-
ings on morphological imaging procedures [10]. However,
many authors have reported that 11C-choline detection rates
depend upon the PSA level of the patient, with rising PSA
levels paralleling an increase in tumor detection. Most in-
stitutions schedule PET at PSA levels of at least >1 ng/mL.
Other radionuclides include 11C-acetate, 18F-tes -
tosterone, and 11C-methionine. The mechanism of
11C-acetate uptake by malignant cells is lipid synthesis.
Thus, acetate accumulation has been attributed to an
increase in fatty acid synthesis in prostate cancer cells.
Like 11C-choline, acetate is not a tumor-specific radio-
nuclide, since uptake is increased also in benign hyper-
plasia and prostatitis. As with 11C-choline, 11C-acetate
has been found to be of higher diagnostic accuracy than
FDG in the detection of prostate cancer foci [11].
18F-dihydrotestosterone (18FHDT) and 11C-methionine
are further tracers that have been used occasionally in
imaging prostate cancer. 18FHDT has affinity for the
androgen receptor and can be used to test for resistance
to androgen ablation therapy. 11C-methionine is a non-
specific radionuclide for amino acid transport and pro-
tein synthesis; it has been used in prostate cancer imag-
ing only infrequently.
a b
Fig. 3 a-c. Retroperitoneal lymph node metastasis in a patient with seminoma. Un-
clear finding on CT (a) but increased FDG-uptake on PET (b) and PET/CT (c) in-
dicates malignancy
185
Testicular Cancer
In testicular cancer, FDG-PET and FDG-PET/CT can be
used for initial tumor staging, treatment monitoring, and
the detection of recurrent disease. The advantage of
these modalities in tumor staging has been demonstrat-
ed in patients in whom CT alone indicated stage II dis-
ease (lymph node metastases) based on the detection of
morphologically enlarged lymph nodes. Integration of
FDG-PET into the staging algorithm may prevent false-
positive CT findings in this setting by demonstrating
PET-negativity within such lymph nodes [12]. Nonethe-
less, upstaging of not enlarged, CT-negative lymph
nodes based on FDG avidity within these nodes has
been discussed controversially. FDG-PET and PET/CT
have the potential to detect sub-centimeter metastases in
non-enlarged nodes, but the limited spatial resolution of
PET leads to a decrease in sensitivity with decreasing
lymph node size (Fig. 3). In addition, in some histo-
pathological types of germ cell tumors, such as mature
teratoma, glycolysis is not increased. Accordingly,
FDG-PET and PET/CT are currently not recommended
for routine use in the staging of patients with testicular
cancer [13, 14].
While CT is recommended for treatment monitoring
in germ cell tumors, the addition of FDG-PET or
PET/CT may be helpful in patients with residual lymph
186
node enlargement after therapy. In a prospective trial in-
volving a series of patients with seminoma, FDG-PET
was used to differentiate viable tumor tissue from
scar/fibrosis in patients with residual masses >1 cm af-
ter chemotherapy [15]. There were no false-positives re-
ported. All cases of residual disease seen on CT as masses
>3 cm and 95% of those cases in which the residual
masses were <3 cm were detected correctly with FDG-
PET. Based on these results, FDG-PET has been recom-
mended in seminoma patients to assess residual masses
seen on CT for potential residual disease. It must be em-
phasized, however, that other authors [16] have reported
a substantial number of false-positive findings due to in-
flammation after chemotherapy of seminoma. In patients
undergoing radiotherapy, the issue of inflammatory
changes or tissue regeneration may take on additional
importance. FDG-PET and PET/CT are currently not
Fig. 4. Algorithm of diagnosis and treatment of testicular germ cell tumors modified according to the current recommendations. (From [13])
Gerald Antoch
recommended for therapy response assessment in pa-
tients with non-seminomatous tumors, as mature differ-
entiated teratoma typically has no or only mild FDG up-
take and thus may not be reliably differentiated from scar
tissue or fibrosis [17]. Figure 4 shows the currently rec-
ommended algorithm for the diagnosis and treatment of
testicular germ cell tumors.
FDG-PET can be used in patients with suspected re-
current disease as determined by tumor markers but with
negative CT. In this population, FDG-PET or PET/CT
can detect otherwise occult recurrent tumor sites, as dis-
cussed by Hain et al. [18]. In that study, although FDG-
PET was able to identify tumor recurrence more reliably
than CT, FDG-PET was falsely negative in some of the
patients later diagnosed with recurrence. Therefore, close
imaging follow-up must be recommended in patients
with clinical signs of recurrence but negative FDG-PET.
Integrated Imaging in Genitourinary Oncology: PET/CT Imaging 187

Ovarian Cancer
Ovarian cancer is the second most common gynecolog-
ical malignancy and is frequently detected at advanced
tumor stages, resulting in a poor patient prognosis. Even
those patients diagnosed in an operable stage have a
high risk of tumor recurrence postoperatively, requiring
close patient follow-up, including the measurement of
tumor markers (mainly 125Ca) and evaluation using dif-
ferent imaging procedures. While FDG-PET and
PET/CT have been used to diagnose the primary tumor
and in the differential diagnosis of ovarian cancer from
other ovarian lesions, both imaging modalities are limit-
ed in menstruating women by false-positives due to fol-
licular cysts, cystadenomas, schwannomas, endometri-
omas, or inflammation. In addition, low-grade malig-
nancies and borderline tumors may be FDG-PET-nega-
tive, resulting in false-negative findings. Thus, the re-
ported sensitivities and specificities of up to 87% [19]
must be interpreted with caution. In post-menopausal
women, however, the risk of false-positive findings is
lower and any FDG uptake must initiate further work-up
to exclude tumor growth.
CT has been the standard of care in staging patients
with ovarian cancer pre-operatively, whereas there are
only limited data on a potential benefit of FDG-
PET/CT in this clinical setting and no recommendation
can currently be made. The main focus of hybrid imag-
ing in ovarian cancer has been the diagnosis of recur-
rent ovarian cancer. CT has been the imaging modality
of choice to localize tumor recurrence in patients with
rising tumor markers, but may be supported with func-
tional data in patients with equivocal findings on CT
alone (Fig. 5). Picchio et al. [20] reported an increase
in the sensitivity for detection of ovarian cancer recur-
rence from 70 to 83% and an increase in the specifici-
ty from 83 to 92% when FDG-PET was added to CT.
Similar sensitivities and specificities were reported by
Bristow et al. [21]. Based on the available literature, the
role of FDG-PET and PET/CT in patients with clini-
cally suspected ovarian cancer recurrence is currently
limited to those patients without tumor detection on CT.
Cervical Cancer
Cervical cancer is the third most common malignancy
in women [7]. These tumors manifest as an aggressive
local growth, and the infiltration of adjacent organs is
not infrequent at the time of diagnosis. Metastases of
cervical cancer are primarily found in locoregional lymph
nodes while hematogenous metastases to the lungs, liver,
or bone will be found in rather later stages of the disease.
The role of FDG-PET or FDG-PET/CT in cervical can-
cer has yet to be defined clearly. Although both tech-
niques have been used to identify the primary tumor

a b

Fig. 5 a-c. a CT detected ascites, an indirect sign of peritoneal carcinomatosis, in this

patient with rising tumor markers after therapy for ovarian cancer. b FDG-PET and
c PET/CT show increased FDG uptake inter-intestinally, confirmed as peritoneal
carcinomatosis on further follow-up
188
a b
Fig. 6 a-c. Detection of cervical cancer in a patient with an equivocal CT (a) with
FDG-PET (b) and FDG-PET/CT (c)
(Fig. 6), diagnosis of the primary is most frequently
achieved by Papanicolaou (PAP) test. It is generally ac-
cepted that MRI is the imaging procedure of choice
when assessing a potential invasion of the primary tumor
into adjacent organs. Thus, FDG-PET/CT has not been
used frequently to assess primary cervical tumors. Tu-
mor staging, however, has been considered an indication
for FDG-PET and PET/CT in cervical cancer. The re-
ported sensitivities and specificities vary widely. Sironi
et al. [22] obtained promising results: a sensitivity of
72% in lymph nodes <5 mm and 100% in those >5 mm,
both coupled with high specificities. Nevertheless, there
are more data suggesting a rather low sensitivity for de-
tection of lymph node metastases. For FDG-PET, Wright
et al. [23] reported a sensitivity of 53% for pelvic nodes
and an even lower sensitivity for retroperitoneal/para-
aortic lymph nodes. Even though the addition of mor-
phological data with PET/CT may be considered helpful
over PET alone, Choi et al. [24] reported similar sensi-
tivities for FDG-PET/CT in local lymph nodes (57%).
Therefore, the role of FDG-PET and FDG-PET/CT in
the staging of cervical cancer has not been established.
Tumor recurrence, by contrast, can be detected reliably
with either FDG-PET or PET/CT. Chung et al. [25] re-
ported a sensitivity of 90% and a specificity of 81% for
the detection of cervical cancer recurrence. Following
PET/CT, patient management was altered from that cho-
sen based on a conventional work-up in 24% of patients.
In conclusion, currently available data support the limit-
ed use of FDG-PET and PET/CT in cervical cancer,
mainly in the detection of tumor recurrence.
Gerald Antoch
Endometrial Cancer
The diagnosis of endometrial cancer is established from
biopsy. Even though these tumors are typically FDG-avid,
the role of FDG-PET and PET/CT in assessment of the
primary tumor is limited due to false-positives caused by
normal endometrial FDG uptake, which varies according
to the endometrial cycle. If potential local invasion of the
tumor into adjacent structures needs to be assessed pre-
operatively, this is typically done with MRI, because of its
superior soft-tissue contrast. Despite reports on the use of
FDG-PET and PET/CT for staging local lymph nodes and
distant organs [26], CT currently remains the imaging
procedure of choice for N-staging and M-staging. Further
data are required to define the role of hybrid imaging pro-
cedures in endometrial cancer.
References
1. Jemal A, Thomas A, Murray T, Thun M (2002) Cancer statis-
tics 2002. CA Cancer J Clin 52:23-47
2. Haberkorn U, Schoenberg SO (2001) Imaging of lung cancer
with CT MRI and PET. Lung Cancer 34 Suppl 3:S13-S23
3. Toloza EM, Harpole L, McCrory DC (2003) Noninvasive stag-
ing of non-small cell lung cancer: a review of the current evi-
dence. Chest 123(Suppl 1):S137-S146
4. Adams S, Baum RP, Stuchensen T et al (1998) Prospective
comparison of 18F-FDG PET with conventional imaging
modalities (CT MRI US) in lymph node staging of head and
neck cancer. Eur J Nucl Med 25:1255-1260
5. Marom EM, McAdams HP, Erasmus JE et al (1999) Staging
non-small cell lung cancer with whole-body PET. Radiology
212:803-809
Integrated Imaging in Genitourinary Oncology: PET/CT Imaging
6. van Tinteren H, Hoekstra O, Smit E et al (2002) Effectiveness
of positron emission tomography in the preoperative assess-
ment of patients with suspected non-small-cell lung cancer:
the PLUS multicentre randomised trial. Lancet 359:1388-1393
7. Jemal A, Murray T, Samuels A et al (2003) Cancer statistics
2003. CA Cancer J Clin 53:5-26
8. Sutinen E, Nurmi M, Roivainen A et al (2004) Kinetics of
[(11)C]choline uptake in prostate cancer: a PET study. Eur J
Nucl Med Mol Imaging 31:317-324
9. Yamaguchi T, Lee J, Uemura H et al (2005) Prostate cancer: a
comparative study of 11C-choline PET and MR imaging com-
bined with proton MR spectroscopy. Eur J Nucl Med Mol
Imaging 32:742-748
10. de Jong IJ, Pruim J, Elsinga PH et al (2003) Preoperative stag-
ing of pelvic lymph nodes in prostate cancer by 11C-choline
PET. J Nucl Med 44:331-335
11. Fricke E, Machtens S, Hofmann M, Van Den J (2003) Positron
emission tomography with 11C-acetate and 18F-FDG in prostate
cancer patients. Eur J Nucl Med Mol Imaging 30:607-611
12. Albers P, Bender H, Yilmaz H, Schoeneich G (1999) Positron
emission tomography in the clinical staging of patients with
Stage I and II testicular germ cell tumors. Urology 53:808-811
13. Krege S, Beyer J, Souchon R et al (2008) European consensus
conference on diagnosis and treatment of germ cell cancer: a
report of the second meeting of the European Germ Cell Can-
cer Consensus group (EGCCCG): part I. Eur Urol 53:478-496
14. Krege S, Beyer J, Souchon R P et al (2008) European consen-
sus conference on diagnosis and treatment of germ cell cancer:
a report of the second meeting of the European Germ Cell Can-
cer Consensus Group (EGCCCG): part II. Eur Urol 53:497-513
15. De Santis M, Becherer A, Bokemeyer C et al (2004) 2-18flu-
oro-deoxy-D-glucose positron emission tomography is a reli-
able predictor for viable tumor in postchemotherapy semino-
ma: an update of the prospective multicentric SEMPET trial. J
Clin Oncol 22:1034-1039
16. Lewis DA, Tann M, Kesler K et al (2006) Positron emission
tomography scans in postchemotherapy seminoma patients
with residual masses: a retrospective review from Indiana Uni-
versity Hospital. J Clin Oncol 24:e54-e55
189
17. Stephens AW, Gonin R, Hutchins GD et al (1996) Positron
emission tomography evaluation of residual radiographic ab-
normalities in postchemotherapy germ cell tumor patients. J
Clin Oncol 14:1637-1641
18. Hain SF, O’Doherty MJ, Timothy AR et al (2000) Fluo-
rodeoxyglucose positron emission tomography in the evalua-
tion of germ cell tumours at relapse. Br J Cancer 83:863-869
19. Kawahara K, Yoshida Y, Kurokawa T et al (2004) Evaluation
of positron emission tomography with tracer 18-fluo-
rodeoxyglucose in addition to magnetic resonance imaging in
the diagnosis of ovarian cancer in selected women after ultra-
sonography. J Comput Assist Tomogr 28:505-516
20. Picchio M, Sironi S, Messa C et al (2003) Advanced ovarian
carcinoma: usefulness of [(18)F]FDG-PET in combination
with CT for lesion detection after primary treatment. Q J Nu-
cl Med 47:77-84
21. Bristow RE, del Carmen MG, Pannu HK et al (2003) Clinical-
ly occult recurrent ovarian cancer: patient selection for sec-
ondary cytoreductive surgery using combined PET/CT. Gynecol
Oncol 90:519-528
22. Sironi S, Buda A, Picchio M et al (2006) Lymph node metas-
tasis in patients with clinical early-stage cervical cancer: de-
tection with integrated FDG PET/CT. Radiology 238:272-279
23. Wright JD, Dehdashti F, Herzog T et al (2005) Preoperative
lymph node staging of early-stage cervical carcinoma by
[18F]-fluoro-2-deoxy-D-glucose-positron emission tomogra-
phy. Cancer 104:2484-2491
24. Choi HJ, Roh JW, Seo SS et al (2006) Comparison of the ac-
curacy of magnetic resonance imaging and positron emission
tomography/computed tomography in the presurgical detection
of lymph node metastases in patients with uterine cervical car-
cinoma: a prospective study. Cancer 106:914-922
25. Chung HH, Jo H, Kang WJ et al (2007) Clinical impact of in-
tegrated PET/CT on the management of suspected cervical
cancer recurrence. Gynecol Oncol 104:529-534
26. Horowitz NS, Dehdashti F, Herzog TJ et al (2004) Prospective
evaluation of FDG-PET for detecting pelvic and para-aortic
lymph node metastasis in uterine corpus cancer. Gynecol On-
col 95:546-551

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging
Thomas F. Hany
Department of Radiology, Clinic and Policlinic of Nuclear Medicine, University Hospital Zurich, Zurich, Switzerland
Introduction
The basic principle of positron emission tomography
(PET) is the use of positron-emitting-isotope-labeled
pharmaceuticals that are integrated into a metabolic path-
way. Positron-emitting isotopes are characterized by a be-
ta plus-decay, in which a positron is emitted. This
positron collides with and annihilates any of the many
shell electrons in the neighboring atoms, thereby produc-
ing two 511-keV gamma rays (photons) which are de-
tected in coincidence by the PET scanner. The additional
integration of a computed tomography scanner (PET/CT)
allows the acquisition of PET and CT images of the pa-
tient in the same imaging session. The clinically and most
widely evaluated positron-emitting isotope labeled phar-
maceutical is fluorine-18 fluoro-2-deoxy-D-glucose
(FDG). This glucose analogue is transported into the cell
by specific transporters and phosphorylated by hexoki-
nase to FDG-6-phosphate. As FDG-6-phosphate is inert
to further metabolic processing or to transmembrane
back-transport outside the cell, it accumulates within the
cells. The physical half-life of FDG is around 110 min.
The compound is used as a metabolic marker in oncolo-
gy, cardiology, neurology and inflammation imaging.
One of the most important advantages of this imaging
technique is the extensive anatomical coverage. Thus, im-
ages from the head to the thighs are acquired routinely,
allowing evaluation of the entire body. All of the current-
ly available data indicate that PET/CT is more sensitive
and specific than either of its constituent imaging meth-
ods. With PET/CT, the most relevant additional effect is
that the CT data frequently add specificity to the FDG-
PET data [1].
Malignant Primary Liver Tumors
Hepatocellular Carcinoma
Hepatocellular carcinoma (HCC) is frequently multifocal
such that at the time of exploration multiple sites in both
liver lobes are involved. Pre-operative assessment should
include the search for extrahepatic metastases, because
IDKD 2010-2013
their presence will preclude a curative approach. Inter-
estingly, extrahepatic metastasis of HCC occurs relative-
ly late in the clinical course and therefore does not sig-
nificantly shorten the survival time. Whenever the stage
of disease allows, surgical treatment should be pursued.
Torizuka et al. evaluated patients with known HCC by
using dynamic and static FDG-PET data acquisition [2].
High degrees of correlation were found between the his-
tological grade and the kinetic rate constants as well as
hexokinase activity. There have been only a few studies
involving PET/CT imaging. The largest of these, by Park
et al., consisted of a prospective evaluation of FDG and
11C-acetate PET/CT in the detection of primary and
metastatic hepatocellular carcinoma [3]. In an analysis
based on biopsied lesions, the sensitivity for primary HCC
of FDG-PET/CT was 64.4% while that of 11C-acetate
PET/CT was 84.4%. The overall sensitivities of FDG,
11C-acetate, and dual-tracer PET/CT for 35 metastatic
HCCs were 85.7, 77.0, and 85.7%, respectively. The addi-
tion of 11C-acetate to FDG-PET/CT increased the overall
sensitivity for the detection of primary HCC but not for the
detection of extrahepatic metastases. FDG, 11C-acetate,
and dual-tracer PET/CT had a low sensitivity for the
detection of small primary HCC, but FDG-PET/CT had a
relatively high sensitivity for the detection of extrahepatic
metastases of HCC. Thus, only in selected cases, in which
there is known moderately to poorly differentiated HCC,
is FDG-PET/CT useful in a pre-therapy setting to detect
additional intra- or extrahepatic lesions. The same holds
true for the evaluation of recurrent disease after therapy.
Cholangiocarcinoma
There is limited knowledge on the use of FDG-PET or
PET/CT in the evaluation of intrahepatic cholangiocarci-
noma. This type of hepatic cancer is associated with pri-
mary sclerosing cholangitis. In the study by Kluge et al.,
both the sensitivity and the specificity of FDG-PET and
PET/CT in the detection of primary lesions were >90% [4];
however, FDG-PET showed a poor performance in the
detection of locoregional metastases. In the study by
Fritscher-Ravens et al., there were several false-negatives,
especially in patients with mucinous adenocarcinoma.
Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging 191

Otherwise, the same results regarding locoregional and Pancreas

distant metastases were achieved with FDG-PET and
PET/CT [5]. In the study by Petrowsky et al. [6], 61 pa-
tients with malignancies of the biliary tract, proven by his-
tology or cytology, were evaluated by FDG-PET/CT and
contrast-enhanced CT. PET/CT detected all gallbladder
cancers (n = 14). Overall, 45 of 61 tumors were correctly
identified with PET/CT (sensitivity 74%) and 40 by
contrast-enhanced CT scan (sensitivity 66%). All 12 distant
metastases were detected by PET/CT, but only 3 out of
the 12 by CT (p <0.001) (Fig. 1). The detection of extra-
hepatic cholangiocarcinoma was significantly lower than
that of intrahepatic cholangiocarcinoma for both tech-
niques; also, regional lymph node metastases were detected
by PET/CT in only 12% of the cases and by CT in only
24%. PET/CT findings resulted in a change of management
in 17% of patients with tumors deemed resectable after
the standard workup. Therefore, PET/CT seems particu-
larly valuable in detecting unsuspected distant metastases,
which are not diagnosed by standard imaging.
Exocrine Pancreatic Carcinoma
The staging of pancreatic cancer includes a histological
workup and proof of the typical histological features of
exocrine pancreatic cancers, mostly adenocarcinoma.
However, T-staging essentially is only possible by evalua-
tion of the arterial vasculature, since tumors that show in-
vasion/encasement of the celiac trunk and superior mesen-
teric artery are unresectable (T4). The most important in-
formation from any imaging modality is, therefore, arter-
ial vascular involvement of the primary tumor, the pres-
ence of peritoneal carcinomatosis, and the detection of
distant metastases. Regarding the performance of
PET/CT, the data are thus far rather sparse. In a study by
Farma et al., 82 patients with assumed resectable pancre-
atic cancer underwent staging with non-contrast-enhanced
PET/CT and contrast-enhanced CT of the chest and ab-
domen [7]. The sensitivity and specificity of PET/CT in

a b e

c f

Fig. 1 a-g. A 56-year-old male patient

with a histologically proven intrahep-
atic cholangiocarcinoma. a In the
maximum intensity projection image d g
of the FDG-PET/CT, the primary tu-
mor is well demarcated, with multiple
foci of pathological uptake. In the ax-
ial images at the level of the liver
(from top to bottom: b PET contrast-
enhanced, c CT, d fused image), the
intrahepatic tumor is highly FDG
avid. In the mid-thoracic region (from
top to bottom: e PET, f contrast-en-
hanced CT, g fused image), a medi-
astinal lymph node metastasis is eas-
ily demarcated paraesophageally
192
diagnosing pancreatic itself cancer were 89 and 88%, re-
spectively. The sensitivity of detecting metastases was
61% for PET/CT, 57% for contrast-enhanced CT, and
87% for the two side by side. PET/CT findings influenced
the clinical management of seven patients (11%), all of
whom had distant metastases. As a major drawback of this
study, the CT component of PET/CT was not performed
using contrast-enhanced triple-phase CT. In a study by
Strobel et al., 50 patients with biopsy-proven pancreatic
cancer (adenocarcinoma) were evaluated by integrated
triple-phase contrast-enhanced PET/CT regarding re-
spectability and overall staging [8]. The criteria for irre-
sectability were distant metastases, peritoneal carcino-
matosis, arterial infiltration, or infiltration of neighboring
organs other than the duodenum. Histology, intraoperative
findings, and follow-up CT together with the clinical find-
ings were used as the standard of reference. Accordingly,
27 patients had unresectable disease because of distant
metastases (n = 17), peritoneal carcinomatosis (n = 5), or
local infiltration (n = 5). In the assessment of resectability,
PET alone had a sensitivity of 100%, a specificity of 44%,
an accuracy of 70%, a positive predictive value of 61%,
and a negative predictive value of 100%; unenhanced
PET/CT had respective values of 100, 56, 76, 66, and
100%; the corresponding values for enhanced PET/CT
were 96, 82, 88, 82, and 96%. In five patients, unre-
sectability was missed by all imaging methods and was
only diagnosed intraoperatively. Enhanced PET/CT was
significantly superior to PET alone (p = 0.035), with a
trend for the superiority of enhanced over unenhanced
PET/CT (p = 0.070). Based on all of these data, it is clear
that non-contrast-enhanced PET/CT is able to detect dis-
tant metastases but not local extent in a large proportion
of cases. Furthermore, in a rather considerable proportion
of patients, only intraoperative findings are able to reveal
surgery-precluding factors, such as deep retroperitoneal
infiltration, small liver metastases, and peritoneal involve-
ment. It seems that the most favorable approach consists
of the use of intravenous contrast-enhanced CT in the
PET/CT protocol; however, this does not reflect clinical
practice, since most patients with a suspected pancreatic
lesion rapidly undergo contrast-enhanced CT or magnetic
resonance imaging (MRI) evaluation.
Endocrine Pancreatic Cancer
Neuroendocrine tumors (NETs) of the pancreas account
for <5% of all malignant pancreatic tumors. PET imaging
using FDG has limited value, since these tumors are often
slow-growing, with an accordingly low metabolism [9]. In
addition to morphological imaging modalities, including
contrast-enhanced CT and MRI, somatostatin receptor
scintigraphy (SRS) is used to localize and characterize
NETs in general and such tumors in the pancreas in par-
ticular. Nonetheless, the interpretation of SRS findings
can be challenging due to the difficulty of distinguishing
tumor from intestinal structures and to the variable densi-
ty of somatostatin receptors on the different tumors. Over-
Thomas F. Hany
all, a significant rate of false-negative results – with sen-
sitivities ranging between 50 and 78% – in the detection
of NET (depending on the localization) has been reported
[10]. PET using the catecholamine precursor 6-(fluoride-
18)-fluoro-dopa (18F-DOPA) has been proposed as a valu-
able imaging option for NET [11]. This tracer highlights
the tumor’s intracellular decarboxylase activity and in the
context of PET imaging provides higher spatial resolution
than obtained with SRS. A major drawback of 18F-DOPA
is its physiological uptake by normal pancreas, which in
certain cases obscures detection of the primary tumor. A
rather newly developed PET tracer uses the basic princi-
ple of SRS, i.e., somatostatin labeling, but with 68Ga. In a
study by Ambrosini et al., 18F-DOPA and 68Ga were com-
pared in a rather small group of 13 patients [12]. 68Ga-
DOTA-NOC (tetra-azycyclododecanetetra-acetic acid-[1-
Nal3]-octreotide) was found to be an accurate tracer for
the assessment of NETs and was better than 18F-DOPA in
the detection of primary NETs – especially those of the
pancreas – and metastases. The authors concluded that
since the pancreas is the most frequent site of NETs, the
routine use of 68Ga-DOTA-NOC is more appropriate.
Here, larger, comparative studies that include morpholog-
ical imaging modalities are needed to determine the role
of PET/CT in NETs of the pancreas.
Colorectal Carcinoma
In the western world, colorectal carcinoma is the most
important cause of death due to cancer, after bronchial
carcinoma [13]. About 70% of patients have curable re-
sectable tumor at initial diagnosis and are treated with cu-
rative intent. Approximately 50% of colon cancer patients
will present with hepatic metastases, either at the time of
initial diagnosis or as a result of recurrence [14]. From a
diagnostic perspective, colon cancer and rectal cancer are
often evaluated as a single group; however, especially
deep rectal cancer has a clearly different pathway of lo-
coregional and distant metastases (Fig. 2).
Initial Staging
Two studies using FDG-PET alone for initial staging
demonstrated the high sensitivity of this modality in the
detection of both primary tumor (100 and 96%) and dis-
tant metastases (87 and 78%) and confirmed its low sen-
sitivity (29 and 29%) in lymph node staging [15, 16]. In a
study by Veit-Haibach et al., 47 patients underwent whole-
body PET/CT colonography one day after colonoscopy [17].
Compared with optimized abdominal CT staging alone,
PET/CT colonography was significantly more accurate in
defining TNM stage (difference: 22%; 95% CI: 9-36%;
p = 0.003), which was mainly based on a more accurate
definition of the T-stage. Differences were not detected
for defining N-stage between PET/CT colonography and
CT alone with a threshold of 0.7 cm for malignant nodes
but were detected with a threshold of 1 cm. There were no
Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging
a b
c d
Fig. 2 a-d. A 63-year-old female patient with a
histologically proven deep rectal cancer. Ax-
ial FDG-PET/CT images at the level of the
primary tumor (a PET, b contrast-enhanced
CT, c fused images) reveal the FDG-avid pri-
mary tumor as well as enlarged and FDG-
avid bilateral inguinal lymph node metas-
tases. The lymph node spread is typical for
deep rectal cancer invading the anal canal.
d The same findings are seen in the contrast-
enhanced, T1-fat suppressed MRI images
differences in defining M-stage separately or when the
accuracy of PET/CT colonography was compared to
that of CT + PET. PET/CT colonography affected
consecutive therapy decisions in four patients (9%; 95%
CI: 2.4-20.4%) compared with conventional staging
(CT alone and colonoscopy). Therefore, the combination
of FDG PET/CT in conjunction with a dedicated contrast-
enhanced CT protocol may be of interest as a possible
single-step staging procedure.
Recurrent Disease
The standard patient workup for the detection of recur-
rence and metastases in colorectal cancer includes regular
clinical examinations, CT scans, colonoscopy, and usual-
ly the measurement of tumor markers such as CEA. How-
ever, this approach lacks specificity and may result in di-
agnostic and therapeutic delays due to several pitfalls:
1. while serological tumor markers are useful, CEA lev-
els have only a 60-70% sensitivity for the detection of
colorectal cancer recurrence [18];
2. the morphology based information provided by CT
does not permit distinction between post-surgical
changes and tumor recurrence nor can it detect tumor
involvement of normal-sized lymph nodes [19];
3. colonoscopy is only useful in the detection of local re-
currence.
The suitability of FDG-PET in detecting recurrence
and metastases has been shown in several studies. While
there is obviously a clear advantage of PET/CT over PET
alone, a dedicated contrast-enhanced CT is required often
by clinicians. The study by Soyka et al. showed that con-
trast-enhanced PET/CT, as a single-step examination, has
the same diagnostic confidence and impact as a sequen-
tial approach consisting of contrast-enhanced CT followed
by non-contrast-enhanced PET/CT [20]. Although the
193
lesion detection rate on contrast-enhanced CT images is
high, evaluation solely by this approach can be challeng-
ing because of the likelihood of inconclusive results that
require further diagnostic evaluation (56% of our patient
population). The reason for this is predominantly related
to the specificity of the structural abnormalities that may
be identified by this modality. Thus, nowadays, patients
with inconclusive contrast-enhanced CT findings are be-
ing frequently referred for further evaluation with
18F-FDG PET/CT. More importantly, the study of Soyka
et al. showed that in 21% of the patients with apparently
conclusive findings on contrast-enhanced CT, the addition
of non-contrast-enhanced PET/CT information led to ap-
propriate changes in therapy. In clinical routine, however,
those patients in whom contrast-enhanced CT findings
were regarded as conclusive would not routinely be re-
ferred for further evaluation with FDG-PET/CT. If con-
trast-enhanced PET/CT had served as the initial imaging
modality, 65% of these patients would have had a clear
benefit, including changes in management as well as in
diagnostic confidence. Therefore, it could be argued that
contrast-enhanced PET/CT should be performed as the
first-line diagnostic tool in the re-staging of colorectal
cancer. Conversely, in 35% of the patients the associated
radiation exposure would be futile, in addition to the need-
less expense of the procedure. However, this analysis
holds true only if contrast-enhanced CT and non-contrast-
enhanced PET/CT are performed within 2-4 weeks. In re-
ality, surgeons generally insist on contrast-enhanced CT
studies not older than 4 weeks before taking a patient in-
to the operating room. Thus, another additional scan with
contrast enhancement (CT or PET/CT) would be needed
in the majority of patients.
Post-surgical and radiotherapy changes in the small
pelvis are the most challenging for morphological im-
aging studies in recurrent rectal cancer, since tumor
194
recurrence cannot be differentiated from benign scar tissue.
In a study by Even-Sapir et al., PET/CT was used to dis-
tinguish benign and malignant pre-sacral abnormalities.
The sensitivity, specificity, positive predictive value, and
negative predictive value were 100, 96, 88, and 100%, re-
spectively. PET/CT findings were clinically relevant in
47% of 62 patients [21]. However, there was no compar-
ison to other, conventional imaging studies. In our own
study, the diagnostic value of contrast-enhanced CT and
non-enhanced PET/CT was prospectively evaluated and
compared in 76 patients referred for pre-operative evalu-
ation for liver resection for metastatic colorectal cancer
[22]. Extrahepatic disease was missed by contrast-
enhanced CT in one-third of the cases (sensitivity 64%),
while PET/CT failed to detect extrahepatic lesions in
only 11% (sensitivity 89%; p = 0.02). New findings on
PET/CT resulted in a change in the therapeutic strategy
in 21% of the patients. This study also demonstrated the
well known limitation in PET imaging’s spatial resolution
at ~4-6 mm, since small tumors (<5 mm) were often not
detected. Also, in patients who had received chemothera-
py within the month prior to PET/CT there was a high in-
cidence of false-negative results. This effect, however,
might be used as a predictor of success in neoadjuvant
chemotherapy before resection. In summary, the above-
described studies clearly illustrate the advantages of
PET/CT imaging in colorectal cancer.
Stomach and Small Bowel
Gastric Cancer
In a review by Dassen et al. regarding the pre-operative
diagnostic utility of FDG-PET/CT in gastric cancer, the
authors concluded, that FDG-PET has no role in the pri-
mary detection of gastric cancer due to its low sensitivi-
ty [23]. FDG-PET, however, is slightly better than CT in
the evaluation of lymph node metastases in gastric cancer
and therefore might have a role in pre-operative staging.
Improvements in accuracy could be achieved by using
PET/CT or PET tracers other than FDG, but these strate-
gies need further investigation. Nonetheless, FDG-PET is
able to adequately detect therapy responders at an early
stage following neoadjuvant chemotherapy.
Small Bowel
Adenocarcinoma of the small bowel comprises a group of
rare tumors, with an annual incidence of around 3.7 per
million people. The pathological (development from ade-
nomatous polyps), genetic, and epidemiological features
are very similar to those of adenocarcinomas of the
large bowel but the prognosis is poor, with reported
5-year-survival rates of 15-35%. Capsule endoscopy, CT
enteroclysis, and contrast-enhanced CT are mainly used
in the diagnosis and workup of the disease. While these
tumors show clear FDG avidity and might therefore be
Thomas F. Hany
evaluated by FDG-PET/CT, there are no comprehensive
data on the use of this modality.
Carcinoid tumors, as malignant tumors of non-epithe-
lial origin, have a similar annual incidence as adeno-
carcinomas (3.8 per million people). The different sub-
types have a natural behavior ranging from benign to
high-grade malignancies. Similar to NETs of the pan-
creas, the diagnostic workup consists of nuclear medicine
studies including SRS combined with contrast-enhanced
CT (SPECT/CT) or 68Ga-DOTA-NOC, since >80% of
carcinoid tumors express somatostatin receptors. The
detection rates are therefore similar to those of NET of
the pancreas [12].
Gastrointestinal stromal tumors (GIST) are mesenchy-
mal tumors that in approximately 90% of patients origi-
nate in the stomach and small intestine. Unlike contrast-
enhanced CT, FDG-PET is able to show early effects in
patients undergoing treatment with imatinib mesylate
(Glivec; Novartis, Switzerland) [24]. In two recent stud-
ies, it was shown that patients without FDG uptake after
the start of treatment had a better prognosis than patients
with residual activity not demonstrated with contrast-
enhanced CT [17, 25]. Furthermore, lesions were better
defined on PET/CT than by PET and CT performed side-
by-side. This is relevant information for clinical decision-
making.
Conclusions
In the work-up of several abdominal malignancies of the
gastrointestinal tract, FDG-PET/CT imaging is becoming
increasingly well established. Its main advantage lies in
the comprehensive evaluation of the patient, including all
body compartments, and therefore in the detection of piv-
otal, therapy-deciding lesions. In the evaluation of prima-
ry liver tumors (cholangiocarcinoma and poorly differen-
tiated HCC), FDG-PET/CT offers high sensitivity in the
detection of distant metastases. Secondary liver tumors,
such as metastases from the gastrointestinal tract, are de-
tected by FDG-PET/CT at a high rate, making this imag-
ing technology a primary tool in the evaluation of patients
with suspicion of recurrent colon cancer. Further, full in-
tegration of contrast-enhanced CT protocols improves di-
agnostic confidence and reduces the sometimes cumber-
some diagnostic pathway for patients. New tracers such as
68Ga-DOTA-TATE or 18F-DOPA will bring significantly
improved diagnostic confidence in the notoriously diffi-
cult evaluation of patients with neuroendocrine tumors.
References
1. von Schulthess GK, Steinert HC, Hany TF (2006) Integrated
PET/CT: current applications and future directions. Radiology
238:405-422
2. Torizuka T, Tamaki N, Inokuma T et al (1995) In vivo assess-
ment of glucose metabolism in hepatocellular carcinoma with
FDG-PET. J Nucl Med 36:1811-1817
Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging
3. Park JW, Kim JH, Kim SK et al (2008) A prospective evalua-
tion of 18F-FDG and 11C-acetate PET/CT for detection of
primary and metastatic hepatocellular carcinoma. J Nucl Med
49:1912-1921
4. Kluge R, Schmidt F, Caca K et al (2001) Positron emission to-
mography with (18)F]fluoro-2-deoxy-D-glucose for diagnosis
and staging of bile duct cancer. Hepatology 33:1029-1035
5. Fritscher-Ravens A, Bohuslavizki KH, Broering DC et al
(2001) FDG PET in the diagnosis of hilar cholangiocarcino-
ma. Nucl Med Commun 22:1277-1285
6. Petrowsky H, Wildbrett P, Husarik DB et al (2006) Impact of
integrated positron emission tomography and computed to-
mography on staging and management of gallbladder cancer
and cholangiocarcinoma. J Hepatol 45:43-50
7. Farma JM, Santillan AA, Melis M et al (2008) PET/CT fusion
scan enhances CT staging in patients with pancreatic neo-
plasms. Ann Surg Oncol 15:2465-2471
8. Strobel K, Heinrich S, Bhure U et al (2008) Contrast-enhanced
18F-FDG PET/CT: 1-stop-shop imaging for assessing the re-
sectability of pancreatic cancer. J Nucl Med 49:1408-1413
9. Pasquali C, Rubello D, Sperti C et al (1998) Neuroendocrine
tumor imaging: can 18F-fluorodeoxyglucose positron emis-
sion tomography detect tumors with poor prognosis and ag-
gressive behavior? World J Surg 22:588-592
10. Gibril F, Reynolds JC, Doppman JL et al (1996) Somatostatin
receptor scintigraphy: its sensitivity compared with that of oth-
er imaging methods in detecting primary and metastatic gas-
trinomas. A prospective study. Ann Intern Med 125:26-34
11. Koopmans KP, de Vries EG, Kema IP et al (2006) Staging of
carcinoid tumours with 18F-DOPA PET: a prospective, diag-
nostic accuracy study. Lancet Oncol 7:728-734
12. Ambrosini V, Tomassetti P, Castellucci P et al (2008) Compar-
ison between 68Ga-DOTA-NOC and 18F-DOPA PET for the
detection of gastro-entero-pancreatic and lung neuro-en-
docrine tumours. Eur J Nucl Med Mol Imaging 35:1431-1438
13. Bade MA, Ohki T, Cynamon J, Veith FJ (2001) Hypogastric
artery aneurysm rupture after endovascular graft exclusion
with shrinkage of the aneurysm: significance of endotension
from a “virtual”, or thrombosed type II endoleak. J Vasc Surg
33:1271-1274
14. Clarke MP, Kane RA, Steele G Jr et al (1989) Prospective com-
parison of preoperative imaging and intraoperative ultrasono-
graphy in the detection of liver tumors. Surgery 106:849-855
15. Abdel-Nabi H, Doerr RJ, Lamonica DM et al (1998) Staging
of primary colorectal carcinomas with fluorine-18 fluoro-
195
deoxyglucose whole-body PET: correlation with histopatho-
logic and CT findings. Radiology 206:755-760
16. Kantorova I, Lipska L, Belohlavek O et al (2003) Routine
(18)F-FDG PET preoperative staging of colorectal cancer:
comparison with conventional staging and its impact on treat-
ment decision making. J Nucl Med 44:1784-1788
17. Veit-Haibach P, Kuehle CA, Beyer T et al (2006) Diagnostic
accuracy of colorectal cancer staging with whole-body
PET/CT colonography. JAMA 296:2590-2600
18. Zervos EE, Badgwell BD, Burak WE Jr et al (2001) Fluo-
rodeoxyglucose positron emission tomography as an adjunct to
carcinoembryonic antigen in the management of patients with
presumed recurrent colorectal cancer and nondiagnostic radi-
ologic workup. Surgery 130:636-643; discussion 643-644
19. Goldberg RM, Fleming TR, Tangen CM et al (1998) Surgery
for recurrent colon cancer: strategies for identifying resectable
recurrence and success rates after resection. Eastern Coopera-
tive Oncology Group, the North Central Cancer Treatment
Group, and the Southwest Oncology Group. Ann Intern Med
129:27-35
20. Soyka JD, Veit-Haibach P, Strobel K et al (2008) Staging path-
ways in recurrent colorectal carcinoma: is contrast-enhanced
18F-FDG PET/CT the diagnostic tool of choice? J Nucl Med
49:354-361
21. Even-Sapir E, Parag Y, Lerman H et al (2004) Detection of re-
currence in patients with rectal cancer: PET/CT after ab-
dominoperineal or anterior resection. Radiology 232:815-822
22. Selzner M, Hany TF, Wildbrett P et al (2004) Does the novel
PET/CT imaging modality impact on the treatment of patients
with metastatic colorectal cancer of the liver? Ann Surg
240:1027-1034; discussion 1035-1036
23. Dassen AE, Lips DJ, Hoekstra CJ et al (2009) FDG-PET has
no definite role in preoperative imaging in gastric cancer. Eur
J Surg Oncol 35:449-455
24. Joensuu H, Roberts PJ, Sarlomo-Rikala M et al (2001) Effect
of the tyrosine kinase inhibitor STI571 in a patient with a
metastatic gastrointestinal stromal tumor. N Engl J Med
344:1052-1056
25. Goerres GW, Stupp R, Barghouth G et al (2004) The value of
PET, CT and in-line PET/CT in patients with gastrointestinal
stromal tumours: long-term outcome of treatment with ima-
tinib mesylate. Comparison of PET, CT, and dual-modality
PET/CT imaging for monitoring of imatinib (STI571) therapy
in patients with gastrointestinal stromal tumors. Eur J Nucl
Med Mol Imaging 4:4
NUCLEAR MEDICINE SATELLITE COURSE
“DIAMOND”

