You are on page 1of 262

Tim Leiner

Mathias Goyen

Martin Rohrer
Stefan Schnberg (Eds.)

Clinical Blood Pool MR Imaging

Tim Leiner
Mathias Goyen
Martin Rohrer
Stefan Schnberg (Eds.)

Clinical Blood
Pool MR Imaging
With 331 Figures and 26 Tables

Tim Leiner Martin Rohrer
Department of Radiology Bayer Schering Pharma AG
Maastricht University Medical Center European Business Unit Diagnostic Imaging
Peter Debijelaan 25 Muellerstrasse 178
6229 HX Maastricht 13353 Berlin
The Netherlands Germany

Mathias Goyen Stefan O. Schoenberg

University Medical Center Department of Clinical Radiology and Nuclear Medicine
Hamburg-Eppendorf University Hospital Mannheim
Martinistrasse 52 Medical Faculty Mannheim University of Heidelberg
20251 Hamburg Theodor-Kutzer-Ufer 1-3
Germany 68167 Mannheim

Clinical Blood Pool MR Imaging

The Vasovist Product Monograph

Marketing Approval Number EU1.DI.06.2008.0017

ISBN 978-3-540-77860-8 Springer Medizin Verlag Heidelberg

Bibliografische Information der Deutschen Bibliothek

The Deutsche Bibliothek lists this publication in Deutsche Nationalbibliographie;
detailed bibliographic data is available in the internet at

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, reuse of illustrations, recitation, broadcasting, reproduction on
microfilms or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted
only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission
for use must always be obtained from Springer-Verlag. Violations are liable to prosecution under the German
Copyright Law.

Springer Medizin Verlag
Springer Medizin Verlag Heidelberg 2008

The use of general descriptive names, registered names, trademarks, etc. in this publications 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 publishers 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

Design: deblik Berlin

Typesetting: TypoStudio Tobias Schaedla, Heidelberg
Cover illustrations: courtesy of Dr. Alexander Huppertz, Imaging Science Institute, Charit, Berlin, Germany,
Dr. Henrik Michaely, Medical Faculty Mannheim, University of Heidelberg, Germany and Dr. Regina Beets-Tan,
Maastricht University Medical Center, The Netherlands
Printer: Grosch! Druckzentrum GmbH, Heidelberg

SPIN: 12215607

18/5135 5 4 3 2 1 0


Magnetic resonance angiography has made great strides, with continuing improvements in
hardware, pulse sequencing, and know-how allowing ever-increasing speed, resolution, and
suppression of artifacts. However, an inherent physical barrier has always been limited SNR.
Gadolinium contrast agents help to increase SNR by facilitating T1 relaxation, but they can
be injected only at a finite rate and at a limited molar dose, and there is a rapid drop in con-
centration following the brief arterial phase due to redistribution into the extracellular fluid
compartment. With its sixfold increase in T1 relaxivity, blood pool distribution, and longer
serum half-life, Vasovist represents a new breakthrough which promises to revolutionize
MRA image quality once again.
This excellent treatise on Vasovist, created by a team of exceptional faculty who are
pioneers in MR angiography, covers the basic techniques, safety, efficacy, image processing,
and pharmacoeconomic details, to successfully implement a new level of MRA image quality
with this new contrast agent. In addition to improving all the usual arterial phase MRA ap-
plications, the blood pool distribution opens up new possibilities, including detecting internal
bleeding and imaging stent graft endoleaks, which are reviewed in detail. In the complex,
competitive field of cardiovascular imaging, this book articulates the cutting edge in imaging
vascular disease.

Martin R. Prince, MD, PhD

Professor of Radiology
Weill Medical College of Cornell University
Columbia College of Physicians and Surgeons
New York, NY, USA


Almost two decades ago, Martin Prince was able to demonstrate that the limitations of non-
contrast MRA techniques could be overcome by injection of contrast agent. Subsequently,
contrast-enhanced-MRA established itself in clinical practice as the standard for non-invasive
depiction of almost all blood vessels. MR manufacturers have addressed the demands for
faster acquisition speed to allow higher resolution imaging during the finite and relatively
short imaging window of first pass MRA by a combination of faster gradients, parallel imaging
techniques, and novel K-space sampling strategies. However, a perceived limit for improve-
ment in spatial resolution, coupled with the negative impact of faster acquisition on contrast-
to-noise ratios, has led to the development of Vasovist, the first contrast agent tailored
to the vascular tree.
With its high relaxivity and unique pharmacokinetics, Vasovist opens up new horizons in
vascular diagnostics with a prolonged imaging window, enhanced topographic information,
and unrivaled new visualization options. The editors and authors have made groundbreaking
contributions towards establishing MR angiography in various investigative settings, render-
ing it more precise and applying it for diverse indications. The work presented here is founded
upon the extensive experience of the editors, and it includes a broad range of experience from
other scientific working groups.
This book presents the applications of Vasovist-enhanced angiography; its potential
advantages, such as the change in signal-to-noise ratio and intravascular distribution, are
discussed systematically, thus giving a comprehensive overview of the basic principles and
imaging techniques. Presentation of the various clinical fields is well-structured and is il-
lustrated with excellent image material that addresses the essential questions concerning
vascular diagnostics. This includes imaging of the intracranial and supra-aortic vessels and
visualization of the coronary arteries, as well as of the renal and visceral vessels. Key chapters
cover MR angiography of the aortoiliac and peripheral vessels. Whole-body MR angiography
represents a special challenge for angiography. The new options offered by Vasovist-enhanced
MR angiography are also discussed. All in all, this monograph presents the ideal opportunity
to gain relevant information, read either as a review or as a detailed account of the increasing
scientific potential offered by this method of vascular MR diagnosis.
What can we predict for the future? Two decades following Dr. Princes (then) heretical
thesis that contrast-agent injection was required for dramatically improved MRA, we are now
equipped with a tailored vascular contrast agent. This development parallels improvements in
scanner performance, satisfies a demand for higher spatial resolution, and opens up a whole
new perspective on the benefit of additional information available from the steady state images
as a routine part of the study.

Prof. Dr.Dr.h.c. Maximilian F. Reiser

Department of Clinical Radiology
Director of the Department
Ludwig-Maximilians-University of Munich
Marchioninistr. 15
81377 Munich


The successful introduction of extracellular gadolinium-based contrast agents for contrast-

enhanced MR angiography, and their wide acceptance today, raise the question of what part an
intravascular contrast agent might play in diagnostic imaging. The answer lies in the capacity
of an intravascular agent to give us high-level diagnostic information from first pass arterial
imaging and, at the same time, to yield additional diagnostic value by allowing delayed imag-
ing from the same contrast injection.
Clinical experience gathered since the introduction of Vasovist (Gadofosveset) appears to
provide the answer yes: not only is Vasovist useful for first pass arterial imaging, but it also
provides high intravascular enhancement that lasts much longer and is significantly greater
than that afforded by conventional extracellular agents. Taking advantage of this effect, one
can now acquire additional high-resolution images in the steady state which lead to much
better delineation of vessel pathology. Steady state imaging offers the possibility of depicting
the entire vascular system without relevant extravasation of the contrast medium from the
intravascular space.
The extended diagnostic window of Vasovist makes the examination more convenient
because it is less dependent upon bolus dynamics. Imaging of a Gadofosveset bolus missed
in the first pass examination does not require an additional injection of contrast agent. For
these reasons, Vasovist may enable physicians to detect systemic vascular disease earlier and
to optimize the evaluation of therapeutic options, including percutaneous intervention and
vascular surgery. In addition, imaging of the vascular system and surrounding tissues in the
delayed phase appears to promise new contrast mechanisms that may improve the detection
of inflammatory or malignant changes.
In summary, Vasovist has the potential to open new horizons in diagnostic MR angiogra-
phy by increasing the spatial resolution and the robustness of MRA examinations and facilitat-
ing the examination of multiple vascular beds. Vasovist was first approved in 2005, and we
are now looking at an expanded spectrum of clinical applications that has rapidly evolved and
addresses the majority of clinical questions in vascular medicine and related fields. Therefore,
this monograph is subdivided into chapters on technology, followed by a detailed review of
the clinical fields for MR angiography with Vasovist. With this steady increase of applications
and clinical experience it is necessary to review not only the technical feasibility and reliability
of the method, but also the potential additional benefit for the patient. Therefore, aspects of
patient management are also analyzed, with the aim of deriving more effective and compre-
hensive imaging standards.
We would like to thank all of the authors for their valuable contributions and dedicated
collaboration, which made this current compilation of essential aspects of Vasovist-enhanced
MR imaging possible. Also, we gratefully acknowledge the contributions of our publisher,
Springer, and Mr. Eric Henquinet for his constructive, friendly and patient collaboration.