Lymphoma: Diagnostic and Therapeutic Applications
of Radiopharmaceuticals
Angelika Bischof Delaloye
Service de Médecine Nucléaire, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland
Introduction
Lymphoma is a heterogeneous family of diseases of the
lymphatic system. There are two main entities, Hodgkin’s
disease (HD) and non-Hodgkin’s lymphoma (NHL), the
latter emerging either from B cells or T cells. Positron
emission tomography (PET) with 18F-fluorodeoxyglu-
cose (FDG) assisted by in-line CT (PET/CT) plays a piv-
otal role in patient management. HD shows high FDG
uptake in close to 100% of the patients whereas FDG
avidity greatly varies among patients with NHL subtypes.
Aggressive lymphomas, such as diffuse large B cell lym-
phoma (DLBCL), the most current form of NHL, and
mantle cell lymphomas (MCL) as well as Burkitt lym-
phomas and even plasmocytomas take up FDG very avid-
ly. Follicular lymphomas (FL) also show high uptake in
more than 90% of the cases whereas extranodal margin-
al zone lymphomas (MZL), such as splenic or mucosa-
associated lymphoid tissue (MALT) MZL, are positive in
only 67 and 54%, respectively [1].
It has been stipulated that the intensity of FDG uptake
is an expression of disease aggressiveness [2-4]. This is
probably true, at least partially, but is not really important
in clinical practice. In fact, a great number of FL, with no
signs of transformation, show high FDG uptake despite be-
ing considered indolent. The intensity of FDG uptake can
therefore not be reliably used for diagnosing transforma-
tion in FL. Due to possible inconsistencies of uptake in-
tensity versus aggressiveness and the great diversity of
NHL, thorough histopathological characterization is nec-
essary to determine the most appropriate treatment
scheme. This step cannot be replaced by FDG-PET studies.
Particularities of Abdominal Lymphoma
Lymphomas of the abdomen and pelvis most often simply
correspond to subdiaphragmatic nodal involvement by
one of the most frequent lymphoma types (Hodgkin’s dis-
ease, DLBCL, FL), usually characterized by high FDG
uptake. However, in the abdomen other types, such as
nodal and extranodal MZL, including MALT lymphomas,
with much less consistently increased uptake need to be
IDKD 2010-2013
considered [5, 6]. Subtypes of MALT lymphomas in par-
ticular have different uptake patterns. Most patients with
MALT in which there is plasmacytic differentiation show
FDG uptake whereas uptake is much less consistent in
typical MALT [7]. Besides histological subtype, location
is another factor that determines uptake patterns. Accord-
ingly, the sensitivity of FDG-PET is lower in gastric
(25-39%) than in non-gastric MALT (46-75%) [8, 9].
Gastric and extragastric MALT lymphomas frequently
present as multifocal disease that is associated with chro-
mosomal translocations or trisomy 18. Dissemination be-
yond the gastrointestinal tract occurs in about 10% of pa-
tients with gastric MALT while significantly more pa-
tients (close to 50%) with extragastric MALT tend to pre-
sent with involvement of other MALT organs [9].
Involvement of the gastrointestinal tract occurs in
10-30% of patients with NHL. The affected sites, in de-
creasing order of frequency, are the stomach, small bow-
el, large bowel, and esophagus. The appearance of these
tumors varies – from polypoidal masses to wall thicken-
ing, diffuse or focal infiltrations, ulcerations, and some-
times solitary or multiple nodules or extension from
mesenteric nodes – and therefore so do the FDG uptake
patterns [4, 5]. The occurrence of normally increased up-
take in the gastrointestinal tract or of non-lymphomatous
pathological conditions further complicates the diagnosis
with FDG-PET [5, 6, 10].
Primary colorectal lymphomas are rare, representing
10-20% of gastrointestinal lymphomas and 0.2-0.6% of
all large bowel malignancies. They are usually FDG-avid
B cell lymphomas, with the most common sites being the
cecum and rectosigmoid [11].
Focal lesions in the liver are frequently observed in pa-
tients with NHL. Civardi et al. studied the prevalence, clin-
ical significance, and role of hepatitis C virus (HCV) in-
fection in a population of 414 consecutive patients with
NHL [12] and 129 focal liver lesions, 76 detected at dis-
ease onset and 53 during follow-up. The prevalence of fo-
cal liver involvement by NHL at disease onset was 7%.
Among the focal liver lesions detected at diagnosis,
30 (39%) were due to lymphoma. In addition, ten cases
of focal liver disease due to other malignancies were
found, seven hepatocellular carcinomas (HCC), and three
200
metastases of another tumor. Of the focal liver lesions,
58% were benign. Conversely, at follow-up, 74% of the
liver lesions were due to NHL and 15% to another malig-
nancy. Focal liver lesions were more often the result of un-
derlying disease (26%) in patients with aggressive lym-
phoma than in those with indolent lymphoma (4%). The
incidence of focal lesions was similar in HCV-negative and
HCV-positive patients, whereas HCC only occurred in the
latter group.
Another particular form of abdominal lymphoma is
enteropathy-associated T cell lymphoma, which is seen in
patients with refractory celiac disease. FDG-PET has
been shown to be superior to CT alone in detecting en-
teropathy-associated lymphoma as well as its distant
spread [13]. Finally, post-transplant lymphoproliferative
disorders (PTLD) need to be mentioned. Immuno-
suppression is an important risk factor in the develop-
ment of PTLD. The WHO has classified PTLD into mor-
phological categories: early lesions, polymorphic PTLD,
or monomorphic PTLD (various types of B and T cell
lymphomas), Hodgkin’s lymphoma, and Hodgkin’s-
lymphoma-like PTLD. The risk that a transplant patient
will develop lymphoma over a 10-year period is 11.8
times greater than in the general population. Early lesions
are most often seen in children and young adults and oc-
cur within the first year of transplantation while the other
types appear later, post-transplantation [14]. PTLD is the
most common malignancy observed in children after
transplantation. Overall, it has been observed in up to
2.3% of kidney transplants, 2.8% of liver transplants,
6.3% of heart transplants, and 20% of small bowel trans-
plants. The clinical presentation is similar to that of non-
transplant-related lymphomas (lymphadenopathy, fever,
weight loss, abdominal pain, splenomegaly). PTLD usu-
ally involves extranodal sites, in particular the gastroin-
testinal tract, and the allograft itself. Treatment consists
of modifying the immunosuppression regimen, chemo-
therapy, and immunotherapy, alone or in combination.
Adoptive T cell immunotherapy is under investigation.
Indications of FDG-PET/CT
FDG-PET is indicated in the staging, response assess-
ment, and re-staging of lymphoma. At diagnosis, the as-
sessment of FDG avidity is probably more important than
the mere staging that is usually performed based on con-
trast enhanced CT. FDG-PET only occasionally leads to
upstaging, by revealing distant sites of involvement in pa-
tients with stage I or II Hodgkin’s lymphoma or NHL, and
therefore to management changes. In advanced disease,
upstaging is less relevant, as the same first-line treatment
is usually administered in stage III and IV disease.
Response Assessment
The most important indication is assessment of the pa-
tient’s response to therapy already after a few treatment
Angelika Bischof Delaloye
courses (chemo- or immunochemotherapy). There are a
consistent number of publications showing that early con-
version to a negative FDG-PET scan, irrespective of the
presence of residual masses on CT, indicates a much
more favorable outcome than in patients showing persis-
tent FDG uptake in their lymphomatous lesions [15-19].
This observation has led to the establishment of new re-
sponse criteria (Table 1); these were originally meant to
be used as surrogate markers for clinical trials only but
are currently also applied in daily patient management
[20-22]. This classification not only eliminates the CRu
(complete response unconfirmed) category but these
combined criteria provide a more accurate response clas-
sification than International Workshop Criteria (IWC)
criteria alone. Indeed, in a multivariate analysis, it repre-
sented the only significant independent predictor of pro-
gression free survival (PFS) [22]. Interestingly, the pre-
dictive power of early FDG-PET (after 2-3 treatment
courses) is higher than that of end of treatment PET. This
suggests that the early disappearance of FDG uptake is
related to the chemosensitivity of the tumor whereas end
of treatment FDG-PET more likely indicates the presence
or absence of viable tumor cells [23]. Figure 1 shows two
FDG-PET scans (MIP, maximal intensity projection) of a
patient with AIDS-related lymphoma; the first, acquired
after the patient had completed chemotherapy, shows
residual uptake in the stomach and spleen. The patient was
treated by more aggressive chemotherapy and splenectomy,
neither of which could hinder the fulgurating progression
of the disease, as shown on the second scan.
The IWC criteria include assessment by contrast-
enhanced CT, which is usually performed at diagnosis.
The discussion about the necessity of a full diagnostic CT
at response assessment is not entirely closed. Most
authors tend to agree, however, that contrast-enhanced
CT is not necessary as a routine procedure but may be
helpful in special cases, such as in the characterization of
small nodes when FDG uptake is difficult to estimate
because of partial-volume effects [24, 25].
Table 1. Response criteria based on International Workshop Criteria
(IWC) and FDG-PET
IWC+PET response Criteria
ACR CR, CRu, PR, or SD (IWC)
PET negative
BM negative if positive prior therapy
PR CR, CRu, or PR (IWC)
PET positive
SD SD (IWC)
PET positive
PD PD (IWC)
PET positive corresponding to CT
abnormality
PET, Positron emission tomography; CR, complete response; CRu,
complete response unconfirmed; PR, partial response; SD, sudden
death; BM, bone marrow; PD, progressive disease; CT, computed
tomography
Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals
a b
Fig. 1 a, b. A 41-year-old patient after chemotherapy for AIDS-relat-
ed diffuse large B cell lymphoma. a Partial response with residual
uptake in the stomach and splenomegaly. b Progression 4 months
later, after intensification of chemotherapy and splenectomy
A main difficulty, however, remains: the definition of
FDG-PET negativity. An independent nuclear medicine
evaluation of the Eastern Cooperative Oncology Group
(ECOG) revealed only moderate reproducibility of the in-
terpretation of interim FDG-PET scans [26]. Some faint
diffuse uptake may persist in residual masses and is due
merely to the presence of these masses, without neces-
sarily being indicative of viable tumors. It may, however,
be difficult to call such a site FDG-negative. For this rea-
son, Mikhaeel et al. proposed an intermediate category,
minimal residual uptake (MRU) [17]. In a series of 121
patients with aggressive lymphoma, a 5-year PFS of 89,
59, and 16% was observed for patients with no, minimal,
and positive residual uptake, respectively, after therapy.
However, this additional category is not widely used; in-
stead, in most reports the results have been designated as
negative or positive. An International Harmonization Pro-
ject (IHP) was convened to discuss the standardization of
clinical-trial parameters in lymphoma based on consen-
sus recommendations derived from the published PET lit-
erature and the collective expertise of its members [27].
These experts recommended visual assessment for deter-
mining PET positivity/negativity. Others could show that
a reduction of the maximum SUV (standardized uptake
value) by 65.7% allowed better separation of patients
with improved event-free survival (EFS) (2-year EFS: 21
vs. 79% in the groups with ≤65.7% >SUVmax decrease,
respectively) than achieved with visual analysis (2-year
EFS: 51 vs. 79% in the PET-positive and PET-negative
groups, respectively), mainly by characterizing indeter-
minate uptake as negative [28].
201
IHP Criteria of Visual FDG-PET Interpretation
• PET is considered positive in case of focal or diffuse
FDG uptake above background in a region where no
physiological uptake should occur.
• Mild and diffusely increased FDG uptake at the site of
a residual mass >2 cm in diameter regardless of local-
ization, with an intensity less than mediastinal blood
pool structures, is considered negative.
• With respect to partial volume effect, uptake above
surrounding background in residual masses <2 cm or
in normal lymph nodes should be considered positive.
• New lung nodules >1.5 cm in a patient with no evi-
dence of lymphoma before therapy should be consid-
ered positive for lymphoma only if their uptake ex-
ceeds that of mediastinal blood pool, except in patients
in whom a complete response is determined in all
known lymphoma sites. In such patients, these “new
lung lesions” most often correspond to infectious or
inflammatory changes.
• Residual hepatic or splenic lesions >1.5 cm should be
considered positive if their uptake exceeds that of liv-
er or spleen, respectively, and negative if their uptake
is equal or lower than that of the surrounding organ.
Diffusely increased splenic uptake (> normal liver) is
compatible with splenic involvement except <10 days
after treatment with cytokines, namely, granulocyte-
colony stimulating factors [29].
• Clearly increased (multi)focal bone marrow uptake
should be considered positive for lymphoma but needs
to be compared to baseline PET (multifocal involve-
ment?). Diffusely increased bone marrow uptake is usu-
ally due to bone marrow stimulation and should not be
considered as indicating lymphoma.
These criteria were essentially developed for patients
with Hodgkin’s lymphoma or aggressive NHL, in which
a management change in patients who did not respond
or only partially responded to first-line therapy might
improve outcome. Persistence of FDG-uptake in most
cases points to therapy-resistant disease, with a high re-
currence rate and reduced PFS (10-50% at 1 year com-
pared with 79-100% in FDG-PET responders) [30, 31].
However, it remains uncertain whether more aggressive
treatment administered at an earlier time point (after 2-3
first-line treatment courses) will be able to improve out-
come to a greater extent than achieved when such treat-
ment is given at first clinical relapse or as consolidation
at the end of first-line therapy. Several trials are currently
under way to study this essential question but no definite
response is available so far. On the contrary, the absence
of residual FDG uptake has a high negative predictive
value in patients with aggressive NHL as well as in those
with Hodgkin’s disease [31, 32]. Therefore, it may well
be possible to reduce the number of treatment courses
and/or involved field irradiation in excellent responders
in order to avoid late side effects, such as other organ
tumors (increased incidence of unilateral and bilateral
breast cancer in female patients previously treated for
202
mediastinal Hodgkin’s disease), or the increased inci-
dence of heart, mainly coronary artery disease.
In patients with primary gastric lymphoma, DLBCL,
or MALT MZL, persistent FDG uptake is not necessari-
ly a sign of residual lymphoma. In a study of 24 patients
examined using FDG-PET and endoscopy at follow-up
after treatment, 11 patients with ulcerative or mucosal le-
sions on endoscopy showed residual FDG uptake while
no lymphoma cells were evidenced by histopathology [4].
Pitfalls in Interpretation
In addition, to those already mentioned above (faint up-
take in residual masses, inflammatory changes in treated
lesions and lung infiltrates, bone marrow stimulation,
normal stomach and bowel uptake, urinary tract activity)
all other well known pitfalls have to be considered
(brown adipose tissue uptake, thymic rebound, granulo-
matous disease, muscle/bone uptake, contaminations,
etc.). A noteworthy situation may be found in patients
treated with immuno-chemotherapy, in whom prolonged
recruitment of inflammatory cells to the tumor may oc-
cur. In vaccination trials, increased FDG uptake in re-
sponding tumors was also observed and must be consid-
ered a favorable sign of response to treatment.
Radioimmunotherapy of Lymphomas
Radiolabeled antibodies directed against the surface anti-
gens CD20 (tositumomab, ibritumomab tiuxetan) and
CD22 (epratuzomab) of B cells have successfully been
used for the treatment of relapsing/resistant NHL, for first-
line treatment as well as for consolidation after first-line
chemotherapy [33-36]. In prior trials, the usual end point
was defined as response to treatment (complete, CR, or par-
tial, PR), with higher response rates, including CR, observed
with radiolabeled than with unlabeled antibodies [36].
Higher response rates were observed in patients with FL
but have also been reported for other, more aggressive
types of lymphoma, such as DLBCL and MCL, as well as
in transformed FL. Radioimmunotherapy is usually well
tolerated, the only major side effects are related to bone
marrow depression, which appears later than with conven-
tional chemotherapy and is marked by a nadir 5-8 weeks
after treatment. However bone marrow depression is most
often easily managed. The incidence of myelodysplastic
syndrome/acute myeloid leukemia (MDS/AML) is compa-
rable to that observed with conventional chemotherapies,
in particular those containing anthracyclines. In patients
achieving CR, very long PFS times have been noted
[37-39]. Thus, it can be expected that a subgroup of pa-
tients with advanced FL, a disease considered incurable
today, might eventually be cured by a combined approach
of chemo(immuno)therapy and radioimmunotherapy [40].
In a recent randomized trial comparing consolidation
with 90Y-ibritumomab tiuxetan (Zevalin, Bayer Schering
Angelika Bischof Delaloye
Pharma, Germany) to controls followed-up by watchful
waiting, median PFS was 2 years longer in treated pa-
tients [41]. Consequently, the indication of consolidation
has been added by most medical regulatory agencies
(FDA, EMEA, Swissmedic). Figure 2 shows the whole-
body scan obtained with 111In-ibritumomab tiuxetan
(Zevalin) 3 days after injection in a patient who underwent
first-line chemotherapy for FL and who was considered
to be in partial remission according to IWC criteria. Radio-
immunoscintigraphy clearly shows uptake in residual
tumor sites of the abdomen. After radioimmunotherapy,
this patient achieved CR, including molecular remission,
which has been ongoing for more than 6 years.
It is also interesting that a significant dose-efficiency
relationship between the dose delivered to the whole
body and bone marrow and PFS has been observed. This
finding underlines the importance of further refining
dosimetry to improve the overall success of radionuclide
therapy [42]. Additional studies are necessary to evaluate
the respective roles of consolidation with Zevalin and
maintenance with non-labeled antibodies such as ritux-
imab.
Furthermore, available data suggest that radioimmuno-
therapy, associated with high-dose chemotherapy, can
be successfully used in conditioning before autologous
stem cell transplantation (ASCT). The data reported
so far regarding reduced-intensity conditioning before al-
logeneic stem cell transplantation have also indicated
Fig. 2. Anterior whole-body scan
obtained 3 days after injection
of 111Indium-tositumomab tiux-
etan (Zevalin) in a patient who
underwent first-line chemother-
apy for follicular non-Hodgkin’s
lymphoma. According to the In-
ternational Workshop Criteria
classification, this patient was
in partial remission. The scan
clearly shows uptake of the la-
beled antibody in the residual
abdominal mass. Consolidation
with 90Yttrium-ibritumomab tiu-
xetan (Zevalin) allowed her to
achieve complete remission, in-
cluding molecular remission,
which has been ongoing for
more than 6 years
Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals
positive results. There is also evidence that radioimmuno-
therapy combined with ASCT can improve clinical out-
come (fewer relapses) without added toxicity; this thera-
peutic strategy therefore represents a very interesting
treatment option for elderly patients, who account for the
majority of the NHL population [43].
Conclusions
Nuclear medicine plays a pivotal role in the management
of patients with Hodgkin’s and NHL, particularly by al-
lowing response assessment at an early stage of treatment
(2-3 courses). The results have high prognostic implica-
tions, with complete metabolic responders having a high-
ly significant better outcome than non-responders.
Radioimmunotherapy using labeled anti-CD20 anti-
bodies combined with chemo(immuno)therapy seems to
represent an efficient treatment option, allowing pro-
longed PFS when administered as consolidation after
first-line therapy. It is also of interest as a conditioning
regimen before ASCT, and potentially in patients under-
going allogeneic stem cell transplantation or in those with
relapsed/resistant NHL.
References
1. Weiler-Sagie M, Bushelev O, Epelbaum R et al (2010) 18F-
FDG avidity in lymphoma readdressed: A study of 766 pa-
tients. J Nucl Med 51:25-30
2. Hutchings M, Loft A, Hansen M et al (2006) Different
histopathological subtypes of Hodgkin’s lymphoma show sig-
nificantly different levels of FDG uptake. Hematol Oncol
24:146-150
3. Schoder H, Noy A, Gonen M et al (2005) Intensity of 18fluo-
rodeoxyglucose uptake in positron emission tomography dis-
tinguishes between indolent and aggressive non-Hodgkin’s
lymphoma. J Clin Oncol 23:4643-4651
4. Yi JH, Kim SJ, Choi JY (2009) (18)F-FDG uptake and its clini-
cal relevance in primary gastric lymphoma. Hematol Oncol Epub
5. Lee WK, Lau EW, Duddalwar VA et al (2008) Abdominal
manifestations of extranodal lymphoma: spectrum of imaging
findings. AJR Am J Roentgenol 191:198-206
6. Tateishi U, Terauchi T, Inoue T, Tobinai K (2009) Nodal status
of malignant lymphoma in pelvic and retroperitoneal lymphat-
ic pathways: PET/CT. Abdom Imaging Epub
7. Hoffmann M, Wohrer S, Becherer A et al (2006) 18F-Fluoro-
deoxy-glucose positron emission tomography in lymphoma of
mucosa-associated lymphoid tissue: histology makes the dif-
ference. Ann Oncol 17:1761-1765
8. Perry C, Herishanu Y, Metzer U et al (2007) Diagnostic accu-
racy of PET/CT in patients with extranodal marginal zone
MALT lymphoma. Eur J Haematol 79:205-209
9. Raderer M, Wohrer S, Streubel B et al (2006) Assessment of
disease dissemination in gastric compared with extragastric
mucosa-associated lymphoid tissue lymphoma using extensive
staging: a single-center experience. J Clin Oncol 24:3136-
3141
10. Takahashi H, Ukawa K, Ohkawa N et al (2009) Significance of
(18)F-2-deoxy-2-fluoro-glucose accumulation in the stomach
on positron emission tomography. Ann Nucl Med 23:391-397
11. Wong MT, Eu KW (2006) Primary colorectal lymphomas.
Colorectal Dis 8:586-591
203
12. Civardi G, Vallisa D, Berte R et al (2002) Focal liver lesions in
non-Hodgkin’s lymphoma: investigation of their prevalence,
clinical significance and the role of Hepatitis C virus infection.
Eur J Cancer 38:2382-2387
13. Hadithi M, Mallant M, Oudejans J et al (2006) 18F-FDG PET
versus CT for the detection of enteropathy-associated T-cell lym-
phoma in refractory celiac disease. J Nucl Med 47:1622-1677
14. Zafar SY, Howell DN, Gockerman JP (2008) Malignancy after
solid organ transplantation: an overview. Oncologist 13:769-778
15. Hutchings M, Loft A, Hansen M et al (2006) FDG-PET after two
cycles of chemotherapy predicts treatment failure and progres-
sion-free survival in Hodgkin’s lymphoma. Blood 107:52-59
16. Jerusalem G, Beguin Y, Fassotte MF et al (2000) Persistent tu-
mor 18F-FDG uptake after a few cycles of polychemotherapy
is predictive of treatment failure in non-Hodgkin’s lymphoma.
Haematologica 85:613-618
17. Mikhaeel NG, Hutchings M, Fields PA et al (2005) FDG-PET
after two to three cycles of chemotherapy predicts progression-
free and overall survival in high-grade non-Hodgkin’s lym-
phoma. Ann Oncol 16:1514-1523
18. Mikhaeel NG, Timothy AR, O’Doherty MJ et al (2000) 18-
FDG-PET as a prognostic indicator in the treatment of ag-
gressive Non-Hodgkin’s Lymphoma-comparison with CT.
Leuk Lymphoma 39:543-553
19. Spaepen K, Stroobants S, Dupont P et al (2002) Early restag-
ing positron emission tomography with (18)F-fluorodeoxyglu-
cose predicts outcome in patients with aggressive non-
Hodgkin’s lymphoma. Ann Oncol 13:1356-1363
20. Brepoels L, Stroobants S, De Wever W et al (2007) Aggressive
and indolent non-Hodgkin’s lymphoma: response assessment
by integrated international workshop criteria. Leuk Lymphoma
48:1522-1530
21. Cheson BD, Pfistner B, Juweid ME et al (2007) Revised re-
sponse criteria for malignant lymphoma. J Clin Oncol 25:
579-586
22. Juweid ME, Wiseman GA, Vose JM et al (2005) Response as-
sessment of aggressive non-Hodgkin’s lymphoma by integrat-
ed International Workshop Criteria and fluorine-18-fluo-
rodeoxyglucose positron emission tomography. J Clin Oncol
23:4652-4661
23. Engles JM, Quarless SA, Mambo E et al (2006) Stunning and
its effect on 3H-FDG uptake and key gene expression in breast
cancer cells undergoing chemotherapy. J Nucl Med 47:
603-608
24. Rodriguez-Vigil B, Gomez-Leon N, Pinilla I et al (2006)
PET/CT in lymphoma: prospective study of enhanced full-
dose PET/CT versus unenhanced low-dose PET/CT. J Nucl
Med 47:1643-1648
25. Schaefer NG, Hany TF, Taverna C et al (2004) Non-Hodgkin’s
lymphoma and Hodgkin’s disease: coregistered FDG PET and
CT at staging and restaging- do we need contrast-enhanced
CT? Radiology 232:823-829
26. Horning SJ, Juweid ME, Schoder H et al (2009) Interim
positron emission tomography (PET) scans in diffuse large
B-cell lymphoma: an independent expert nuclear medicine
evaluation of the Eastern Cooperative Oncology Group E3404
study. Blood Epub
27. Juweid ME, Stroobants S, Hoekstra OS et al (2007) Use of
positron emission tomography for response assessment of lym-
phoma: consensus of the Imaging Subcommittee of Interna-
tional Harmonization Project in Lymphoma. J Clin Oncol
25:571-578
28. Lin C, Itti E, Haioun C et al (2007) Early 18F-FDG PET for
prediction of prognosis in patients with diffuse large B-cell
lymphoma: SUV-based assessment versus visual analysis. J
Nucl Med 48:1626-1632
29. Sugawara Y, Zasadny KR, Kison PV et al (1999) Splenic flu-
orodeoxyglucose uptake increased by granulocyte colony-
stimulating factor therapy: PET imaging results. J Nucl Med
40:1456-1462
204
30. Brepoels L, Stroobants S, Verhoef G (2007) PET and PET/CT
for response evaluation in lymphoma: current practice and de-
velopments. Leuk Lymphoma 48:270-282
31. Hutchings M, Barrington SF (2009) PET/CT for therapy
response assessment in lymphoma. J Nucl Med 50 Suppl 1:
21S-30S
32. Kobe C, Dietlein M, Franklin J et al (2008) Positron emission
tomography has a high negative predictive value for progres-
sion or early relapse for patients with residual disease after
first-line chemotherapy in advanced-stage Hodgkin’s lym-
phoma. Blood 112:3989-3994
33. Kaminski MS, Tuck M, Estes J et al (2005) 131I-tositumomab
therapy as initial treatment for follicular lymphoma. N Engl J
Med 352:441-449
34. Morschhauser F, Illidge T, Huglo D et al (2007) Efficacy and
safety of yttrium-90 ibritumomab tiuxetan in patients with re-
lapsed or refractory diffuse large B-cell lymphoma not appro-
priate for autologous stem-cell transplantation. Blood 110:54-58
35. Sharkey RM, Brenner A, Burton J et al (2003) Radioim-
munotherapy of non-Hodgkin’s lymphoma with 90Y-DOTA
humanized anti-CD22 IgG (90Y-Epratuzumab) do tumor tar-
geting and dosimetry predict therapeutic response? J Nucl
Med 44:2000-2018
36. Witzig TE, Gordon LI, Cabanillas F et al (2002) Randomized
controlled trial of yttrium-90-labeled ibritumomab tiuxetan ra-
dioimmunotherapy versus rituximab immunotherapy for pa-
tients with relapsed or refractory low-grade, follicular, or
transformed B-cell non-Hodgkin’s lymphoma. J Clin Oncol
20:2453-2463
Angelika Bischof Delaloye
37. Buchegger F, Antonescu C, Bischof Delaloye A et al (2006)
Long-term complete responses after 131I-tositumomab thera-
py for relapsed or refractory indolent non-Hodgkin’s lym-
phoma. Br J Cancer 94:1770-1776
38. Fisher RI, Kaminski MS, Wahl RL et al (2005) Tositumomab
and iodine-131 tositumomab produces durable complete re-
missions in a subset of heavily pretreated patients with low-
grade and transformed non-Hodgkin’s lymphomas. J Clin On-
col 23:7565-7573
39. Witzig TE, Molina A, Gordon LI et al (2007) Long-term re-
sponses in patients with recurring or refractory B-cell non-
Hodgkin’s lymphoma treated with yttrium 90 ibritumomab
tiuxetan. Cancer 109:1804-1810
40. Buchegger F, Press OW, Delaloye AB, Ketterer N (2008) Ra-
diolabeled and native antibodies and the prospect of cure of
follicular lymphoma. Oncologist 13:657-667
41. Morschhauser F, Radford J, Van Hoof A et al (2008) Phase III
trial of consolidation therapy with yttrium-90-ibritumomab tiux-
etan compared with no additional therapy after first remission in
advanced follicular lymphoma. J Clin Oncol 26:5156-5164
42. Bischof Delaloye A, Antonescu C, Louton T et al (2009)
Dosimetry of 90Y-ibritumomab tiuxetan as consolidation of
first remission in advanced-stage follicular lymphoma: results
from the international phase 3 first-line indolent trial. J Nucl
Med 50:1837-1843
43. Gisselbrecht C, Vose J, Nademanee A et al (2009) Radioim-
munotherapy for stem cell transplantation in non-Hodgkin’s
lymphoma: in pursuit of a complete response. Oncologist 14
Suppl 2:41-51

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal
and Genitourinary Tract Disorders
Ariane Boubaker
Service de Médecine Nucléaire, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland
Introduction
Conventional nuclear medicine investigations of the
gastrointestinal (GI) tract are not commonly performed in
routine clinical practice, although prevalence of disorders
affecting the upper gastrointestinal tract is quite high
ranging from 15% to 40% in European countries. Most
diagnostic tests used to differentiate organic from nonor-
ganic cause are invasive (endoscopy, manometry, pH
monitoring) and may be not well tolerated by some
patients. Radionuclide procedures are non-invasive and
allow characterizing functional and motility abnormalities
of the esophagus and stomach. They are easy to perform,
widely available and deliver low radiation burden to the
patient at low cost. Despite the lack of standardization,
both esophageal transit studies and gastric emptying
scintigraphy have shown excellent diagnostic results and
are helpful complementary tools for the clinician at
diagnosis and/or follow-up after surgery or conservative
management.
Conventional nuclear medicine investigations of the
urinary system are part of the daily routine clinical prac-
tice and procedures are well established and standardized,
being a common investigation in children with congeni-
tal abnormalities and/or recurrent urinary tract infections.
In adults, dynamic renal scintigraphy with/without ACE
inhibitor is useful to detect haemodynamically significant
renal artery stenosis in patients with suspicion of reno-
vascular hypertension. Preoperative assessment of renal
function (hydronephrosis, live kidney donor, renal tumor)
and follow-up of renal transplanted patients are also
common indications.
Gastrointestinal Tract
Salivary Glands
The main indication to perform salivary-gland scintigra-
phy is xerostomia and suspicion of Sjögren disease. It
may help the clinician in assessing parenchymal dys-
function due to inflammatory disease and/or fibrosis. A
30-minute dynamic acquisition is started after injection
IDKD 2010-2013
of 150 to 200 MBq of Tc-99m pertechnetate with lemon
juice given at 15 min, allowing studying both the extrac-
tion and secretion phase. Quantitative parameters (capta-
tion index, extraction ratio) are helpful to confirm low
parenchymal extraction.
Esophageal Transit
Esophageal transit scintigraphy is useful for the diagno-
sis and evaluation of response to treatment for achalasia,
sclerodermia, esophageal spasm, or dysfunction due to
gastroesophageal reflux disease (GERD) [1-3]. The pa-
tient can be examined in sitting or supine position: po-
sition and gravity will affect the transit-time, the upright
position being more physiologic. Having the patient ly-
ing supine may increase the sensitivity of the procedure.
The patient must have been fasting for at least 3 hours
before examination. He should practice first with un-
labeled liquid to avoid false-positive abnormal transit by
moistening the esophageal mucosa, and get used to the
test. Esophageal motility varies with viscosity and volume
of the bolus: the ejection force of the pharynx is sufficient
to propel a liquid from mouth to stomach, whereas a
semisolid bolus requires more peristalsis from the two
lower thirds of the esophagus, thus increasing the sensi-
tivity of the test. A large field of view cameras is used to
image the entire esophagus including mouth and stom-
ach. If anterior and posterior projections cannot be ac-
quired simultaneously, the anterior projection should be
preferred. The typical procedure consists of a high-rate
dynamic acquisition (0.5 s/frame for 2 min) started just
before swallowing of 10-20 mL of orange juice contain-
ing 5-10 MBq of 99mTc-sulfur colloids, or any radio-
pharmaceutical not absorbed by the gastrointestinal
tract. Both activity and volume have to be adjusted when
examining children. Qualitative evaluation should be
done in cine mode to identify abnormal pattern such as
oral/pharyngeal retention, bolus fragmentation, tracheal
aspiration and gastroesophageal reflux. Esophageal
transit time is quantified by drawing ROIs encompass-
ing the upper, middle and lower third of the esophagus
and the stomach. Normal transit time (>10% activity has
cleared from esophagus) is <15 s.
206
Gastroesophageal Reflux Disease
Clinical presentation of GERD differs considerably in
children when compared to adults [4-6]. In adults, the
main symptom is heartburn. Common manifestations of
reflux in children include respiratory symptoms and
failure to thrive. Gastroesophageal reflux (GER) is
physiologic in infants and resolves spontaneously at 7-
9 months of age, a clinically important reflux being gen-
erally evident by 2 months of age. Serious complica-
tions of GERD may be esophagitis, bleeding, perfora-
tion, Barrett’s esophagus, cancer, stricture or recurrent
pulmonary infections/asthma. The radionuclide method
is the most sensitive non-invasive diagnostic test to de-
tect GER. Because esophageal transit time is an impor-
tant factor that may maintain or facilitate GER, the pro-
cedure should include an esophageal transit scintigra-
phy if possible. Half of the liquid (milk, water, or orange
juice) is labeled with 1 MBq 99mTc-sulfur colloids/kg
(minimum 3 MBq), the other half is used to wash radio-
activity from mouth and esophagus and to complete
the feeding. In case of dual phase study (liquid and sol-
id), one third of the total activity to be administered is
used for the liquid, and two thirds to label the solid
(eggs). The child/patient is placed supine with the cam-
era in the back with stomach and mouth in the field of
view. A 30- to 60-min dynamic acquisition at a rate of
5-10 sec/frame is recommended. A high sensitivity col-
limator should be used to increase the sensitivity of the
test. In older children and adults, GER can be provoked
by an increased abdominal pressure (cough, Valsalva
maneuver). Data should be analyzed in cine mode in or-
der to describe the importance and frequency of the re-
flux. GER can be quantified as a fraction of total initial
gastric activity and/or an overall reflux index. The study
should be completed by 5-minute anteroposterior im-
ages of the thorax to look for possible aspiration. If neg-
ative, 10-15 min additional views of the thorax should
be repeated at 2 and 24 h [7].
Gastric Emptying Scintigraphy
Both rapid and delayed gastric emptying (GE) can cause
the same symptoms [8]. Causes of rapid GE are post-
operative (pyloroplasty, hemigastrectomy), functional
dyspepsia, hyperthyroidism and Zollinger-Ellison syn-
drome. Gastroparesis is defined as a delayed gastric emp-
tying without mechanical outlet obstruction. The most
common causes are diabetes, post-surgical and idiopathic.
The true prevalence of gastroparesis is unknown (around
5%), women being affected more frequently (4:1, females:
males) [9]. Gastric emptying scintigraphy (GES) is con-
sidered as the reference method as it is simple, largely
available and non-invasive and provides reproducible
results of GE measurement [10]. Major efforts have been
made to standardize the procedure because the GE rate is
influenced by the volume and meal content, patient’s con-
dition and position, as well as drug interference [11].
Ariane Boubaker
Once a standardized protocol has been defined, normal
results should be determined by each centre. Solid GES
is generally affected first, being more sensitive for the di-
agnosis than the liquid GES. Liquid emptying is rapid,
begins immediately after ingestion without any lag phase,
and the normal clearance curve pattern is exponential
(normal half-emptying <20 minutes). A normal solid
gastric emptying is characterized by an initial lag-phase
with no emptying (normal 5-25 minutes), during which
the solid material is converted into chime by acids and
mechanical grinding, followed by a linear constant-rate
clearance. 99mTc-sulfur colloids are used because they are
not absorbed in the gastrointestinal system. The meal
should be labeled by cooking the radiopharmaceutical
with proteins (i.e., egg albumin) to avoid elution from the
solid to the liquid phase which would result in erro-
neously shortened GE. After a fast of 4 hours, the patient
must eat the meal within 10-15 min and be positioned im-
mediately after the end of the meal. The use of simulta-
neous antero-posterior projections using a double-headed
camera will allow correcting for movements and geomet-
ric variations in counting. The total duration of the study
is 3 h. Images should be displayed to describe the pro-
gression of the solid within the different portions of the
stomach and the time of activity appearance in the duo-
denum. Additional events like gastroesophageal reflux,
retention in esophagus/hiatal hernia should be described
(Fig. 1). A time activity curve is generated using ROIs
and 50% emptying time is calculated.
Gastrointestinal Bleeding
Lower GI bleeding accounts for one third of all acute
bleeding events, is more frequent in men, and increases in
incidence with age with a mortality of about 4% [12]. The
most common cause of acute lower bleeding is diverticu-
losis, followed by angiodysplasia. Because GI bleeding is
often intermittent and possibly occurs at a slow rate,
scintigraphy is particularly suitable for the accurate
diagnosis having the ability to detect bleeding flow rates
<0.1 mL/min. 99mTc-sulfur colloids can be used to local-
ize the source of bleeding if it occurs during the first
minute after injection. Scintigraphy with 99mTc labeled
red blood cells (RBCs) is the best method in patients with
intermittent, low rate GI bleeding and has been proven to
be superior to barium enema, angiography, computer
tomography (CT) scan and even colonic endoscopy in this
clinical setting [13]. It is recommended to use the in vitro
labeled RBC to minimize elution of 99mTc from the RBCs,
thus avoiding false-positive results due to secretion of free
pertechnetate by the kidneys, and gastric and colonic mu-
cosa [14]. Scintigraphy gives the possibility to perform
successive acquisitions till 24 h after injection and contin-
uous monitoring if bleeding does not occur during the ini-
tial acquisition period. Single photon emission CT
(SPECT) or SPECT/CT may help to localize the bleeding
site. The diagnosis is made by detection of an extravasation
site of RBCs that increases with time. Both anterograde
Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders 207

a 6’ 7’ 8’ 9’ 10’

26’ 27’ 28’ 29’ 30’

Fig. 1 a-c. Solid gastric emp-

tying scintigraphy per- b
formed in a 69-year-old
woman presenting signs
and symptoms of gastro-
paresis (early satiety, bloat-
ing and post-prandial ab- 1h 1.5h 2h 3h
dominal fullness). a Con-
secutive 1-minute posterior
projections showing gastro-
esophageal reflux at 6-7
and 28-30 min. b Anterior
and posterior images at 1 h, c 100
1.5 h, 2 h and 3 h after in-
Gastric emptying (%)

gestion of the 99mTc la-

beled solid meal reveal re- 80
tention in the fundus and
body of the stomach. c The
time-activity curve correct- 60
ed for radioactive decay
confirms a delayed gastric 40
emptying with a dediffer-
entiation of the initial lag-
phase when compared to 20
normal values (dashed lines 60 120 180
representing ± 1 SD) Time (min)

and retrograde movements may occur in the intestinal lu-
men (Fig. 2). Dynamic imaging is essential to correctly
identify the active source of bleeding. Colonic bleeding is
seen at the periphery of the abdomen whereas bleeding in
the small intestine is more central with rapid visualisation
of curvilinear progression of the activity.
Meckel’s Diverticulum and Heterotopic Gastric Mucosa
99mTc pertechnetate (99mTcO4–) scintigraphy has been used
since the 70s to diagnose heterotopic gastric mucosa
[12-15]. Mucin-secreting cells of the stomach and proximal
small bowel are responsible for the uptake and secretion of
99mTcO – via the NIS (sodium iodide symporter system).
4
Unexplained gastrointestinal bleeding and/or recurrent ab-
dominal pain are the main indication to perform 99mTcO4–
scintigraphy to look for Meckel’s diverticulum (MD) or
heterotopic gastric mucosa (HGM). The prevalence of MD
in the general population is estimated to be 1-4%, with
HGM being present in about 50% of MD. Reported sensi-
tivity of scintigraphy range from 50% to 92% and phar-
macological preparation with pentagastrin, glucagon, or
histamine type-2 (H2) receptor antagonists (cimetidine,
ranitidine) have been used to increase the sensitivity of the
test [16]. Pentagastrin is administered subcutaneous 15-20
min prior to scintigraphy (6 μg/kg) to increase the uptake
99mTcO – by the gastric mucosa; H receptor blockers in-
4 2
crease the uptake in the mucosa by blocking the secretion
from the cells to the lumen; ranitidine (2 mg/kg in children,
150 mg in adults) should be preferred to cimetidine be-
cause it has fewer side-effects. Glucagon relaxes the
smooth muscle of the gastrointestinal system and decreas-
es peristalsis (50-6 μg/kg intravenous 10 min after
99mTcO – injection). Omeprazole is a proton pump in-
4
hibitor which has been also reported to increase the sensi-
tivity of the MD scintigraphy. Patient should be fasting for
3-4 hours, and barium enema studies should not be
performed during the 3-4 days prior to scintigraphy. The
camera is positioned in anterior projection with stomach
and bladder in the field of view. A 60×60 sec/frame
dynamic acquisition is started after injection of 2 MBq/kg
(10-150 MBq) of 99mTcO4– and additional static views
(post void, erect, oblique, SPECT) are obtained. A bladder
full of non-radioactive urine prior to injection of the
208 Ariane Boubaker

15” 30” 45” 1’ 1’15”

16’ 45’ 1h20’ 1h45’ 2h05’

Fig. 2 a, b. 99mTc labeled red blood cells (RBCs) bleeding scintigraphy performed in a 71-year-old woman presenting acute lower gastroin-
testinal bleeding. Colonic endoscopy performed two days before scintigraphy did not allow localizing the source of active bleeding. a Ini-
tial dynamic acquisition started immediately after injection of 920 MBq of 99mTc-RBCs shows a diffuse activity in the periphery of the ab-
domen (arrow) at 45 sec that increases with time. b Consecutive anterior static views show both anterograde and retrograde progression of
the activity in the transverse and sigmoid colon confirming the left descending colon to be the source of active bleeding

radiopharmaceutical may delay the renal excretion of
pertechnetate in the urinary tract and avoid false-positive
results during the dynamic acquisition. The diagnostic cri-
terion for HGM is the appearance of a focal uptake of
99mTcO – at the same time as the gastric mucosa (Fig. 3).
4
MD is usually seen as a focal increased activity in the peri-
toneal cavity most frequently in the right lower part of the
abdomen. Pitfalls in interpretation and causes for false-
positive and false-negative results are listed in Table 1.
Protein-Losing Enteropathy
An excessive loss of protein in the gastrointestinal system
may be due to lymphatic obstruction (intestinal lym-
phangiectasia, cirrhosis), inflammatory bowel disease
(Crohn’s disease, ulcerative colitis) gastrointestinal ma-
lignancy (gastric cancer, lymphoma) and increased per-
meability (celiac disease, infections). Serial abdominal
images acquired after intravenous injection of 99mTc la-
beled human serum albumin may show tracer extravasa-
tion and accumulation in the intestine helping to localize
the site of excessive loss of protein.
Hepatobiliary Scintigraphy
Hepatobiliary scintigraphy (HBS) is an imaging tech-
nique performed using 99mTc labeled iminodiacetic acids
derivatives, the most commonly used being mebrofenin
(Bridatec, GE Healthcare, The Netherlands) [17, 18]. Af-
ter intravenous injection, the radiopharmaceutical is
rapidly extracted at the vascular pole of the hepatocyte
and secreted at the biliary pole of the cell. Although not
frequently performed in clinical routine practice, HBS
may be useful in a variety of clinical situations. In neonates
past the age of 2 weeks with persistent jaundice and
hyperbilirubinemia, HBS may be used to distinguish the
treatable causes of hepatitis (hypothyroidism, sepsis,
panhypopituitarism, galactosemia) from extrahepatic
biliary atresia, as early surgery performed before the age
of 2 months offers a better prognosis. The child has to fast
for at least 2 hours before injection and should receive a
phenobarbital pre-treatment (5 mg/kg/day) 3 to 5 days
before HBS in order to stimulate excretion of the radio-
pharmaceutical and avoid false-negative results. A dynamic
acquisition of 60×1 min/frame in anterior projection is
started immediately after intravenous injection of 2-7
MBq 99mTc-mebrofenin/kg (minimal recommended
activity 20 MBq), with consecutive static views per-
formed at 2, 4, 6 and 24 hours. A normal HBS includes
a rapid extraction in the liver with no significant cardiac
blood-pool activity at 5-10 minutes post injection and
the visualization of activity in the small bowel at 60
minutes. In the neonate the intrahepatic main bile ducts
are usually not seen. The diagnostic value of HBS relies
Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders 209

1’ 2’ 3’ 4’

Fig. 3 a, b. Meckel’s divertic- 5’ 6’ 7’ 8’

ulum scintigraphy perform-
ed in a 2-year-old child 24 b
hours after a first episode
of acute gastrointestinal
bleeding. a Initial dynamic
acquisition demonstrates a
focus of activity in the low-
er left quadrant of the ab-
domen that appears simul-
taneously to gastric mucosa
(arrows). b Left and right
lateral and antero-posterior
static views obtained 1 hour
after injection of 60 MBq of
99mTc-pertechnetate show
persistent focal uptake high-
ly suggestive of a Meckel’s
diverticulum containing
gastric mucosa as the
source of acute gastroin-
testinal bleeding

Table 1. Pitfalls and sources of error in the interpretation of Meckel’s on its high-negative predictive value in case of free
diverticulum scintigraphy excretion in the small bowel, whereas an absence of
False-negative results intestinal activity at 24 hours is not specific for biliary
1. Procedures atresia and may be due to other causes of hepatocellular
Barium studies (3-4 days prior to scintigraphy) dysfunction. Pitfalls in interpretation are mostly related
Administration of potassium perchlorate (thyroid blocade) to the urinary excretion of the tracer, posterior projec-
2. Anatomical causes
Ischemia, necrosis, ulceration tion or SPECT (SPECT/CT) being helpful to avoid
Lack of heteropic gastric mucosa false-positive results. Reported sensitivity and speci-
Obscured by urinary tract or full stomach ficity of HBS for the diagnosis of biliary atresia are 83
False-positive results to 100%, 33 to 100%, respectively [19]. The gold stan-
1. Procedures dard is liver biopsy with a sensitivity of 89-99% and a
Endoscopy
Laxative specificity of 83-98%. Common indications for HBS in
2. Anatomical causes adults are acute cholecystitis, biliary leak/extravasation
Urinary tract activity (hydronephrosis, ectopic kidney, bladder after surgery and follow-up after liver transplantation. A
diverticulum) normal HBS is characterized by a diffuse activity in the
Small bowel obstruction (intussusception, volvulus)
Neoplasm (carcinoid, lymphoma, colic adenocarcinoma)
liver beginning 6-8 s after spleen and kidneys (75% of
Inflammation (Crohn’s disease, appendicitis, peptic ulcer) blood supply to the liver supported by the portal vein),
Other site of heterotopic gastric mucosa (duplication, Barrett’s a rapid blood pool clearance (5-10 min), visualization of
esophagus) common bile duct and gallbladder 10-30 min after in-
Vascular abnormalities (angiodysplasia, hemangioma, aneurysm) jection. Transit from the biliary ducts to the small bowel
210 Ariane Boubaker

occurs within 30-45 min. Non-visualization of the gall-
bladder is a diagnostic criteria for acute cholecystitis, but
may be related to an insufficient (<2 h) or too long (>24 h)
fasting period, bile duct obstruction and severe hepatocel-
lular dysfunction. A partial biliary obstruction may be sus-
pected in case of delayed biliary to bowel transit beyond 60
min: other causes are opiates drugs, chronic cholecystitis,
dysfunction of the sphincter of Oddi. Some medications
are known to decrease gallbladder contraction and should
be interrupted before HBS: morphine, atropine, nifedipine,
indomethacin, benzodiazepine, octreotide.
Urinary Tract
Urinary Tract Dilation
With the increased use of prenatal ultrasound, the num-
ber of neonates diagnosed with unilateral or bilateral
mild to moderate pelvic dilation has tremendously in-
creased during the past years. The recommendation al-
gorithm for post-natal examinations aimed first at con-
firming the dilation with US performed at 2 and 7 days
of life, and to exclude other renal abnormalities such
as urethral valves, vesicoureteric reflux, multicystic dys-
plasia, duplex kidney and primary mega-ureter [20].
Unilateral hydronephrosis due to pelviureteric junction
stenosis is the most frequent congenital malformation. It
is a benign disease that will spontaneously regress in up
to 70% of the cases, and surgery should be done in se-
lected children with renal function deterioration and/or
complications (recurrent abdominal pain, infections)
[21]. Diuretic renography using tubular tracers (99mTc-
MAG3, 123I-hippuran) is up to now the only non-invasive
diagnostic modality that gives crucial information on re-
nal function and urinary flow during a single procedure,
even in neonates and young infants. It is simple, repro-
ducible and there is no need to insert an intravenous can-
ula or a bladder catheter. In a well-hydrated child, the si-
multaneous injection of radiotracer and furosemide
(1mg/kg) will enable to have the child voiding at least
one time during the examination, which is mandatory in
assessing urinary flow (Fig. 4). The procedure is well
standardized and can be used from 3-4 weeks of life, and
repeated at 1-month interval during the first 6 months
when needed [22]. There is no consensus on the use of
camera-based methods to measure absolute renal func-
tion, and the EANM recommendation is to use isotopic
clearance methods to measure glomerular filtration rate
(GFR) and/or effective renal plasma flow (ERPF) when
absolute function has to be checked [23]. In adults, a
unilateral hydronephrosis may be diagnosed by chance

1’ 2’ 5’ 10’ 20’
a

100 50

80 40
Count rate (Cts/s)

60 30

40 20

20 10
b c d
0 0
0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200

e f Time (s)

Fig. 4 a-f. 123I-hippuran F0 dynamic renography performed during follow-up in a 9 months old infant with congenital left pelviureteric junction
stenosis (PUJS) that was treated conservatively. a Consecutive 1-minute posterior views obtained 1, 2, 5, 10 and 20 minutes after injection of
9 MBq of 123I-hippuran and 1 mg/kg of furosemide show a normal right kidney and a preserved parenchymal extraction by the left dilated
kidney. Bladder activity is present at 5 minute. The clearance from the right kidney is normal, whereas there is clearly a delayed urinary out-
put from the left kidney. Delayed posterior static views obtained at 20 min (b), after micturition (c), and at 1 hour after injection (d) show a
progressive dilution of radioactive urine in the dilated renal pelvis that persists unchanged despite change of position and micturition. Left kid-
ney time-activity curve (e) is cumulative and confirms the preserved renal function of the parenchyma. The time-activity curve of the normal
contralateral kidney (f) demonstrates both rapid extraction and secretion of the radiotracer with a time-to-peak less than 3 minutes
Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders 211

on ultrasound or CT-scan performed for other reason and
diuretic renography is valuable to assess renal function
and urinary flow.
Urinary Tract Infection and Vesicoureteric Reflux
Urinary tract infection (UTI) is an important routine clin-
ical problem in paediatrics: its prevalence in children with
fever ≥38.5 ranges from 1% to 20% depending mainly on
sex and age, being more frequent in girls and in children
aged less than 2 years [24]. Symptoms and signs are often
non specific, the major challenge being to differentiate
cystitis from acute pyelonephritis (APN). The risk for per-
manent renal damage is related to the delay between the
onset of infection and treatment. 99mTc-DMSA scintigra-
phy is the gold standard for cortical renal imaging. It is
easy to perform, does not require sedation or special pa-
tient preparation and has a sensitivity ranging from 80%
to 100% to detect parenchymal defects due to infection,
but does not allow to distinguish between APN and renal
scars: a normal 99mTc-DMSA scintigraphy during the
acute phase has a high negative predictive value for late
scarring, whereas an abnormal scan during the acute
phase is not predictive of long-term outcome [25]. Vesico-
ureteric reflux (VUR) is a common cause of recurrent UTI
and can be treated either conservatively in neonates and
young infants (spontaneous resolution in 45-70% of cas-
es) or surgically. Voiding cystourethrography (VCUG) is
the reference method for diagnosis and grading of VUR.
Direct radionuclide cystography delivers a lower radiation
dose and is more sensitive than VCUG because of contin-
uous acquisition during filling and voiding phases [26].
The disadvantages are the lack of anatomical information
and the invasiveness and non physiologic condition due
to bladder catheterisation. Indirect radionuclide cysto-
graphy is performed after a dynamic renography with
99mTc-MAG3 or 123I-hippuran [27]. It is less sensitive and
specific than direct cystography, but is more physiologic
and non-invasive.
Acute Renal Failure
Dynamic renography may help the clinician in patients
presenting acute renal failure (ARF) by assessing the po-
tential for renal function recovery. ARF due to acute
tubular necrosis (ATN) has an overall good prognosis
when compared to cortical necrosis or loss of nephronic
mass due to recurrent cholesterol embolism for example. In
case of ATN, diuretic renography will demonstrate pro-
gressive and symmetric tracer uptake in both kidney and
delayed parenchymal retention (Fig. 5). Post-renal cause
of ARF may be difficult to exclude in anuric patients,

a 1’ 2’ 5’ 10’ 20’
220
200
180
Count rate (Cts/s)

160
140
120
100
80
60
40
20
b c d
0
0 200 400 600 800 1000 1200
Time (s)

Fig. 5 a-d. Acute renal failure in a 46-year-old man after acute rhabdomyolisis. 123I-hippuran dynamic renography was performed to evalu-
ate the potential for renal functional recovery. a Consecutive 1-minute posterior views obtained 1, 2, 5, 10 and 20 minutes after injection
of 46 MBq of 123I-hippuran show a symmetric and heterogeneous renal parenchyma with increase of activity in the renal cortex despite
appearance of radioactive urine in the bladder at 5 min pi. There is no urine retention in the renal pelvis or ureters. Static views obtained
at 20 min (b) and 8 hours (c) after injection show persistent retention of the radiotracer in the renal parenchyma, typical of acute tubular
necrosis (ATN). Time-activity curves of both left and right kidneys (d) show preserved initial parenchymal extraction with delayed secre-
tion consistent with a good potential for renal function recovery
212
because ultrasound may not show significant dilation. In
such patients, dynamic renography can be a sensitive tool
by allowing delayed acquisition until 24 h pi and possi-
bly demonstrate urinary retention in renal pelvis and/or
ureters even at very low rate of urine output.
Renovascular Hypertension
Although the prevalence of renovascular hypertension
(RVH) in non selected patients is less than 1%, 15-45%
among patients referred to a specialty center, for refracto-
ry hypertension will have RVH. When a stenosis becomes
hemodynamically significant, the glomerular filtration
rate (GFR) of the affected kidney decreases and the renin-
angiotensin system is activated in order to maintain the
GFR by vasoconstriction (angiotensin II) of the efferent
arterioles. Giving an angiotensin converting enzyme in-
hibitor (ACEI) will block the compensatory mechanism in
the affected kidney, and provoke a decrease of GFR. The
ACE-inhibitor dynamic renography will show typically a
shift of the time-to-peak, a delayed production of urine
and parenchymal retention of the tracer [28]. Oral capto-
pril given 1 hour prior to renography has been used for
many years, but variable absorption may occur. Intra-
venous enalapril (40 μg/kg, maximum 2.5 mg) infused
over 3-5 minutes is more reliable. The test accuracy can
be improved by administration of 20 mg furosemide dur-
ing renography. A normal ACEI renography should be re-
ported as low-probability for RVH disease, and a baseline
study is not mandatory. An equivocal result should be re-
ported when the baseline renography is abnormal and
there is no significant change under ACEI. A high proba-
bility study (significant change under ACEI when com-
pared to baseline) is a strong indicator of potential im-
provement after angioplasty or surgery. The accuracy of
the test is significantly decreased in patients with poor
renal function and/or a small shrunken kidney.
Renal Transplantation
Live Kidney Donor
Live kidney donation has become more frequent over the
last years to cover the increasing numbers of patients
waiting for renal graft. The first objective is not to harm
a previously healthy person, the preoperative assessment
of the potential donor is mandatory including psycho-
logical and biological examinations [29]. Ablation of a
kidney will usually lead to a loss of about 25% of renal
function if both kidneys are equally participating to the
global function. The role of conventional nuclear medi-
cine procedure in the live kidney donor is to assess renal
function and to select the best functioning kidney to be
left in the donor. Isotopic clearance using 51Cr-EDTA for
GFR and 99mTc-MAG3 or 123I-hippuran for tubular ex-
traction rate/effective renal plasma flow (ERPF) mea-
surement have not been used extensively so far, but may
be used in case of non conclusive creatinin clearance
Ariane Boubaker
measurements or in patients with borderline renal func-
tion [30]. Dynamic renography with tubular tracers is
useful to evaluate the renal parenchyma, to calculate rel-
ative renal function and to assess urinary flow. When
ultrasound and angio-CT do not reveal any significant
anatomical abnormality, the choice of the donated kid-
ney is usually based on the results of the renography: the
kidney with the best function will be left in the donor.
Evaluation of the Renal Graft
The most common causes of early and late complications
after renal transplantation are listed in Table 2. Dynamic
renography with tubular tracers can be performed during
the first 24 to 48 hours after transplantation and repeated
if necessary, as no contrast medium injection, potentially
nephrotoxic, is required. It allows verification of
parenchymal function: signs of acute tubular necrosis (cu-
mulative curve due to parenchymal retention of the trac-
er) will be present in most patients. Renal function should
be quantified either by concomitant plasma clearance
(GFR, ERPF) or by quantification based on the renogram.
The procedure should be standardized in order to allow
comparison of the results during follow-up. Acute rejec-
tion may be difficult to distinguish from acute tubular
necrosis in the first days after operation: a worsening of
the vascular phase, decrease of renal parenchymal extrac-
tion and worsening of parenchymal retention will be eas-
ier to diagnose if a baseline study is available. Renogra-
phy is a very sensitive method to diagnose a urinary leak
even at very low urine flow. If 99mTc-MAG3 is used, de-
layed acquisitions may be difficult to interpret because of
hepato-biliary excretion: bowel activity may mask urinary
activity. Whereas arterial thrombosis is a rare complica-
tion (<1%), renal artery stenosis (RAS) has been reported
in up to 23% of renal allografts [1, 24]. ACE-inhibitor
scintigraphy is useful to determine if systemic hyperten-
sion is dependent on the renin-angiotensin system, thus
allowing proper clinical management (Fig. 6). When ob-
struction is suspected, the use of diuretic renography may
help in a similar manner as in native kidneys.
Table 2. Common complications after renal transplantation
Early complications (delayed graft function or failure to improve)
Acute tubular necrosis
Acute rejection
Vascular occlusion
Urinary leak/urinoma
Hematoma
Drug toxicity
Late complications (decrease of renal function)
Chronic allograft nephropathy (chronic rejection)
Drug toxicity
Renovascular hypertension (renal artery stenosis)
Urinary tract infection
Vesicoureteric reflux
Obstructive uropathy (intrinsic, lymphocoele)
Bladder dysfunction (postvoiding residue, small bladder volume)
Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders 213

a
1’ 2’ 5’ 10’ 20’

b
120 500

100
400
Count rate (Cts/s)

80
300

60

200
40

100
20

0 0
0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200
c Time (s) Time (s)
280 280

240 240
Count rate (Cts/s)

200 200

160 160

120 120

80 80

40 40

0 0
0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200
d Time (s) Time (s)

Fig. 6 a-d. Baseline and ACE-inhibitor 123I-hippuran dynamic renography performed 3 weeks after renal transplantation in a 53-year-old
woman presenting a severe systemic hypertension and episodes of acute left cardiac failure. a At baseline, the renal transplant shows pre-
served parenchymal extraction (1 min pi) with prompt urine output (5 min) and rapid washout (10 and 20 min). b Under ACE-inhibitor
(2.5 mg of enalapril intravenously over 5 minutes) there is a clear delay in urinary output (faint bladder activity at 10 min) and significant
retention of the tracer in the renal cortex. c Time-activity curves obtained at baseline study show normal pattern of the transplanted kid-
ney (grey curve), and rapid bladder filling (dotted curve). d Under ACE-inhibitor, time-activity curve of the renal transplant is clearly ab-
normal with preserved initial extraction phase and clearly delay of secretion (no time-to-peak, lack of descending curve). The bladder time-
activity curve (dotted line) shows no significant activity till 10 min pi and low urine output. Renovascular hypertension was diagnosed and
surgery confirmed a narrowing of renal artery at the site of anastomosis
214
References
1. Ziessman HA, O’Malley JP, Thrall JH (2006) Nuclear medi-
cine: The requisite. Radiology, 3rd edn. Elsevier Mosby,
Philadelphia, PA, USA
2. Mariani G, Boni G, Barreca M et al (2004) Radionuclide gas-
trooesophageal motor studies. J Nucl Med 45:1004-1028
3. Nakajima K, Inali A, Hiramatsu T et al (2009) Esophageal
transit scintigraphy and structured questionnaire in patients
with systemic sclerosis with endoscopically proven reflux
esophagitis. Ann Nucl Med October 24 Epub ahead of print
4. Boz AB, Aydin F, Celmeli F et al (2009) Does gastro-
esophageal reflux scintigraphy correlate with clinical findings
in children with chronic cough? Nulc Med Commun 30:
802-806
5. Morigeri C, Bhattacharya A, Mukhopadhyay K et al (2008)
Radionuclide scintigraphy in the evaluation of gastrooe-
sophageal reflux in symptomatic and asymptomatic pre-term
infants. Eur J Nucl Med Mol Imaging 35:1659-1665
6. International Pediatric Endosurgery Group (IPEG) Standard
and Safety Committee (2009) IPEG guidelines for the surgical
treatment of pediatric gastrooesophageal reflux disease
(GERD). J Laparoendosc Adv Surg Tech A 19:x-xiii
7. Ravelli AM, Panarotto MB, Verdoni L et al (2006) Pulmonary
aspiration shown by scintigraphy in gastrooesophageal reflux-
related respiratory disease. Chest 130:1520-1526
8. Abell TL, Camilleri M, Donohoe K et al (2008) Consensus
recommendations for gastric emptying scintigraphy: a joint re-
port of the American Neurogastroenterology and Motility So-
ciety and the Society of Nuclear Medicine. Am J Gastroenterol
103:753-763
9. Waseem S, Moshiree B, Draganov PV (2009) Gastroparesis:
current diagnostic challenges and management considerations.
World J Gastroenterol 15:25-37
10. Szarka LA, Camilleri M (2009) Methods for measurement of
gastric motility. Am J Physiol Gastrointest Liver Physiol
296:461-475
11. Donohoe KJ, Maurer AH, Ziessman HA et al (2009) Procedure
guideline for adult solid-meal gastric-emptying study 3.0. J
Nucl Med Technol 3:196-200
12. Mariani G, Pauwels EKJ, AlSharif A et al (2008) Radionuclide
evaluation of the lower gastrointestinal tract. J Nucl Med
49:776-787
13. Howarth DM (2006) The role of nuclear medicine in the de-
tection of acute gastrointestinal bleeding. Semin Nucl Med
36:133-146
14. Ford PV, Bartold SP, Fink-Bennett DM et al (1999) Society of
Nuclear Medicine procedure guideline for gastrointestinal
Ariane Boubaker
bleeding and Meckel’s diverticulum scintigraphy 1.0. J Nucl
Med 40:1226-1232
15. Kiratli PO, Aksoy T, Bozkurt MF et al (2009) Detection of ec-
topic gastric mucosa using 99mTc pertechnetate: review of the
litterature. Ann Nucl Med 23:97-105
16. Rerksuppaphol S, Hutson JM, Oliver MR (2004) Ranitidine-
enhanced 99m technetium pertechnetate imaging in children
improves the sensitivity of identifying heterotopic gastric mu-
cosa in Meckel’s diverticulum. Pediatr Surg Int 20:323-325
17. Balon HR, Brill DR, Fink-Bennett DM et al (2001) Society of
Nuclear Medicine procedure guideline for hepatobiliary
scintigraphy 3.0. J Nucl Med 38:1654-1657
18. Nadel HR (1996) Hepatobiliary scintigraphy in children.
Semin Nucl Med 26:25-42
19. Yang JG, Ma DQ, Peng Y et al (2009) Comparison of differ-
ent diagnostic methods for differentiating biliary atresia from
idiopathic neonatal hepatitis. Clin Imaging 33:439-446
20. Riccabona M, Avni FE, Blickman JG et al (2009) Imaging rec-
ommendations in paediatric uroradiology. Pediatr Radiol
39:891-898
21. Koff SA (2008) Requirements for accurately diagnosing
chronic partial upper urinary tract obstruction in children with
hydronephrosis. Pediatr Radiol 38(suppl 1):S41-S48
22. Gordon I, Colarinha P, Fettich J et al (2001) Guidelines for
standard and diuretic renogram in children. Eur J Nucl Med
28:BP21-BP30
23. Piepsz A, Colarinha P, Gordon I et al (2001) Guidelines for
glomerular filtration rate determination in children. Eur J Nu-
cl Med 28:BP31-BP36
24. Boubaker A, Prior JO, Meuwly JY et al (2006) Radionuclide
investigations of the urinary tract in the era of multimodality
imaging. J Nucl Med 47:1819-1836
25. Piepsz A, Colarinha P, Gordon I et al (2001) Guidelines on
99m Tc-DMSA scintigraphy in children. Eur J Nucl Med
28:BP37-41
26. Fettich J, Colarinha P, Fischer S et al (2003) Guidelines for di-
rect radionuclide cystography in children. Eur J Nucl Med Mol
Imaging 30:B39-44
27. Gordon I, Colarinha P, Fettich J et al (2001) Guidelines for in-
direct radionuclide cystography. Eur J Nucl Med 28:BP16-20
28. Taylor AT, Blaufox MD, Dubovsky EV et al (1998) Society of
Nuclear Medicine procedure guideline for diagnosis of reno-
vascular hypertension 3.0. J Nucl Med 39:1297-1302
29. A report of the Amsterdam Forum on the Care of the Live Kid-
ney Donor (2005) Data and Medical Guidelines. Transplanta-
tion 79:S53-S66
30. Prigent A (2008) Monitoring renal function and limitations of
renal function tests. Semin Nucl Med 38:32-46