Tim Leiner (Maastricht University Hospital, The Netherlands),

Mathias Goyen (University Medical Center Hamburg-Eppendorf, Germany),
Martin Rohrer (Bayer Schering Pharma AG, Berlin, Germany),
Stefan Schnberg (University Hospital Mannheim, Germany)

Table of Contents
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX

Part I Contrast Agent Properties and Technical Aspects

1 MRI Contrast Media Introduction and Basic Properties of the Blood

Pool Agent Vasovist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Martin Rohrer

2 Contrast Enhanced MRA First Pass and Steady State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Harald H. Quick

Part II Safety and Efficacy

3 Risks and Safety Issues Related to Radiological Imaging, in Particular MRI. . . . . . . . . . 35

Gunnar Brix

4 Summary of Safety of Vasovist at 0.03 mmol/kg Body Weight Dose

Clinical Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Matthias Voth and Andrea Lwe

5 Efficacy of Vasovist: Overview of the Clinical Development Program . . . . . . . . . . . . . . 51

Mathias Goyen

Part III Blood Pool Agents in MRA: Indications, Clinical

Applications and Benefits

6 Head and Neck MRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Marco Essig

7 Pulmonary MRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Christian Fink, Ulrike Attenberger, and Konstantin Nikolaou

8 Abdominal MRA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Henrik Michaely

9 MRA of the Peripheral Vasculature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Tim Leiner and Jeffrey H. Maki

10 Magnetic Resonance Venography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

Giles Roditi
XII Table of Contents

11 Whole-body MRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

Harald Kramer, Maximilian F. Reiser, and Konstantin Nikolaou

12 Endoleak Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Sandra A.P. Cornelissen, Mathias Prokop, Hence J.M. Verhagen, and Lambertus W. Bartels

13 Gastrointestinal Bleeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Joachim Lotz

Part IV New Horizons in Vascular Diagnostics

14 Vasovist for Ischemic and Congenital Heart Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

Sebastian Kelle, Gerald Greil, Reza Razavi, and Eike Nagel

15 Vasovist in Brain Tumor Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

Marco Essig and Clemens Fitzek

16 Vasovist in Lymph Node Imaging: Present Status and Future Development . . . . . . 181
Max J. Lahaye, R. Bert Jan de Bondt, Sanne M.E. Engelen, Geerard L. Beets,
and Regina G.H. Beets-Tan

17 Vasovist for Breast Cancer Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Joan C. Vilanova and Klaus Wasser

18 Vasovist for Multiple Sclerosis Recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

Michael Forsting

19 Vasovist in Interventional MR Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

Michael Bock and Frank Wacker

20 Vasovist in Lymphography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

Christoph U. Herborn

Part V Image Processing

21 Image Post-processing of Blood Pool MR Angiograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

Joachim Graessner

Part VI Pharmacoeconomic Impact

22 Health Economic Assessment of Vasovist Technical, Clinical

and Cost Benefits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Carsten Schwenke, Susanne Kienbaum, and Karsten Bergmann

23 Patient Management and Referrals: Impact of High-Resolution Steady State

MRA with Vasovist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Winfried A. Willinek and Dariusch R. Hadizadeh

List of Contributors

Ulrike Attenberger Gunnar Brix Clemens Fitzek

Department of Clinical Radiology Department of Medical Radiation Neuroradiological Centre
Grosshadern Campus Hygiene and Dosimetry ASKLEPIOS Fachklinikum Brandenburg
Ludwig-Maximilians-University of Federal Office for Radiation Protection Anton-Saefkow-Allee 2
Munich Ingolstdter Landstrasse 1 14772 Brandenburg
Marchioninistrasse 15 85764 Oberschleiheim Germany
81377 Munich Germany
Germany Michael Forsting
Sandra A.P. Cornelissen Department of Diagnostic and
Lambertus W. Bartels Department of Radiology/Image Interventional Radiology and
Image Sciences Institute, QS.459 Sciences Institute, E01.132 Neuroradiology
University Medical Center Utrecht University Medical Center Utrecht University Hospital Essen
Heidelberglaan 100 Heidelberglaan 100 Hufelandstrasse 55
3584 CX Utrecht 3584 CX Utrecht 45147 Essen
The Netherlands The Netherlands Germany

Geerard L. Beets R. Bert Jan de Bondt Frederik L. Giesel

Department of Surgery Department of Radiology German Cancer Research Center (dkfz)
University Hospital Maastricht Maastricht University Medical Center Department of Radiology
PO Box 5800 Peter Debijelaan 25 Im Neuenheimer Feld 280
6202 AZ Maastricht 6229 HX Maastricht 69120 Heidelberg
The Netherlands The Netherlands Germany

Regina G.H. Beets-Tan Sanne M.E. Engelen Mathias Goyen

Department of Radiology Department of Radiology University Medical Center Hamburg-
Maastricht University Medical Center Maastricht University Medical Center Eppendorf
Peter Debijelaan 25 Peter Debijelaan 25 Martinistrasse 52
6229 HX Maastricht 6229 HX Maastricht 20251 Hamburg
The Netherlands The Netherlands Germany

Karsten Bergmann Marco Essig Joachim Graessner

Bayer Schering Pharma AG Department of Radiology Siemens AG
Global Medical Affairs Diagnostic German Cancer Research Center (dkfz) Healthcare Sector
Imaging Im Neuenheimer Feld 280 MED ES BMG MR
Mllerstrasse 178 69120 Heidelberg Lindenplatz 2
13353 Berlin Germany 20099 Hamburg
Germany Germany
Christian Fink
Michael Bock Department of Clinical Radiology and Gerald F. Greil
German Cancer Research Center (dkfz) Nuclear Medicine Kings College London
Department of Medical Physics in University Medical Center Mannheim Division of Imaging Sciences
Radiology Medical Faculty Mannheim - University The Rayne Institute,
Im Neuenheimer Feld 280 of Heidelberg 4th Floor, Lambeth Wing, St. Thomas
69120 Heidelberg Theodor-Kutzer-Ufer 1-3 Hospital
Germany 68167 Mannheim London SE1 7EH
Germany United Kingdom
XIV List of Contributors

Dariusch R. Hadizadeh Andrea Lwe Mathias Prokop

Department of Radiology Bayer Schering Pharma AG Department of Radiology, E01.132
University of Bonn Global Medical Affairs Diagnostic University Medical Center Utrecht
Sigmund-Freud-Strasse 25 Imaging Heidelberglaan 100
53105 Bonn Mllerstrasse 178 3584 CX Utrecht
Germany 13353 Berlin The Netherlands
Christoph U. Herborn Harald H. Quick
Medical Prevention Center Hamburg Joachim Lotz Department of Diagnostic and
(MPCH) Magnetic Resonance Imaging Interventional Radiology and
University Medical Center Hamburg- Department of Radiology Neuroradiology
Eppendorf Hannover Medical School University Hospital Essen
Falkenried 88 Carl-Neuberg-Strasse 1 Hufelandstrasse 55
20251 Hamburg 30625 Hannover 45122 Essen
Germany Germany Germany

Sebastian Kelle Jeffrey H. Maki Reza Razavi

German Heart Institute Berlin Department of Radiology Kings College London
Department of Internal Medicine Puget Sound Veterans Division of Imaging Sciences
Division of Cardiology Adminstration Health Care System Floor 5 Thomas Guy House
Augustenburger Platz 1 Seattle, Washington, Guys Hospital
13353 Berlin USA London SE1 9RT
Germany UK
Henrik J. Michaely
Susanne Kienbaum Department of Clinical Radiology and Giles Roditi
Bayer Schering Pharma AG Nuclear Medicine Department of Radiology
Global Health Economics, Outcomes & University Medical Center Mannheim Glasgow Royal Infirmary
Reimbursement Medical Faculty Mannheim - University 16 Alexandra Parade
Diagnostic Imaging of Heidelberg Glasgow G31 2ER
Muellerstrasse 178 Theodor-Kutzer-Ufer 1-3 UK
13353 Berlin 68167 Mannheim
Germany Germany Martin Rohrer
Bayer Schering Pharma AG
Harald Kramer Eike Nagel European Business Unit Diagnostic
Institute for Clinical Radiology Kings College London Imaging
University Hospitals Munich Division of Imaging Sciences Muellerstrasse 178
Grosshadern The Rayne Institute, 4th Floor Lambeth 13353 Berlin
Marchioninistr. 15 Wing Germany
81377 Munich St. Thomas Hospital
Germany London SE1 7EH Carsten Schwenke
Max J. Lahaye Zeltinger Strasse 58g
Department of Radiology Konstantin Nikolaou 13465 Berlin
Maastricht University Medical Center Department of Clinical Radiology Germany
Peter Debijelaan 25 Grosshadern Campus
6229 HX Maastricht Ludwig-Maximilians-University of Stefan O. Schoenberg
The Netherlands Munich Department of Clinical Radiology and
Marchioninistrasse 15 Nuclear Medicine
Tim Leiner 81377 Munich University Hospital Mannheim
Department of Radiology Germany Medical Faculty Mannheim - University
Maastricht University Medical Center of Heidelberg
Peter Debijelaan 25 Theodor-Kutzer-Ufer 1-3
6229HX Maastricht 68167 Mannheim
The Netherlands Germany
List of Contributors

Hence J.M. Verhagen

Department of Vascular Surgery
Suite H-993
Erasmus University Medical Center
PO Box 2040
3000 CA Rotterdam
The Netherlands

Joan C. Vilanova
Department of Magnetic Resonance
& CT
Clnica Girona
Lorenzana, 36
17002 Girona

Matthias Voth
Bayer Schering Pharma AG
Global Medical Affairs Diagnostic
Muellerstrasse 178
13353 Berlin

Frank Wacker
Charit, Campus Benjamin Franklin
Radiology and Nuclear Medicine
Hindenburgdamm 30
12200 Berlin

Klaus Wasser
Department of Clinical Radiology and
Nuclear Medicine
University Hospital Mannheim
Medical Faculty Mannheim - University
of Heidelberg
Theodor-Kutzer-Ufer 1-3
68167 Mannheim

Winfried A. Willinek
Department of Radiology
University of Bonn
Sigmund-Freud-Strasse 25
53105 Bonn

List of Abbreviations

2D Two-dimensional ECG Electrocardiogram

3D Three-dimensional EMEA European Medicines Agency
3D FFT 3D Fast Fourier Transform EMF Electromagnetic field
SSFP Steady state free precession ECS Extracellular space
ADC Apparent diffusion coefficient EUS Endoluminal ultrasonography
ALARA As low as reasonably achievable EVAR Endovascular aortic aneurysm repair
AngioSURF Angiographic System for Unlimited F/P First pass
Rolling Field-of-views FDA US Food and Drug Administration
APAOD Atherosclerotic peripheral arterial FDG 18Fluorodeoxyglucose

occlusive disease FFT Fast Fourier-transformation

ASSET Array Spatial Sensitivity Encoding FLAIR Fluid attentuated inversion recovery
Technique FLASH Fast low-angle shot
AVF Arteriovenous fistulae FMD Fibromuscular Dysplasia
BBB Blood-brain barrier FNAC Fine-needle aspiration cytology
BOLD Blood oxygenation level-dependent FOV Fields-of-view
BPCAs Blood-pool contrast agents GBCA Gd-based contrast agent
CA Contrast agents GCP Good clinical practice
CAD Coronary artery disease Gd Gadolinium
CE Contrast-enhanced GI Gastrointestinal
CE-MRA Contrast-enhanced magnetic resonance GRAPPA Generalized Autocalibrating Partially Parallel
angiography Acquisitions
CENTRA Contrast-enhanced timing robust GRE Gradient recalled echo
angiography H&E Histological examination
CFA Common femoral artery HIFU High-intensity focused ultrasound
CIS Clinically isolated syndrome HNSCC Head and neck squamous cell carcinoma
CKD Chronic kidney disease HSA Human serum albumin
CLI Critical limb ischemia HTA Health technology assessment
CM Contrast medium IA-DSA Intra-arterial X-ray-based digital subtraction
CMR Cardiovascular MR angiography
CNR Contrast-to-noise ratio IC Intermittent claudication
CNS Central nervous system ICH-GCP International Conference on Harmonisation
CSF Cerebrospinal fluid on Good-Clinical-Practice
CT Computed tomography ICNRIP International Commission on Non-ionizing
CTA Computed tomography angiography Radiation Protection
CTEPH Chronic thromboembolic pulmonary IEC International Electrotechnical Commission
hypertension iPAT Integrated Parallel Acquisition Techniques
CVC Central venous catheters IVC Inferior vena cava
cVR Color volume rendering IVUS Intravascular ultrasound
d Diameter KTWS Klippel-Trenaunay-Weber syndrome
Da Daltons LAVA Liver acquisition with volume acquisition
DCE-MRI Dynamic contrast-enhanced MRI LGIB Lower GI bleeding
DEALE Declining Exponential Approximation LITT Laser-induced thermal therapy
of Life Expectancy LNT Linear non-threshold
DKG-NT Deutsche Krankenhausgesellschaft MAPCAs Major aorto-pulmonary collateral arteries
Nebenkostentarif MBF Myocardial blood flow
DOR Diagnostic odds ratio MDCT Multidetector computed tomography
DSA Digital subtraction angiography MIP Maximum intensity projection
DVT Deep venous thrombosis MPR Multiplanar reconstructions
E/P Equilibrium phase MR Magnetic resonance
ECCM Extracellular contrast media MRA Magnetic resonance angiography
XVIII List of Abbreviations

MRCA Magnetic resonance coronary TIPS Transjugular intrahepatic portosystemic

angiography shunts
MRI Magnetic resonance imaging TOF Time-of-flight
MRV Magnetic resonance venography TR Repetition Time
MS Multi-slice TREAT Time-resolved echoshared angiography
mSENSE Modified SENSE technique
MSI Maximal-signal-intensity TTP Time-to-peak
MTT Mean-transit-time UGIB Upper GI bleeding
NSF Nephrogenic systemic fibrosis USg-FNAC Ultrasound-guided fine-needle aspiration
PAD Peripheral artery disease cytology
PAH Pulmonary arterial hypertension USPIO Ultrasmall super paramagnetic iron oxide
PAOD Peripheral arterial obstructive disease VESPA Venous-emhanced substracteed peak arterial
PAT-factor Parallel acquisition technique factor VIBE Volumetric interpolated breath-hold
PC Phase-contrast examination
PE Pulmonary embolism VQ scan Ventilation-perfusion scintigraphy
PET Positron emission tomography VRT Volume rendering technique
PR Perfusion reserve VSOP Very small superparamagnetic iron oxide
PTA Percutaneous transluminal angioplasty XRA X-ray angiography
QALY Quality-adjusted-life-year m Average time
RARE Rapid acquisition with relaxation
RAS Renal artery stenosis
RES Reticuloendothelial system
RF Radiofrequency
R-factor Acceleration factor
RIME Receptor-induced magnetization
RVT Renal vein thrombosis
SAE Serious adverse events
SAR Severe adverse reactions
SENSE Sensitivity encoding
SI Signal intensity
SLE Systemic lupus erythematosus
SLN Sentinel lymph node
SMA Superior mesenteric artery
SMASH Simultaneous acquisition of spatial
SNR Signal-to-noise ratio
SPECT Single photon emission computed
SPGR Spoiled gradient recalled echo
SPIO Superparamagnetic iron oxide particles
SR Surface rendering
SSD Surface-shaded display
SSFP Steady state free precession
STD Standard deviation
STIR Short tau inversion recovery
SWI Susceptibility-weighted imaging
T1-SE T1-spin echo
T2-FSE T2-fast spin echo
TAO Thromboangiitis obliterans
TE Echo time
THRIVE T1-weighted high-resolution isotropic volume

Part I Contrast Agent Properties

and Technical Aspects

Chapter 1 MRI Contrast Media Introduction and Basic Properties

of the Blood Pool Agent Gadofosveset (Vasovist) 3
Martin Rohrer

Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass

and Steady State 17
Harald H. Quick

MRI Contrast Media Introduction

and Basic Properties of the Blood Pool Agent
Gadofosveset (Vasovist)
Martin Rohrer

1.1 Introduction 4
1.1.1 Basic Mode of Action of MRI Contrast Media 4
1.1.2 Metal Complexes (Gd-Chelates) 4
1.1.3 Iron Oxide Nanoparticles 5
1.1.4 Overview of Currently Marketed MRI Contrast Media 5

1.2 Blood Pool Contrast Media for MRI 5

1.2.1 Cellular Target Component: Erythrocyte-bound or Liposomal Systems 6
1.2.2 Aqueous Target Component: Macromolecules Polymers, Iron-oxide Particles
and Covalently Bound Metal Complexes 6
1.2.3 Proteins as Target Component: Non-covalently Bound Metal Complexes 6

1.3 Molecular Structure and Physicochemical Properties 6

1.4 Biophysical Properties Reversible Protein Binding 8

1.4.1 Pharmacokinetics Prolonged Retention Time 8
1.4.2 Pharmacodynamics Relaxivity 10
1.4.3 Causes for Increased Relaxivity Due to Protein Binding 11

1.5 Implications for Clinical Application 12

1.5.1 Relevant Implications for Application in the First Pass 12
1.5.2 Relevant Implications for Application in the Steady State 13

References 14
4 Chapter 1 MRI Contrast Media Introduction and Basic Properties of the Blood Pool Agent Gadofosveset (Vasovist)

1.1 Introduction local proton density the most important physical pa-
1 rameters in MRI to obtain spatially resolved differences
Twenty years ago, the first MRI contrast medium (MRI- in signal intensity (SI) and hence the prerequisites for any
CM) was introduced to the market: in 1988, Gd-DTPA soft-tissue and other contrasts in MRI. Taking effect at
(Magnevist, Bayer Schering Pharma AG, Berlin, Ger- this most basic level, the option to additionally enhance
many) received market approval for clinical use in the contrast in MRI by introducing T1- and T2(*)-shortening
United States, Europe and Japan. In the years to follow, ap- contrast media has made contrast-enhanced MRI such an
plication of MRI-CM became a widely established, pow- important and often indispensable diagnostic tool.
erful tool in MRI for improved diagnosis in approximately Furthermore, not only does the use of MRI-CM pro-
30% of all MRI examinations worldwide [1,2,3,4,5]. vide an exclusive opportunity to directly modulate con-
During these two decades, only a few novel contrast trasts at the most basic biophysical level of MRI; because
media concepts have successfully stepped out of research MRI-CM significantly shortens relaxation times, e.g. in T1-
laboratories to undergo clinical development and eventu- weighted MRA, there is another important and basic ad-
ally to receive market approval. These include: vantage of contrast media use compared with unenhanced
Coated iron oxide nanoparticles, often referred to as MRI procedures: contrast-enhanced MRI allows not only
SPIO (superparamagnetic iron oxide particles) with an for additional contrasting with a minimum of artifacts, but
average core particle size in the nanometer range. The also for substantially accelerated data acquisition due to
SPIO-based MRI-CM Ferumoxide (Feridex, Bayer much shorter repetition times (TR) and ultimately shorter
HealthCare Pharmaceuticals, Wayne, NJ, USA) and scan times at a given spatial resolution. This is particularly
Ferucarbotran Resovist (Ferucarbotran, Bayer Scher- the case for MRA, where the T1 shortening achieved by
ing Pharma AG, Berlin, Germany) are taken up pre- contrast media reduces the native longitudinal relaxation
dominantly by the reticuloendothelial system (RES) in time (T1) of blood from over 1 sec down to the millisecond
the liver. range, as will be discussed below in more detail.
Tissue-specific Gd- or Mn-based approaches. Mar- For all classes of MRI-CM, shortening of proton spin
keted products are the liver-specific MRI-CM Gd- relaxation times is achieved by the paramagnetic proper-
EOB-DTPA (Primovist, Bayer Schering Pharma AG, ties of the contrast media. Paramagnetism is based on un-
Berlin, Germany), Gd-BOPTA (MultiHance, Bracco, paired electrons, resulting in strong and fluctuating local
Milan, Italy) and Mn-DPDP (Teslascan, GE Health- magnetic field distortions in the vicinity of the contrast
care, Chalfont St. Giles, U.K.). medium molecules. These local magnetic field distortions
Highly concentrated contrast media solutions, such as are capable of destroying the much weaker proton spin
the 1.0 molar Gd-concentrated Gadobutrol (Gadovist order in the external, static magnetic field. Consequently,
1.0, Bayer Schering Pharma AG, Berlin, Germany). shortening of spin-lattice (T1) and spin-spin (T2) relax-
ation times is obtained in the tissue or liquid containing
Most recently, another class of MRI contrast agents has a sufficiently high concentration of contrast medium
been introduced to the market, mainly to overcome cur- molecules. A more detailed and quantitative explanation
rent limitations of MR angiography (MRA): highly in- of these effects is provided in Sect. 4.2.
travascular, slow-clearing blood pool contrast agents for
MRI. Vasovist (Gadofosveset, Bayer Schering Pharma
AG, Berlin, Germany) the first intravascular contrast 1.1.2 Metal Complexes (Gd-Chelates)
agent approved for use with MRA in the European Union
is based on non-covalent transient protein binding, As a result of their high number of unpaired electron
which leads to both an extended imaging window and spins, ions of transition metals and of lanthanides have
strongly decreased Gd-dosage requirements. been found to be well-suited as paramagnetic atoms in
MRI-CM. The Gd3+ cation with seven unpaired electron
spins (S = 7/2) and its large magnetic moment became the
1.1.1 Basic Mode of Action of MRI Contrast most important atom for use in MRI-CM. To assure the
Media stability and biodistribution required for safety, Gd3+ ions
are chelated in different chemical complexes, which also
Contrast media for MRI are well-known to strongly in- modulate their magnetic and pharmacokinetic proper-
fluence proton spin relaxation times, represented in vivo ties to some extent. (Please also refer to the chapters in
mainly by the vastly available hydrogen nuclei from water this book related to safety.) The interaction between the
molecules in organic liquids and tissues. unpaired electron spins of the metal ion and the nuclear
The tissue-specific longitudinal (T1) and transverse spins of protons from the surrounding water molecules is
(T2(*)) relaxation times are besides the less important dominated by the short-distance, anisotropic dipol-dipol
1.2 Blood Pool Contrast Media for MRI
5 1