PET in Hepatobiliary-Pancreatic Tumors
Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri
Department of Nuclear Medicine, University of Bologna, Bologna, Italy
Overview
Primary liver cancer is a relatively uncommon entity, ac-
counting for only 1-2% of malignant tumors, while
cholangiocarcinoma is even rarer. However, in parts of
Africa and Asia these tumors may constitute up to 20-
30% of all malignancies. Other benign and malignant
hepatobiliary-pancreatic tumors are very rare; they in-
clude hepatoblastoma in infancy and early childhood,
hemangiomas (the most common benign tumors), and
angiosarcomas (vascular tumors). Hepatic adenomas,
although also rare, may occur particularly in women
taking oral contraceptives. In the USA, pancreatic can-
cer has an incidence of 12 per 100,000 men and women
per year, and cancer of the pancreas is the third most
common malignancy of the gastrointestinal tract and the
fifth most common cause of cancer-related mortality.
Islet cell tumors constitute 5-10% of pancreatic tumors
and are usually referred to as neuroendocrine tumors of
the pancreas.
It is noteworthy that metastatic malignant lesions of
the liver are common in clinical practice, with an inci-
dence at least 20 times greater than that of primary car-
cinoma. By contrast, true cancers of the liver may escape
clinical recognition, as they often occur in patients with
cirrhosis (60-75% of cases) and the symptoms may ini-
tially suggest a progression of the underlying liver dis-
ease. The course of hepatocellular carcinoma and of
cholangiocarcinoma is fatal and usually rapid, with sur-
gical resection offering the only chance of cure. Nonethe-
less, the 5-year survival is low and only very few patients
have resectable tumors at the time of presentation.
Pancreatic carcinoma also has a very poor prognosis,
and early diagnosis is extremely difficult. The physical
examination rarely confirms the presence of localized
pancreatic cancer. The initial symptoms that prompt fur-
ther investigations are painless jaundice (in case of bile
duct obstruction) and weight loss, but they depend on the
site of the tumor within the pancreas and the degree of
organ involvement. In advanced disease, palpable lymph
nodes, hepatomegaly (due to metastasis), splenomegaly,
ascites, and peripheral edema (portal vein obstruction), or
an abdominal mass may be seen.
IDKD 2010-2013
Common laboratory findings in liver cancers are ane-
mia as well as elevated alkaline phosphatase and α-feto-
protein (AFP); in particular, very high levels of AFP oc-
cur in 70% of patients with hepatocellular carcinoma.
Laboratory findings in pancreatic cancer are non-specific,
since prolonged biliary obstruction may alter liver en-
zymes and cause abnormalities in vitamin-K-dependent
clotting factors (due to malabsorption of fat-soluble vita-
mins), while pancreatic duct obstruction may result in
pancreatic atrophy and subsequent impaired glucose
tolerance or frank diabetes. Carbohydrate antigen 19-9
(CA 19-9) is the most commonly employed serological
marker in pancreatic cancer. Although it has a sensitivity
of 80%, serum levels are increased in many benign and
malignant gastrointestinal conditions, such that the speci-
ficity for detecting pancreatic cancer is low.
Current Role of Imaging
A number of imaging procedures are used to detect liver
tumors, including ultrasound (US), computed tomography
(CT), magnetic resonance imaging (MRI), and angiogra-
phy. US has been suggested in the screening of high-risk
populations. It is widely available, reasonably accurate,
and does not involve radiation exposure. It therefore
should be considered as the first imaging procedure to use
when hepatocellular carcinoma or cholangiocarcinoma is
suspected. Although the use of contrast-enhanced US is
recommended, and contrast enhancement is mandatory for
CT, MRI is used with increasing frequency and is usually
the most accurate diagnostic approach. All these imaging
modalities can be used for primary diagnosis, staging pur-
poses, and to identify tumor recurrence.
In pancreatic cancer, US, CT, and MRI are widely used
as well, but they are supplemented by MR-cholangio-
pancreatography (MRCP) and endoscopic US (EUS).
Positron Emission Tomography
Positron emission tomography (PET) imaging allows the
in vivo study of tissue metabolism, and thus demonstrates
216
malignant tumors as hypermetabolic lesions based on
their increased tracer uptake. Nowadays, fluorine-18
radiolabeled fluorodeoxyglucose (18F-FDG) is the tracer
compound of choice in clinical practice. FDG is initially
carried into the cell by glucose transporters (GLUT-1),
just as normal glucose, then is rapidly phosphorylated
and trapped in the cells. Cancer cells are known to have
increased anaerobic glycolytic activity and to express
higher numbers of glucose transporters.
Despite the wide clinical application of FDG, not all tu-
mors show significantly increased metabolic activity on
FDG-PET imaging. In particular, prostate cancer and neu-
roendocrine tumors may be difficult to study with FDG-
PET, as the exam lacks sensitivity. Therefore, in addition
to FDG, several other tracers have been proposed, some of
which are already used for clinical applications. For exam-
ple, 68Ga-DOTA-somatostatin analogues have been suc-
cessfully applied in the study of neuroendocrine tumors,
both pancreatic and extra-pancreatic, while 11C-acetate has
been proposed for hepatocellular carcinoma PET.
In the past, the main limitation of PET imaging was its
failure to provide anatomical data; however, with the
introduction of a PET-CT hybrid system, morphological
and metabolic imaging can be performed in a single ses-
sion, thereby reducing false-positive findings and incon-
clusive studies and increasing diagnostic accuracy.
Role of FDG-PET in Liver Cancer
The success achieved with FDG-PET in studies of hepato-
cellular carcinoma has been limited by a false-negative
rate of 40-50%, with poor reliability especially in evalu-
ations of well-differentiated cancers. Of the few studies
available in the literature, most are retrospective and all
of them report an inadequate sensitivity of FDG-PET in
the detection of primary hepatocellular carcinoma. The
true-positive rate of FDG was better in poorly differenti-
ated tumors, with an increase in FDG uptake correlating
with lower survival. While FDG-PET might be useful in
the evaluation of extra-hepatic metastases, data support-
ing this possibility are limited. Instead, alternative tracers
have been proposed in conjunction with FDG; however,
the use of FDG alone may provide important prognostic
information, especially in patients who are candidates for
liver transplantation.
FDG-PET has also been proposed for patients with
cholangiocarcinoma, mainly in disease diagnosis and tu-
mor staging. In the former, PET seems to be helpful in
discriminating between malignant and benign lesions.
However, the accuracy of FDG-PET is dependent on the
lesion’s anatomical location, growth pattern, and patho-
logical characteristics. For this reason, its application is
limited to the detection of extra-hepatic, infiltrating, and
mucinous cholangiocarcinomas. Moreover, due to its low
sensitivity, PET provides complementary rather than con-
firmative information in the diagnosis of regional lymph
node metastasis. FDG-PET, however, has shown high ac-
Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri
curacy in detecting unsuspected distant metastases. Its
role in detecting cancer recurrence, monitoring treatment
response, and predicting prognosis is still controversial.
Role of Acetate PET in Liver Cancer
11C-acetate is used by cells as a precursor of membrane
fatty acids but it can also be transformed into acetyl-CoA,
entering the tricarboxylic acid cycle. It is thus processed
as an intermediate in glucose catabolism and in membrane
synthesis. The original application of 11C-acetate was not
in oncology but in cardiology, because accumulation of
the labeled compound in the myocardium is proportional
to the level of fatty acids oxidation and thus reflects car-
diac energy metabolism. Initially, the oncological applica-
tion of 11C-acetate was as a choline analogue for the de-
tection of prostate cancer; later, its use was expanded to
the evaluation of liver masses, together with 18F-FDG
studies. Preliminary results showed that 11C-acetate has
good sensitivity in the detection of low-grade but not
high-grade hepatic cancer, while FDG has the opposite
behavior. In our experience, acetate-based PET is the pre-
ferred approach to study most hepatocellular carcinomas,
especially in the differential diagnosis of masses not iden-
tified by conventional imaging and not suitable for biop-
sy, and for suspected recurrence of a hepatocellular carci-
noma that was previously treated surgically.
Role of PET with Other Tracers in Liver Cancer
11C-choline is another tracer suggested for studies of
hepatocellular carcinoma; the results have been similar to
those achieved with acetate. As only a very few studies
have been published, a role for 11C-choline remains to
be confirmed, but as for acetate, it may be useful in the
evaluation of low-grade tumors. Finally, use of the tracer
18F-fluorothymidine has been proposed to assess the
proliferation of hepatocellular carcinoma and cholangio-
carcinoma.
Role of FDG-PET in Pancreatic Cancer
Most of the literature describing the use of PET in pan-
creatic cancer refers to the tracer FDG. Normal pancreas
has low glucose utilization, whereas in pancreatic cancer
GLUT-1 transporters are over-expressed compared with
normal tissue; therefore, the tumor/background FDG up-
take ratio is high.
The most important step in the initial approach to a
patient suspected to have pancreatic carcinoma is to de-
cide whether the lesion is benign or malignant. The ma-
jor limitation of morphological imaging techniques is
their inability to confidently characterize small as well
as cystic lesions. In this setting, PET/CT may be helpful
due to its high sensitivity (85-100%) and moderate
PET in Hepatobiliary-Pancreatic Tumors
specificity (67-90%). Several studies have reported that
FDG-PET is more accurate than CT (the average sensi-
tivity and specificity for CT is 82 and 75%, respective-
ly). Regarding characterization of the lesion, the princi-
pal cause of false-positive findings at PET is inflamma-
tion due to chronic pancreatitis. However, the distribu-
tion of areas of avid FDG uptake within the parenchyma
can guide the diagnosis, as diffuse high uptake in the
whole pancreas is more often due to inflammation while
focal uptake is a feature of pancreatic cancer. Unfortu-
nately, morphological analysis is not specific because
pancreatic cancer is sometimes accompanied by pancre-
atitis, with FDG uptake in tumor tissue likely related to
the presence of inflammatory cells. In such cases it is es-
sentially impossible to clearly separate the two compo-
nents. Furthermore, cancer cells can diffusely infiltrate
the entire pancreas, with pancreatic cancer manifesting
as diffuse, high FDG uptake throughout the organ. By
contrast, benign lesions such as autoimmune pancreati-
tis can assume a pattern of focal FDG uptake. Semi-
quantitative approaches are not very helpful due to the
wide overlap in standardized uptake values (SUVs) be-
tween inflammation and malignant pancreatic disease.
False-positive findings can also occur due to recent
surgery or endoscopy, tissue inflammation after irradia-
tion, abscess, autoimmune pancreatitis, massive lympho-
cyte infiltration, retroperitoneal fibrosis, hemorrhage in
pancreatic pseudocysts, inflammatory pseudotumors,
pancreatic tuberculosis, and focal high-grade dysplasia.
Most false-negative results occur in cases involving
both tumors of small size and elevated serum glucose lev-
els. It should be noted that many such patients suffer from
pancreatic insufficiency and diabetes; consequently, the
high serum glucose levels compete with FDG for glucose
transporters sites, reducing the sensitivity of FDG-PET in
the detection of malignant lesions. Poor tumor cellulari-
ty, characteristic of scirrhous-type and cystic-type tumors
as well as those featuring a desmoplastic reaction, is also
an important cause of false-negative findings. The pauci-
ty of cells in these not-rare forms of pancreatic cancer is
seen even in fairly large tumors. False-negatives on FDG-
PET studies also arise from pancreatic tumors such as
mucinous or neuroendocrine tumors (NETs), which do
not have high glucose metabolism.
Regarding the detection accuracy of FDG-PET/CT, a
recent meta-analysis suggested that although the addition
of FDG-PET to the diagnostic work-up may enhance the
diagnosis of pancreatic malignancy, the usefulness of this
combined approach will vary depending upon the pre-test
probability of the tumor, the results of CT, and the
provider’s testing thresholds. Average sensitivity and
specificity shift from 92 and 68% after a positive CT re-
port, to 73 and 86% after a negative CT report, and 100
and 68% after an indeterminate CT report.
In conclusion, the greatest benefit of PET in the dif-
ferentiation of benign from malignant lesions is the pos-
sibility of excluding cancer without the need for biopsy
or surgery, either of which may increase morbidity.
217
Regarding the use of FDG-PET for staging, about 40%
of pancreatic cancers determined to be resectable by pre-
operative imaging turned out to be non-resectable at the
time of surgery. Therefore, correct staging is the princi-
pal aim of imaging in pancreatic malignancies, to deter-
mine the appropriate management and the prognosis of
the disease. For T staging, the poor spatial resolution of
FDG-PET limits its utilization: anatomical imaging
modalities such as multidetector CT, EUS, and MRCP
are better suited to demonstrate the relationship between
the tumor and the adjacent organs or vascular structures.
At present, there are no data to support the usefulness of
the hybrid-modality PET/CT in local T staging. Lymph
node metastasis is one of the most important aspects of
clinical management, providing an independent prognos-
tic indicator for patients with pancreatic cancer. Un-
fortunately, both CT and PET/CT are of low sensitivity
(30-40%) for lymph node detection, perhaps due to the
strong radioactive scatter from the main tumor to peri-
pancreatic small lymph nodes and to the low number of
cancers cells in small metastatic lymph nodes. The prin-
cipal cause of false-positive lymph nodes at FDG-PET is
the presence of reactive locoregional lymphadenopathies
following biliary instrumentation.
After the diagnosis of pancreatic cancer has been es-
tablished and local resectability of the tumor confirmed,
the main objective of staging a pancreatic cancer is to
identify those patients with distant metastasis because
this group is currently excluded from surgical treatment.
In this regard, the capability of whole-body scanning with
a single examination at a single session is evidently an
advantage of FDG-PET over other imaging modalities.
Whole-body FDG-PET detection of distant metastasis or
unexpected lesions changes patient management, is cost-
saving, and improves the patient’s quality of life by avoid-
ing unnecessary surgery. Several studies have reported
better diagnostic accuracy in the detection of distant
metastasis using whole-body FDG-PET rather than other
modalities, such as CT or US. The sensitivity of PET was
between 80 and 90% and thus better than CT; the posi-
tive predictive value is very high for both modalities.
Regarding metastatic disease, liver is the commonest
organ to be affected, followed by the lungs and bone mar-
row. Direct tumor spread into the peritoneum is also not
uncommon and often missed on conventional anatomical
imaging. The accuracy of PET in detecting liver metasta-
sis is almost the same as obtained with conventional
imaging (94 and 90%, respectively). False-positive find-
ings can occur due to intrahepathic cholestasis or in some
types of inflammation, such as abscess or intrahepatic
bile duct infection, mainly due to percutaneous trans-
hepatic cholangiodrainage. False-negative PET findings
may occur with lesions of small size but also because of
the heterogeneity of hepatic parenchymal FDG uptake
and respiratory motion artifacts.
PET has also been suggested in the evaluation of re-
sponse to therapy, and it may allow an earlier therapeutic
response assessment than is possible with conventional
218
imaging. However, only very preliminary data are avail-
able, and larger prospective studies are necessary to con-
firm the role of PET in treatment response and to assess
the correct time between therapy and PET post-treatment
examination.
With respect to recurrence, an elevated CA 19-9 has a
positive predictive value of only 69% for pancreatobiliary
malignancy. This means that >30% of patients with ele-
vated CA 19-9 may have another tumor originating in an-
other organ, or they may have no tumor at all. False-pos-
itive results have been associated with other pancreato-
biliary disorders, such as gallstones, pancreatitis, inflam-
matory bowel disease, other liver disorders, pulmonary
diseases such as pneumonia, and hydronephrosis. There-
fore, if CA 19-9 is elevated despite negative findings at
CT, then FDG-PET may have a role in detecting sites of
recurrence, either locally or as metastases in the liver,
lungs, peritoneum, and distant lymph nodes. FDG-PET
might also be appropriate for excluding the presence of
recurrence in patients with indeterminate findings using
other imaging modalities.
Suggested Reading
Bang S, Chung HW, Park SW et al (2006) The clinical usefulness
of 18-fluorodeoxyglucose positron emission tomography in
the differential diagnosis, staging, and response evaluation af-
ter concurrent chemoradiotherapy for pancreatic cancer. J Clin
Gastroenterol 40:923-929
Breitenstein S, Apestegui C, Clavien PA (2008) Positron emission
tomography (PET) for cholangiocarcinoma. HPB (Oxford)
10:120-121
Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri
Dierckx R, Maes A, Peeters M, Van De Wiele C (2009) FDG PET
for monitoring response to local and locoregional therapy in
HCC and liver metastases. Q J Nucl Med Mol Imaging
53:336-342
Eckel F, Herrmann K, Schmidt S et al (2009) Imaging of prolifer-
ation in hepatocellular carcinoma with the in vivo marker 18F-
fluorothymidine. J Nucl Med 50:1441-1447
Higashi T, Saga T, Nakamoto Y et al (2003) Diagnosis of pancre-
atic cancer using fluorine-18 fluorodeoxyglucose positron
emission tomography (FDG PET) – usefulness and limitations
in “clinical reality”. Ann Nucl Med 17:261-279
Kauhanen SP, Komar G, Seppänen MP et al (2009) A prospective
diagnostic accuracy study of 18F-fluorodeoxyglucose positron
emission tomography/computed tomography, multidetector
row computed tomography, and magnetic resonance imaging
in primary diagnosis and staging of pancreatic cancer. Ann
Surg 250:957-963
Kornberg A, Küpper B, Thrum K et al (2009) Increased 18F-FDG
uptake of hepatocellular carcinoma on positron emission to-
mography independently predicts tumor recurrence in liver
transplant patients. Transplant Proc 41:2561-2563
Lee TY, Kim MH, Park do H et al (2009) Utility of 18F-FDG
PET/CT for differentiation of autoimmune pancreatitis with
atypical pancreatic imaging findings from pancreatic cancer.
Am J Roentgenol 193:343-348
Pakzad F, Groves AM, Ell PJ et al (2006) The role of positron emis-
sion tomography in the management of pancreatic cancer.
Semin Nucl Med 36:248-256
Salem N, Kuang Y, Wang F et al (2009) PET imaging of hepato-
cellular carcinoma with 2-deoxy-2[18F]fluoro-D-glucose,
6-deoxy-6[18F] fluoro-D-glucose, [1-11C]-acetate and
[N-methyl-11C]-choline. Q J Nucl Med Mol Imaging 53:
144-156
Seo S, Hatano E, Higashi T et al (2008) Fluorine-18 fluorodeoxy-
glucose positron emission tomography predicts lymph node
metastasis, P-glycoprotein expression, and recurrence after re-
section in mass-forming intrahepatic cholangiocarcinoma.
Surgery 143:769-777

PET in Tumors of the Digestive Tract
Thomas F. Hany
Department of Radiology, Clinic and Policlinic of Nuclear Medicine, University Hospital Zurich, Zurich, Switzerland
Introduction
The basic principle of positron emission tomography
(PET) is the use of pharmaceuticals labeled with
positron-emitting isotopes. These agents are such that
they can be integrated into one of the body’s metabolic
pathways. Positron-emitting isotopes are characterized by
a beta plus-decay, in which a positron is emitted. This
positron collides with any of the numerous shell electrons
of neighboring atoms and the resulting annihilation pro-
duces two 511-keV gamma rays. These two photons are
detected in coincidence by the PET scanner. Additional-
ly, the use of an integrated PET/computed tomography
(CT) machine allows PET and CT images of the patient
to be acquired in the same imaging session. The clinical-
ly and most widely evaluated of the labeled pharmaceu-
ticals is fluorine-18 fluoro-2-deoxy-D-glucose (18F-FDG).
This glucose analogue is transported into the cell by spe-
cific transporters and phosphorylated by hexokinase to
18F-FDG-6 phosphate. The latter is inert to further meta-
bolic processing or to transmembrane back-transport out-
side the cell and therefore accumulates intracellularly.
The long physical half-life of FDG, around 110 min,
makes this compound amenable for use as a metabolic
marker, with applications in oncology, cardiology, neu-
rology, and inflammation imaging. One of the most im-
portant advantages of PET/CT imaging is its broad
anatomical coverage, since images of the patient extend-
ing from the head to the thighs are acquired routinely. All
of the currently available data indicate that PET/CT is
more sensitive and specific than either of its constituent
imaging methods. With PET/CT, the most relevant effect
is that the CT data frequently add specificity to the FDG-
PET data [1].
Esophagus, Stomach, and Small Bowel
Esophagus
The incidence of esophageal cancer has been increasing
worldwide, particularly for adenocarcinomas of the lower
esophagus and gastroesophageal junction [2]. Squamous
IDKD 2010-2013
cell carcinoma is associated with dietary factors whereas
adenocarcinoma is related to reflux. This distinction ex-
plains why the former occurs more proximally in the
esophagus and the latter more distally. The major goals of
the pre-operative evaluation of patients with esophageal
cancers are to exclude distant metastases and to decide
whether complete resection of the primary tumor and
its lymphatic drainage (R0 resection) can be achieved
(Fig. 1). Evaluation of locoregional lymph nodes is
important in the determination of prognosis, since
patients with peritumoral lymph node metastases have a
significantly reduced overall survival, but the presence
of such lymph nodes has no influence on either tumor
resectability or therapeutic strategy [3].
Endoscopic ultrasonography (EUS) is considered the
most accurate modality for local staging of esophageal
cancer. Most studies show that the accuracy of EUS is ap-
proximately 85% for T staging and 75% for N staging. CT
and PET/CT scanning are not as accurate as EUS for local
staging, because neither technique can assess the layers of
the esophageal wall, nor can small lymph nodes be reliably
distinguished from adjacent tumor. Further improvement
of locoregional staging is achieved by EUS used in com-
bination with fine-needle aspiration (EUS-FNA). EUS-
FNA is reported to be more sensitive (83 vs. 29%;
p <0.001) and more accurate (87 vs. 51%; p <0.001) than
CT or EUS (87 vs. 74%; p=0.012) for nodal staging [4].
For initial staging, a meta-analysis comparing EUS,
CT, and FDG-PET demonstrated sensitivities for region-
al lymph node metastases of 80, 50, and 57%, respec-
tively, and specificities of 70, 83, and 85%, respectively.
Interestingly, diagnostic performance did not differ sig-
nificantly across these tests. For distant metastases, sen-
sitivity and specificity were, respectively, 71 and 93% for
FDG-PET and 52 and 91% for CT. The diagnostic per-
formance of FDG-PET for distant metastases was shown
to be significantly higher than that of CT, which was not
significantly affected by study type and patient charac-
teristics. For the detection of regional lymph node metas-
tases, EUS is the most sensitive modality, whereas CT
and FDG-PET are more specific. For the evaluation of
distant metastases, FDG-PET has probably a higher sen-
sitivity than CT [5].
220
a b
Curative treatment can be achieved by radiochemo-
therapy and/or surgery. FDG-PET/CT has demonstrated a
high negative predictive value. In a study by Lordick et
al., the effect of neo-adjuvant chemotherapy prior to
surgery after 2 weeks of chemotherapy was compared to
the initial scanning. An improved median survival in so-
called metabolic responders compared to non-responders
was demonstrated, whereas response was defined as a
≥35% decrease in standard uptake values (SUV). Re-
markably, major histological remissions (<10% residual
tumor) were noted in 58% of metabolic responders, but
no histological response was seen in metabolic non-
responders. Accordingly, FDG-PET has a relatively high
negative predictive value after treatment and vital tumor
tissue is very likely to be present in the unchanged FDG
uptake of the primary tumor [6]. These facts have found
their way into the clinical practice guidelines of the
National Comprehensive Cancer Network (NCCN),
which recommends PET/CT for initial staging and after
neo-adjuvant treatment.
Stomach: Malignant Disease
In a review by Dassen et al. regarding the pre-operative
diagnostic utility of FDG-PET/CT in gastric cancer, the
Thomas F. Hany
Fig. 1 a, b. Initial staging in a 49-year-old
patient with an advanced distal esophageal
cancer. In the maximum-intensity-projection
image (a), FDG uptake is seen in the distal
esophagus and in two sites in projection to
the right upper abdomen. These two lesions
can be localized easily to an enlarged celiac
trunk lymph node metastasis and an adrenal
metastasis (b)
authors concluded that the technique has no role in the
primary detection of gastric cancer due to its low sensi-
tivity [7]. FDG-PET shows, however, slightly better re-
sults than CT in the evaluation of lymph node metastases
in gastric cancer and could therefore have a role in pre-
operative staging of the tumor. Improvements in the ac-
curacy of FDG-PET could be achieved by using PET/CT
or PET tracers other than FDG, but these approaches
need further investigation. The role of FDG-PET/CT is
likewise limited in the detection of gastric cancer recur-
rence after curative tumor resection. In a study by Sim et
al., in which 52 patients underwent restaging by PET/CT
and contrast-enhanced CT (ceCT), the sensitivity was
68.4% (26/38) for PET/CT and 89.4% (34/38) for ceCT
(p=0.057). The specificity was 71.4% (10/14) and 64.2%
(9/14), respectively (p=1.0). Contrast-enhanced CT was
more sensitive than PET/CT (p=0.039) in the detection of
peritoneal seeding. Additional PET/CT combined with
ceCT showed no further increase of positive predictive
value regardless of tumor site. PET/CT was as sensitive
and specific as ceCT in detecting a recurrence of gastric
cancer, except in the case of peritoneal seeding. However,
additional PET/CT combined with ceCT did not increase
diagnostic accuracy in the detection of recurrent gastric
cancer [8]. Accordingly, further studies are warranted to
PET in Tumors of the Digestive Tract
validate the role of PET/CT in the detection of gastric
cancer recurrence; however, this approach adequately de-
tects therapy responders at an early stage following neo-
adjuvant chemotherapy.
Regarding primary non-epithelial tumors, lymphoma
is an important disease beside gastrointestinal stroma tu-
mors (discussed below). Primary gastric lymphoma
(PGL) is díagnostically challenging due to the physiolog-
ical activity of FDG in the stomach and variability in the
degree of the tracer’s uptake in tumors of different histo-
logical subtypes [9]. In a study by Radan et al., PET/CT
studies of 62 newly diagnosed PGLs were reviewed: 24
of low-grade mucosa-associated lymphoid tissue (MALT)
type and 38 consisting of aggressive non-Hodgkin’s lym-
phoma (AGNHL). FDG avidity was present in 89% of the
PGLs, including all AGNHL, but only in 71% of MALT-
type lymphomas. Especially in AGNHL-PGL, FDG uptake
can be differentiated from physiological tracer activity by
its intensity but not by its pattern.
Small Bowel: Malignant Disease
Adenocarcinomas of the small bowel are rare tumors,
with an incidence of around 3.7 per million people each
year. The pathological (development from adenomatous
polyps), genetic, and epidemiological features are very
similar to those of adenocarcinomas of the large bowel
but the prognosis is poor (reported 5-year-survival of
15-35%). Capsule endoscopy, CT enteroclysis, and
ceCT are mainly used in disease diagnosis and work-up.
These tumors do show clear FDG avidity and might be
better evaluated by FDG-PET/CT, although no compre-
hensive date on the use of FDG-PET/CT are yet available.
Carcinoid tumors, as malignant tumors of non-
epithelial origin, have a similar incidence as adenocar-
cinomas (3.8 per million people each year). The different
subtypes have a natural behavior ranging from benign to
high-grade malignancies. Similar to neuroendocrine
tumors (NET) of the pancreas, the diagnostic work-up
consists of nuclear medicine studies, including somato-
statin receptor scintigraphy combined with contrast-
enhanced CT (SPECT/CT) or Ga-68 octreotide (68Ga-
DOTA-NOC) labeling, since more than 80% of carcinoid
tumors express somatostatin receptors. Detection rates
are therefore similar to those of NET of the pancreas [10].
Gastrointestinal Stromal Tumors
Gastrointestinal stromal tumors (GIST) are mesenchymal
tumors that in approximately 90% of patients originate in
the stomach and small intestine. FDG-PET is able to
show early effects in patients undergoing treatment with
imatinib mesylate (Glivec; Novartis, Switzerland) [11].
In two recent short-term follow-up studies, patients
without FDG uptake after the start of treatment had a
better prognosis than those with residual activity not
demonstrated with ceCT [12, 13]. Furthermore, lesions
221
were better defined on PET/CT than on either PET or CT,
when compared side-by-side. Normally, Response Evalu-
ation Criteria in Solid Tumors (RECIST) criteria are used
in the evaluation of therapeutic response, based on the
change in tumor size. In a modified RECIST analysis,
carried out in time intervals of 2 months for up to 28
months, Choi et al. used ceCT studies to evaluate changes
in tumor size and density [14]. When PET, RECIST, and
modified RECIST criteria were used in the analysis, a de-
crease in tumor size >10% or a decrease in tumor density
>15% on CT was demonstrated to have a sensitivity of
97% and a specificity of 100% in identifying PET re-
sponders vs. 52 and 100% by RECIST. Good responders
on CT at 2 months had significantly longer time to pro-
gression than patients who did not respond (p = 0.01).
Therefore, the search for small changes in tumor size or
density on CT may be a sensitive and specific method to
assess the response of GISTs.
Colorectal Carcinoma
Colorectal carcinoma is the most important cause of
death due to cancer in the western world, after bronchial
carcinoma [15]. About 70% of patients have curable re-
sectable tumor at initial diagnosis and are treated with cu-
rative intent. Approximately 50% of colon cancer patients
will present with hepatic metastases, either at the time of
initial diagnosis or as a result of recurrence [16]. From a
diagnostic perspective, colon cancer is often evaluated
together with rectal cancer as a single group; however,
especially deep rectal cancer has a clearly different path-
way of locoregional and distant metastases. In deep rectal
cancer without invasion of the anal canal, lung metastases
are more prevalent than metastases to the liver (Fig. 2).
In very deep rectal cancer, lymphatic drainage is no
longer towards para-rectal/meso-rectal lymph node sta-
tions but to inguinal sites, analogous to anal squamous
cancer. In these cases, the inguinal lymph nodes and
possibly, the external iliac region have to be evaluated,
since clinical inspection, ultrasound, and fine-needle as-
piration cost-effectively lead to correct staging.
Initial Staging
Two studies in which FDG-PET alone was used for ini-
tial staging demonstrated the high sensitivity of this ap-
proach in the detection of primary tumor (100 and 96%)
and distant metastases (87 and 78%) but a low sensitivi-
ty (29 and 29%) in lymph node staging [17,18]. In a study
by Veit-Haibach et al., 47 patients underwent whole-body
PET/CT colonography one day after colonoscopy [13].
Compared with optimized abdominal CT staging alone,
PET/CT colonography was significantly more accurate in
defining TNM stage (difference, 22%; 95% CI, 9-36%;
p = 0.003), mainly based on a more accurate definition of
the T stage. Differences in defining N stage were not de-
tected between PET/CT colonography and CT alone
222
a b
when the threshold for malignant nodes was 0.7 cm but
were detected at a threshold of 1 cm. Differences were
not detected in defining M stage separately or when the
accuracies of PET/CT colonography were compared with
CT + PET. PET/CT colonography affected consecutive
therapy decisions in 4 patients (9%; 95% CI, 2.4-20.4%)
compared with conventional staging (CT alone and
colonoscopy). The combination of FDG-PET/CT in con-
junction with a dedicated ceCT protocol could be of in-
terest as a single-step staging procedure.
Recurrent Disease
Standard patient work-up for the detection of recurrence
and metastases in colorectal cancer includes regular clin-
ical examinations, CT scans, colonoscopy, and, usually,
the measurement of tumor markers such as CEA (Fig. 3).
However, this approach lacks specificity and may result
in diagnostic and therapeutic delays. Serological tumor
markers are useful, although it has been shown that the
serum CEA level has only 60-70% sensitivity for the de-
tection of colorectal cancer recurrence [19]. The mor-
Thomas F. Hany
Fig. 2 a, b. A 64-year-old male patient with a
histologically proven deep rectal cancer. In
the maximum-intensity-projection image (a),
focal uptake is seen just below the bladder.
Unexpectedly, additional uptake is seen is in
the right thorax. This lesion was identified as
a singular lung metastasis, typical for deep
rectal cancer without evidence of liver metas-
tases (b)
phology-based information obtained with CT does not
permit a distinction between post-surgical changes and
tumor recurrence, nor can it detect tumor involvement of
normal-sized lymph nodes [20]. Colonoscopy is only use-
ful in the detection of local recurrence. The suitability of
FDG-PET in identifying recurrence and metastases has
been confirmed in several studies. Despite the obvious
advantage of PET/CT over PET alone, a dedicated ceCT
is often requested by clinicians. Soyka et al. found that
cePET/CT, as a single-step examination, has the same di-
agnostic confidence and impact as a sequential approach,
with ceCT first and non-cePET/CT afterward [21]. Al-
though the lesion detection rate on ceCT images is high,
evaluation by ceCT alone can be challenging because of
the possibility of inconclusive results that require further
diagnostic evaluation (56% of our patient population).
The reason for this is predominantly related to specifici-
ty issues regarding the structural abnormalities depicted
by this modality. Consequently, patients with inconclu-
sive ceCT findings are now frequently referred for fur-
ther evaluation with 18F-FDG-PET/CT. More important-
ly, the same study showed that in 21% of the patients with
PET in Tumors of the Digestive Tract
a b
Fig. 3 a, b. A 71-year-old male patient with a
history of colon cancer in whom an increase
in serum CEA was detected. In the maximum-
intensity-projection image (a), focal uptake is
seen in projection to the liver. The axial con-
trast-enhanced CT and fused PET/CT image
demonstrate a singular liver metastasis in the
right liver lobe, segment VII (b)
apparently conclusive findings on ceCT, the addition of
non-cePET/CT information led to appropriate changes in
therapy. In clinical routine, in those cases in which ceCT
was judged to be conclusive, the patient would not rou-
tinely be referred for further evaluation with 18F-FDG-
PET/CT. However, if cePET/CT had been used as the ini-
tial imaging modality, 65% of the patients would have
had a clear benefit, including changes in management
and in diagnostic confidence. Therefore, one could argue
that cePET/CT should be the first-line diagnostic tool in
the restaging of colorectal cancer. Nonetheless, one could
also argue that in 35% of the patients both the radiation
exposure and the costs of the procedure would have been
futile. However, the former argument holds true only if
ceCT and non-cePET/CT are performed within 2-4
weeks. In general, surgeons insist on ceCT studies not
older than 4 weeks before taking a patient into the oper-
ating room. Thus, another, additional scan with contrast
enhancement (ceCT or cePET/CT) would be needed in
the majority of patients.
Post-surgical and radiotherapy-induced changes in
the small pelvis are the most challenging for morpho-
logical imaging studies in recurrent rectal cancer, since
tumor recurrence cannot be differentiated from benign
scar tissue. In a study by Even-Sapir et al., PET/CT was
used to distinguish benign from malignant pre-sacral
abnormalities. The sensitivity, specificity, positive pre-
dictive value, and negative predictive value were 100,
96, 88, and 100%, respectively, and PET/CT findings
were clinically relevant in 47% of 62 patients [22]. A
223
comparison with other “conventional” imaging studies
was not performed. In our own study, by Seltzner et al.,
the diagnostic value of ceCT and non-enhanced
PET/CT was prospectively evaluated and compared in
76 patients referred for pre-operative evaluation for liv-
er resection for metastatic colorectal cancer [23]. Extra-
hepatic disease was missed by ceCT in one-third of the
patients (sensitivity 64%), while PET/CT failed to de-
tect extrahepatic lesions in only 11% (sensitivity 89%;
p=0.02). New findings derived from PET/CT resulted
in a change in the therapeutic strategy in 21% of the pa-
tients. This study also demonstrated the well known
limitation in spatial resolution of around 4-6 mm of
PET imaging, since small tumours (e.g., <5 mm) were
often not detected. Also, patients who underwent
chemotherapy within the month prior to PET/CT had a
high incidence of false-negative results. Alternatively,
this effect might be used as a predictor of success
in neo-adjuvant chemotherapy before resection. The
above-mentioned studies clearly demonstrate the ad-
vantages of PET/CT imaging in colorectal cancer.
Therapy Response Assessment
In general, a decrease or reduction to normal of FDG
uptake levels in tumor tissue is correlated with response
to treatment. Systematic reviews have only been per-
formed in patients with rectal cancer before and after
neo-adjuvant radio-chemotherapy. In a study by Kalff et
al., the prognostic information obtained from the degree
224
of change in tumor FDG uptake induced by chemoradi-
ation before radical curative surgery was evaluated in
patients with T3/T4 rectal cancer. In 34 consecutive pa-
tients with T3/T4 Nx M0 rectal cancer, FDG-PET was
performed at baseline and after radiochemotherapy be-
fore planned curative surgery. The change in FDG up-
take was measured by SUV as well as by visual grading
as complete (CMR), partial (PMR), or no metabolic re-
sponse. Histopathological findings were available in 30
patients. After an estimated median 3.1 years of follow-
up, all 17 CMR patients were free of disease. The PET
response was highly significantly associated with over-
all survival duration (p<0.0001) and time to progression
(p<0.0001). Pathological complete response was the
only other statistically significant prognostic factor
(p<0.03). The percentage of maximum SUV change af-
ter chemoradiation was not predictive of survival in
PMR patients. Based on a simple qualitative assess-
ment, post-chemoradiation 18F-FDG-PET provides
good medium-term prognostic information in patients
with advanced rectal cancer undergoing radical surgery
with curative intent [24].
Conclusions
FDG-PET/CT imaging is becoming more established in
the work-up of several abdominal malignancies of the
gastrointestinal tract. The main advantage lies in its com-
prehensive evaluation of the patient, including all body
compartments, and therefore the detection of pivotal, ther-
apy-deciding lesions. The performance of FDG-PET/CT
in the evaluation of primary tumors of the gastrointestinal
tract is characterized by a high sensitivity in the detection
of distant metastases. Secondary liver tumors such as
gastrointestinal metastases are detected by FDG-PET/CT
at a high rate, making this imaging technology a primary
tool in the evaluation of patients with suspicion of recur-
rent colon cancer. Furthermore, full integration of ceCT
protocols improves diagnostic confidence and reduces the
sometimes cumbersome diagnostic pathway for patients.
FDG-PET/CT can be reliably used for therapy response
assessment. New tracers, such as Ga-68-DOTA-TATE or
18F-DOPA, will bring significantly improved diagnostic
confidence in the notoriously difficult evaluation of pa-
tients with NET of the small bowel.
References
1. von Schulthess GK, Steinert HC, Hany TF (2006) Integrated
PET/CT: current applications and future directions. Radiology
238:405-422
2. Wei JT, Shaheen N (2003) The changing epidemiology of
esophageal adenocarcinoma. Semin Gastrointest Dis 14:
112-127
3. Lerut T, Coosemans W, Decker G et al (2001) Cancer of the
esophagus and gastro-esophageal junction: potentially curative
therapies. Surg Oncol 10:113-122
Thomas F. Hany
4. Vazquez-Sequeiros E, Wiersema MJ, Clain JE et al (2003) Im-
pact of lymph node staging on therapy of esophageal carcino-
ma. Gastroenterology 125:1626-1635
5. van Vliet EP, Heijenbrok-Kal MH, Hunink MG et al (2008)
Staging investigations for oesophageal cancer: a meta-analy-
sis. Br J Cancer 98:547-557
6. Lordick F, Ott K, Krause BJ et al (2007) PET to assess early
metabolic response and to guide treatment of adenocarcinoma
of the oesophagogastric junction: the MUNICON phase II tri-
al. Lancet Oncol 8:797-805
7. Dassen AE, Lips DJ, Hoekstra CJ et al (2009) FDG-PET has
no definite role in preoperative imaging in gastric cancer. Eur
J Surg Oncol 35:449-455
8. Sim SH, Kim YJ, Oh DY et al (2009) The role of PET/CT in
detection of gastric cancer recurrence. BMC Cancer 9:73
9. Radan L, Fischer D, Bar-Shalom R et al (2008) FDG avidity
and PET/CT patterns in primary gastric lymphoma. Eur J Nucl
Med Mol Imaging 35:1424-1430
10. Ambrosini V, Tomassetti P, Castellucci P et al (2008) Compar-
ison between 68Ga-DOTA-NOC and 18F-DOPA PET for the
detection of gastro-entero-pancreatic and lung neuro-
endocrine tumours. Eur J Nucl Med Mol Imaging 35:1431-
1438
11. Joensuu H, Roberts PJ, Sarlomo-Rikala M et al (2001) Effect
of the tyrosine kinase inhibitor STI571 in a patient with a
metastatic gastrointestinal stromal tumor. N Engl J Med
344:1052-1056
12. Goerres GW, Stupp R, Barghouth G et al (2004) The value of
PET, CT and in-line PET/CT in patients with gastrointestinal
stromal tumours: long-term outcome of treatment with ima-
tinib mesylate. Comparison of PET, CT, and dual-modality
PET/CT imaging for monitoring of imatinib (STI571) therapy
in patients with gastrointestinal stromal tumors. Eur J Nucl
Med Mol Imaging 4:4
13. Veit-Haibach P, Kuehle CA, Beyer T et al (2006) Diagnostic
accuracy of colorectal cancer staging with whole-body
PET/CT colonography. JAMA 296:2590-2600
14. Choi H, Charnsangavej C, Faria SC et al (2007) Correlation of
computed tomography and positron emission tomography in
patients with metastatic gastrointestinal stromal tumor treated
at a single institution with imatinib mesylate: proposal of new
computed tomography response criteria. J Clin Oncol 25:
1753-1759
15. Bade MA, Ohki T, Cynamon J, Veith FJ (2001) Hypogastric
artery aneurysm rupture after endovascular graft exclusion
with shrinkage of the aneurysm: significance of endotension
from a “virtual,” or thrombosed type II endoleak. J Vasc Surg
33:1271-1274
16. Clarke MP, Kane RA, Steele G Jr et al (1989) Prospective
comparison of preoperative imaging and intraoperative ultra-
sonography in the detection of liver tumors. Surgery 106:849-
855
17. Abdel-Nabi H, Doerr RJ, Lamonica DM et al (1998) Staging
of primary colorectal carcinomas with fluorine-18 fluo-
rodeoxyglucose whole-body PET: correlation with histopatho-
logic and CT findings. Radiology 206:755-760
18. Kantorova I, Lipska L, Belohlavek O et al (2003) Routine
(18)F-FDG PET preoperative staging of colorectal cancer:
comparison with conventional staging and its impact on treat-
ment decision making. J Nucl Med 44:1784-1788
19. Zervos EE, Badgwell BD, Burak WE Jr et al (2001) Fluoro-
deoxyglucose positron emission tomography as an adjunct to
carcinoembryonic antigen in the management of patients
with presumed recurrent colorectal cancer and nondiagnostic
radiologic workup. Surgery 130:636-643; discussion 643-
634
20. Goldberg RM, Fleming TR, Tangen CM et al (1998) Surgery
for recurrent colon cancer: strategies for identifying resectable
recurrence and success rates after resection. Eastern Coopera-
tive Oncology Group, the North Central Cancer Treatment
PET in Tumors of the Digestive Tract
Group, and the Southwest Oncology Group. Ann Intern Med
129:27-35
21. Soyka JD, Veit-Haibach P, Strobel K et al (2008) Staging path-
ways in recurrent colorectal carcinoma: is contrast-enhanced
18F-FDG PET/CT the diagnostic tool of choice? J Nucl Med
49:354-361
22. Even-Sapir E, Parag Y, Lerman H et al (2004) Detection of re-
currence in patients with rectal cancer: PET/CT after ab-
dominoperineal or anterior resection. Radiology 232:815-822
225
23. Selzner M, Hany TF, Wildbrett P et al (2004) Does the novel
PET/CT imaging modality impact on the treatment of patients
with metastatic colorectal cancer of the liver? Ann Surg
240:1027-1034; discussion 1035-1026
24. Kalff V, Duong C, Drummond EG et al (2006) Findings on
18F-FDG PET scans after neoadjuvant chemoradiation pro-
vides prognostic stratification in patients with locally ad-
vanced rectal carcinoma subsequently treated by radical
surgery. J Nucl Med 47:14-22