hyperfine interaction, as well as by contributions from 1.2 Blood Pool Contrast Media for MRI
the isotropic Fermi-contact interaction. Longitudinal and
transverse relaxation times are comparably affected, lead- Several classes of MRI contrast media have been briefly
ing to ratios r2/r1 between 1 and 2 [6]. introduced in the previous section. This section takes a
closer look at the characteristic properties of blood pool
contrast media for MRI.
1.1.3 Iron Oxide Nanoparticles Intravascular contrast agents show less extravasation
into the interstitium than conventional extracellular con-
As mentioned above, so-called superparamagnetic par- trast agents. Depending on their mode of action, they also
ticles have also been shown to be suitable as MRI-CM. have a longer retention time in the vascular system due to
They are often referred to as superparamagnetic iron slower blood clearance. Both characteristics can be real-
oxide particles (SPIO) and have been investigated with ized only with special physicochemical and biochemical
different coatings (e.g. carboxydextran or dextran), dif- properties. For this reason, these properties, as well as the
ferent core particle diameters, and hydrodynamic size resulting typical pharmacokinetics, will be examined first
distributions. Their basic magnetic properties are deter- in general, and then specifically for Vasovist.
mined not only by the interactions of nuclear spins with If we take a closer look, additional subgroups within
single paramagnetic ions in a chelate, but also by the bulk the class of intravascular MR contrast agents can be
effects of the much larger spin ensembles of the unpaired distinguished, and it becomes clear that the terms intra-
electrons from the Fe2O3 and Fe3O2 molecules in the iron vascular and blood pool do not have to be synonyms
oxide nanoparticles. For the larger SPIO particles and in the strictly literal sense. According to the dominant
for SPIO clusters in particular, these effects add up to effect reduced extravasation, decelerated resorption, or
almost macromolecular conditions and local ferromag- glomerular excretion and the resulting pharmacokinet-
netic properties (superparamagnetism). Consequently, ics, the terms intravascular and blood pool will be
depending on SPIO core diameters or SPIO clustering used differently.
effects, their influence on proton spin relaxation times ap- According to a proposal by Bogdanov and Weissleder
plies over larger distances and is characterized by strong [7,8], the components circulating in the blood vessel sys-
susceptibility effects and larger-scale magnetic field inho- tem may be divided into the following three groups:
mogeneities. Naturally, the transverse relaxation processes 1. Cellular components (predominantly erythrocytes)
dominate, which makes the SPIOs often better suited for 2. Water
T2 and T2*-weighted MRI sequences. Resulting r2/r1 ratios 3. Proteins (mainly albumin as the most common plasma
can be up to the order of one hundred. protein)

1.1.4 Overview of Currently Marketed MRI Contrast Media

Table 1.1. Contrast media for MRI with market approval in at least either Europe, Japan or the USA

Short name or internal development code Generic name (INN)1) Trade name(s) Company
Gd-DTPA Gadopentetate dimeglumine MAGNEVIST Bayer Schering Pharma AG
Gd-DO3A-butrol Gadobutrol GADOVIST Bayer Schering Pharma AG
MS-325 Gadofosveset trisodium VASOVIST Bayer Schering Pharma AG
Gd-EOB-DTPA Gadoxetic acid, disodium PRIMOVIST Bayer Schering Pharma AG
SH U 555 A Ferucarbotran RESOVIST Bayer Schering Pharma AG
AMI-25 Ferumoxide FERIDEX / ENDOREM Bayer Healthcare / Guerbet
Gd-HP-DO3A Gadoteridol PROHANCE Bracco
Gd-BOPTA Gadobenate dimeglumine MULTIHANCE Bracco
Gd-DTPA-BMA Gadodiamide OMNISCAN General Electric Healthcare
Mn-DPDP Mangafodipir trisodium TESLASCAN General Electric Healthcare
Gd-DOTA Gadoterate meglumine DOTAREM Guerbet
Gd-DTPA-BMEA Gadoversetamide OPTIMARK Tyco Healthcare

) International Nonproprietary Name
6 Chapter 1 MRI Contrast Media Introduction and Basic Properties of the Blood Pool Agent Gadofosveset (Vasovist)

Quite analogously, the optional modes for intravascular In terms of very small superparamagnetic iron oxide
1 contrast agents may be presented according to these tar- particles (ultrasmall superparamagnetic iron oxide (US-
get components. PIO), very small superparamagnetic iron oxide (VSOP)),
reference is made to VSOP-C184 (Ferropharm GmbH)
[16], ferumoxtran-10 (Sinerem, Guerbet, Combidex,
1.2.1 Cellular Target Component: Advanced Magnetics) [17,18], and Ferucarbotran (Su-
Erythrocyte-bound or Liposomal pravist, SH U 555 C, Bayer Schering Pharma AG)
Systems [19,20]. These compounds are partially also being in-
vestigated for other than MRA indications, such as lym-
So far it has not been possible to advance research strate- phography.
gies on intravascular MR contrast agents based on direct
coupling or on cytoplasmatic loading of erythrocytes
through to clinical development, owing to the exces- 1.2.3 Proteins as Target Component:
sive quantities of contrast agent that were required [9]. Non-covalently Bound Metal Complexes
However, a vascular circulation pattern similar to that in
natural cells has been observed in liposomal, cell-like sys- This group of innovative MR contrast agents includes
tems, whereby liposomes could be formed with compart- Gadofosveset (Vasovist) as the only blood pool con-
ments separated by a double lipid layer similar to a cell trast agent currently (January 2008) approved in the
membrane. Iron, manganese, or gadolinium complexes 27 member states of the European Union and several
could be trapped in their aqueous inner space and thus other countries, including Switzerland, Norway, Turkey,
transported into the circulatory system. However, the Canada and Australia. An increase in retention time in
compounds investigated so far were not stable over a long the vascular system is achieved by reversible binding
period and became rapidly metabolized by the reticuloen- of the magnetically active metal complex specifically
dothelial system (RES) [10]. to the most prevalent albumin [human serum albumin
(HSA)] in the plasma. An essential feature is the non-
covalent type of binding between the contrast agent
1.2.2 Aqueous Target Component: molecule and the plasma protein, which distinguishes
Macromolecules Polymers, this type of intravascular contrast agent from those
Iron-oxide Particles and Covalently molecules which bind permanently to large proteins
Bound Metal Complexes or cells. Nevertheless, a high fraction of binding can
be achieved with this concept, and both a significantly
The systems within this subcategory are characterized longer retention time and greatly reduced extravasation
by their inherent or, in the case of covalently bound sys- combined with generally good tolerance are therefore
tems, their effective macromolecular size, which is why ensured [21].
they are subject to extravasation to only a small degree, For the bound fraction in blood, the concentration
if at all. Only large Gd-based systems and coated iron- of the contrast agent molecule is observed to be largely
oxide particles are discussed here, as some of these com- independent of the relative concentrations of proteins in
pounds are currently undergoing preclinical or clinical the dynamic equilibrium, which is quickly attained after
development, respectively. An example of the first group venous injection. This and other characteristics will be
is Gadomer [11], a dendritic macromolecule, in which examined in detail below, based on the example of Gado-
24 Gd3+ ions are bound. Its molecular weight is approxi- fosveset.
mately 17 kDa. This was the origin of the previous name,
Gadomer-17, for this dendrimeric concept developed by
Bayer Schering Pharma AG. The investigational drug 1.3 Molecular Structure and Physicochemical
successfully underwent preclinical evaluation for vari- Properties
ous indications, such as coronary angiography [12] and
lymphography [13]. Another example is gadomeritol The molecular structure of Gadofosveset is shown in
(Vistarem, P792, Guerbet) [14], whose intravascular Fig. 1.1 and in a 3-D model in Fig. 1.2. The molecule
properties are also based on molecular size. Gadomeritol can be divided into two functional entities. The first, MR-
also shows rapid clearance through glomerular filtration active, part is formed by the stable chelate structure of the
[15]. Gadomer and gadomeritol are highly intravascular gadolinium acid, gadolinium diethylenetriaminepentaa-
contrast agents in the strict sense with limited blood cetic acid, Gd-DTPA [22]. The cation Gd3+ bound within
pool characteristics (fast- clearing intravascular contrast this structure and the dipole-dipole interactions of its un-
agents). paired electron spin (S = 7/2) lead to strong decreases in
1.3 Molecular Structure and Physicochemical Properties
7 1

nuclear spin relaxation times, especially of proton spins (I

= ) of the H2O in the immediate molecular vicinity.
Ph The innovative, additional properties of Vasovist,
Ph on the other hand, are achieved by the second functional
O O entity, a diphenylcyclohexyl group covalently bound with
P the Gd complex via a phosphate diester bridge. This
O O-
enhances the hydrophilic character of the molecule and
allows reversible, non-covalent binding of the molecule
N with albumin.
O Gd O O Further properties are provided in Table 1.2 [6, 23].
O O O The active substance consists of the metal complex as
N a3 O O O O described (stable up to 282C), which, as a white powder,
H H has a low solubility for organic solvents but a very high
water solubility.
The administered formula of Vasovist consists of a
250-mM (244 mg/ml) aqueous solution of the active sub-
Fig. 1.1. Molecular structure of Gadofosveset stance with a 0.1% excess (by mass) of the ligand.

Fig. 1.2. Three-dimensional model of the active substance of Gadofosveset. The chelate structure with the Gd3+ atom (gray) may be seen on
the right.

Table 1.2. Physicochemical properties of Vasovist

Viscosity at 20C 3.0 mPas

Density at 25C 1.12 g/ml

Osmolarity at 37C 825 mOsm/kg

pH Ready-to-use solution 6.58.0

Relaxivity r1 in H2O at 37C, 1.5 T, 0.5 mM (5.2 0.3) l mmol-1 s-1

Relaxivity r1 in bovine plasma at 37C, 1.5 T, 0.5 mM (19 ) l mmol-1 s-1

8 Chapter 1 MRI Contrast Media Introduction and Basic Properties of the Blood Pool Agent Gadofosveset (Vasovist)

1.4 Biophysical Properties Reversible under the title Receptor-induced magnetization enhance-
1 Protein Binding ment (RIME) [24].
The special characteristics of the pharmacokinetics
As described in the previous section, the diphenylcyclo- and pharmacodynamics are interdependent owing to their
hexyl group (linked to the Gd complex via a phosphate common origins via protein binding, which is described
diester bridge) leads to the specific, reversible binding of in the following section.
the molecule to albumin [22,23].
A spatial depiction of the molecule reversibly bound to
a protein binding site is shown in Fig. 1.3. It is not a 1.4.1 Pharmacokinetics Prolonged
chemical process in the strict sense; rather the temporary Retention Time
link is based on structural molecular properties and on
the formation of weak hydrogen bridges between the pro- Different retention times are shown schematically in
tein and the diphenylcyclohexyl group from the contrast Fig. 1.4 for both a conventional extracellular and a
agent molecule. slow-clearing intravascular contrast agent with extended
The key effects of these unique binding properties on retention time in the vascular system. The representation
the contrast agent effects are well-known phenomenologi- qualitatively illustrates that an initial bolus phase exists in
cally: they are first to be found in modified pharmacoki- both cases; in the case of an intravascular contrast agent
netics compared with conventional extracellular contrast with extended retention time, this is followed by an equi-
agents, which result in a significantly longer retention librium phase or steady state.
time for the active substance in blood circulation. In the Let us take a closer look at the special pharmacokinet-
bound state, the contrast agent is not subject to glomeru- ics of Gadofosveset: Following injection, there is initially
lar filtration in the kidneys, i.e. only a small part accord- a high local concentration of Gadofosveset in blood. Due
ing to the non-bound fraction at the time is filtered out to its specific binding affinity to albumin, the fraction of
of circulation in the kidneys. Extravasation is also largely the bound Gadofosveset molecules depends both on the
reduced, as this is limited to the non-bound fraction of local albumin concentration and on the local Gadofosve-
the active substance in the blood. set concentration. At very high Gadofosveset concentra-
The second key effect lies in the influence on phar- tions (excess Gadofosveset), the binding fraction is, of
macodynamics and consists essentially of a considerably course, lower than at low Gadofosveset concentrations
increased relaxivity compared with conventional extracel- (excess albumin). This is shown in Fig. 1.5 for plasma
lular contrast agents. This principle of increasing relaxi- solutions under different example conditions [25]: Ac-
vity through reversible protein binding was first described cordingly, below the respective minimum concentrations
there is a plateau range with an almost constant fraction
of bound molecules.
For the pharmacokinetics in the initial phase after
injection this means that the relative binding share has to
rise from a small initial value (with very high local Ga-
enhancement (a.u.)

(slow clearing)


first pass steady state time

Fig. 1.3. A binding site of a protein structure (extract from the much
larger total protein) with the non-covalently bound contrast agent mo-
lecule Gadofosveset. The phosphate groups are located in the binding Fig. 1.4. Schematic representation of the available signal enhance-
pocket of the protein (center of the image). The chelate structure contai- ments for conventional extracellular (solid line) and slow-clearing intra-
ning the Gd ion (gray) may be seen on the right. vascular contrast agents (dashed line).
1.4 Biophysical Properties Reversible Protein Binding
9 1

dofosveset concentration) to an equilibrium state. This is bound fraction is subject to extravasation, into the inter-
shown in Fig. 1.6 with measurements from a preclinical stitial space or by passage of a defect blood-brain barrier.
investigation in primates [26], especially well-suited for The two half-lives t and t must be taken into con-
illustration purposes due to the relatively high dose used sideration for an accurate, quantitative description of the
of 0.1 mmol/kg. The initial rise in the albumin-bound pharmacokinetics of Gadofosveset; these characterize the
percentage fraction and the arrival at the plateau at ap- typical initial (t = 0.48 0.11 h) and the subsequent time
proximately 8085% for this species can be clearly seen. course (t = 16.3 2.6 h) of the Gadofosveset concentra-
Obviously, the time course of the protein-bound frac- tion in blood. In Fig. 1.7 this is depicted on the basis of
tion has a direct influence on the pharmacokinetics: the results of a clinical licensing study over a period of
Because only the unbound fraction is subject to renal 48 h after injection [27]. From the logarithmic scale of
filtration, elimination from the vascular system is faster at the parameter axes (plasma concentration in mmol/l) the
first until an approximately homogeneous distribution is typical, averaged half-lives t and t may be recognized
attained in the total blood volume. Furthermore, the un- in the early and the later time points, respectively.

Fig. 1.5. Fraction of protein-bound Gadofos-

veset as a function of concentration in serum
albumin from different species: Human,
Rabbit, Pig, Rat. Derived from [25]

protein-bound / non protein-bound

fraction of G adofosveset




0 Fig. 1.6. Time course of the protein-bound

0 20 40 60 80 100 120 fraction following injection. Example from a
time [min] preclinical study following high-dose injection
of 0.1 mmol/kg in primates [26]
10 Chapter 1 MRI Contrast Media Introduction and Basic Properties of the Blood Pool Agent Gadofosveset (Vasovist)

1.00 agents, which have little or no interaction with plasma

1 proteins. The difference in the relaxation rates (1/Ti 1/
plasma Gadofosveset concentration [mmol/L]

Ti(0)) is often abbreviated as 1/Ti. 1/Ti represents the

actual observed reduction in relaxation times under
the given conditions quantitatively and is therefore one
0.10 of the most important parameters to describe contrast
agent effects in MRI. The 1/Ti measurement values are
often entered in a graph as a function of concentration
[CM]. For non-protein-binding contrast agents a good
approximation to a linear fit is obtained, whose constant
0.01 slope (1/Ti)/[CM] corresponds to the respective re-
laxivity ri.
However, for MR contrast media which interact ap-
preciably with proteins and whose relaxivity is influenced
by protein binding which is the case for Vasovist and
0 12 24 36 48 some other MR contrast agents there is also a clear
time [h] dependence of the reduction in relaxation time on the
relative concentration of the respective plasma proteins as
Fig. 1.7. Logarithmic presentation of the bi-exponential pharma- well as on the concentration of the contrast agent. The re-
cokinetics of Gadofosveset. Early time points can be assigned to the laxivity can therefore no longer be described by a straight
initial distribution phase (t), whereas the much longer distribution line over a large concentration range in this case [29,30].
phase (t) is reflected by the straight section of the graph
This is shown in Fig. 1.8 for Vasovist on the basis of
results from a clinical study [27].
As is apparent in Fig. 1.8, the slope (1/Ti)/[CM]
The half-life labeled here with disposition phase and of the curve falls with higher concentrations as antici-
characterized by (t) is also known as the (terminal) pated, because the relative fraction of bound molecules
plasma elimination time. It has to be distinguished from as described above is lower at higher Gadofosve-
the terminal elimination time tterm, with which the active set concentrations. At the same time, Fig. 1.8 also
substance is eliminated not only from the vascular system, shows that the deviation from linearity in the concentra-
but also from the organism. For Vasovist this is tterm = tion range shown is moderate overall and especially at
18.5 3 h [27]. low concentrations below 0.1 mmol/l is negligible. This
means that relaxation values measured at concentrations
up to 0.1 mmol/l cannot necessarily be extrapolated to
1.4.2 Pharmacodynamics Relaxivity very high concentrations. According to international
recommendations for the nomenclature of MR contrast
The reduction in the longitudinal (T1) and transversal agent parameters [28], the concentration should also be
(T2) relaxation times of the proton spin detected by mag- specified when specifying the relaxivity in such cases.
netic resonance imaging with an MR contrast agent is The good agreement with a linear relationship of 1/Ti
quantitatively described by relaxivity values (r1 and r2). versus concentration [CM] in the concentration range
The definition of ri (with i = 1, 2) is given by the differ- up to 0.5 mmol/l within the statistical measurement ac-
ence in the relaxation rates (inverse relaxation times 1/Ti) curacy allows good comparability in this range, also with
of a solution with and without contrast agent (1/Ti(0)). In relaxivity values of non-protein-binding contrast agents.
the definition, there is also division by the contrast agent It is important in comparative analyses that the measured
concentration, which yields independence of the relaxiv- concentration ranges, as well as other measurement con-
ity values from the concentration in the first approxima- ditions, are always identical.
tion [28]: Moreover, relaxation values are generally dependent
on additional conditions: Both the magnetic field strength
1 1 1
ri = i = 1, 2 Eq. (1) and the temperature have an influence on the measure-
T T [CM ] ment values. Hence, these conditions should also be
i i(0)

specified for the respective solution environment of the

The contrast medium concentration (in mmol/l) is termed contrast agent (e.g. blood, plasma, albumin concentra-
[CM]. tion, water). Table 1.3 compares several values for Ga-
Equation (1) is obviously a very expedient definition dofosveset and a weak protein-binding contrast agent
for the relaxivities, especially for conventional contrast (Gd-BOPTA, MultiHance, Bracco, Milan, Italy).
1.4 Biophysical Properties Reversible Protein Binding
11 1