Tumors of the Adrenergic System: Imaging and Therapy
Cornelis A. Hoefnagel
Department of Nuclear Medicine, The Netherlands Cancer Institute, Amsterdam, The Netherlands
Targeting of Neuroendocrine Tumors
Neuroendocrine tumors (NET) include pheochromocy-
toma, neuroblastoma, carcinoid, paraganglioma, chemo-
dectoma, medullary thyroid carcinoma, islet cell tumors,
gastrinoma, small cell lung cancer, melanoma, and Merkel
cell tumor. NET vary considerably in their clinical presen-
tation, location, and histology but share a common embry-
onic tissue origin, i.e., the neural crest. In addition, all of
these tumors express several unique characteristics that
may be exploited in the targeting of radiopharmaceuticals
for diagnostic as well as therapeutic purposes [1]. The spe-
cific targeting of NET may be achieved via the metabolic
route (meta-iodo-benzylguanidine, MIBG), through recep-
tor binding (peptides), or immunologically (antibodies).
In metabolic targeting, 123I- or 131I-MIBG and 111In-
pentetreotide are sensitive and highly specific tracers,
and therefore the most widely used. In addition, compar-
ative studies have demonstrated a complementary role for
these procedures [1]. An active uptake mechanism at the
cell membrane and by neurosecretory storage granules in
the cytoplasm of neural crest tumors is responsible for the
uptake and retention of 123I- or 131I-MIBG, respectively.
Although the radiopharmaceutical may be released from
the granules, its specific re-uptake maintains a prolonged
intracellular concentration; this is in contrast to non-
adrenergic tissues, in which uptake relies on passive dif-
fusion only. The high selective accumulation results in
high tumor/non-tumor ratios. However, it should be kept
in mind that a number of drugs may interfere with the up-
take and/or the retention of MIBG [2].
Somatostatin analogs target peptide receptors on the
cell surface and therefore can be used in the diagnosis
and therapy not only of several NET but also in other tu-
mors in which peptide receptors have been demonstrated
autoradiographically [3]. Unlike MIBG- and antibody-
based targeting, the peptides are not specific for neural
crest tumors. 111In-pentetreotide is mostly used for diag-
nostic scintigraphy, while 90Y-labeled octreotide or lan-
reotide and 177Lu-labeled octreotate are administered for
targeted therapy. Other peptides used for diagnostic
scintigraphy include 111In-lanreotide, 99mTc-depreotide,
and 123I-vasoactive intestinal peptide.
IDKD 2010-2013
Radiolabeled monoclonal antibodies target antigens on
the cell surface of some NET. Examples are the murine
monoclonal antibodies 131I-UJ13A and 131I-3F8, used in
the 1980s, and, more recently, chimeric antibodies, such
as 131I-chCE7, in the treatment of neuroblastoma. Ra-
dioiodinated anti-CEA antibodies/fragments may target
CEA-producing medullary thyroid carcinoma (MTC).
More recently, bispecific anti-DTPA/anti-CEA immuno-
conjugates have been shown to improve tumor/non-tumor
ratios and to prolong retention in MTC [4].
Non-specific tumor-seeking agents, such as 67Ga-cit-
rate, 201Tl-chloride, 99mTc-pentevalent DMSA, and
99mTc-sestamibi, are also used for diagnostic scintigra-
phy. Although these tracers may be sensitive, they lack
specificity and therefore cannot be used therapeutically.
New specific tracers for positron emission tomogra-
phy (PET) imaging have recently been developed, for ex-
ample, 124I-MIBG, 11C-hydroxyephedrin, 6-[18F]-fluo-
rodopamine, and 68Ga-octreotide (the latter from a gen-
erator). These agents combine great sensitivity and
specificity with the high-quality hybrid imaging provid-
ed by PET/computed tomography (CT). These tech-
niques will significantly improve the quality and relia-
bility of diagnostic imaging of NET. While 18F-fluo-
rodeoxyglucose (18F-FDG) does not have this high de-
gree of specificity, it may be used to detect dedifferenti-
ating, rapidly growing tumors [5].
As the role of radiolabeled peptides for imaging is dis-
cussed elsewhere in this volume, the remainder of this
chapter focuses on the diagnosis and therapy of NET us-
ing 123I-MIBG or 131I-MIBG.
Techniques
Currently, there are several nuclear medicine procedures
involving the use of MIBG in the diagnosis and therapy
of NET.
Diagnostic scintigraphy of the whole body uses either
123I-MIBG (γ-emitter, t
1/2fys = 13 h, photon energy 159
Kev) or 131I-MIBG (β/γ-emitter, t1/2fys = 8 days, photon
energy 364 KeV). 123I-MIBG scintigrams are of better
quality and the results are more readily available, where-
as 131I-MIBG enables delayed imaging over several days.
Tumors of the Adrenergic System: Imaging and Therapy
Single photon emission tomography (SPECT) or SPECT/
CT using 123I-MIBG provides improved detection in ad-
dition to accurate localization of NET sites by hybrid (fu-
sion) imaging. Positron emission tomography (PET) or
PET/CT using novel, specific PET tracers (see above) is
currently the most accurate diagnostic modality for NET.
Following re-injection of 123I-MIBG, an intraoperative
gamma-probe can be used to guide the surgical resection
of NET. Finally, radionuclide therapy with high doses of
131I-MIBG is an option for tumors retaining a high con-
centration of the radiopharmaceutical for a prolonged
period of time (as demonstrated on the diagnostic scinti-
gram). The procedure can be monitored by post-therapy
total-body scintigraphy with or without 131I-MIBG
SPECT/CT fusion imaging.
Diagnostic Imaging
Pheochromocytoma
The role of 123I/131I-MIBG-scintigraphy in the diagnosis of
pheochromocytoma is not as a screening test. Instead it is
the best initial procedure in patients who, on the basis of a
clinical or familial history, are suspected of having
pheochromocytoma and with high plasma levels or urinary
excretion rates of catecholamines and catecholamine
metabolites. The cumulative sensitivity of 123I/ 131I-MIBG
scintigraphy in potential pheochromocytoma patients is 88%
[1]. Although CT and magnetic resonance imaging of adren-
al masses provide the surgeon with better anatomical detail,
a positive 123I/131I-MIBG scan is a highly specific finding.
The scintigraphic technique is superior for localizing extra-
adrenal, recurrent, multifocal, and malignant disease [6]. Al-
though cumulative results of 111In-pentetreotide scintigraphy
also show a sensitivity for this technique of 88% in
pheochromocytoma, a disadvantage in the detection of an
adrenal tumor is the renal, hepatic, and splenic accumulation
of the tracer. A recent prospective multicenter evaluation of
123I-MIBG in 150 patients with confirmed or suspected
pheochromocytoma or paraganglioma demonstrated a sen-
sitivity of 82-88% and a specificity of 82-84%. In this se-
ries, the addition of SPECT hardly affected these values [7].
Neuroblastoma
The cumulative findings of 131I-MIBG scintigraphy re-
ported in the literature [1] indicate that 92% of neuro-
blastomas concentrate MIBG. 123I/131I-MIBG imaging
allows the detection of primary tumors, residual or
recurrent disease, and metastases, regardless of localiza-
tion, in a single procedure. When used together with
urinalysis for catecholamine metabolites, MIBG imaging
is the most sensitive and highly specific indicator of
neuroblastoma [8]. The uptake of MIBG is so tissue-
specific that in a child presenting with a tumor of un-
known origin 123I/131I-MIBG-scintigraphy can non-
invasively establish the diagnosis of neuroblastoma and
227
rule out Wilms tumor, Ewing sarcoma, rhabdomyo-
sarcoma, osteosarcoma, and malignant lymphoma [9].
A recent prospective trial of 123I-MIBG scintigraphy in
100 neuroblastoma patients showed an overall sensitivity
of 88%, which increased slightly, to 91%, by the addi-
tion of SPECT. In patients with a recent diagnosis of
neuroblastoma, the sensitivity of the procedure was 93%
and its specificity 92% [10]. At present, 123I/131I-MIBG
scintigraphy has an established role in the staging of
disease and as a parameter in the response criteria.
Discrepant findings of MIBG and bone scintigrams have
been described, in favor of either the former or the lat-
ter. Since a positive finding on an MIBG scintigram is
the more specific one, the initial use of 123I/131I-MIBG
is preferred, but complementary bone scintigraphy may be
indicated. Radioimmunoscintigraphy with 131I-3F8 [11]
and, more recently, radioiodinated chimeric antibodies
directed against neuroblastoma have yielded results com-
plementary to those obtained with MIBG imaging [12].
Carcinoid Tumors
The cumulative sensitivity of 111In-pentetreotide scintig-
raphy in patients with carcinoid (86%) is higher than that
of 131I-MIBG scintigraphy (70%) [1], thus favoring use
of the former in the initial diagnosis. 131I-MIBG scintig-
raphy should not be used as a screening test for the ini-
tial diagnosis of carcinoid, nor can it be relied upon to ex-
clude disease. However, the combined use of the two
techniques are possibly of therapeutic interest. A positive
111In-pentetreotide scintigram may predict a response to
palliative octreotide therapy and indicate the feasibility of
90Y-octreotide or 177Lu-octreotate therapy, whereas in the
work-up of patients with proven carcinoid 131I-MIBG
scintigraphy allows the selection of those patients who
may benefit from 131I-MIBG therapy.
Other Neuroendocrine Tumors
In MTC, radioimmunoassays of serum calcitonin and
CEA levels are currently the most sensitive parameters in
diagnosis and follow up, but in recent years many nuclear
medicine procedures have emerged that allow the disease
to be localized. This is especially the case for the detec-
tion of liver metastases and adrenal pheochromocytomas.
Total-body scintigraphy with SPECT using 201Tl-chlo-
ride, 99mTc-pentavalent DMSA, and/or 99mTc-sestamibi
and PET using 18F-FDG are best used initially. These are
all relatively non-specific procedures, but their sensitivi-
ties are in the range of 80-90%. 111In-pentetreotide and
radiolabeled anti-CEA antibodies, with sensitivities of
60-70%, may have a complementary role. 131I-MIBG has
the lowest sensitivity (35%) and thus should only be used
once MTC metastases have been confirmed, to evaluate
its potential therapeutic role [1,13].
A comparison of the results of 131I-MIBG and 111In-
pentetreotide in a variety of other neural crest tumors [1]
showed highest sensitivities for 111In-pentetreotide in
228
paraganglioma (97%), small cell lung cancer and, to a
lesser degree, endocrine gastroenteropancreatic (GEP)
tumors, Merkel cell tumor, melanoma, and functioning
pituitary tumors. 123I/131I-MIBG scintigraphy is useful
for the detection of ganglioneuroma, paraganglioma, and
chemodectoma; it is of limited use in pancreatic islet cell
tumors, retinoblastoma, schwannoma, and Merkel cell tu-
mors. It has no place in the diagnosis of small cell lung
cancer and melanoma [1].
Although high sensitivities for 111In-pentetreotide
scintigraphy have been reported in non-neural crest tu-
mors, e.g., non-small cell lung cancer, brain tumors, and
lymphomas, as well as in granulomatous and auto-
immune diseases, 131I-MIBG scintigraphy, as a highly
specific procedure for neural crest tumors, is virtually
always negative in non-neural crest tumors [1, 9].
Rationale for Using MIBG and Somatostatin-Receptor
Imaging Procedures
Scintigraphy using 111In-pentetreotide is the best initial
procedure in patients with carcinoid, endocrine gastroen-
teropancreatic tumors, and (benign) paraganglioma.
131I-MIBG can be reserved to evaluate the feasibility of
therapy and for radionuclide treatment of these tumors.
123I/131I-MIBG scintigraphy remains the best initial pro-
cedure for pheochromocytoma, neuroblastoma, and ma-
lignant paraganglioma, because of its high sensitivity/
specificity as well as its effective therapeutic application
in these conditions. Both tracers play a modest role in
MTC, i.e., complementary to the more sensitive but
non-specific tracers and in the evaluation of the various
therapeutic options.
Radionuclide Therapy
Indications/Contraindications for Therapy
Any malignant neural crest tumor showing sufficient up-
take and prolonged retention of 131I-MIBG on a diagnos-
tic tracer study (ideally >1% of the administered dose,
depending on tumor volume) is a candidate for radio-
nuclide therapy. Apart from tracer concentration, the
Table 1. Pooled results of 131I-MIBG therapy in neural crest tumors (EANM Radionuclide Therapy Committee Workshop, Barcelona,
October 1999)
Disease Patients (N) Objective response:
tumor volume (%)
Pheochromocytoma 77
Paraganglioma 34
Neuroblastoma 229
Medullary thyroid carcinoma 29
Carcinoid 159
Other 6 2/6
Total 534
Cornelis A. Hoefnagel
availability and feasibility of other treatment modalities
as well as the patient’s condition determine the indication.
The principle indications for 131I-MIBG therapy are ma-
lignant pheochromocytoma and paraganglioma, neuro-
blastoma stage III and IV, MTC, and symptomatic,
metastatic carcinoid tumors [14].
Contraindications for radionuclide therapy in general
are: pregnancy, continued breast feeding, myelosuppres-
sion, and renal failure. In addition, relative contraindica-
tions apply to those patients whose condition is unstable
or who fail to understand or cooperate with the radiation
protection guidelines, or if isolation facilities are lacking.
Malignant Pheochromocytoma and Paraganglioma
The objective of 131I-MIBG therapy includes objective tu-
mor volume reduction (complete or partial response), tumor
arrest (stabilization of previously progressive disease), a re-
duction of the tumor’s metabolic function (as the prognosis
in pheochromocytoma may depend on the long-term con-
sequences of catecholamine hypersecretion, this may
actually prolong survival), and palliation of symptoms (e.g.,
hypertension, bone pain, sweating, constipation) [14].
In 1991, the results of 131I-MIBG therapy in 117 pa-
tients with pheochromocytoma treated in 14 centers
worldwide were pooled [15]. An objective response, de-
fined as a >50% decrease in catecholamine excretion, a
>50% reduction of tumor volume, or, if lesions could not
be measured, significant scintigraphic improvement, was
determined in 56% of the patients. The response of soft-
tissue metastases was better than that of skeletal metas-
tases. In addition, a subjective improvement of symp-
toms, decrease in blood pressure, and pain relief were
achieved in >60% of the patients. Long-lasting objective
responses have been reported also in malignant para-
ganglioma, in secreting and in non-secreting types [16].
These tumors may be treated either with 131I-MIBG or
90Y-/177Lu-labeled octreotide/octreotate.
At a European Association of Nuclear Medicine
(EANM) Radionuclide Therapy Committee workshop on
131I-MIBG therapy, held in 1999, treatment results were
gathered for 534 patients with neural crest tumors, in-
cluding 77 patients with malignant pheochromocytoma
and 34 with paraganglioma (Table 1). The cumulative
objective response rates with respect to tumor volume
Objective response: Subjective response:
biochemical (%) palliation (%)
51 68 68
48 51 70
51 NA Most patients
23 60 60
8 24 60
NA NA
Tumors of the Adrenergic System: Imaging and Therapy
were 51 and 48%, respectively; a >50% decrease in cat-
echolamine excretion was observed in 68 and 51%, re-
spectively, while symptomatic palliation occurred in 68%
of the patients. These results compare favorably with the
best reported results of combination chemotherapy and
were attained with a treatment that is non-invasive and as-
sociated with minimal side effects.
Recently published results in a group of 20 patients
with malignant pheochromocytoma or paraganglioma
who were treated with moderate administered doses
(7.4 GBq) at the Netherlands Cancer Institute (objective
response 47%, metabolic response 67%, subjective re-
sponse 89%) [17] compare well with those reported by
the group at Duke University (Durham, NC, USA), who
treated 18 patients with moderate doses (7.4 GBq) and 15
with high doses (18.5 GBq). The objective response was
38%, metabolic response 60%, and subjective response
86% [18]. Moreover, both a metabolic response and a
subjective response were suggested to have an important
influence on survival and quality of life, even in the ab-
sence of an objective volume response.
Neuroblastoma
Since 1984, therapeutic doses of 131I-MIBG have been
administered to children with metastatic or recurrent neu-
roblastoma that failed to respond to conventional treat-
ment. In 1991, the pooled results of the major centers
(273 patients) indicated an objective response rate of
35% [15]; more recently, the response rate increased to
51% (Table 1). Most of these patients had stage IV, pro-
gressive, and intensely pre-treated disease, and were ad-
ministered 131I-MIBG only after other treatment modali-
ties had failed. Both 131I-MIBG therapy and isolation are
generally well tolerated by children; however, hematolog-
ical side effects may occur. Apart from the objective re-
sponse, the palliative effect was often impressive. Thus,
for patients with recurrent and progressive disease after
conventional treatment 131I-MIBG therapy is probably the
best palliative treatment, as its invasiveness and toxicity
compare favorably with that of chemotherapy and exter-
nal beam radiotherapy [19].
Some groups have combined 131I-MIBG therapy with
chemotherapy and/or total-body irradiation, accepting
more toxicity, as well as with myeloablative chemotherapy
requiring autologous bone marrow or stem-cell rescue [20].
Voûte et al. [21] combined 131I-MIBG therapy with
oxygen treatment under hyperbaric conditions. Their aim
was to improve survival in patients with recurrent stage
IV neuroblastoma by adding the toxic effect of hydroxyl
radicals to the radiation effect. Subsequently, high-dose
vitamin C therapy was added to this regimen.
More recently, 131I-MIBG therapy has been integrated
in the treatment protocol as the initial therapy instead
of its use in pre-operative combination chemotherapy in
children presenting with advanced/inoperable neuro-
blastoma. The objective is to reduce the tumor volume,
thereby enabling adequate surgical resection, and to avoid
229
toxicity and the early induction of drug resistance.
Chemotherapy is reserved for the post-operative treat-
ment of minimal residual disease. Initial results have
demonstrated the feasibility and effectiveness of this ap-
proach, i.e., a higher objective response rate (>70%) and
considerably less toxicity than obtained with 131I-MIBG
therapy after conventional treatment [22]. By 2001, re-
sults in 56 patients showed that 131I-MIBG is as effective
as chemotherapy in attaining operable neuroblastoma:
43 of 56 evaluable patients (77%) had complete or >95%
resection of the primary tumor or did not require surgery
at all. At follow-up (13-144 months), 5-year survival was
37%. Based upon these results, two new multicenter stud-
ies have been initiated in which 131I-MIBG therapy is
integrated up-front in the treatment protocol of neuro-
blastoma. Patients with favorable parameters receive a
less aggressive therapy than before, consisting of two
cycles of 131I-MIBG followed by surgery, whereas in
patients with unfavorable parameters (high-risk group)
131I-MIBG therapy is intensified and combined with the
topoisomerase I inhibitor Topotecan to enhance radiation-
induced cytotoxicity.
Medullary Thyroid Carcinoma
In the abdomen, medullary thyroid carcinoma (MTC)
may present with liver metastases. Results of combina-
tion chemotherapy are disappointing whereas radio-
nuclide therapy using 131I-MIBG or 131I-anti CEA anti-
bodies may provide both tumor regression and palliation.
Pooled results in 29 patients with MTC treated with
131I-MIBG (Table 1) showed that an objective response
rate occurred in only 23% and tumor marker response in
60%; nevertheless, palliative effects, which may be quite
meaningful, were achieved in 60% of the patients. How-
ever, only a minority of patients demonstrated sufficient
uptake of 131I-MIBG.
More patients may be amenable to radioimmuno-
therapy. In a phase I/II study of treatment using bi-
specific anti-DTPA/anti-CEA immunoconjugates fol-
lowed by 131I-hapten in a two-step procedure, 26 MTC
patients showed mixed responses. Stabilization of dis-
ease and palliation were attained with limited hemato-
logical toxicity, but a HAMA (human anti-mouse anti-
body) response was reported in more than half of the
patients [23]. As patients may require several such
treatments, the use of chimeric or humanized immuno-
conjugates would be more appropriate.
Carcinoid Tumors
Palliative treatments for metastatic carcinoid tumors
include long-acting somatostatin analogs (Sandostatin),
α-interferon, hepatic artery embolization, 131I-labeled and
unlabeled MIBG, and 90Y- or 177Lu-labeled octreotide
therapy. The cumulative results of 131I-MIBG therapy in
159 patients with symptomatic, metastatic disease showed
an objective response rate of only 8% and a >50% decrease
230
in 5-hydroxyindoleacetic acid (5-HIAA) excretion in 24%
(Table 1). Despite the absence of an objective response,
palliation was achieved in 60% of patients and without
significant side effects [24]. In view of the often indolent
character of this disease, the value of a prolonged sympto-
matic response should not be underestimated. In a study
at Duke University Medical Center, 98 patients with
metastatic carcinoid were treated with 131I-MIBG. In this
group, a subjective response was found to correlate with
prolonged survival [25].
In patients with carcinoid tumors not qualifying for
131I-MIBG therapy because of no or insufficient uptake
by the tumor, palliative treatment with high doses of un-
labeled MIBG also proved beneficial in 60% of the cas-
es, albeit with a shorter mean duration [26]. Improved
biochemical and palliative effects of 131I-MIBG treatment
due to enhanced tumor/non-tumor ratios by pre-dosing
with non-labeled MIBG have also been reported [27]. A
combination of higher doses of 131I-MIBG and unlabeled
MIBG is used for therapy whenever comparative scintig-
raphy demonstrates a >20% increase of the tumor/non-
tumor-ratio following the addition of unlabeled MIBG.
References
1. Hoefnagel CA (1994) Metaiodobenzylguanidine and somatos-
tatin in oncology: role in the management of neural crest tu-
mours. Eur J Nucl Med 21:561-581
2. Khafagi FA, Shapiro B, Fig LM et al (1989) Labetalol reduces
Iodine-131 MIBG uptake by pheochromocytoma and normal
tissues. J Nucl Med 30:481-489
3. Reubi JC (1995) Neuropeptide receptors in health and disease: the
molecular basis for in vivo imaging. J Nucl Med 36:1825-1835
4. Bardiès M, Bardet S, Faivre-Chauvet A et al (1996) Bispecif-
ic antibody and Iodine-131-labeled bivalent hapten dosimetry
in patients with medullary thyroid or small-cell lung cancer. J
Nucl Med 37:1853-1859
5. Goldsmith SJ (2009) Update on nuclear medicine imaging of
neuroendocrine tumors. Future Oncol 5:75-84
6. Troncone L, Rufini V, Montemaggi P et al (1990) The diagnostic
and therapeutic utility of radioiodinated metaiodobenzylguani-
dine (MIBG). 5 years experience. Eur J Nucl Med 16:325-335
7. Wiseman GA, Pacak K, O’Dorisio MS et al (2009) Usefulness
of 123I-MIBG scintigraphy in the evaluation of patients with
known or suspected primary or metastatic pheochromocytoma
or paraganglioma: results from a prospective multicenter trial.
J Nucl Med 50:1448-1454
8. Hoefnagel CA, De Kraker J (2004) Pediatric tumors. In: Ell PJ
and Gambhir SS (eds) Nuclear medicine in clinical diagnosis
and treatment, 3rd edition. Churchill Livingstone, Edinburgh,
pp 195-206
9. Leung A, Shapiro B, Hattner R et al (1997) The specificity of
radioiodinated MIBG for neural crest tumors in childhood. J
Nucl Med 38:1352-1357
10. Vik TA, Pfluger T, Kadota R et al (2009) (123)I-mIBG scintig-
raphy in patients with known or suspected neuroblastoma: re-
sults from a prospective multicenter trial. Pediatr Blood Can-
cer 52:784-790
Cornelis A. Hoefnagel
11. Yeh SDJ, Larson SM, Burch L et al (1991) Radioimmunode-
tection of neuroblastoma with Iodine-131-3F8: correlation
with biopsy, Iodine-131-Metaiodobenzylguanidine and stan-
dard diagnostic modalities. J Nucl Med 32:769-776
12. Hoefnagel CA, Rutgers M, Buitenhuis CKM et al (2001) A
comparison of targetting neuroblastoma with mIBG and anti
L1-CAM antibody mAB chCE7: therapeutic efficacy in a neu-
roblastoma xenograft model and imaging of neuroblastoma pa-
tients. Eur J Nucl Med 28:359-368
13. Hoefnagel CA, Delprat CC, Zanin D, van der Schoot JB
(1988) New radionuclide tracers for the diagnosis and therapy
of medullary thyroid carcinoma. Clin Nucl Med 13:159-165
14. Hoefnagel CA and Lewington VJ (2004) MIBG therapy.
In: Ell PJ and Gambhir SS (eds) Nuclear medicine in clinical
diagnosis and treatment, 3rd edn. Churchill Livingstone,
Edinburgh, pp 445-457
15. Troncone L, Galli G (1991) Proceedings International Work-
shop on The Role of [131I]metaiodobenzylguanidine in the
Treatment of Neural Crest Tumors. J Nucl Biol Med 35:177-
362
16. Baulieu J-L, Guilloteau D, Baulieu F et al (1988) Therapeutic
effectiveness of Iodine-131 MIBG metastases of a nonsecret-
ing paraganglioma. J Nucl Med 29:2008-2013
17. Gedik GK, Hoefnagel CA, Bais E, Olmos RA (2008) 131I-
MIBG therapy in metastatic phaeochromocytoma and para-
ganglioma. Eur J Nucl Med Mol Imaging 35:725-733
18. Safford SD, Coleman RE, Gockerman JP et al (2003) Iodine-
131 metaiodobenzylguanidine as an effective treatment for
malignant pheochromocytoma and paraganglioma. Surgery
134:956-962
19. Hoefnagel CA (1999) Nuclear medicine therapy of neuroblas-
toma. Q J Nucl Med 43:336-343
20. Yanik GA, Levine JE, Matthay KK et al (2002) Pilot study of
iodine-131 metaiodobenzylguanidine in combination with
myeloablative chemotherapy and autologous stem-cell support
for the treatment of neuroblastoma. J Clin Oncol 20:2142-
2149
21. Voûte PA, van der Kleij AJ, de Kraker J et al (1995) Clinical
experience with radiation enhancement by hyperbaric oxygen
in children with recurrent neuroblastoma stage IV. Eur J Can-
cer 31A:596-600
22. Hoefnagel CA, de Kraker J, Valdés Olmos RA, Voûte PA
(1994) 131I-MIBG as a first-line treatment in high-risk neur-
oblastoma patients. Nucl Med Commun 15:712-717
23. Kraeber-Bodéré F, Bardet S, Hoefnagel CA et al (1999) Ra-
dioimmunotherapy in medullary thyroid cancer using bispe-
cific antibody and iodine-131-labeled bivalent hapten: Prelim-
inary results of a phase I/II clinical trial. Clin Cancer Res
5:3190s-3198s
24. Zuetenhorst H, Taal BG, Boot H et al (1999) Longterm pallia-
tion in metastatic carcinoid tumours with various applications
of meta-idobenzylguanidine: pharmacological MIBG, 131I-la-
beled MIBG and the combination. Eur J Gastroenterol Hepa-
tol 11:1157-1164
25. Safford SD, Coleman RE, Gockerman JP et al (2004) Iodine-
131 metaiodobenzylguanidine treatment for metastatic carci-
noid. Results in 98 patients. Cancer 101:1987-1993
26. Taal BG, Hoefnagel CA, Valdés Olmos RA et al (1996) Pal-
liative effect of Metaiodobenzylguanidine in metastatic carci-
noid tumors. J Clin Oncol 14:1829-1838
27. Taal BG, Hoefnagel CA, Boot H et al (2000) Improved effect
of 131I-MIBG treatment by predosing with non-radiolabeled
MIBG in carcinoid patients, and studies in xenografted mice.
Ann Oncol 11:1437-1443
 IDKD 2010-2013

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy

Dik J. Kwekkeboom
Department of Nuclear Medicine, Erasmus Medical Center, Rotterdam, The Netherlands

Introduction between a successful localizing study and a disappointing

one. For details of the scanning protocol, the reader is re-
The development of peptide receptor scintigraphy in ferred to the procedural guidelines for somatostatin re-
combination with radioiodinated somatostatin analogues ceptor scintigraphy with [111In-DTPA0]octreotide, pub-
allowed the in vivo demonstration of somatostatin-receptor- lished by the Society of Nuclear Medicine [2].
positive tumors in patients [1]. Later, other radiolabeled
somatostatin analogues were developed, two of which [111In-DTPA0]Octreotide Scintigraphy: Normal Scintigraphic
subsequently became commercially available. With the Findings and Artifacts
advent, over the past decade, of positron emission to-
mography (PET) tracers for somatostatin receptor imag- Normal scintigraphic features include visualization of the
ing, superior image quality and increased sensitivity in thyroid, spleen, liver, and kidneys, and in some patients the
tumor site detection have become possible, as confirmed pituitary gland (Fig. 1). In addition, the urinary bladder and
by several research groups. In the 1990’s, attempts at
treatment with radiolabeled somatostatin analogues were
undertaken in patients with inoperable and/or metasta-
sized neuroendocrine tumors. Improvements in the pep-
tides (higher receptor affinity) and the available radionu-
clides (β instead of γ emission), together with precautions
to limit the radiation dose to the kidneys and bone mar-
row, led to better results and a virtually negligible per-
centage of serious adverse events.