(1/T1)observed (s-1)

(1/T1)observed (s-1)


0.00 0.02 0.04 0.06 0.08
30 [Gadofosveset], (mmol/L)


Fig. 1.8. Dependence of observed relaxation

rates (1/T1)observed on the Gadofosveset concen-
0.0 0.5 1.0 1.5 2.0
tration. Note the high linearity at low concen-
[Gadofosveset], (mmol/L)
tration as depicted separately

Table 1.3. Relaxivity values1 for Gadofosveset and a weak protein-binding MR contrast agent measured in bovine plasma at concentra-
tions up to 0.5 mmol/l. Temperature at 1.54.7 T: 37C [6]

0.47 T 2) 1.5 T 3T 4.7 T

r1 r2 r1 r2 r1 r2 r1 r2

Vasovist 28 40 19 34 9.9 60 6.9 60

(27-29) (38-42) (18-20) (32-36) (9.4-10.4) (56-64) (6.6-7.2) (57-63)

MultiHance 9.2 12.9 6.3 8.7 5.5 11.0 5.2 10.8

(8.7-9.7) (12.2-13.6) (6.0-6.6) (7.8-9.6) (5.2-5.8) (10.0-12.0) (4.9-5.5) (10.1-11.5)
1values in l mmol-1 s-1; 2measured at 40 C

1.4.3 Causes for Increased Relaxivity Due a local disturbance or an additional, fluctuating, and
to Protein Binding highly inhomogeneous magnetic field for the nuclear
Within the theoretical framework given here, a simplified, The precise mechanisms on a molecular level are
illustrative description of the biophysical relationships is divided into contributions based on interactions with
provided to make the increased relaxivity due to protein protons in direct coordination with the metal ion of the
binding comprehensible. Mathematical presentations will complex inner-sphere effects and those based on
not be offered; instead reference is made to some reviews more distant protons above the second coordination level
on relaxation theory [31,32,33,34]. and also diffusing protons outer-sphere effects (see
As with other gadolinium-based MR contrast agents, Fig. 1.9) [31,32].
magnetic interactions take place between unpaired elec- For paramagnetic metal complexes, the contributions
tron spins (= paramagnetic center) of the Gd3+ cation can largely be described mathematically with the Solo-
bound to the complex (S = 7/2) and unpaired nuclear mon-Bloembergen-Morgan equation [35]. The magnetic
spins from the solution environment of the molecule. interactions between electron and nuclear spins are de-
The nuclear spins are essentially attributable to protons scribed here with anisotropic dipole-dipole interactions
(hydrogen nuclei) from neighboring H2O molecules. and scalar Fermi-contact interactions. The contributions
The known reduction in nuclear spin relaxation times to longitudinal relaxation of the first coordination level
arises from the much larger magnetic moment of elec- (inner sphere), which usually dominate, depend on the
tron spins compared with proton spins. The magnetic number of transiently bound water molecules (hydration
moment of the unpaired electrons can be viewed as number) and their relaxation time (T1m), as well as on
12 Chapter 1 MRI Contrast Media Introduction and Basic Properties of the Blood Pool Agent Gadofosveset (Vasovist)

1.5 Implications for Clinical Application

second coordination sphere
1 water
H The product properties of Vasovist as described above
reveal important aspects about how the product should
be used to optimally exploit its extraordinary diagnostic
free water molecule
potential in clinical applications. It must be stressed,
however, that although the basic properties certainly lead
to essential, trend-setting information for optimizing se-
inner coordination sphere quence parameters and injection protocols, successive
identification and exploitation of the new imaging meth-
ods with Gadofosveset finally depend on radiological
experience and expertise.
Fig. 1.9. Representation of three classes of water molecules with Some summaries of initial application recommenda-
regard to the Gd complex. The molecules from the first coordination tions were published early [36,37,38] and are based both
level (inner sphere) are directly coordinated with Gd(III); in the second on the experiences of early-stage clinical studies and on
coordination level (2nd sphere) there are hydrogen bridges to the theoretical considerations. Special attention should be
drawn to the chapters of this book describing clinical
applications as well as to the previous publications on
selected studies [39,40,41], which also examine technical
application aspects. However, some fundamental aspects
the average time (m) for the coordination of these water of first pass and steady state applications are described in
molecules with the metal ion. the following sections.
The latter factor, the average time (m) for the co-
ordination of individual water molecules on the metal
complex (which corresponds to the inverse exchange rate 1.5.1 Relevant Implications for Application
kex) is worthy of closer examination. Intuitively, a longer in the First Pass
average period of presence of the directly influenced wa-
ter molecule in the immediate vicinity of the Gd complex As a result of its considerably higher relaxivity, a sig-
can also contribute to a more pronounced reduction in nificantly lower gadolinium dosage of 0.03 mmol/kg body
overall relaxation times. It is similarly intuitive that m weight compared with conventional contrast agents could
is influenced by the respective characteristic times of be found for MRA with Gadofosveset. This results in
molecular movements, such as rotation, vibration and around one third of the standard dose of conventional con-
diffusion: If there are slower molecular movements (e.g., trast agents of 0.1 mmol/kg body weight. Bearing in mind
caused by Brownian motion due to lower temperature, the concentration of 0.25 mol/l of the ready-to-use solution
by higher viscosity of the solution environment, and in (compared with 0.5 mol/l for most other contrast agents),
larger molecules) the respective values for m are cor- the volume-related dose is 0.12 ml/kg body weight.
respondingly larger and the relaxation time reduction As with any other conventional contrast agent, Vaso-
achieved increases. vist can be intravenously administered manually or with
The molecular movements are characterized by cor- an automatic injection system for first pass MRA. All the
relation times; the rotation correlation time is abbrevi- well-established techniques regarding bolus dynamics and
ated by r for example. The molecular dimensions of correct timing are the same as with first pass MRA using
proteins, such as serum albumin, are several orders of conventional contrast agents. The situation in a patient to
magnitude above those of low-molecular-weight Gd be examined can be determined with a test bolus of 1 ml
complexes. Hence, the contrast agent molecule experi- volume prior to the measurement itself. The parameters
ences a strong increase in its effective rotational cor- covered here are: the individual synchronization of bolus
relation time of around two orders of magnitude (factor injection, arrival time of the bolus in the target region,
100) as a result of protein binding. Correspondingly, and the start of measurement. Alternatively, the respective
the much slower motions in the protein-bound state are manufacturers fluoroscopic methods of bolus detection
expressed by a decreased rotation rate [31]. The aver- with fast 2D imaging of suitable target regions can be car-
age period of presence m of the water molecules co- ried out in real-time (for example BolusTrack, CareBolus
ordinated with the Gd complex is thereby significantly or SMARTprep). A saline flush following the injection is
increased as a consequence, accompanied by the strong obligatory.
increase in the relaxivity of Gadofosveset through pro- To obtain a comparable length of the contrast agent
tein binding. bolus and, at the same time, to use the tested relation-
1.5 Implications for Clinical Application
13 1

ship between the period of injection, scan delay, and data approximate temporal equilibrium. This applies particu-
acquisition as established with 0.5-molar contrast agents, larly to the period of time following the actual distri-
the Vasovist injection rate can be adapted as follows: bution phase, whose start is very much dependent on
Example case: A protocol using the standard dose of the individual circulatory parameters of the respective
0.1 mmol/kg of a 0.5-molar contrast agent has become es- patient. For patients with normal cardiovascular func-
tablished for a specific MRA application. For a patient with tion, an approximately homogeneous distribution may be
75 kg body weight, a typical volume of 15 ml was injected at assumed after just a few circulation cycles within the first
a rate of 3 ml/s over a period of 5 s. According to the stan- 35 min after injection, whereas in patients with cardiac
dard dosage for a 75-kg patient, the Vasovist contrast agent insufficiency or other relevant diseases of the circulatory
volume is 9 ml. To reproduce the same injection period with system, complete distribution can take up to 10 min,
the lower volume, Vasovist requires a roughly 1/3 reduced according to current knowledge. These precursory com-
injection rate in this case. Consequently, the volume of 9 ml ments are in no way intended to imply that acquisition
should be applied at 2 ml/s in 4.5 s in this case. prior to the attainment of a homogeneous distribution
Comparison with the common application of a double should be generally avoided. Rather, it should be pointed
or triple dose of conventional MR contrast agent yields out that, dependent on the individual circulatory param-
even more pronounced variations. For instance, compar- eters, relative fluctuation in arterial and venous contrast
ing the double dose of 0.2 mmol/kg body weight of a enhancement is still to be expected during the later dis-
conventional MR contrast agent with Vasovist, the Gd tribution phase.
standard dose is reduced by 85% and the injection volume Assuming a homogeneous distribution of the protein-
by 70% (9 ml instead of 30 ml). binding contrast agent in the entire vascular system in
In principle, the specific pharmacodynamics caused the steady state, the blood pool agent is generally further
by protein binding (including higher relaxivity and con- diluted compared with the initially higher concentration
centration dependence) described in the last section could of a bolus in the first pass. This dilution effect also leads to
also be taken into account for the initial bolus phase to somewhat longer relaxation times compared with the first
optimize the sequence parameters. However, the lower pass, even for the very high relaxivity of Gadofosveset.
binding fraction in the initial bolus is offset by the very The anticipated order of magnitude of T1 is therefore esti-
high initial concentration of Vasovist, so that, all in all, mated in the following simple example calculation [36].
it can be safely assumed that this relative effect is over-
compensated by the effect of the absolute very high local Estimation of T1 in the steady state on the basis of an ex-
concentration of Vasovist to produce the actual reduction ample calculation:
in T1 relaxation time. Body weight: 75 kg
Hence, this aspect appears rather insignificant for Total blood volume: 5.5 l
most clinical applications and therefore adaptation of se- Vasovist standard dose: 0.03 mmol / kg
quence parameters (e.g. flip angle) to improve image qual- Individual Gd dose: 2.25 mmol
ity would not appear worthwhile in most practical cases. Vasovist injected concentration: 0.25 mmol/l
This is supported by quantitative estimates as well as by Vasovist injected volume: 9 ml
current experience, which shows the use of established Calculated maximum concentration
sequence parameters with the common, fast 3D gradient [CM]max: 0.45 mmol/l
echo sequences to produce excellent results in first pass Estimated concentration following
MRA using Vasovist. initial distribution phase 80% of
[CM]max; [CM]: 0.36 mmol/l

1.5.2 Relevant Implications for Application Entering typical values for the non-enhanced longitudinal
in the Steady State relaxation time in blood T1(0) and for the relaxivity r1 of
Vasovist under the given standard conditions:
The situation in the later acquisition phases, i.e., ad-
vanced distribution throughout the entire vascular sys- T1(0)-1 = 0.8 s-1 (equivalent to 1.25 s = T1(0) of blood)
tem, differs from the well-known first pass, as described r1 = 19 l mmol-1 s-1 (measured in whole blood at 37 C, 1.5
previously. The situation in the steady state can certainly T, 0.5 mM concentration) [6],
be viewed as generally far simpler and easier to calculate,
as both the spatial and temporal dynamics have changed and using the general relationship between relaxivity, con-
from the contrast agent bolus rapidly propagating in centration and relaxation times (Eq. 1)
the vascular system with inhomogeneous concentration
distribution into a now homogeneous distribution and T1-1 = T1(0)-1 + r1[CM]
14 Chapter 1 MRI Contrast Media Introduction and Basic Properties of the Blood Pool Agent Gadofosveset (Vasovist)

The estimated relaxation rate T1-1 in blood is derived as: this without sacrificing signal-to-noise ratio. One of
1 these is to increase the repetition time (for example in
9.35 s-1 = 0.8 s-1 + 19 l mmol-1 s-1 0.36 mmol l-1 the 8- to 15-ms range), to select the flip angles smaller
(approximately 1525), to keep TE at the minimum,
and the corresponding reciprocal value for the estimated and to reduce the detection bandwidth, as appropriate. A
relaxation time T1 of approx. 130 ms. Dependent on the complementary strategy would be to improve the signal-
exact time of observation and on widely varying individual to-noise ratio by increasing the number of acquisitions,
physiological conditions, T1 can be used to calculate relax- while maintaining TR short in the usual range (for ex-
ation times of approx. 100200 ms. ample 47 ms), but to lower the flip angle by around 10
to account for the longer T1. The latter method can be
In comparison, T1 relaxation times of 20 ms and below considered, for example, for body regions that are subject
can exist in the bolus for first pass MRA due to the to respiratory motions, and may be applied in conjunc-
higher contrast agent concentrations here. It is quite tion with respiratory triggering and advanced navigator
obvious that the common protocols for fast 3D gradient techniques.
echo sequences established for contrast-enhanced MRA, Specific descriptions and sequence parameters found
for example, allow leeway for optimization in the steady for various regions of examination are described in the
state with Vasovist: The sequence protocols imple- following chapters.
mented for MRI scanners are generally optimized for
shorter first pass relaxation times, especially in regard
to the signal intensity as a function of the parameters
flip angle, echo time (TE), repetition time (TR), and References
relaxation times.
Accordingly, the parameter changes in the steady state 1 Claussen C, Laniado M, Schorner W, et al (1985) Gadolinium-DTPA
in MR imaging of glioblastomas and intracranial metastases. AJNR
tend towards smaller flip angles and longer repetition
Am J Neuroradiol 6:669674
times, whereby optimization of the flip angle is calculated 2 Weinmann H-J, Brasch RC, Press WR, et al (1984) Characteristics of
with the Ernst angle : gadolinium-DTPA complex: a potential NMR contrast agent. AJR
Am J Roentgenol 142:619624
= arccos(exp(-TR/T1)) Eq. (2) 3 Carr DH, Brown J, Bydder GM, et al (1984) Intravenous chelated
gadolinium as a contrast agent in NMR imaging of cerebral tu-
The option also arises of acquiring images with Vasovist
mours. Lancet 1:484486
in the steady state over an extended period of time with 4 Runge VM, Carollo BR, Wolf CR, et al (1989) Gd DTPA: a review of
a considerable increase in spatial resolution. There are clinical indications in central nervous system magnetic resonance
different ways for 3D gradient echo sequences to achieve imaging. Radiographics 9:929958

Take home messages

Based on unpaired electrons and their paramagnetic fully undergone clinical development and received
properties, MRI contrast media shorten proton spin market authorization.
relaxation times in vivo. Thereby enhancing the The mode of action of Gadofosveset is characterized
most important contrast mechanism in MRI, they basically by the two effects of the reversible protein
provide a valuable option for obtaining additional binding: 1. significantly increased relaxivity in the
diagnostic information. protein-bound state; 2. vastly prolonged retention
The most frequently applied class of MRI-CM con- time in the vascular system, providing much longer
sists of 0.5-molar, Gd-based metal complexes with imaging windows of up to 1 h (slow clearing, intravas-
extracellular biodistribution and fast glomerular cular).
filtration via the kidneys. Other concepts include In addition to the well-known first pass imaging, the
a highly concentrated 1.0-molar contrast medium, blood pool properties and the much longer period of
coated superparamagnetic iron oxide (SPIO) par- T1 shortening (steady state) of Gadofosveset overcome
ticles, and liver-specific Gd-based MRI-CM with weak the limitations of conventional extracellular MRI-CM,
protein binding. for example poor resolution in peripheral MRA.
Blood pool MRI-CM have been developed, but so Application of Gadofosveset for both first pass and
far only godofosveset, based on non-covalent and steady state MRA is easily achieved, and is optimized
reversible binding to serum albumin, has success- in the steady state by minor protocol adaptations.
15 1

5 Knopp MV, Balzer T, Esser M, Kashanian FK, Paul P, Niendorf HP man serum albumin and its effect on proton relaxation rates. J Am
(2006) Assessment of utilization and pharmacovigilance based on Chem Soc 124:31523162
spontaneous adverse event reporting of gadopentetate dimeglu- 24 Lauffer RB (1991) Targeted relaxation enhancement agents for
mine as a magnetic resonance contrast agent after 45 million ad- MRI. Magn Reson Med 22:339342
ministrations and 15 years of clinical use. Invest Radiol 41:491499 25 Eldredge HB, Spiller M, Chasse JM, Greenwood MT, Caravan P
6 Rohrer M, Bauer H, Mintorovitch J, Requardt M, Weinmann H-J (2006) Species dependence on plasma protein binding and relax-
(2005) Comparison of magnetic properties of MRI contrast me- ivity of the gadolinium-based MRI contrast agent MS-325. Invest
dia solutions at different magnetic field strengths. Invest Radiol Radiol 41:229243
40:715724 26 Parmelee DJ, Walovitch RC, Ouellet HS, Lauffer RB (1997) Preclini-
7 Bogdanov Jr. A, Lewin M, Weissleder R (1999) Approaches and cal evaluation of the pharmacokinetics, biodistribution, and elimi-
agents for imaging the vascular system. Adv Drug Delivery Rev nation of MS-325, a blood pool agent for magnetic resonance
37:279-293 imaging. Invest Radiol 32:741747
8 Jacques V, Desreux JF (2002) New classes of MRI contrast agents. 27 Clinical Study Report MS-325-06, Clinical Summary 2.7.2. Schering
Top Curr Chem 221:123164 AG
9 Nunn AD, Liner KE, Tweedle MF (1997) Can receptors be imaged 28 EMRF (European Magnetic Resonance Forum Foundation) (1997)
with MRI agents? Q J Nucl Med 41:155-162 Recommendations for the nomenclature of MR imaging contrast
10 Fossheim SL, Colet J-M, Mansson S, Fahlvik AK, Muller RN, Klave- agent terms. Acta Radiol 38:5
ness J (1998) Paramagnetic liposomes as magnetic resonance im- 29 Port M, Corot C, Violas X, et al (2005) How to compare the
aging contrast agents: assessment of contrast efficacy in various efficiency of albumin-bound and nonalbumin-bound contrast
liver models. Invest Radiol 33:810821 agents in vivo: the concept of dynamic relaxivity. Invest Radiol
11 Misselwitz B, Schmitt-Willich H, Ebert W, Frenzel T, Weinmann H-J 40:565573
(2001) Pharmacokinetics of gadomer-17, a new dendritic mag- 30 de Haen C, Calabi L, La Ferla R (2002) The problematic determina-
netic resonance contrast agent. Mag Reson Materials Physics, Biol tion of proton magnetic relaxation rates of protein-containing
Med 12 128134 solutions. Acad Radiol 9 [Suppl 1]:S24
12 Li D, Zheng J, Weinmann H-J (2001) Contrast-enhanced MR imag- 31 Lauffer RB (1987) Metal complexes as water proton relax-
ing of coronary arteries: comparison of intra- and extravascular ation agents for NMR imaging: Theory and design. Chem Rev
contrast agents in swine. Radiology 218:670678 87:901927
13 Misselwitz B, Schmitt-Willich H, Michaelis M, Oellinger JJ (2002) 32 Caravan P, Ellison JJ, McMurry TJ, et al (1999) Gadolinium(III)
Interstitial magnetic resonance lymphography T1 contrast agent: Chelates as MRI contrast agents: structure, dynamics, and applica-
initial experience with gadomer-17. Invest Radiol 37:146151 tions. Chem Rev 99:22932352
14 Port M, Corot C, Rousseaux O, et al (2001) P792: a rapid clearance 33 Banci L, Bertini I, Luchinat C (1991) Nuclear and electron relax-
blood pool agent for magnetic resonance imaging: preliminary ation: the magnetic nucleus-unpaired electron coupling in solu-
results. MAGMA 12:121127 tion. Weinheim: VCH
15 Port M, Corot C, Raynal I, et al (2001) Physicochemical and bio- 34 Bertini I, Luchinat C (1996) NMR of paramagnetic substances.
logical evaluation of P792, a rapid clearance blood pool agent for Coord Chem Rev 150:1292
magnetic resonance imaging. Invest Radiol 36:445454 35 Bloembergen N, Morgan LO (1961) Proton relaxation times in
16 Taupitz M, Wagner S, Schnorr J, Kravec I, Pilgrimm H, Bergmann- paramagnetic solutions. Effects of electron spin relaxation. J Chem
Fritsch H, Hamm B (2004) Phase I clinical evaluation of citrate- Phys 34:843850
coated monocrystalline very small superparamagnetic iron oxide 36 Hartmann M, Wiethoff AJ, Hentrich H-R, Rohrer M (2006) Initial
particles as a new contrast medium for magnetic resonance imag- imaging recommendations for Vasovist angiography. Eur Radiol
ing. Invest Radiol 39:394405 Suppl 16 [Suppl 2] B15B23
17 Anzai Y, Prince MR, Chenevert TL, et al (1997) MR angiography with 37 Goyen M, Shamsi K, Schnberg SO (2006) Vasovist-enhanced MR
an ultrasmall superparamagnetic iron oxide blood pool agent. J angiography. Eur Radiol 16[Suppl 2] B9B14
Magn Reson Imaging 7:209214 38 Wang MS, Haynor DR, Wilson GJ, Leiner T, Maki JH (2007) Maximiz-
18 Kellar KE, Fujii DK, Gunther WH, et al (2000) NC100150 injection, ing contrast-to-noise ratio in ultra-high resolution peripheral MR
a preparation of optimized iron oxide nanoparticles for positive- angiography using a blood pool agent and parallel imaging. J
contrast MR angiography. J Magn Reson Imaging 11:488494 Magn Reson Imaging 26:580588
19 Clarke SE, Weinmann H-J, Dai E, et al (2000) Comparison of two 39 Nikolaou K, Kramer H, Grosse C, Clevert D, Dietrich O, Hartmann
blood pool contrast agents for 0.5-T MR angiography: experimen- M, Chamberlin P, Assmann S, Reiser MF, Schoenberg SO (2006):
tal study in rabbits. Radiology 214:787794 High-spatial-resolution multistation MR angiography with parallel
20 Tombach B, Reimer P, Bremer C, Allkemper T, Engelhardt M, Mahler imaging and blood pool contrast agent: initial experience. Radiol-
M, Ebert W, Heindel W (2004) First pass and equilibrium-MRA of ogy 241: 861872
the aortoiliac region with a superparamagnetic iron oxide blood 40 Rapp JH, Wolff SD et al (2005) Aortoiliac occlusive disease in pa-
pool MR contrast agent (SH U 555 C): results of a human pilot tients with known or suspected peripheral vascular disease: safety
study. NMR Biomed 17:500506 and efficacy of gadofosveset-enhanced MR angiography multi-
21 Steger-Hartmann T, Graham PB, Mller S, Schweinfurth H (2006) center comparative phase III study. Radiology 236:7178
Preclinical safety assessment of Vasovist (gadofosveset trisodium), 41 Grist TM et al (1998) Steady state and dynamic MR angiography
a new magnetic resonance imaging contrast agent for angiogra- with MS-325: initial experience in humans. Radiology 207:539
phy. Invest Radiol 41:449459 544
22 Farooki A, Narra V, Brown J (2004) Gadofosveset EPIX/Schering.
Curr Opin Invest Drugs 5:967976
23 Caravan P, Cloutier NJ, Greenfield MT, McDermid SA, Dunham SU,
Bulte JWM, Amedio JC, Looby RJ, Supkowski RM, Horrocks WDeW,
McMurry TJ, Lauffer RB (2002) The interaction of MS-325 with hu-