Somatostatin-Receptor-Based Radionuclide Imaging

The many drawbacks of [123I,Tyr3]octreotide, the radio-
iodinated somatostatin analogue first used for imaging in
patients, resulted in its replacement by the chelated and
111In-labeled somatostatin analogue [111In-DTPA0]oc-
treotide (OctreoScan, Covidien, Petten, The Netherlands),
which is commercially available and is now the most
commonly used agent for somatostatin receptor imaging
(SRI). The preferred dose of [111In-DTPA0]octreotide
(containing at least 10 mg of the peptide) is about
200 MBq. This dose is appropriate for single photon
emission computed tomography (SPECT), which shows
increased sensitivity in the detection of somatostatin-
receptor-positive tissues and provides better anatomical
delineation than planar views. The acquisition of suffi- Fig. 1. Normal distribution in somatostatin receptor imaging (SRI).
cient counts per view and the generation of spot images Variable visualization of the pituitary and thyroid (arrows, upper
panels). Faint breast uptake can sometimes be seen in women (right
with a sufficient counting time – as opposed to the low middle panel, arrow). Normal uptake in the liver, spleen, and kid-
count density of whole-body scans – are other important neys, and also some bowel activity is seen in the lower panel. Gall-
advantages of this approach and may make the difference bladder visualization in the lower right panel (arrow). Anterior views
232
Table 1. Pitfalls and causes of potential misinterpretation of positive
results with [111In-DTPA0]octreotide scintigraphy
Radiation pneumonitis
Accessory spleen
Focal collection of stools
Surgical scar tissue
Gallbladder uptake
Nodular goiter
Ventral hernia
Bacterial pneumonia
Respiratory infections
Common old (nasal uptake)
Cerebrovascular accident
Concomitant granulomatous disease
Diffuse breast uptake
Adrenal uptake
Urine contamination
Concomitant second primary tumor
Table 2. Causes of potential misinterpretation of negative results
with [111In-DTPA0]octreotide scintigraphy
The presence of unlabeled somatostatin, because of octreotide
therapy or due to the production of somatostatin by the tumor itself,
may lower tumor detectability
Different somatostatin receptor subtypes have different affinities
for the radioligand; variable tumor differentiation and receptor
expression also influences tumor detectability. This may be impor-
tant especially in patients with insulinomas and medullary thyroid
carcinomas
Liver metastases of neuroendocrine tumors may appear isointense
because of a similar degree of tracer accumulation by the normal
liver. Correlation with anatomical imaging and/or SPECT imaging
may be helpful
the bowel are usually visualized to variable degrees. Visu-
alization of the pituitary, thyroid, and spleen is due to re-
ceptor binding whereas uptake in the kidneys is for the
most part due to re-absorption of the radiolabeled peptide
by the renal tubular cells after glomerular filtration. While
there is predominant renal clearance of the somatostatin
analogue, hepatobiliary clearance via the bowel also oc-
curs, thus necessitating the administration of laxatives in
order to facilitate the interpretation of abdominal images.
False-positive results of SRI with [111In-DTPA0]oc-
treotide have been reported in virtually all cases; howev-
er, the term “false-positive” is a misnomer because it in-
cludes somatostatin-receptor-positive lesions unrelated to
the pathology for which the investigation was performed
(see the review by Gibril et al. [3]). The most common of
these are listed in Table 1. The potential causes of a false-
negative study interpretation are given in Table 2.
Imaging Results of [111In-DTPA0]Octreotide Scintigraphy
in Neuroendocrine and Other Tumors
Somatostatin receptors have been identified in vitro in a
large number of human neoplasias, in particular, neu-
roendocrine tumors (NETs) have a high incidence and
Dik J. Kwekkeboom
density of somatostatin receptors. NETs comprise a
group of tumors that includes pituitary adenoma, pancre-
atic islet cell tumor, carcinoid, pheochromocytoma, para-
ganglioma, medullary thyroid cancer, and small cell lung
carcinoma [4]. Tumors of the nervous system, including
meningioma, neuroblastoma, and medulloblastoma, also
very often express a high density of somatostatin recep-
tors, as do tumors not classically originating from en-
docrine or neural cells, such as lymphoma, breast cancer,
renal cell cancer, hepatocellular cancer, prostate cancer,
sarcoma, and gastric cancer. In the majority of these tu-
mors, somatostatin receptor (SR) subtype-2 is predomi-
nantly expressed, although low amounts of other SR sub-
types may be concomitantly present [5]. It should also be
emphasized that selected non-tumoral lesions may ex-
press SRs. For instance, SRs are expressed on the epithe-
lioid cells of active granulomas in sarcoidosis and by in-
flamed joints in active rheumatoid arthritis, especially the
proliferating synovial vessels [6]. Therefore, SR expres-
sion is not specific for tumoral pathologies.
The most common indication for [111In-DTPA0]oc-
treotide scintigraphy is the detection and localization of
gastroenteropancreatic neuroendocrine tumors (GEP-
NETs) and their metastases, the staging of these patients,
the follow-up of patients with known disease, and, lastly,
the selection of patients with inoperable and/or metastat-
ic tumors for peptide receptor radionuclide therapy
(PRRT) [7-12].
Newer Ligands for Somatostatin Receptor Imaging
99mTc-depreotide (Neotect) is a commercially available so-
matostatin analogue that has been approved specifically for
use in the detection of lung cancer in patients with pul-
monary nodules [13]. Due to the relatively high abdominal
signal background and the impossibility to perform de-
layed imaging because of the tracer’s short half-life, it is
less suited for the detection of abdominal NETs [14].
Analogues that are used for PET or hybrid PET/CT
imaging are of particular interest because of two advan-
tages that they have over γ-emitting analogues. First, many
of them have a better affinity for SR subtype-2, the subtype
most commonly expressed by NETs; likewise, there are
some analogues that better target other SR subtypes and are
therefore more appropriate for visualizing the respective tu-
mors. Second, PET and the combined anatomical and func-
tional information obtained with PET/CT provide images
with high spatial resolution, which results in a higher sen-
sitivity of this type of scanning. However, based on a re-
view of the results obtained with these newer analogues,
there are also causes for concern. Importantly, in many
studies these newer analogues were compared to [111In-
DTPA0]octreotide scintigraphy using inadequate scanning
protocols or comparisons were made between two or
more new analogues such that a validated reference method
was lacking. Also, the multitude of newly available PET
analogues has led to a situation in which each center has to
accumulate its own results on the normal findings and
Neuroendocrine Tumors of the Abdomen: Imaging and Therapy
artifacts of their scanning methods. This hampers the ex-
change of data and their shared interpretation. It is, howev-
er, likely that one of the new PET analogues, [68Ga-DOTA0,
Tyr3]octreotide or [68Ga-DOTA0, Tyr3]octreotate, will be-
come the new standard for SRI using PET. This is due to
the fact that these somatostatin analogues have a high affin-
ity for SR subtype-2, and 68Ga is a generator-produced
rather than a cyclotron-produced product, such that label-
ing of the compound is simplified [15]. An additional rea-
son favoring the use of [68Ga-DOTA0,Tyr3]octreotide or
[68Ga-DOTA0, Tyr3]octreotate as the standard analogue for
PET imaging is that the 90Y or 177Lu-labeled counterparts
of these compounds are administered for PRRT; thus, the
peptide used in diagnostic imaging closely mimics the one
that is used for therapy. Other radionuclide-coupled ligands
that do not rely on the presence of SRs for tumor visual-
ization have also been tested in patients with GEPNETs.
The oldest of these, 123I-MIBG, performs poorly compared
to [111In-DTPA0]octreotide scintigraphy [16].
PET scanning with (18F)2-fluoro-2-deoxy-D-glucose
18
( F-FDG) has gained importance for tumor staging and
the evaluation of treatment response for a number of tu-
mor types. The method is based on glucose consumption
by the tumors, such that fast-growing tumors usually
show high tracer uptake. However, 18F-FDG PET is less
suited for GEPNETs, because of the slow growing nature
of these tumors. Therefore, the technique is recommend-
ed only in patients with negative SRI findings [17], a sit-
uation that usually correlates with more aggressive tumor
behavior and faster tumor growth.
Newer PET radioligands that have been clinically
tested in patients with GEPNETs include 18F-DOPA and
11C-5-hydroxy-tryptophan [18]. PET with these ligands
has been reported to be more sensitive than SRI with
[111In-DTPA0]octreotide. These PET ligands, however,
have a short half-life and therefore have to be synthesized
in the close vicinity of or in the hospital where they are
to be administered. Also, both 18F-DOPA and 11C-5-
hydroxy-tryptophan, unlike the radiolabeled somatostatin
analogues used in PET, lack a sequel in PRRT.
Somatostatin-Receptor-Based Radionuclide Therapy
As noted above, the functioning and non-functioning en-
docrine pancreatic tumors and carcinoids that make up
Table 3. Tumor responses in patients with GEPNETs, treated with different radiolabeled somatostatin analogues
Center Tumor response ligand Patient Complete
(reference) number remission remission
Rotterdam [20] [111In-DTPA0]octreotide 26
New Orleans [21] [111In-DTPA0]octreotide 26
Milan [22] [90Y-DOTA0,Tyr3]octreotide 21
Basel [23, 24] [90Y-DOTA0,Tyr3]octreotide 74
Rotterdam [25] [90Y-DOTA0,Tyr3]octreotide 58
Rotterdam [29] [177Lu-DOTA0,Tyr3]octreotate 310
233
the GEPNETs are usually slow-growing. The treatment
of tumor metastases with somatostatin analogues results
in reduced hormonal overproduction and symptomatic re-
lief in most cases. Treatment with somatostatin analogues
is, however, seldom successful in terms of tumor size re-
duction [19].
A new treatment modality for patients with inoperable
or metastasized endocrine GEPNETs is the use of radio-
labeled somatostatin analogues. The majority of en-
docrine GEP tumors possess SRs and can therefore be vi-
sualized with SRI. A logical sequence to tumor visual-
ization in vivo is to then treat these patients with radiola-
beled somatostatin analogues. In the early phases of this
approach, virtually all patients considered as candidates
for PRRT had well-differentiated GEPNETs. At the time
of these early studies, in the mid- to late 1990s, no other
chelated somatostatin analogues labeled with β-emitting
radionuclides were available, such that [111In-DTPA0]oc-
treotide was used for PRRT. The results of these studies,
in which high doses of the radionucleotide were admin-
istered to patients with metastasized NETs, were encour-
aging with regard to symptom relief but partial remis-
sions (PRs) were exceptional [20, 21] (Table 3).
The next generation of SR-mediated radionuclide thera-
py was based on the use of the modified somatostatin ana-
logue [Tyr3]octreotide, which has a higher affinity for SR
subtype-2, and a different chelator, DOTA instead of DTPA,
in order to ensure a more stable binding of the intended
β-emitting radionuclide, 90Yttrium (90Y). The resulting
compound (90Y-DOTATOC; OctreoTher, Novartis, Switzer-
land), was used in several phase-1 and phase-2 PRRT trials
[22-25] (Table 3) but renal insufficiency and myelodys-
plastic syndrome were reported as serious adverse events
(SAEs). The incidence of these SAEs could, however, be
dramatically reduced through adequate renal protection,
achieved by the co-infusion of amino acids. Consequently,
SAEs have become relatively rare, occurring in <10% of
patients [26]. Despite differences in protocols, the rate of
complete remission (CR) and PR reported by most of the
different studies with [90Y-DOTA0,Tyr3]octreotide are in the
same range, between 10 and 30%, which is better than the
rates obtained with [111In-DTPA0]octreotide.
Reubi et al. [27] reported a nine-fold increase in so-
matostatin receptor subtype-2 affinity for [DOTA0,
Tyr3]octreotate vs. [DOTA0,Tyr3]octreotide, and a six to
seven-fold increase in the affinity for the Yttrium-loaded
Partial Minimal Stable Progressive Complete
response disease disease + partial
remissions
0 0 5 (19%) 11 (42%) 10 (38%) 0%
0 2 (8%) NA 21 (81%) 3 (12%) 8%
0 6 (29%) NA 11 (52%) 4 (19%) 29%
3 (4%) 15 (20%) NA 48 (65%) 8 (11%) 24%
0 5 (9%) 7 (12%) 33 (61%) 10 (19%) 9%
5 (2%) 86 (28%) 51 (16%) 107 (35%) 61 (20%) 29%
234 Dik J. Kwekkeboom

counterparts of this compound. In addition, a comparison
carried out in patients showed that the uptake of radioac-
tivity, expressed as percentage of the injected dose of
[177Lu-DOTA0,Tyr3]octreotate, was comparable to that
after [111In-DTPA0]octreotide for kidneys, spleen, and
liver, but was 3- to 4-fold higher for four of five tumor
types [28]. Therefore, [177Lu-DOTA0,Tyr3]octreotate po-
tentially represents an important improvement because of
the higher absorbed doses that can be achieved in most
tumors and the essentially equal doses to potentially
dose-limiting organs. Moreover, the lower tissue penetra-
tion range of 177Lu vs. 90Y may be especially important
for small tumors. These findings support the use of
177Lu-octreotate as the radiolabeled somatostatin ana-
logue of choice in PRRT.
The side effects and treatment outcome of [177Lu-
DOTA0,Tyr3]octreotate therapy have been analyzed by
our group in 504 and 310 patients with GEPNETs, re-
spectively [29]. In the 504 patients, acute side effects oc-
curring within 24 h after administration of the radio-
pharmaceutical included nausea (following 25% of admin-
istrations), vomiting (10%), and abdominal discomfort or
pain (10%). Six patients were hospitalized within 2 days
of administration of the radiopharmaceutical because of
hormone-related crises [30]. All patients recovered after
adequate care. Subacute, hematological toxicity of WHO
toxicity grade 3 or 4 occurred 4-8 weeks after 3.6% of ad-
ministrations, or, expressed in a patient-based manner, af-
ter at least one of several treatments in 9.5% of patients.
Serious delayed toxicities were observed in 9 out of 504
patients. There were two cases of renal insufficiency,
both of which were probably unrelated to 177Lu-octreo-
tate treatment. Three patients showed serious liver toxic-
ity, in two of these cases probably treatment-related. Last-
ly, myelodysplastic syndrome occurred in four patients
and was probably treatment-related in three.
Treatment responses according to tumor type at
3 months after the last therapy cycle were analyzed in
310 patients. The overall objective tumor response rate,
comprising CR, PR, and minimal response (MR), was
46% (for an example, see Fig. 2). Prognostic factors pre-
dicting tumor remission, i.e., CR, PR, or MR, as treat-
ment outcome were uptake on the OctreoScan (p <0.01)
and Karnofsky performance score (KPS) >70 (p <0.05). A
small percentage of patients who had either stable disease
(SD) or MR at their first two evaluations after therapy
had a further improvement in categorized tumor response
at 6 and 12 months follow-up, occurring in 4% and 5%

3-2007 5-2007 11-2007 2-2008 11-2008

CT

SRI

Post Tx 1 Post Tx 4
Fig. 2. Serial CT scans, SRI, and post peptide receptor radionuclide therapy (PPRT) scans in a patient with metastatic neuroendocrine tu-
mor with unknown primary. Month and year are indicated in the top row. Notice the ongoing tumor regression on CT, and also the im-
provement on SRI. The last post-therapy scan shows less tumor uptake than in the first; the difference is even more impressive if the same
scaling is used (but then, due to the impressive tumor uptake, in May 2007 the tumors would appear as one large hot spot occupying most
of the abdomen). This diminishing tumor uptake on subsequent post-therapy scans usually implies tumor shrinkage
Neuroendocrine Tumors of the Abdomen: Imaging and Therapy
of patients, respectively. Three of four patients with clin-
ically non-functioning neuroendocrine pancreatic tumors,
that were judged inoperable before treatment with 177Lu-
octreotate and who had PR, successfully underwent
surgery 6-12 months after their last treatment; the fourth
patient died of postoperative complications. The median
time to progression was 40 months from the start of treat-
ment. Median overall survival in our 310 GEP tumor
patients was 46 months (median follow-up 19 months;
101 deaths). Median disease-related survival was >48
months (median follow-up 18 months; 81 deaths). Median
progression-free survival was 33 months. The most impor-
tant factor predicting survival was treatment outcome.
Low KPS and liver involvement were also very signifi-
cant predictors.
Comparison of Survival Data
Since, in our study, treatment with [177Lu-DOTA0,Tyr3]
octreotate is still open for new patients and the median
follow-up in relation to survival is still relatively short,
we analyzed our local, Dutch patients separately, and
specifically those patients with longer follow-up. The re-
sults from these analyses suggest that both overall and
disease-specific survival times are consistently at or
above 48 months. These numbers compare favorably to
those reported in the literature. A comparison of survival
data for our patients, either from time of diagnosis or
from time of referral, with data from different epidemio-
logical studies or studies pertaining to a specific inter-
vention, and limiting our data to similar subgroups of pa-
tients, showed a benefit in overall survival for patients
treated with 177Lu-octreotate, which ranged from 40 to 72
months from the time of diagnosis [29]. We are aware
that comparisons with historical controls should be inter-
preted with caution, but we also think that this consistent
difference with many other reports in similar patient
groups cannot be ignored, and is most probably caused by
a real difference in survival.
Timing of Treatment
In patients treated with [177Lu-DOTA0,Tyr3]octreotate,
median overall survival was shorter in patients with a
poor performance score and in those with extensive liver
involvement by the tumor. This implies that treatment
with [177Lu-DOTA0,Tyr3]octreotate should preferably be
started early in the disease’s evolution. Since GEPNETs
can be clinically stable for years, however, it is, in our
opinion, good clinical practice to wait for signs of disease
progression if the tumor load is moderate. Such signs
should not be restricted to CT-assessed tumor growth but
should also include rises in serum tumor markers, in-
crease in symptoms, or involuntary weight loss. In pa-
tients with very limited tumor load and in whom cure is
potentially possible, treatment should be initiated without
further delay. The same holds true for patients with ex-
tensive tumor load, hepatomegaly, or those with signifi-
235
cant weight loss, when waiting for formally assessed tu-
mor progression would place these patients in an unfa-
vorable starting position for treatment or would even
make them ineligible for treatment.
Conclusions
The use of [111In-DTPA0]octreotide in SRI has a proven
role in the diagnosis and staging of GEPNETs. Newer
radiolabeled somatostatin analogues that can be used in
PET imaging, and which have a higher SR affinity, espe-
cially for subtype-2, have been developed. It would be de-
sirable, however, if one radiolabeled analogue became the
new standard for PET imaging, as the current application
of a multitude of analogues implies a fragmented knowl-
edge regarding image interpretation.
Treatment with radiolabeled somatostatin analogues
is a promising new tool in the management of patients
with inoperable or metastasized NETs. The results thus
far obtained with [ 90Y-DOTA 0,Tyr 3]octreotide and
[177Lu-DOTA0,Tyr3]octreotate have been very encour-
aging in terms of tumor regression. Also, if kidney-
protective agents are used, the side effects of this form
of therapy are few and mild, and the duration of the
therapy response for both radiopharmaceuticals may be
longer than 30 months. Lastly, compared to historical
controls, there appears to be a benefit in overall sur-
vival of several years from the time of diagnosis in pa-
tients treated with [ 177Lu-DOTA 0,Tyr 3]octreotate.
These data compare favorably with the limited number
of alternative treatment approaches. If more wide-
spread use of PRRT can be guaranteed, it may well be-
come the therapy of choice in patients with metasta-
sized or inoperable GEPNETs.
References
1. Krenning EP, Bakker WH, Breeman WA et al (1989) Local-
ization of endocrine related tumors with radioiodinated ana-
logue of somatostatin. Lancet 1:242-245
2. Balon HR, Goldsmith SJ, Siegel BA et al (2001) Procedure
guideline for somatostatin receptor scintigraphy with 111In-
pentetreotide. J Nucl Med 42:1134-1138
3. Gibril F, Reynolds JC, Chen CC et al (1999) Specificity of so-
matostatin receptor scintigraphy: a prospective study and ef-
fects of false-positive localizations on management in patients
with gastrinomas. J Nucl Med 40:539-553
4. Reubi JC (1997) Regulatory peptide receptors as molecular tar-
gets for cancer diagnosis and therapy. Q J Nucl Med 41:63-70
5. Reubi JC, Waser B, Schaer JC, Laissue JA (2001) Somato-
statin receptor sst1-sst5 expression in normal and neoplastic
human tissues using receptor autoradiography with subtype-
selective ligands. Eur J Nucl Med 28:836-846
6. Reubi JC, Waser B, Krenning EP et al (1994) Vascular so-
matostatin receptors in synovium from patients with rheuma-
toid arthritis. Eur J Pharmacol 271:371-378
7. Krenning EP, Kwekkeboom DJ, Bakker WH et al (1993) So-
matostatin receptor scintigraphy with [111In-DTPA-D-Phe1]-
and [123I-Tyr3]-octreotide: the Rotterdam experience with
more than 1000 patients. Eur J Nucl Med 20:716-731
236
8. Kwekkeboom DJ, Krenning EP, Bakker WH et al (1993) So-
matostatin analogue scintigraphy in carcinoid tumors. Eur J
Nucl Med 20:283-292
9. Westlin JE, Janson ET, Arnberg H et al (1993) Somatostatin
receptor scintigraphy of carcinoid tumours using the [111In
DTPA D Phe1] octreotide. Acta Oncol 32:783-786
10. De Kerviler E, Cadiot G, Lebtahi R et al (1994) Somatostatin
receptor scintigraphy in forty eight patients with the Zollinger
Ellison syndrome. Eur J Nucl Med 21:1191-1197
11. Gibril F, Reynolds JC, Doppman JL et al (1996) Somatostatin
receptor scintigraphy: its sensitivity compared with that of oth-
er imaging methods in detecting primary and metastatic gas-
trinomas. A prospective study. Ann Intern Med 125:26-34
12. Lebtahi R, Cadiot G, Sarda L et al (1997) Clinical impact of
somatostatin receptor scintigraphy in the management of pa-
tients with neuroendocrine gastroenteropancreatic tumors. J
Nucl Med 38:853-858
13. Menda Y, Kahn D (2002) Somatostatin receptor imaging of non-
small lung cancer with 99mTc depreotide. Semin Nucl Med
32:92-96
14. Lebtahi R, Le Cloirec J, Houzard C et al (2002) Detection of
Neuroendocrine Tumors: (99m)Tc-P829 Scintigraphy Compared
with (111)In-Pentetreotide Scintigraphy. J Nucl Med 43:889-895
15. Hofmann M, Maecke H, Börner R et al (2001) Biokinetics and
imaging with the somatostatin receptor PET radioligand (68)Ga-
DOTATOC: preliminary data. Eur J Nucl Med 28:1751-1757
16. Quigley AM, Buscombe JR, Shah T et al (2005) Intertumour-
al variability in functional imaging within patients suffering
from neuroendocrine tumours. An observational, cross-sec-
tional study. Neuroendocrinology 82:215-220
17. Belhocine T, Foidart J, Rigo P et al (2002) Fluorodeoxyglucose
positron emission tomography and somatostatin receptor
scintigraphy for diagnosing and staging carcinoid tumours:
correlations with the pathological indexes p53 and Ki-67. Nu-
cl Med Commun 23:727-734
18. Koopmans KP, de Vries EG, Kema IP et al (2006) Staging of
carcinoid tumours with 18F-DOPA PET: a prospective, diag-
nostic accuracy study. Lancet Oncol 7:728-734
19. Arnold R, Benning R, Neuhaus C et al (1993) Gastroen-
teropancreatic endocrine tumours: effect of Sandostatin on tu-
mour growth. The German Sandostatin Study Group. Diges-
tion 54(Suppl 1):72-75
Dik J. Kwekkeboom
20. Valkema R, de Jong M, Bakker WH et al (2002) Phase I study
of peptide receptor radionuclide therap1y with [111In-DT-
PA0]Octreotide: the Rotterdam experience. Semin Nucl Med
32:110-122
21. Anthony LB, Woltering EA, Espanan GD et al (2002) Indium-
111-pentetreotide prolongs survival in gastroenteropancreatic
malignancies. Semin Nucl Med 32:123-132
22. Bodei L, Cremonesi M, Zoboli S et al (2003) Receptor-medi-
ated radionuclide therapy with 90Y-DOTATOC in association
with amino acid infusion: a phase I study. Eur J Nucl Med Mol
Imag 30:207-216
23. Waldherr C, Pless M, Maecke HR et al (2001) The clinical val-
ue of [90Y-DOTA]-D-Phe1-Tyr3-octreotide (90Y-DOTATOC)
in the treatment of neuroendocrine tumours: a clinical phase II
study. Ann Oncol 12:941-945
24. Waldherr C, Pless M, Maecke HR et al (2002) Tumor response
and clinical benefit in neuroendocrine tumors after 7.4 GBq
(90)Y-DOTATOC. J Nucl Med 43:610-616
25. Valkema R, Pauwels S, Kvols LK et al (2006) Survival and re-
sponse after peptide receptor radionuclide therapy with [90Y-
DOTA0,Tyr3]octreotide in patients with advanced gastroen-
teropancreatic neuroendocrine tumors. Semin Nucl Med
36:147-156
26. Valkema R, Pauwels SA, Kvols LK et al (2005) Long-term fol-
low-up of renal function after peptide receptor radiation ther-
apy with 90Y-DOTA0,Tyr3-octreotide and 177Lu-DOTA0,
Tyr3-octreotate. J Nucl Med 46(Suppl 1):83S-91S
27. Reubi JC, Schaer JC, Waser B et al (2000) Affinity profiles for
human somatostatin receptor sst1-sst5 of somatostatin radio-
tracers selected for scintigraphic and radiotherapeutic use. Eur
J Nucl Med 27:273-282
28. Kwekkeboom DJ, Bakker WH, Kooij PP et al (2001) [177Lu-
DOTA0Tyr3]octreotate: comparison with [111In-DTPA0]oc-
treotide in patients. Eur J Nucl Med 28:1319-1325
29. Kwekkeboom DJ, de Herder WW, Kam BL et al (2008) Treat-
ment with the radiolabeled somatostatin analog [177Lu-
DOTA0,Tyr3]octreotate: toxicity, efficacy, and survival. J Clin
Oncol 26:2124-2130
30. De Keizer B, van Aken MO, Feelders RA et al (2008) Hor-
monal crises following receptor radionuclide therapy with the
radiolabeled somatostatin analogue [177Lu-DOTA0,Tyr3]
octreotate. Eur J Nucl Med Mol Imag 35:749-755
PEDIATRIC SATELLITE COURSE
“KANGAROO”

Imaging Cystic Kidneys in Children
Fred E. Avni
Department of Radiology, University Clinics of Brussels, Erasme Hospital, Brussels, Belgium
Introduction
Renal cystic diseases may be discovered or suspected at
any gestational week during fetal life or at any age in
childhood. They encompass a large number of conditions
that can be separated according to whether or not they are
hereditarily transmitted (Table 1). Imaging, mainly ultra-
sound (US), but also computed tomography (CT) and
magnetic resonance imaging (MRI) in selected indica-
tions, plays an important role in differentiating between
the various types of cystic diseases, by showing the char-
acteristics of renal involvement as well as associated
anomalies [1-5].
Cystic Renal Diseases in the Fetus and in the Perinatal
Period
In the fetus (and perinatal period), cystic renal disease
should be suspected whenever bilateral hyperechoic kid-
neys or cysts (uni- or bilateral) are discovered during an
obstetrical ultrasound examination. The imaging ap-
proach for the diagnosis should be based on a detailed
sonographic analysis that includes measurement of renal
length, the presence or absence of corticomedullary dif-
ferentiation (CMD), and the presence, number, size and
location of the cysts. This evaluation should be complet-
ed through an analysis of the entire fetus, looking for
Table 1. Classification of cystic diseases of the kidney in the fetus and in children
Genetic diseases
Non-genetic diseases (congenital or acquired)
IDKD 2010-2013
associated anomalies. The timing of detection is impor-
tant as is the amount of amniotic fluid, since both are im-
portant features for the final diagnosis and prognosis.
A pre-existing familial history of any renal cystic dis-
ease or of any syndrome with renal involvement in a fe-
tus in which abnormal kidneys have been detected should
raise strong suspicion. Finally, the evaluation should in-
clude a detailed clinical inquiry, looking for toxic causes
or maternal diseases that may influence fetal health.
Hyperechoic Kidneys in the Perinatal Period
This group includes many diseases that are genetically
transmitted or acquired during fetal life. Their evaluation
requires a step by step approach. Most will be discovered
during the second and third trimester.
Sonographic Criteria
Renal cortical hyperechogenicity is determined by com-
paring the renal cortex to the adjacent liver or to the spleen.
While hyperechogenicity is in most cases visually obvious,
there are four objective criteria indicative of cystic disease:
1. the renal cortex should not be hyperechoic compared
to the liver or spleen during the third trimester;
2. the appearance of the CMD, which can be normal, in-
creased, absent or reversed (due to a hyperechoic
medulla);
Autosomal recessive polycystic kidney disease (ARPKD)
Autosomal dominant polycystic kidney disease (ADPKD)
Glomerulocystic kidney diseases (including TCF2 anomalies,
nephronophitisis/medullary sponge kidney complex)
Cystic dysplasia
Medullary cystic dysplasia associated with syndromes
Renal obstructive dysplasia (associated with urinary tract malformations)
Multicystic/dysplastic kidney (some cases with genetic transmission)
Localized cystic dysplasia
Simple cyst
Multilocular cyst
Cystic tumor
Cysts associated with chronic dialysis
240
3. renal size: markedly increased (>+4 SD), moderately
increased (>+2 SD), or normal or small (<–2 SD) [6];
4. the presence of renal cysts.
Finally, any familial history or associated findings
constitute important factors for the differential diagnosis
[2, 6-9].
Differential Diagnosis
In case of markedly enlarged (>+4SD) hyperechoic kid-
neys diagnosed during the late first and early second
trimesters, Meckel-Gruber syndrome should be consid-
ered first, especially if the medulla appears enlarged and
hypoechoic and if polydactyly and cerebral anomalies are
associated. If the condition is detected during the second
and third trimesters, the main diagnosis to be added to the
differential would be autosomal recessive polycystic kid-
ney disease (ARPKD) and Bardet-Biedl syndrome
(BBS). In ARPKD, CMD may be partially absent, com-
pletely absent, or even reversed. A few visible cysts are
rare but may be seen in utero. Oligohydramnios is a fre-
quent finding and is associated with pulmonary hypopla-
sia, which confers a very poor prognosis [7-10].
In BBS, the kidneys are enlarged and hyperechoic and
there is post-axial polydactyly. The other symptoms of
the disease will develop after birth. Cysts can be ob-
served already in utero or appear after birth [7-12].
In case of moderately enlarged hyperechoic kidneys
(+2 SD), three diagnoses have to be considered first:
1. TCF2 mutation associated nephropathy;
2. ARPKD;
3. autosomal dominant polycystic kidney disease (ADPKD).
An anomaly of TCF2 (leading to HNF-1β-related mor-
phological anomalies) was recently shown to represent
the main cause of fetal hyperechoic kidneys [11]. This
mutation is associated with a wide spectrum of renal
morphological and structural anomalies that histological-
ly include glomerulocystic-type changes, cystic dyspla-
sia, and renal agenesis. Hepatic ductular plate anomalies
are commonly associated findings. In such kidneys, be-
sides renal hyperechogenicity, CMD may or not be visi-
ble. Cysts may be detected already in utero or, more of-
ten, after birth; they are located in the subcortical area. A
familial history of diabetes is a frequent finding.
The involvement and extent of the kidney lesions related
to ARPKD can vary from 10 to 90%, with the US appear-
ances varying accordingly. In cases with mild involvement,
the kidneys may be moderately enlarged, with a hyper-
echoic cortex and a few small cysts mainly within the pyra-
mids. After birth, cysts may also develop, throughout the
medulla first and within the cortex thereafter. Fetuses with
mildly enlarged kidneys have a better prognosis for sur-
vival than those with massive enlargement [9].
Already in utero, ADPKD may be suspected based on a
marked hyperechoic renal cortex that increases CMD. The
kidneys are usually normal in size or slightly enlarged.
Such finding should prompt familial inquiry. Cysts may be
observed in utero but usually develop after birth [13].
Fred E. Avni
Table 2. Causes of moderately enlarged hyperechoic kidneys in the
fetus
TCF2 mutation
Autosomal recessive polycystic kidney disease
Autosomal dominant polycystic kidney disease
Maternally related diseases
Infection
Ischemia
Metabolic diseases
Dysplasia
Nephrotic syndromes
“Transient”
Once these three diagnoses are excluded, there is a
wide spectrum of other diseases that can lead to hyper-
echoic kidneys; clinical inquiry may suggest the diagnosis
[14, 15]. Complementary examinations, such as chromo-
somal analysis, are directed at searching for infectious,
toxic, maternally related, or ischemic causes and will help
to reach a diagnosis (Table 2).
Renal Cyst(s) Discovered in the Perinatal Period
A unilocular, single renal cyst occurring in otherwise
normal-appearing kidneys can be detected in utero or after
birth. It should be differentiated, especially if the cyst is
septated, from a cystic tumor, segmental cystic dysplasia,
a dysplastic upper-pole of a duplex kidney, or a urinoma.
Associated urinary tract malformation and dilatation may
help make the diagnosis. Noteworthy is the fact that
ADPKD may start asymmetrically (with a single cyst) [16].
Whenever multiple cysts are detected, the first criteri-
on for the differential diagnosis is uni- or bilateral in-
volvement. Multiple cysts detected in one kidney only
most often correspond to a multicystic dysplastic kidney
(MCDK), which usually has a straightforward US ap-
pearance: multiple cysts of various sizes without inter-
connection, no recognizable normal renal parenchyma,
and no central renal pelvis. MDCK should be differenti-
ated from obstructive dysplasia (associated with urinary
tract obstructive malformation), in which the dilated uri-
nary tract is recognizable. The disease can also occur in
the upper pole of a duplex kidney. MDCK evolves such
that in most cases the kidney will eventually shrink. This
can be followed by US [17, 18].
Bilateral multiple renal cysts can be visualized in a large
number of isolated renal or syndromic diseases (Table 3)
[3, 4] and may or may not be associated with global renal
Table 3. Bilateral multiple cysts in the perinatal period
Bilateral multicystic dysplastic kidney disease
Bilateral obstructive dysplasia (+urinary tract dilatation)
Autosomal dominant polycystic kidney disease
Autosomal recessive polycystic kidney disease
Glomerulocystic disease (subcortical cysts)
Syndromes with cystic dysplasia, including (but not limited to)
Ivemark syndrome, Zellweger syndrome, Meckel Gruber syn-
drome, Bardet-Biedl syndrome, tuberous sclerosis complex
Imaging Cystic Kidneys in Children
Table 4. Bilateral macrocysts
Tuberous sclerosis complex
Simpson-Golabi-Behmel
TCF2 mutation syndrome
Bilateral multicystic dysplastic kidney disease
hyperechogenicity (see above). A step-by-step approach,
including detailed US analysis, familial history, and com-
plementary examinations, will lead to the diagnosis. Am-
niotic fluid volume and associated morphological or chro-
mosomal anomalies are mandatory for the prognosis. An
interesting subgroup includes diseases with macro-cysts
that are already obvious at birth [19, 20] (Table 4).
Many renal cystic diseases present with some type of
hepatic ductular anomaly. However, these are rarely de-
tected by perinatal US. Furthermore, in some patients it
may be of interest to perform fetal MRI in order to bet-
ter characterize the renal involvement [21].
Cystic Renal Diseases in Childhood
Renal cystic diseases in children are usually discovered
during an US examination performed: (1) in the follow-
up of a known, perinatally diagnosed disease, (2) during
the workup of syndromes diagnosed after birth, or
(3) during screening in an at-risk family. They may also
be detected as an incidental finding in an US examina-
tion performed for other reasons.
The main sonographic features of renal cystic diseases
are obviously the presence of visible cysts uni-or bilateral-
ly [1, 3]. The sonographic approach is similar to the one de-
scribed in the fetus and is based on the characteristics of the
cortex, medulla, and CMD as well as the size, location, and
number of cysts. Certain features will orient the diagnosis
[22, 23]: subcapsular cysts are indicative of glomerulocystic
disease; medullary cysts suggest medullary cystic dysplasia
(ARPKD); diffuse cortical cysts point to ADPKD, and
macro-cysts to ADPKD or tuberous sclerosis complex.
The role of US is in the follow-up of these patients once
the disease has been diagnosed, including monitoring dis-
ease evolution and the potential development of compli-
cations (hemorrhage, lithiasis). Another important role for
US is to verify the occurrence of hepatic complications of
the diseases and the development of portal hypertension
[24-26]. For this purpose, MRI may provide additional
information important for patient management [27].
Nephronophthisis (NPS) deserves special mention. This
genetically transmitted, autosomal recessive disease is be-
ing increasingly recognized due to improved genetic map-
ping. The symptoms of NPS include anemia, polyuria, and
polydysia, along with chronic renal failure. The disease
can be part of a syndrome (i.e., Joubert syndrome) or an
isolated finding. At US examination, the kidneys are rela-
tively small and cysts will develop at the corticomedullary
junction [28, 29].
241
Conclusions
The differential diagnosis of renal cystic disease is a dif-
ficult challenge. Cystic disease is suspected based on the
discovery of hyperechoic kidney or/and cysts. It can be
identified by approaching the diagnosis in a step-by-step
manner that includes US analysis, familial history, and
clinical evaluation.
References
1. De Bruyn R, Gordon R (2000) Imaging in cystic renal disease.
Arch Dis Child 83:401-407
2. Avni EF, Garel L, Cassart M et al (2006) Perinatal assessment
of hereditary cystic renal diseases: the contribution of sono-
graphy. Pediatr Radiol 35:405-414
3. Rizk D, Chapman AB (2003) Cystic and inherited kidney dis-
eases. Am J Kidn Dis 42:1305-1317
4. Deshpande C, Hennekam RCM (2008) Genetic syndromes and
prenatally detected renal anomalies. Semin Fetal Neonat Med
13:171-180
5. Winyard P, Chitty LS (2008) Dysplastic kidneys. Semin Fetal
Neonat Med 13:142-151
6. Cohen HL, Cooper J, Eisenberg P et al (1991) Normal length
of fetal kidneys. AJR Am J Roentgenol 157:545-548
7. De Bruyn R, Marks SD (2008) Post-natal investigation of fe-
tal renal disease. Semin Fetal Neonat Med 12:133-141
8. Tsatsaris V, Gagnadoux MF, Aubry MC et al (2002) Prenatal di-
agnosis of bilateral isolated fetal hyperechogenic kidneys. It is
possible to predict long-term outcome? BJOG 109:1388-1393
9. Chaumoitre K, Brun M, Cassart M et al (2006) Differential di-
agnosis of fetal hyperechogenic cystic kidneys unrelated to re-
nal tract anomalies. Ultrasound Obstet Gynecol 28:911-917
10. Ickowicz V, Eurin D, Maugey-Laulom B et al (2006) Meckel-
Gruber syndrome: sonography and pathology. Ultrasound Ob-
stet Gynecol 27:296-300
11. Decramer S, Parant O, Beaufils S et al (2007) Anomalies of
the TCF2-Gene are the main cause of fetal bilateral hyper-
echogenic kidneys. J Am Soc Nephrol 18:923-929
12. Cassart M, Eurin D, Didier F et al (2004) Antenatal renal sono-
graphic anomalies and post-natal follow-up of renal involve-
ment in Bardet-Biedl syndrome. Ultrasound Obstet Gynecol
24:51-54
13. Brun M, Maugey-Laulom B, Eurin D et al (2004) Prenatal
sonographic patterns in autosomal dominant polycystic kidney
disease. Ultrasound Obstet Gynecol 24:55-61
14. Slovis TL, Bernstein J, Gruskin A (1993) Hyperechoic kidneys
in the newborn and young infant. Pediatr Nephrol 7:294-302
15. Nortier JL, Debiec H, Tournay Y et al (2006) Neonatal disease
in NEP alloimmunization: lessons for immunological moni-
toring. Pediatr Nephrol 21:1399-1405
16. McHugh K, Stringer DA, Hebert D, Babiak GA (1991) Sim-
ple renal cysts in children: diagnosis and follow-up with US.
Radiology 178:383-385
17. Kuwertz-Broeking E, Brinkmann OA, VanLengerke HJ et al
(2004) Unilateral MDK: experience in children. BJU Internat
93:388-392
18. Aslam M, Watson AR (2006) Unilateral MDK: long-term out-
comes. Arch Dis Child 91:820-823
19. Glazier DB, Fleisher MH, Cummings KB, Barone JG (1996)
Cystic renal disease and TS in children. Urology 48:613-615
20. Newmann HPH, Schwarzkopf G, Henske EP (1998) Renal an-
giomyolipomas, cysts and cancer in TSC. Semin Pediatr Neu-
rol 5:269-275
21. Cassart M, Massez A, Metens T et al (2004) Complementary
role of MRI after US in assessing bilateral UT anomalies in
the fetus. AJR Am J Roentgenol 182:684-695
242
22. Jaim M, Lequesne GW, Bourne AJ, Henning P (1997) High-res-
olution US in the differential diagnosis of cystic diseases of the
kidney in infancy and childhood. J Ultrasound Med 16:235-240
23. Traubici J, Daneman A (2005) High-resolution renal sonography
in children with ARPKD. AJR Am J Roentgenol 184:1630-1633
24. Lipschitz B, Berdon WE, Defelice AR, Levy J (1993) Associ-
ation of congenital hepatic fibrosis with ADPKD. Pediatr Ra-
diol 23:131-133
25. Premkumar A, Berdon WE, Levy J et al (1988) Emergence of
hepatic fibrosis and portal HT in ARPKD. Pediatr Radiol
18:123-129
Fred E. Avni
26. Avni EF, Guissard G, Hall M et al (2002) Hereditary polycys-
tic kidney diseases in children: changing sonographic patterns
through childhood. Pediatr Radiol 32:169-174
27. Turkbey B, Ocak I, Daryanani K et al (2009) ARPKD and con-
genital hepatic fibrosi. Pediatr Radiol 39:100-111
28. Salomon R, Saunier S, Niaudet P (2009) Nephronophtisis. Pe-
diatr Nephrol 24:2333-2344
29. Blowey DL, Querfeld U, Geary D et al (1996) US findings in
juvenile nephronophtisis. Pediatr Nephrol 10:22-24

Understanding Duplication Anomalies of the Kidney
Jeanne S. Chow
Departments of Urology and Radiology, Children’s Hospital, Boston, MA, USA
Introduction
Learning the basic rules governing duplication anomalies
of the kidney is particularly rewarding because the fun-
damental rules can be applied reliably to the great variety
of cases and using all different imaging modalities. Al-
though ultrasound is typically the starting point of imag-
ing of the urinary tract in infants and children, renal du-
plex anomalies are also studied by micturating cysto-
urethrography (MCU), also called voiding cystourethro-
graphy (VCUG), intravenous pyelography (IVP), mag-
netic resonance urography (MRU), computed tomogra-
phy (CT), MAG-3 lasix renogram, and DMSA scan.
While more and more duplex anomalies are being dis-
covered in utero, most people with duplex anomalies are
asymptomatic and thus may never be studied at all.
Embryology
The ureteric bud must meet the metanephros in order for
the kidney to form. Without this interaction, kidney forma-
tion is not induced. If two ureteric buds meet the
metanephric blastema, then the kidney becomes duplex and
the collecting system and ureter become duplicated (Fig. 1).
Ureteral duplication may be complete or incomplete.
a Mesonephric Duct
Fig. 1 a, b. Duplex Kidney: embryology of the ectopic pathway.
a Embryology of a complete ureteral duplication showing that
the accessory ureter joins the upper pole of the metanephros.
b As the kidney continues to develop, the bladder insertion of
the lower-pole ureter ascends to the normal (orthotopic) posi-
tion in the bladder trigone. The upper-pole ureter descends so
that its orifice in the bladder is medial and inferior to the lower-
pole ureter
IDKD 2010-2013
In incomplete ureteral duplication, a single ureteric
bud, which is derived from the the mesonephric (Wolffian)
duct, bifurcates and meets the metanephros during
approximately week 5-6 of gestation. The two branches
of the ureter may join at the level of the renal pelvis (bifid
pelvis) or the proximal, middle or distal ureter (bifid
ureter) and terminate in a single distal ureter that inserts
orthotopically into the bladder. Rarely, one of the ureter-
al buds may be blind-ending and never appear to “reach”
the kidney (blind ending-ureteric duplication). Since the
upper and lower poles of the duplex kidney with a bifid
ureter have a common distal ureter, the upper and lower
poles typically appear similar (and normal). If there is
reflux or obstruction at the end of the common ureter,
both the upper and the lower pole will be affected.
In complete ureteral duplication, two separate ureteric
buds arise from the mesonephric duct to meet the
metanephric blastema. The lower-pole ureter is the ana-
logue of the normal single-system ureter and has a nor-
mally located ureteral orifice in the corner of the trigone
of the bladder. The upper-pole ureter is the “accessory
ureter” and inserts medially and inferiorly to the normal
ureteral orifice (Weigert-Meyer rule). Since the lower
pole of the duplex kidney is analogous to the normal
single-system kidney, abnormalities of the lower pole are
b
Accessory
Ureter
Metanephros
Ureter
244 Jeanne S. Chow

Fig. 2. Comparison of the calyceal axes

b
similar to those of a single (non-duplex) collecting sys-
tem, such as vesicoureteral reflux and obstruction at the
ureteropelvic or ureterovesical junction.
Reflux into the lower-pole ureter and intrarenal col-
lecting system can be easily distinguished from reflux in-
to a single (non-duplex) system kidney by carefully not-
ing the axis of the calyces (Fig. 2). Normally, the axis of
the calyces (the line drawn from the lowest to the highest
calyx) of a single-system kidney is toward the contralat-
eral shoulder. Since only the lower calyces are opacified
in lower-pole reflux, the axis of the visualized calyces is
altered and lies toward the ipsilateral shoulder [1].
Ureteral obstruction of the lower pole ureter can occur
at the level of either the renal pelvis or the insertion of
the ureter into the bladder. Ureteropelvic junction ob-
struction (UPJO) of the lower pole of the kidney is visu-
alized by ultrasound [2]. Occasionally, when the upper Fig. 3 a, b. The ectopic pathway in boys (a) shows that the ectopic
pole is dysplastic, lower-pole UPJO can be mistaken for ureter may insert from just below the trigone of the bladder, to the
posterior wall of the uretha as low as the veromontanum, and the
obstruction of a single collecting system. Ureterovesical ejaculatory duct and its branches. The ectopic pathway in the girl
junction obstruction (UVJO) can occur due to primary (b) shows that the ectopic ureter may insert from just below the
mega-ureter, or it may be secondary to the effect of a di- trigone of the bladder down to the posterior wall of the urethra,
lated obstructed upper-pole ureter. In these cases, when to the vulva and vagina
the upper-pole ureter is decompressed, the lower pole ob-
struction also resolves.
The upper pole ureter inserts ectopically, medially, and
inferiorly, to the normal ureteral orifice into any
mesonephric duct derivative. In addition to forming the Ectopic ureters are often obstructed, usually at the
ureter, the Wolffian (mesonephric) duct contributes to the level of the ureterovesical junction, but rarely reflux.
formation of the trigone of the bladder, the urethra, and If the ectopic ureter inserts into the urethra at the level of
the vagina in females, and the posterior urethra and gen- the urinary sphincter, it is both obstructed and refluxes,
ital ducts in males. The “ectopic pathway” follows the depending on whether the sphincter is closed or open, the
pathway created by the Wolffian duct (Fig. 3). so-called sphinteric ectopic ureter (Fig. 5) [4]. The more
Girls with an ectopic ureter inserting into the vagina distal the ureteral insertion, the more dysplastic and
or perineum may present with constant urinary drib- dysfunctional is the associated renal parenchyma that
bling [3] (Fig. 4). In boys, the ureter can terminate in it drains.
Wolffian duct derivatives, including the seminal vesi- A ureterocele is the dilated submucosal terminal seg-
cles and vas deferens. However, ectopic ureters in boys ment of the ureter. It is associated with the upper-pole
never terminate below the urinary sphincter and thus ureter of a double collecting system in girls. In boys,
never cause incontinence. ureteroceles are rare, but when they do occur they are
Understanding Duplication Anomalies of the Kidney 245

Prolapse
Fig. 7. Prolapsing ectopic ureterocele

Fig. 4. Vaginal ectopic ureter

Sphincter closed Sphincter open

a b
(Voiding)
Fig. 5. Sphincteric ectopic ureter Fig. 8 a, b. Ureterocele disproportion. a A typical ectopic, obstruct-
ed, left upper pole ureter ending in ureterocele. b Ureterocele dis-
proportion

a b

imperceptible (efface) (Fig. 6a). When the intravesical

Effacement of ureterocele
“Intussusception”
of
ureterocele
Fig. 6 a, b. Effacing (a) and everting (b) ureterocele
most commonly associated with single-system kidneys
and their orifices are orthotopic. These are associated
with varying degrees of ureteral obstruction and pelvi-
calyceal dilatation.
Although ureteroceles protrude into the lumen of the
bladder, when the intravesical pressure equals that of
the ureterocele, the ureterocele can flatten and become
pressure exceeds that of the ureterocele, the ureterocele
everts or intussuscepts into its ureter (Fig. 6b). Everting
ureteroceles are often confused for periureteral diverticula.
Ectopic ureteroceles involving the bladder neck can
prolapse into the urethra and cause obstruction of the
bladder outlet (Fig. 7).
Typically, ureteroceles of upper pole ureters are asso-
ciated with hydroureteronephrosis. Rarely, the upper
pole is diminutive and dysplastic, and the ureterocele is
relatively large. This combination is described as urete-
rocele disproportion (Fig. 8). The condition can be mis-
taken for single ectopic ureters [5]. Ureteroceles can al-
so be associated with multicystic dysplastic kidneys. As
the fluid in the multiple cysts resolves, these develop the
appearance of ureterocele disproportion. Very rarely,
three complete ureters or three incompletely separated
ureters form, resulting in complete or incomplete in
ureteral triplication [6].
246
In paired organs, such as the kidney, when there is a
congenital anomaly in one and there is something wrong
with the other, it is almost always the same anomaly, but
often different in degree. Thus, if one kidney is duplex,
the other is more likely to be duplex as well.
Conclusions
These are the fundamental principles to remember:
1. The ureters are duplicated but the kidney is called “du-
plex”.
2. Ureteral duplication may be incomplete or complete.
3. The Weigert-Meyer rule applies to complete ureteral
duplication and states that the upper-pole ureteral ori-
fice is ectopic. When the ectopia is slight, the upper
pole is normal. When the ectopia is moderate or se-
vere, the upper pole is abnormal.
4. The lower pole is the analogue of a single-system kidney.
5. Lower-pole reflux can be distinguished from reflux in-
to a single-system kidney by the axis of the calyces.
6. Ectopic ureters terminate along the ectopic pathway
and can cause incontinence in girls, but never in boys.
7. Ureteroceles are caused by obstruction; they do not
cause obstruction.
Jeanne S. Chow
8. Ureteroceles associated with duplex kidneys and ec-
topic upper-pole ureters are more common in girls
than in boys. In boys, ureteroceles are usually associ-
ated with single (non-duplex) kidneys.
9. Ureteroceles are dynamic.
References
1. Claudon M, Ben-Sira L, Lebowitz RL (1999) Lower pole re-
flux in children: uroradiologic appearance and pitfalls. AJR
172:795-801
2. Fernbach SK, Zawin JK, Lebowitz RL (1995) Complete du-
plication of the ureter with ureteropelvic junction obstruction
of the lower pole of the kidney: imaging findings. AJR 164:
701-704
3. Carrico C, Lebowitz RL (1998) Incontinence due to an in-
frasphincteric ectopic ureter: why the delay in diagnosis and
what the radiologist can do about it. Pediatric Radiology
28:942-949
4. Wyly JB, Lebowitz RL (1984) Refluxing urethral ectopic
ureters: recognition by the cyclic voidng cystourethrogram.
AJR 142:1263-1267
5. Share JC, Lebowitz RL (1989) Ectopic ureterocele without
ureteral and calyceal dilatation (ureterocele disproportion):
findings on urography and sonography. AJR 152:567-571
6. Gill RD (1952) Triplication of the ureter and renal pelvis. J
Urol 68:140-147

Malrotation: Techniques, Spectrum of Appearances, Pitfalls,
and Management
Alan Daneman
Department of Radiology, University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada
Introduction
The midgut is that part of the bowel supplied by the su-
perior mesenteric artery (SMA) and it extends from the
mid-portion of the second part of the duodenum to the
distal third of the transverse colon. The term “malrota-
tion” is broadly used to describe the spectrum of devel-
opmental abnormalities of the midgut that are associated
with abnormal rotation and/or fixation [1-3]. Malrotation
may occur as an isolated entity. However, it is also asso-
ciated with congenital defects in the development of the
abdominal wall (e.g., omphalocele and gastroschisis) or
diaphragm (e.g., congenital diaphragmatic hernia). Fur-
thermore, it may also be associated with abnormalities of
visceral situs and with certain syndromes.
Embryology
During embryological development, the midgut under-
goes a complicated process of growth and lengthening,
herniation, rotation, reduction, and fixation [1]. Growth
and lengthening of the midgut result in its herniation in-
to the umbilical cord along the axis of the SMA, such that
the apex of the herniated midgut is located at the level of
the omphalomesenteric duct. This herniated bowel under-
goes a 270° counterclockwise rotation, with subsequent
reduction of the midgut back into the abdomen before the
end of the first trimester. The proximal loop of herniated
bowel (including the small bowel up to the level of the
omphalomesenteric duct) is the first to reduce, followed
by reduction of the distal loop (including the distal small
bowel and colon to the level of the distal third of the
transverse colon). The final phase involves fixation of the
bowel into its final anatomical position. The cecum, how-
ever, is usually initially reduced into the upper abdomen
on the right. This precedes elongation of the proximal
colon until the cecum finally reaches the right lower
quadrant. The latter process may only be finally achieved
during early infancy.
This complicated sequence is essential for the midgut
to assume its normal position in the abdomen. One im-
portant aspect is the development of the normal duodenal
IDKD 2010-2013
loop. Initially, the duodenum rotates to the right, then
posteriorly, and finally to the left of the SMA, thereby
completing the normal duodenal loop. The third part of
the duodenum crosses the midline from right to left in the
angle between the SMA (anterior to the duodenum) and
the aorta (posterior). The fourth part of the duodenum as-
cends to the left of the spine to reach the duodeno-jeju-
nal flexure (D-J flexure) at the ligament of Treitz. The po-
sition of the D-J flexure has been traditionally used as a
landmark for documenting normal rotation of the midgut
on contrast examinations of the upper gastrointestinal
(GI) tract. The other landmark used is the position of the
cecum in the right iliac fossa. When the D-J flexure and
cecum are in their normal positions, the small bowel
mesentery has a long base extending from the left upper
quadrant (D-J flexure) obliquely down to the right lower
quadrant (cecum). This long mesenteric base protects
against the development of volvulus.
Abnormalities due to the arrest of rotation and/or fix-
ation can occur at any phase of the above-described
process and may involve only a part or all of the midgut.
This leads to many variations of malrotation and/or mal-
fixation and accounts for the spectrum of clinical pre-
sentations and radiological appearances.
Causes of Symptoms
The majority of these variations of malrotation and /or
malfixation are associated with clinical symptoms that
usually present within the first few months of life and can
be life-threatening [1, 2]. Others may be associated with
few or no symptoms, with the latter type often found on-
ly incidentally.
Abnormalities of rotation and/or fixation usually lead
to a situation in which the cecum and duodenum lie clos-
er to each other than normal, such that the base of the
small bowel mesentery is much shorter than normal. The
obstruction occurs primarily in the duodenum (Figs. 1, 2)
and is most often due to peritoneal (Ladd) bands that an-
chor the cecum to the retroperitoneum across the duode-
num in the right upper quadrant. However, even more im-
portantly, obstruction may also be due to volvulus as a re-
248 Alan Daneman

Fig. 1. Series of anteroposterior fluoroscopic images from a contrast examination of the upper GI tract in a young infant with malrotation.
In the initial image (top left), before contrast was injected into the nasogastric tube, there is a non-specific bowel gas pattern with dimin-
ished gas in the mid and right abdomen. Contrast administration is followed by intermittent dilatation of the second and proximal third
parts of the duodenum; initially, no dilatation is visible but subsequent images show varying degrees of dilatation. The dilated third part of
the duodenum reaches only to the level of the right pedicles and does not cross the midline. The distal duodenum is on the right. The nor-
mal D-J flexure is absent. The findings are diagnostic of midgut malrotation with obstruction of the duodenum by Ladds bands and sug-
gest a volvulus, which was confirmed at surgery. This case illustrates the potential non-specificity of the abdominal radiograph in the pres-
ence of duodenal obstruction and the intermittent nature of the duodenal dilatation

sult of the narrow base of the small bowel mesentery

(Fig. 2). Much less commonly, there may be an associat-
ed internal hernia (Fig. 3). The clinical picture and imag-
ing appearance depend on the nature and degree of the
obstruction (which may be intermittent) as well as on the
presence or absence of vascular compromise.

Fig. 2. Lateral fluoroscopic view of contrast examination of the up-
per GI tract in a neonate with malrotation and volvulus. The prox-
imal duodenum is dilated and ends in a beak that leads into a
corkscrew pattern of non-dilated small bowel. The beak and
corkscrew pattern are typical of obstruction due to volvulus, which
was confirmed at surgery. The lateral view is often better than the
anteroposterior view shown in Fig. 1 in demonstrating the volvu-
lus but the latter view remains essential to depict the position of
the of D-J flexure
Imaging Modalities
As the clinical findings are often non-specific, pediatri-
cians and surgeons rely on the radiologist to confirm or
exclude the diagnosis [1-3]. The radiologist thus plays an
exceptionally important role in the diagnosis of a malro-
tation and must be able to recognize the spectrum of ap-
pearances of its many variations, as depicted by any imag-
ing modality (Figs. 1-5). Failure to do so may lead to a
delay in treatment and thus potentially to bowel necrosis
(which may require extensive resection) and even death.
On a plain abdominal radiograph, malrotations may
show a wide spectrum of appearances [3]. In contrast to
what might be expected, the finding of duodenal disten-
tion with gas as a typical component of duodenal ob-
struction is often absent. If the duodenum is fluid-filled or
Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management
Fig. 3. Anteroposterior abdominal radiograph from an upper GI se-
ries and follow-through contrast examination in a young teenager
with midgut malrotation and an internal hernia. Contrast outlines
the stomach and duodenal cap and then leads into a very well de-
fined, rounded collection of small bowel loops on the right, repre-
senting a paraduodenal internal hernia. This appearance of con-
tained, well defined small bowel is typical of an internal hernia
Fig. 4. Contrast enema in a young infant with malrotation proven at
surgery. Note that the proximal colon turns back towards the left in
the right upper quadrant, and the cecum and appendix lie in the up-
per abdomen close to the position of the duodenum. As a result of
the proximity of the duodenum and cecum, the small bowel mesen-
tery is much shorter than normal, thus predisposing to develop-
ment of a volvulus
collapsed, it may not be visible and gaseous distension of
the stomach alone may suggest gastric obstruction. A
volvulus may lead to the presence of a soft-tissue mass
due to the fluid-filled bowel, or the abdomen may be al-
most completely gasless. In children with severe vascular
compromise due to volvulus, there is often gaseous
249
Fig. 5 a, b. Transverse sonograms of the upper abdomen in a neonate
with malrotation and volvulus proven at surgery. The volvulus
appears as a characteristic whirlpool in gray-scale (a) imaging and
on color Doppler evaluation (b). The small bowel and mesenteric
veins are coiled around the superior mesenteric artery
distention (with air fluid levels) of the entire small bowel,
resembling a low bowel obstruction or ileus. The appear-
ances of malrotation are indeed often non-specific (Fig. 1)
and may be normal even in the presence of volvulus.
Due to this wide variation of appearances on plain ra-
diographs, the clinician should never rely on plain film
findings to exclude malrotation. Any child in whom there
is a clinical suspicion of malrotation should be studied
with modalities that will directly depict the position of the
bowel and/or the nature of the obstruction. These include
following modalities:
1. Contrast examinations of the gastrointestinal (GI) tract:
An upper GI series (alone or in combination with a fol-
low-through series) and/or a contrast enema (Figs. 1-4)
[4-8]. The most important features of the GI tract to
define are the position of the D-J flexure (which re-
quires a straight A-P view on the upper GI series)
and/or the position of the cecum and proximal colon.
250
2. Cross-sectional imaging of the abdomen, particularly
with sonography (but also with computed tomogra-
phy and magnetic resonance) [9-15]. These modali-
ties may depict a dilated duodenum, malposition of
the D-J flexure, the whirlpool sign indicative of
volvulus (Fig. 5), an internal hernia, or abnormalities
of the relationship of the SMA and superior mesten-
teric vein.
Meticulous attention to technique is critical with each
of these modalities in order to delineate the relevant
structures accurately. However, even when a technically
perfect examination is performed, none of the above fea-
tures are 100% accurate in allowing the confirmation or
exclusion of malrotation. Accordingly, there has been
much debate over the years as to which modality should
be used first and what protocol of subsequent modalities
may be required in order to enable the radiologist to
rapidly make an accurate diagnosis. Independent of
which modality is chosen first, the radiologist should
never hesitate to ask for or perform other types of exam-
inations aimed at obtaining more information that may
increase confidence in the diagnosis – whether malrota-
tion is present or not. As information from the various ex-
aminations is accumulated, the balance of evidence from
all the examinations should be weighed together so that
the correct diagnosis can be determined.
Most institutions still use the upper GI series as the
modality of choice in children in whom malrotation is
suspected clinically [4-8]. This examination is relative-
ly non-invasive, easy to perform, and the position of the
D-J flexure is highly accurate in predicting malrotation
(Fig. 1). Indeed, the presence or absence of malrotation
can be readily made in most children using this modal-
ity alone. However, there will remain a group of chil-
dren in whom the diagnosis of malrotation will be dif-
ficult based on the upper GI series alone, either because
of technical difficulties in some patients or because of
difficulties in differentiating normal variations in duo-
denal anatomy from true abnormalities of rotation. It is
in these children that the radiologist must not be reluc-
tant to extend the GI contrast examination or to perform
another type of examination that will generate further
diagnostic information.
For this purpose, the next most commonly evaluated
factor to determine on contrast examinations of the GI
tract is the position of the cecum [4, 5]. In the acute clin-
ical situation, particularly in neonates and young infants,
the quickest way to achieve this is by performing an im-
mediate contrast enema (Fig. 4), i.e., before too much
contrast from the prior upper GI series fills the small
bowel. The advantage of the contrast enema is that it rel-
atively quickly provides information on the position of
the entire large bowel. However, interpretation of the ce-
cal position may be difficult. It must be remembered that
the position of the cecum and proximal small bowel has
a wide range of normality, particularly in neonates and
young infants, and may well be normal even in patients
with malrotation. Furthermore, the cecum and proximal
Alan Daneman
colon may be easily displaced into positions that simu-
late malrotation by adjacent markedly dilated loops of
small bowel. This is seen particularly in neonates with
congenital obstruction involving the distal small bowel.
In such patients, the presence of a microcolon usually
excludes malrotation, as this combination is rare. In old-
er children, especially if the clinical setting is not acute,
contrast from the original upper GI series can be fol-
lowed with serial plain radiographs to determine the ce-
cal position. This approach may take much longer than
performing a contrast enema and is usually less fruitful
in the neonate and young infant, since in this age group
it is not always possible to clearly depict the position of
the cecum.
In recent years, signs of malrotation have been visu-
alized using sonography and other cross-sectional imag-
ing modalities, but they are still not used as the modali-
ty of choice by most radiologists [9-15]. It is true, how-
ever, that sonography is being used much more fre-
quently than suggested in the literature. Direct visualiza-
tion of a volvulus (whirlpool sign) (Fig. 5) may obviate
the necessity for contrast examination of the GI tract; but
this sign is not present in those children without volvu-
lus but who are symptomatic because of obstruction due
to bands. Furthermore, the sign may be difficult to ap-
preciate in children with volvulus and a large amount of
dilated, gas-filled bowel. Inversion of the superior
mesenteric artery and vein relationship may be present
in normal rotation, and a normal relationship may be
present in children with malrotation. Therefore, a normal
sonogram does not exclude malrotation and, to date,
sonography has not been used as a screening procedure
for this condition. Nevertheless, it is essential that radio-
logists be able to recognize these abnormal signs when
they are detected as an incidental or unexpected finding
on cross-sectional imaging.
Yousefzadeh [15] has drawn attention to the sono-
graphic depiction of the third part of the duodenum (D3)
in the angle between the SMA and the aorta, and has sug-
gested that documentation of the presence of the D3 in
this normal position excludes the presence of malrota-
tion. This is an extremely interesting approach because, if
shown to be accurate, then the sonographic depiction of
D3 in this position would obviate the necessity for doing
a fluoroscopic, contrast examination of the upper GI tract
in those children suspected of having this condition.
However, to date, there are no data to substantiate this
suggestion.
Filling the duodenum with fluid administered orally or
through a feeding tube may facilitate delineation of the
position of the duodenum and D-J flexure on sonography.
This approach has been advocated by some as the tech-
nique of choice, and some groups have used it quite ex-
tensively but have yet to publish their data. Gent and
LeQuesne presented their experience with this technique,
based on over 100 cases, at the World Federation of Ul-
trasound in Medicine and Biology Meeting in Sydney in
2009. These authors illustrated their technique, which
Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management
delineates the anatomy of the duodenum exquisitely. They
have been able to easily confirm the diagnosis of mal-
rotation in the vast majority of patients, with no false-
negatives. In a few patients in whom the diagnosis was
unsure, an upper GI series was required for confirmation.
To the best of this author’s knowledge, there has, to date,
been no large single series comparing the accuracy of this
technique to the routine contrast upper GI series.
Color Doppler sonographic assessment of bowel per-
fusion may be useful in patients in whom bowel perfusion
is in doubt. However, its role is limited as poor perfusion
is usually seen in sicker patients and surgeons are keen to
get these children to the operating room as soon as pos-
sible without necessarily relying on the need for diagnos-
tic imaging.
Conclusions
Accurate diagnosis of malrotation requires (i) a high in-
dex of suspicion, (ii) an understanding of normal varia-
tions of the GI tract, (iii) an appreciation of the spectrum
of appearances of malrotation on imaging modalities, and
(iv) attention to meticulous technique when performing
GI contrast examinations or cross-sectional imaging.
Although the diagnosis of malrotation can be accu-
rately made in many instances based on a single exami-
nation, the radiologist should never be reticent about using
more than one examination (or repeating examinations)
when attempting to confirm or exclude malrotation in
difficult cases. As information is accumulated from the
various diagnostic examinations, the balance of evidence
derived from all of them should be weighed together to
determine the correct diagnosis. If uncertainty remains
after all imaging avenues have been exhausted, as does
happen (though uncommonly), then the decision whether
to perform laparoscopy or to operate should be left to the
surgeon.
251
References
1. Strouse PJ (2000) Disorders of intestinal rotation and fixation
(“malrotation”). Pediatr Radiol 34:837-851
2. Lampl B, Levin TL, Berdon WE, Cowles RA (2009) Malrota-
tion and midgut volvulus: a historical review and current con-
troversies in diagnosis and management. Pediatr Radiol
39:359-366
3. Daneman A (2009) Malrotation: the balance of evidence. Pe-
diatr Radiol 39(2):S164-S166
4. Long FR, Kramer SS, Markowitz RI et al (1996) Intestinal
malrotation in children: Tutorial on radiographic diagnosis in
difficult cases. Radiology 198:775-780
5. Long FR, Kramer SS, Markowitz RI, Taylor GE (1996) Radio-
graphic Patterns of Intestinal Malrotation in Children. Radio-
Graphics 16:547-556
6. Katz ME, Siegel MJ, Shackelford GD, McAlister WH (1987)
The Position and Mobility of the Duodenum in Children. AJR
148:947-951
7. Donnolly LF, Rencken IO, de Lorimier AA, Gooding CA (1996)
Left paraduodenal hernia leading to ileal obstruction. Pediatr
Radiol 26:534-536
8. Manji R, Warnock GL (2000) Left paraduodenal hernia: an un-
usual cause of small-bowel obstruction. Can J Surg 44:455-457
9. Dufour D, Delaet MH, Dassonville M et al (1992) Midgut mal-
rotation, the reliability of sonographic diagnosis. Pediatr Radiol
22:21-23
10. Chao HC, Kong MS, Chen JY et al (2000) Sonographic fea-
tures related to volvulus in neonatal intestinal malrotation. J
Ultrasound Med 19:371-376
11. Shimanuki Y, Aihara T, Takano H et al (1996) Clockwise
whirlpool sign at color Doppler US: an objective and definite
sign of midgut volvulus. Radiology 199:261-264
12. Yoo SJ, Park KW, Cho SY et al (1999) Definitive diagnosis
of intestinal volvulus in utero. Ultrasound Obstet Gynecol
13:200-203
13. Loyer E, Eggli KD (1989) Sonographic evaluation of superior
mesenteric vascular relationship in malrotation. Pediatr Radiol
19:173-175
14. Weinberger E, Winters WD, Liddell RM et al (1992) Sono-
graphic diagnosis of intestinal malrotation in infants: impor-
tance of the relative positions of the superior mesenteric vein
and artery. AJR Am J Roentgenol 159:825-828
15. Yousefzadeh DK (2009) The position of the duodenojejunal
junction: the wrong horse to back on in diagnosing or exclud-
ing malrotation. Pediatr Radiol 39:S172-S177
 IDKD 2010-2013

Pediatric Intestinal Ultrasonography

Simon G. Robben
Department of Radiology, Maastricht University Medical Centre, Maastricht, The Netherlands

Introduction However, increased knowledge, improved technique

(e.g., graded compression), improved hardware (high-
Ultrasonography (US) is the imaging modality of choice frequency transducers), and improved software (adaptive
for the initial evaluation of diseases in children for imaging, compound imaging) have resolved the limita-
many reasons. First, it is relatively inexpensive and pa- tions to US use. Nowadays it is impossible to imagine
tient friendly. Second, it lacks radiation and motion ar- US without intestinal US, especially in pediatrics! The
tifacts. Third, the small size of the child compensates for hallmark of intestinal US is the “gut signature”, i.e.,
the limited penetration of sound waves and facilitates the characteristic appearance of the layers of the gut
the use of high-frequency transducers. Fourth, flow (Table 1 and Fig. 1).
studies are possible using the Doppler mode, while real-
time imaging allows the visualization of movements
such as peristalsis. Moreover, US involves direct contact
with the patient, thus offering a unique opportunity to
ask specific questions and to perform additional physi-
cal examinations, emphasizing the role of the radiolo- Table 1. Gut signature, characterized by alternating hyper- and
gist as a clinician. Accordingly, ultrasonography has be- hypoechoic layers
come the most important imaging technique in children
Layer Echogenicity
and can be considered as the workhorse of pediatric
radiology. Mucosal surface Hyperechoic
Initially, US of the stomach and intestines was not Mucosa Hypoechoic
popular for obvious reasons: bowel gas has the annoying Submucosa Hyperechoic
characteristic of reflecting all sound waves or creating Muscularis propria Hypoechoic
artifacts because of its abnormally low acoustic impedance. Serosal surface Hyperechoic

a b c

Fig. 1 a-c. Gut signature in a stomach, b ileum, and c appendix (between arrows)
Pediatric Intestinal Ultrasonography 253

Gastroesophageal Junction palpation is accurate but not always successful as it de-

The gastroesophageal junction (Fig. 2) can be visualized
with US in 87-95% of children with suspected gastro-
esophageal reflux disease (GERD), and reflux of gastric
content into the esophagus can be demonstrated ultra-
sonographically [1, 2]. A threshold of three reflux periods
within 10 min corresponds to a sensitivity of 82% and a
specificity of 85% for GERD [1]. Farina et al. increased
the sensitivity to 98% using color Doppler US [3]. How-
ever, in premature infants the results differ: the sensi-
tivity is 38% and the specificity is 100% compared to
24-hour pH-metry [4].
US studies have also shown that there is a positive cor-
relation between GERD and the length of the abdominal
esophagus, protrusion of the gastric mucosa, and an in-
creased gastroesophageal angle (angle of His). Therefore,
US can be the initial imaging technique in children sus-
pected of having GERD. Moreover, it can evaluate the
post-operative situation after a fundoplication and gastric
emptying, another important contributing factor to the
pathogenesis of GERD.
pends on factors such as the experience of the examiner,
the presence of gastric distension, and the cooperation of
the infant.
In virtually all patients, US is very accurate in facili-
tating the diagnosis and therefore plays a key role in the
initial care of these infants. It is important that the radio-
logist understands the anatomical changes of the pyloric
channel in affected infants, as demonstrated by US. Py-
loric muscle hypertrophy is shown to a variable degree
during the US examination. In addition, a certain amount
of thickening of the mucosa is present (Fig. 3). A muscle
thickness that is consistently v3 mm is considered to be
diagnostic of IHPS, although some clinicians have stated
that the overall morphological and dynamic impression,
including length of the pyloric canal as well as relaxation
and peristalsis, are just as important. The US examination
can be performed in a very short time with an accuracy
approaching 100%.

Stomach
Gastric emptying in vomiting patients can be evaluated
with US, which depicts dynamic and anatomical abnor-
malities.

Infantile hypertrophic pyloric stenosis (IHPS) [5] is a

condition of unknown etiology that affects young infants
ages 2-8 weeks and with a male-to-female ratio of ap-
proximately 4:1. In IHPS, the antropyloric portion of the
stomach becomes abnormally thickened and manifests as
an obstruction to gastric emptying. Typically, infants with
IHPS are clinically normal at birth but during the first
few weeks of postnatal life, they develop non-bilious
“projectile” vomiting that leads to weight loss, dehydra-
tion and hypochloremic alkalosis, and eventually death. Fig. 3. Hypertrophic pyloric stenosis (between large arrows). Mus-
Surgical treatment is curative. The clinical diagnosis re- cle thickening is indicated by small arrows. A Antrum of stomach
lies on palpation of the thickened pylorus. Abdominal with retained fluid

a b

Fig. 2 a, b. Sagittal slice of a

normal gastroesophageal
junction (arrows) in a 3-
month-old boy. b Same
slice during transit of air.
During real-time ultra-
sonography, the direction
of transit can easily be
appreciated. S Stomach,
L liver
254
Foveolar hyperplasia consists of a polypoid thickening of
the mucosal layer. It may be seen after long-standing
prostaglandin therapy, hypertrophic gastropathy, or cow’s
milk allergy, but it may also be idiopathic. While the con-
dition may simulate pyloric hypertrophy, on closer US
examination the obstruction will appear to be caused by
thickened mucosa instead of thickened muscle.
Rare causes of gastric outlet obstruction include
pylorospasm, (eosinophilic) gastritis, food allergy,
chronic granulomatous disease, hyperlipidemia, du-
plication cysts, ectopic pancreas, benign and malignant
tumors, and bezoars.
Small Bowel
Conventional radiography and US are the initial imag-
ing modalities in children with abdominal pain or ob-
struction. The most important additional value of US
over conventional abdominal radiographs in these chil-
dren is its capability to visualize peristalsis, vascularity,
bowel wall characteristics, dilatation of fluid-filled
loops, and extra-intestinal abnormalities, e.g., ascites
and other fluids.
The jejunum and ileum can be distinguished from the
colon based on anatomical location, caliber, contents,
folds, and peristalsis.
Anatomical location: The colon has a peripheral location
in which the ascending and descending colon lie dorsal-
ly in both flanks and the transverse colon is located ven-
trally in the upper abdomen. The sigmoid colon traverses
the left psoas muscle and courses into the pelvis whereas
the small bowel has a more central position.
Caliber: The diameter of the small bowel is small while,
as its name indicates, the diameter of the large bowel is
relatively large.
a b
Simon G. Robben
Contents: The small bowel is either empty or filled with
liquid contents but little air, whereas the colon is gener-
ally filled with gas-filled bulky stools.
Folds: The folds in the jejunum are more numerous,
longer, thinner, and closer together than the ileal folds. In
the terminal ileum, the mucosa may be thickened due to
hyperplasia of lymphoid tissue. The colon is recognized
by its haustrations.
Peristalsis: The small bowel moves continuously due to
persistaltic waves whereas the colon shows sparse move-
ments.
Baud proposed a systematic US approach for identify-
ing small bowel disease, based on wall thickening [6].
1. Determine wall thickening: normal (f3 mm), mild (3-
6 mm), moderate (6-9 mm), or severe (>9 mm).
2. Determine location (proximal or distal) and extent (fo-
cal 5 cm, segmental 6-40 cm, or diffuse >40 cm).
3. Determine stratification. The bowel wall is stratified
when the hyperechogenicity of the submucosa is pre-
served and the mucosa, submucosa, and muscularis
propria are visible as separate layers. Non-stratifica-
tion implies the absence of distinction between mucosa
and submucosa or between all three layers (Fig. 4).
4. Determine the valvular fold pattern: normal, thick-
ened, thumb-printing, and hyperplastic valvular folds.
In general, thickened small bowel loops show de-
creased peristalsis and contain little air. They are there-
fore easily visualized and measured. At least three pat-
terns can be distinguished.
Stratified thickening of the small bowel is found in infec-
tious ileitis, advanced appendicitis, early Crohn’s disease,
and graft versus host disease.
Non-stratified thickening occurs in Henoch-Schönlein
purpura, advanced Crohn’s disease, tuberculous ileitis,
Fig. 4 a, b. Difference be-
tween a stratified wall
thickening in early Crohn’s
disease and b non-strati-
fied wall thickening in ad-
vanced Crohn’s disease
Pediatric Intestinal Ultrasonography
Fig. 5. Two benign small bowel intussusceptions in a patient with a
malabsorption syndrome
protein-losing enteropathy, hereditary angioedema, is-
chemia, celiac disease, Burkitt’s lymphoma, Kawasaki’s
disease, and viral enteritis.
Non-stratified thickening with hyperplastic valvular folds
can be seen in viral (and sometimes bacterial) lymphoid
hyperplasia and Yersinia ileitis.
Malrotation and midgut volvulus are discussed in the
chapter by Daneman.
Benign small bowel intussusception (BSBI) is a recently
described entity. It differs from the classical symptomatic
ileocolic intussusception in that it occurs predominantly
in the right lower quadrant or periumbilical region, has a
smaller diameter (mean diameter 1.4 vs. 2.5 cm), a thin-
ner outer rim, and does not contain mesenteric lymph
nodes (Fig. 5) [7]. Moreover, peristalsis in the intussus-
cepted loop persists, in contrast with ileocolic intussus-
ception. Often, BSBI is an incidental finding but it occurs
with increased frequency and number in celiac disease. In
general, BSBI does not need immediate reduction be-
cause of its spontaneously resolving nature.
Appendicitis, Meckels diverticulum, necrotizing entero-
colitis, and ileocolic intussusception are discussed else-
where in this volume.
Large Bowel
The differences between the normal small and large bow-
el were described above. In addition to the systematic ap-
proach to the small bowel, Baud proposed an analogous
approach to the colon [8].
1. Determine wall thickening: normal (f3 mm), mild (3-
6 mm), moderate (6-9 mm), or severe (>9 mm).
2. Determine the extent and location of disease: diffuse,
cecum, ascending colon, proximal and distal trans-
verse colon, descending colon, or rectosigmoid.
255
3. Determine stratification, which is best seen on trans-
verse images. The bowel wall is stratified when the
hyperechogenicity of the submucosa is preserved and
the mucosa, submucosa, and muscularis propria are
visible as separate layers. Non-stratification implies
an absence of distinction between the mucosa and
submucosa or between all three layers.
4. Determine the haustral pattern, which is best seen on
longitudinal images; also, confirm the absence or pres-
ence of the haustral folds and their length (normal or
shortened) and aspect.
These criteria distinguish three patterns.
Stratified thickening is found in infectious colitis, ad-
vanced appendicitis, and inflammatory bowel disease (ul-
cerative colitis and Crohn’s disease).
Non-stratified thickening with loss of haustral folds is
found in early hemolytic uremic syndrome (HUS) and
advanced Crohn’s disease.
Non-stratified thickening with preservation of normal
haustral fold length is found in pseudomembranous coli-
tis and neutropenic colitis (typhlitis).
Rectum
In several studies, US was used to measure the trans-
verse diameter of the rectum. This seems to be a reliable
approach to identifying rectal impaction and may replace
digital rectal examination. All children with rectal im-
paction on digital examination were ultrasonographical-
ly shown to have had a rectal diameter >30 mm [9].
Moreover, several studies reported that in children with
constipation the mean diameter of the rectum is signifi-
cantly larger than in normal children. In a series of 225
children, the mean rectal diameter in normal children
was 32 mm (SD 8.2) and in children with constipation
43 mm (SD 9.7) [10]. To overcome the problem of age-
dependency of the rectal diameter in normal children,
Bijos et al. proposed the rectopelvic ratio, defined as the
ratio of the rectal diameter (as determined with US) to
the distance between the anterior superior iliac spines
[10]. A rectopelvic ratio >0.189 corresponds to a sensi-
tivity for rectal impaction of 88% compared to procto-
scopy. The rectal diameter can also be used to monitor
therapy.
Anus
Anal atresia is a relatively frequent congenital abnor-
mality in which the anus is absent and a rectoperineal,
rectovestibular, rectovaginal, rectourethral, or rectovesi-
cal fistula may be identified in almost all cases. Since
the fistula demonstrates an internal sphincter, some clin-
icians instead prefer the term “ectopic anus” or “anorectal
256
malformation”. During pre-operative evaluation, it is im-
portant to assess the type of anal atresia, which may be
high (distal rectal pouch above the puborectal sling), in-
termediate (at the sling), or low (through the sling).
Transperineal US is a good diagnostic modality for
defining the type of anal atresia as it can be used to mea-
sure the distance between the rectal pouch and the per-
ineum (P-P distance). A P-P threshold value of 15 mm
discriminates the low type of atresia from the intermedi-
ate and high types with a sensitivity of 100% and a
specificity of 86% [11]. Moreover, internal fistula can be
correctly identified in 82% of the patients with the high
type of anal atresia [12].
Transperineal sonography is also a useful method for
differentiating between an anteriorly displaced anus,
which is a normal anatomical variant, and a low-type im-
perforate anus with perineal fistula, which is a patholog-
ical developmental abnormality requiring surgical repair
[13]. In adults, transperineal ultrasonography is a simple,
painless, cost-effective and real-time method to detect
and classify perianal fluid collections, abscesses, fistulas,
and sinus tracts [14-16]. These data can probably be ex-
trapolated to the pediatric population, e.g., children with
Crohn’s disease.
Cystic Intestinal Masses
Cystic intra-abdominal masses originating from the ali-
mentary canal are increasingly recognized because of the
advent of routine prenatal US. These masses can be di-
vided into cysts originating from solid organs (mes-
enchymal hamartoma, congenital splenic cyst, pancreatic
pseudocyst, pancreatic cystadenoma, hydronephrosis,
multicystic dysplastic kidney, multilocular cystic nephro-
ma, adrenal hemorrhage, ovarian cysts and cystic neo-
plasms, hematocolpos, urachal cysts, abdominal and
sacrococcygeal teratoma, and cerebrospinal fluid pseudo-
cyst) and those originating from the alimentary canal and
its appendages (hydrops of the gallbladder, choledochal
cyst, mesenteric and omental cysts, gastrointestinal du-
plication cyst, meconium pseudocyst, and appendiceal
abscess) [17] (Table 2).
Table 2. Cystic masses of the gut and gut-related structures
Biliary system
Choledochal cyst
Hydrops of gallbladder
Gastrointestinal tract
Mesenteric cyst/lymphangioma
Enteric/duplication cyst
Omental cyst
Meconium pseudocyst
Miscellaneous
Abscess
Teratoma
Sacrococcygeal teratoma
Simon G. Robben
When a cystic mass is found on US examination, it
should be evaluated for its size, shape, location, relation
to organs, and contents and wall characteristics. In the
majority of cases, US can provide a specific diagnosis or
offer a narrow differential diagnosis [18].
Hydrops of the gallbladder is a rare cause of a right up-
per quadrant mass in children. It has been described in
the absence of stones, infections, or congenital abnor-
malities, in which case it is probably caused by transient
obstruction of the cystic duct or increased mucus secre-
tion with ineffective emptying. It has also frequently been
associated with Kawasaki’s disease. Rarely, it is associat-
ed with childhood infections, Henoch-Schönlein purpura,
Cryptosporidium infection in immunocompromised chil-
dren, Epstein-Barr virus infections, and typhoid fever. US
examination reveals a dilated anechoic elliptical gallblad-
der without wall thickening. The bile ducts are not dilat-
ed and no stones are seen [17, 19].
Choledochal cysts are actually focal cystic dilatations of
the biliary tree. Most patients present in the first decade
of life with symptoms of episodic abdominal pain, mass,
and jaundice. US is the best initial method of evaluating
dilatation of the bile ducts. On US, the choledochal cyst
is located in the porta hepatis, separate from the gall-
bladder, with bile duct(s) leading into or out of it. The en-
tire biliary tree should be evaluated but intrahepatic bile
duct dilatation may be absent. Surgical resection is nec-
essary to prevent the development of ascending cholan-
gitis, stones, or malignant degeneration [17, 18, 20].
Mesenteric cysts are cystic lymphangiomas that are most
often found in the small bowel mesentery, especially the
ileal mesentery but also in the greater and lesser omen-
tum and occasionally in the mesenteric root and retroperi-
toneum. About one-third occur in children younger than
15 years. Newborns present with abdominal distention
and a palpable mass whereas children are much more
likely to present with pain, anorexia, vomiting, or fever.
US can characterize the mass as a typical thin-walled uni-
or multilocular cyst that displaces adjacent structures to
the periphery of the abdomen (Fig. 6). Calcification of
Fig. 6. Multilocular cystic lymphangioma
Pediatric Intestinal Ultrasonography 257

the wall is rare in children. The cyst content is anechoic a

but if hemorrhage has occurred then debris may be seen.
Lymphangiomas may be so large that they are difficult to
distinguish from severe ascites. However, ascites sepa-
rates individual bowel loops and fills the perihepatic and
perisplenic spaces [17, 18].

Gastrointestinal duplication cysts are spherical or tubular

masses adherent to the gastrointestinal (GI) tract and
sometimes communicating with it [17, 18, 21]. These
cysts are lined with intestinal epithelium and contain
smooth muscle within their walls. They may occur any-
where along the alimentary tract but the most common
location is the ileum, followed by the stomach. Most pa-
tients present within the first year of life; their symptoms
include GI obstruction and, less commonly, a palpable
mass, intussusception, and abdominal distension.
The cystic nature of duplication cysts can be easily ap-
preciated on US. The content of the cyst is often anechoic
but there may be debris after hemorrhage or due to mu-
coid material. Rarely, the cyst appears completely hyper-
echoic after hemorrhage. Two signs are virtually diag-
nostic of duplication cysts:
1. a double-layered wall consisting of echogenic mucosa
and hypoechoic muscularis propria (Fig. 7);
2. peristalsis in the cyst. b
In 15-20% of cases, the cyst contains gastric mucosa,
which accounts for the above-mentioned hemorrhages.

Meconium pseudocyst is a manifestation of the cystic

type of meconium peritonitis that results from in utero
bowel perforation. Bowel perforation may be secondary
to intestinal obstruction (meconium ileus, atresia, or
volvulus) or idiopathic. The spilled meconium is encap-
sulated and forms a large meconium-filled (hyperechoic)
cyst that is lined by a thick inflammatory and fibrotic
membrane, often containing calcifications. The perfora-
tion may still communicate with the cyst postnatally. Di-
lated fluid-filled bowel loops may be seen next to the cyst
if obstruction is still present. Distant peritoneal or scrotal
calcifications are additional evidence for meconium peri-
tonitis and may also be demonstrated with US or con-
ventional radiographs [17, 18, 22]. Fig. 7 a, b. Duplication cyst. a Large right-sided cyst on prenatal
magnetic resonance imaging shows no typical features. b Postnatal
Intra-abdominal abscesses often are complications of ap- ultrasound demonstrates gut signature in the wall of the cyst
pendicitis, inflammatory bowel disease, or intra-abdomi- suggestive of duplication cyst
nal surgery. However, they can also occur in patients on
immunosuppressive therapy or in those with AIDS or
chronic granulomatous disease. The typical presentation is Abdominal teratomas usually arise from the retroperi-
intermittent fever, increased white blood cell count or toneum, presacral region, or the ovary, whereas mesen-
C-reactive protein, and abdominal tenderness. US shows teric and gastric teratomas are rare. These tumors occur
an ill-defined, fluid-filled mass with a thickened wall and mainly in children, and 80% are benign. Mature, imma-
irregular inner surface. The fluid (pus) demonstrates sep- ture, and malignant types have been described. Most chil-
tations, debris, or debris-fluid levels. The presence of gas dren present with a palpable abdominal mass that on US
bubbles in the pus often makes differentiation from bow- is seen to be either (multi)cystic or solid. Teratomas typ-
el loops difficult; the presence of peristalsis rules out an ically contain calcifications and fat. The latter is difficult
abscess. If feasible, ultrasonographically guided aspiration to appreciate with US but highly echogenic tissue is sug-
and/or drainage can be performed, if necessary in the ICU. gestive of fatty tissue.
258
Sacrococcygeal teratomas usually present as a large ex-
ternal cystic tumor located at the coccyx but in some
cases consist of a large cystic intra-abdominal mass
without any external mass. These pre-sacral tumors ob-
struct the rectum and bladder. A pre-sacral extension of
an intra-abdominal cyst provides a clue to the diagno-
sis of sacrococcygeal teratoma. Cystic tumors carry a
better prognosis than do solid, hypervascular tumors
[17, 18].
Cerebrospinal fluid pseudocyst (liquor cyst) is a compli-
cation resulting from ventriculoperitoneal shunt, with a
frequency of approximately 3%. Risk factors for pseudo-
cyst formation are related to inflammatory processes and
CNS tumors. Pseudocysts tend to occur within 6 months
of the last abdominal surgical procedure. Children pre-
sent with abdominal pain, distention, or mass. US will
demonstrate a sonolucent, well-defined mass in non-
infected cysts, whereas infected cysts show septa, inter-
nal debris, and fluid-fluid levels. There is no statistically
significant correlation between pseudocyst size and the
presence of infection. It is important to identify the tip of
the shunt within the cyst, producing the characteristic
“railroad sign” [17, 23, 24].
Solid Intestinal Masses
The many types of solid masses (often of neoplastic ori-
gin) that can be found during US examination are sum-
marized in Table 3.
Burkitt’s lymphoma is a fast-growing and aggressive ma-
lignant neoplasm predominantly affecting children and
that may be associated with immunodeficiency. US will
demonstrate bowel wall thickening combined with a
mesenteric mass, predominantly of the cecum and distal
Table 3. Solid masses of the gut and gut-related structures
Lymphoma
Burkitt’s lymphoma
Non-Hodgkin’s lymphoma
Peritoneal, mesenteric, and omental
Lipoma, lipoblastoma, and liposarcoma
Inflammatory myofibroblastic tumor
Fibromatosis (= desmoids)
Neurofibromas (NF1-associated)
Rhabdomyosarcoma
Metastasis
Mesenteric lymphadenitis
Small and large bowel
Juvenile colonic polyps
Hamartomatous small bowel polyps
Polyposis
Carcinoid
Vascular malformations
Simon G. Robben
ileum. The bowel wall thickening is extensive, asymmet-
rical, and poorly stratified whereas the mesenteric in-
volvement is bulky and lobulated and appears to be in
continuity with the bowel wall. Despite the extensive bow-
el wall thickening, the lumen may remain wide. Burkitt’s
lymphoma can lead to intestinal obstruction and intussus-
ception. It can also involve the liver, spleen, kidneys and
pancreas. Extensive involvement of the omentum and
peritoneum is rare [25, 26].
Intra-abdominal lipomatous tumors (lipoma, lipoblastoma,
and the rare liposarcoma) predominantly involve the
mesentery and omentum [27-29]. They show hypo- or
hyperechoic textures and are finely lobulated, homo-
geneous, or with fibrovascular septa. In most cases,
magnetic resonance imaging is necessary for further
evaluation.
Inflammatory myofibroblastic tumor (inflammatory
pseudotumor) most commonly occurs in the mesentery of
children or young adults [29, 30]. The typical complaints
are of fever, malaise, weight loss, or abdominal pain. The
US characteristics of inflammatory pseudotumor are non-
specific: solid, well-defined (sometimes lobulated), with
mixed echo-texture and frequent calcifications. Infiltra-
tion of the adjacent bowel may occur. Prominent vascu-
larity may be shown with Doppler US.
Fibromatosis or abdominal desmoid is part of the clini-
cal-pathological spectrum of deep fibromatoses [29, 30].
The latter encompass a group of benign fibroproliferative
processes that are locally aggressive and have the capac-
ity to infiltrate or recur but not to metastasize. Mesenteric
structures are the most common sites of origin of intra-
abdominal fibromatosis. Other locations are the abdomi-
nal wall, pelvis, and retroperitoneum. Thirteen percent of
patients with mesenteric fibromatosis have familial ade-
nomatous polyposis (FAP), specifically, the Gardner syn-
drome variant. In these patients, prior abdominal surgery
is an important risk factor for the development of mesen-
teric fibromatosis. The US appearance is a solid, well-
circumscribed mass of variable echo-texture and homo-
geneity. Locally aggressive fibromatosis infiltrates the
mesenteric fat.
Neurofibromatous tumors are associated with neurofibro-
matosis type 1 (NF1). Abdominal involvement is found
in 10-25% of patients with NF1, regardless of their age.
Intra-abdominal neurofibromas present as hypoechoic
heterogeneous masses or as multiple rounded, hypo-
echoic, well-circumscribed, variably sized mesenteric
nodules.
Rhabdomyosarcomas are rare intra-abdominal pediatric
tumors involving the mesentery, peritoneum, or omen-
tum, often associated with ascites. Leung et al. described
an omental embryonal rhabdomyosarcoma consisting of
lobulated round masses surrounded by tissue with a
Pediatric Intestinal Ultrasonography 259

cerebriform pattern [31]. Rhabdomyosarcomas may also a

occur in the biliary tree, and are actually the most common
pediatric tumor of the biliary tree. US findings are non-
specific but if a solid tumor in the biliary tree is detected,
it is highly suggestive of rhabdomyosarcoma [32].

Intestinal polyps in the small bowel are usually hamar-

tomatous polyps in patients with Peutz-Jeghers syn-
drome. These polyps occur more often in the jejunum
than in the ileum. US is only able to detect polyps >15
mm and is therefore not a screening tool for the detection
of uncomplicated polyps. However, it is valuable in diag-
nosing complications associated with polyps of the small
bowel, e.g., intussusceptions.
b
Juvenile polyps are the most common neoplasms of the
large bowel in children. The sigmoid and rectum are pref-
erential locations. The presenting symptom is rectal bleed-
ing in over 90% of patients. Occasionally, a colo-colic in-
tussusception is the first manifestation of a juvenile polyp.
US demonstrates a pedunculated spherical nodule 10-25
mm in diameter and containing multiple 2- to 3-mm cysts.
Administration of fluid within the bowel lumen will great-
ly improve the visualization of these polyps; however, the
assessment of rectal polyps is unreliable.

Mesenteric lymphadenitis. Abdominal lymph nodes vary-

ing from 0 to 10 mm in diameter are detected in almost
all asymptomatic children (Fig. 8a). Approximately half
of these lymph nodes (43-54%) are >5 mm in their short
axis, with the number and shape of the nodes being
age-independent [33, 34]. There is a gender dependency
in that boys are more affected than girls [33, 35]. In chil-
dren with recurrent or acute abdominal pain without
known cause, the lymph nodes are significantly larger
than in asymptomatic children, but there is considerable
overlap between the two groups. Children with abdomi-
nal pain and abdominal lymph nodes >10 mm in their
short axis may be considered as having primary mesen-
teric lymphadenitis, if no other acute inflammatory
process is identified [34, 36]. There are few US criteria
that are typical for a specific causative agent. The pres-
ence of calcifications and necrosis suggests tuberculous
lymphadenitis (Fig. 8b).
Neonatal Bowel Obstruction
Many neonatal bowel obstructions are caused by diseases
discussed in other chapters, e.g., duplication cysts, mal-
rotation or volvulus, meconium peritonitis, necrotizing
enterocolitis, anal atresia, and pyloric hypertrophy. Addi-
tional causes of neonatal obstruction include various
types of atresia, annular pancreas, meconium ileus,
Hirschsprung’s disease, and meconium plug syndrome.
These diseases are diagnosed by conventional abdominal
radiographs and/or conventional contrast studies but US
may be of additional value [37].
Fig. 8 a, b. Enlarged mesenteric lymph nodes. a Asymptomatic
child shows lymph nodes with normal texture and ellipsoid shape;
the largest node has a short axis of 8 mm. b Tuberculous mesen-
teric lymph node shows indistinct margins, round shape, and het-
erogeneous texture with hypoechoic areas of necrosis and a diam-
eter of 20 mm
Atresias are congenital interruptions of the lumen of the
alimentary canal caused by a failure of canalization dur-
ing organogenesis. The result is dilatation proximal to the
atresia and complete collapse distal to the atretic segment.
While the clinical and radiographic presentation is typical,
US may be used to demonstrate accompanying pathology:
1. exclusion of associated malrotation in cases of high
obstruction;
2. evaluation of the colon and rectum in cases of high ob-
struction; single atresias are associated with a normal-
sized colon, multiple atresias, with microcolon;
3. evaluation of associated malformations of the heart,
kidneys, and biliary tree;
4. demonstration of extrinsic duodenal compression in
cases of high obstruction, e.g., duplication cysts, pre-
duodenal portal vein, annular pancreas; and
5. demonstration of signs of prenatal meconium peritoni-
tis as a cause of small bowel atresia (meconium cysts,
subtle peritoneal or scrotal calcifications).
260
Meconium ileus is a small bowel occlusion caused by in-
spissated, abnormally tenacious, meconium in the distal
ileum, invariably associated with cystic fibrosis. It occurs
in 10-20% of neonates with cystic fibrosis. US may
demonstrate the microcolon (mean diameter of 4 mm),
small bowel dilatation with echogenic contents (in con-
trast to the hypoechoic contents in ileal atresia), pseudo-
thickening of the bowel wall (caused by a circular layer
of tenacious meconium adherent to the bowel wall), gran-
ular pattern (air bubbles in the tenacious meconium), and
detection of meconium pellets in the distal ileum.
Hirschsprung’s disease is caused by an absence of gan-
glion cells, which results in abnormal gut motility and a
lack of gut relaxation. The length of the aganglionic
segment is variable but always involves the distal end of
the intestinal tract. In a small number of patients, the en-
tire colon and even the ileum and jejunum are involved.
US is of limited value because of the air-artifacts in
dilated bowel loops. However in very early neonatal US
examination, air is not yet present and distention of the
colon in a neonate with distal occlusion suggests
Hirschsprung’s disease.
Meconium plug syndrome can be considered as neonatal
constipation. US demonstrates moderate dilatation of the
entire colon without a transition zone, echogenic colonic
content, and dilatation of the proximal small bowel [37].
Conclusions
Ultrasonography is a reliable initial imaging technique to
evaluate a variety of gastrointestinal pediatric diseases
and malformations. A solid knowledge of the patho-
physiology of these conditions is necessary to understand
their ultrasonographic manifestations.
References
1. Couture A, Baud C, Ferran J et al (2008) Gastrointestinal tract
sonography in fetuses and children. 1st edn. Springer-Verlag,
Berlin-Heidelberg
2. Westra SJ, Derkx HH, Taminiau JA (1994) Symptomatic gas-
troesophageal reflux: diagnosis with ultrasound. J Pediatr Gas-
troenterol Nutr 19:58-64
3. Farina R, Pennisi F, La Rosa M et al (2008) Contrast-enhanced
colour-Doppler sonography versus pH-metry in the diagnosis of
gastro-oesophageal reflux in children. Radiol Med 113:591-598
4. Pezzati M, Filippi L, Psaraki M et al (2007) Diagnosis of gas-
tro-oesophageal reflux in preterm infants: sonography vs. pH-
monitoring. Neonatology 91:162-166
5. Hernanz-Schulman M (2003) Infantile hypertrophic pyloric
stenosis. Radiology 227:319-331
6. Baud C (2008) Small bowel thickening. In: Couture A, Baud C,
Ferran J et al (eds) Gastrointestinal tract sonography in fetuses
and children. 1st edn. Springer-Verlag, Berlin-Heidelberg,
pp 253-296
7. Park NH, Park SI, Park CS et al (2007) Ultrasonographic find-
ings of small bowel intussusception, focusing on differentia-
tion from ileocolic intussusception. Br J Radiol 80:798-802
Simon G. Robben
8. Baud C (2008) Infectious and inflammatory colitis. In: Cou-
ture A, Baud C, Ferran J et al (eds) Gastrointestinal tract
sonography in fetuses and children. 1st edn. Springer-Verlag,
Berlin-Heidelberg, pp 297-339
9. Joensson IM, Siggaard C, Rittig S et al (2008) Transabdomi-
nal ultrasound of rectum as a diagnostic tool in childhood con-
stipation. J Urol 179:1997-2002
10. Bijos A, Czerwionka-Szaflarska M, Mazur A, Romanczuk W
(2007) The usefulness of ultrasound examination of the bowel
as a method of assessment of functional chronic constipation
in children. Pediatr Radiol 37:1247-1252
11. Haber HP, Seitz G, Warmann SW, Fuchs J (2007) Transper-
ineal sonography for determination of the type of imperforate
anus. AJR Am J Roentgenol 189:1525-1529
12. Choi YH, Kim IO, Cheon JE et al (2009) Imperforate anus: de-
termination of type using transperineal ultrasonography. Kore-
an J Radiol 10:355-360
13. Haber HP, Warmann SW, Fuchs J (2008) Transperineal sono-
graphy of the anal sphincter complex in neonates and infants:
differentiation of anteriorly displaced anus from low-type im-
perforate anus with perineal fistula. Ultraschall Med 29:383-387
14. Bonatti H, Lugger P, Hechenleitner P et al (2004) Transper-
ineal sonography in anorectal disorders. Ultraschall Med
25:111-115
15. Maconi G, Ardizzone S, Greco S et al (2007) Transperineal ul-
trasound in the detection of perianal and rectovaginal fistulae
in Crohn's disease. Am J Gastroenterol 102:2214-2219
16. Stewart LK, McGee J, Wilson SR (2001) Transperineal and
transvaginal sonography of perianal inflammatory disease.
AJR Am J Roentgenol 177:627-632
17. Wootton-Gorges SL, Thomas KB, Harned RK et al (2005) Gi-
ant cystic abdominal masses in children. Pediatr Radiol
35:1277-1288
18. Khong PL, Cheung SC, Leong LL, Ooi CG (2003) Ultra-
sonography of intra-abdominal cystic lesions in the newborn.
Clin Radiol 58:449-454
19. Maurer K, Unsinn KM, Waltner-Romen M et al (2008) Seg-
mental bowel-wall thickening on abdominal ultrasonography:
an additional diagnostic sign in Kawasaki disease. Pediatr Ra-
diol 38:1013-1016
20. Kim OH, Chung HJ, Choi BG (1995) Imaging of the chole-
dochal cyst. Radiographics 15:69-88
21. Segal SR, Sherman NH, Rosenberg HK et al (1994) Ultra-
sonographic features of gastrointestinal duplications. J Ultra-
sound Med 13:863-870
22. Yang WT, Ho SS, Metreweli C (1997) Case report: antenatal
sonographic diagnosis of meconium peritonitis and subsequent
evolving meconium pseudocyst formation without peritoneal
calcification. Clin Radiol 52:477-479
23. Pathi R, Sage M, Slavotinek J, Hanieh A (2004) Abdominal
cerebrospinal fluid pseudocyst. Australas Radiol 48:61-63
24. Roitberg BZ, Tomita T, McLone DG (1998) Abdominal cere-
brospinal fluid pseudocyst: A complication of ventriculoperi-
toneal shunt in children. Pediatr Neurosurg 29:267-273
25. Biko DM, Anupindi SA, Hernandez A et al (2009) Childhood
Burkitt lymphoma: abdominal and pelvic imaging findings.
AJR Am J Roentgenol 192:1304-1315
26. Wong S, Sanchez TR, Swischuk LE, Huang FS (2009) Diffuse
peritoneal lymphomatosis: atypical presentation of Burkitt
lymphoma. Pediatr Radiol 39:274-276
27. Moholkar S, Sebire NJ, Roebuck DJ (2006) Radiological-
pathological correlation in lipoblastoma and lipoblastomatosis.
Pediatr Radiol 36:851-856
28. Prando A, Wallace S, Marins JL et al (1990) Sonographic fea-
tures of benign intraperitoneal lipomatous tumors in children-
report of 4 cases. Pediatr Radiol 20:571-574
29. Veyrac C (2008) Intraperitoneal masses. In: Couture A, Baud
C, Ferran J et al (eds) Gastrointestinal tract sonography in fe-
tuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg,
pp 511-544
Pediatric Intestinal Ultrasonography
30. Levy AD, Rimola J, Mehrotra AK, Sobin LH (2006) From the
archives of the AFIP: benign fibrous tumors and tumorlike le-
sions of the mesentery: radiologic-pathologic correlation. Ra-
diographics 26:245-264
31. Leung RS, Calder A, Roebuck D (2009) Embryonal rhab-
domyosarcoma of the omentum: two cases occurring in chil-
dren. Pediatr Radiol 39:865-868
32. Roebuck DJ, Yang WT, Lam WW, Stanley P (1998) Hepato-
biliary rhabdomyosarcoma in children: diagnostic radiology.
Pediatr Radiol 28:101-108
33. Karmazyn B, Werner EA, Rejaie B, Applegate KE (2005)
Mesenteric lymph nodes in children: what is normal? Pediatr
Radiol 35:774-777
261
34. Watanabe M, Ishii E, Hirowatari Y et al (1997) Evaluation of
abdominal lymphadenopathy in children by ultrasonography.
Pediatr Radiol 27:860-864
35. Vayner N, Coret A, Polliack G et al (2003) Mesenteric lym-
phadenopathy in children examined by US for chronic and/or
recurrent abdominal pain. Pediatr Radiol 33:864-867
36. Simanovsky N, Hiller N (2007) Importance of sonographic de-
tection of enlarged abdominal lymph nodes in children. J Ul-
trasound Med 26:581-584
37. Couture A (2008) Bowel obstruction in neonates and children.
In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal
tract sonography in fetuses and children. 1st edn. Springer-
Verlag, Berlin-Heidelberg, pp 131-251