Technical Aspects of Contrast Enhanced MRA

First Pass and Steady State
Harald H. Quick

2.1 Introduction 18

2.2 Hardware 18
2.2.1 Main Magnet 18
2.2.2 Gradient System 18
2.2.3 RF System 18

2.3 Data Acquisition: from k-Space to MRA Images 20

2.3.1 Parallel Imaging Strategies 20

2.4 Data Acquisition and Contrast Kinetics 23

2.4.1 First pass MRA 23
2.4.2 Steady state MRA 24
2.4.3 Suppression of Venous Signal 24

2.5 Extending the Field-of-View 26

2.5.1 Multi-station Peripheral MRA 27
2.5.2 Continuously Moving Table Peripheral MRA 28

2.6 Conclusion 30

References 30
18 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

2.1 Introduction its RF signal homogeneity and signal sensitivity over the
imaging volume ( Fig. 2.1). State-of-the-art MRA impo-
Since its introduction in the mid 1990s [1], contrast-en- ses very specific demands on these system components of
hanced magnetic resonance angiography (CE-MRA) has the scanner hardware.
2 increased enormously in impact and today has replaced
X-ray-based invasive catheter angiography for numerous
diagnostic investigations [2]. From the beginning, MRA 2.2.1 Main Magnet
applications have been very demanding on the hardware
requirements of MR imaging (MRI) scanners. Requiring The magnet of a whole-body MRI scanner should have
state-of-the-art equipment, MRA has not just profited a high main magnetic field strength, B0, to provide suf-
from the wide range of software and hardware develop- ficient equilibrium magnetization and therefore a high
ments of the recent past, but also has been a driving force potential SNR for good image quality. MR scanners with
behind numerous technical innovations. field strengths above 1.0 Tesla or, better yet, 1.5 Tesla are
In addition to superconducting magnets providing currently viewed as the standard and are increasingly being
high field strength for a strong magnetization and thus supplemented with 3-Tesla systems in clinical use. The ho-
a basis for high signal-to-noise ratio (SNR), the gradient mogeneity of the basic magnetic field over the examination
system plays an important role: it must be strong enough volume should be as high as possible to ensure low image
to switch the highest possible gradient amplitudes in the distortion and high signal homogeneity. The homogeneous
shortest possible time to deliver short repetition times examination volume should be as large as possible and
(TR) and echo times (TE). The radiofrequency (RF) coils should be free of artifacts and distortions. A cylindrical de-
have to receive the precious MRI signal with the very sign of the main magnet (solenoid) is conformant with all
minimum of loss. Phased-array surface RF coil techno- these requirements and therefore represents the currently
logy here provides a speed-up in acquisition time and most frequently occurring magnet design ( Fig. 2.1).
enhanced data acquisition flexibility using parallel ima-
ging strategies combined with whole-body surface RF coil
coverage of the patient. Finally, technical developments, 2.2.2 Gradient System
such as stepwise or even continuously moving table tech-
niques in association with new image acquisition and For performing contrast-enhanced MRA, two parameters
reconstruction schemes, expand the constraints of a con- of the gradient system are critical: a fast gradient slew rate
ventional limited field-of-view (FOV) to seamless large- (in [mT/m/ms]) combined with a high gradient amplitude
FOV MRA data sets with whole-body MRA coverage. (in [mT/m]) are the prerequisites for short repetition and
With the clinical availability of intravascular con- echo times (TR and TE) and thus for fast imaging and
trast agents [3,4] such as Vasovist (Gadofosveset, Bayer coverage of a large examination volume in the shortest
Schering Pharma AG, Berlin, Germany) [4-6], contrast- possible time and within the arterial phase of the contrast
enhanced MRA now has reached another degree of fle- agent kinetics. This can be seen as a fundamental prere-
xibility. Beyond conventional first pass MRA, intravascu- quisite for clinically acceptable examination times and for
lar contrast agents with their prolonged blood retention covering the large volumes of multistation or continuo-
times additionally allow for prolonged data acquisition usly moving table MRA. Attainment of a short TR is also
in the steady state of the contrast agent, which potentially directly coupled to achieving the highest possible T1 con-
can be used to obtain increased spatial resolution, exten- trast in MRA. This results in high signal of the vasculature
ded anatomic coverage, or, alternatively, to acquire MR filled with T1-shortening contrast agent while signal from
venograms with high spatial resolution. In this chapter the static tissue is saturated, thus providing high vessel-to-
the general technical aspects of contrast-enhanced MRA background contrast. A high degree of gradient linearity
as well as the specific aspects of MRA using intravascular over a large volume is required by the gradient system to
contrast agents are discussed. keep image distortion in and around the image volume to
a minimum. These requirements stated for the gradient
system can also be best realized with a cylindrical design
2.2 Hardware ( Fig. 2.1).

The usable FOV of an MRI system in general is defined

by three hardware groups and their related parameters: 1) 2.2.3 RF System
the main magnet with its homogeneity over the imaging
volume, 2) the gradient system with its linearity over the The radiofrequency (RF) system of an MRI scanner in
imaging volume, and 3) the radiofrequency system with general consists of a built-in RF transmit coil for RF
2.2 Hardware
19 2

Fig. 2.1AC. A sagittal cross-sectional view through a cylindrically for spin excitation inside the patients body. All three fields (B0, Gxyz,
shaped MRI scanner. Basically, three different cylindrically shaped sets and B1) are specified such that they form an imaging volume in the
of coils are implemented into the scanner housing: (A) the magnet coil isocenter of the magnet bore. Here the fields B0 and B1 are very homo-
producing the homogeneous static magnetic field B0 for spin align- geneous and the gradient fields GXYZ are relatively linear. The MR signal
ment; (B) a set of gradient coils producing the time-varying gradient response emitted from the excited spins in the patient is detected by
fields, Gx, Gy, and Gz, for spatial encoding of the MR signals as well as a set of local surface RF receiver coils that are placed on above and
(C) the transmit radiofrequency (RF) coil that generates an RF field (B1) underneath the anatomic imaging region

excitation of the spins as well as of an RF receiver coil Hence, diverse concepts with local RF surface coils have
system for receiving the weak RF signals being emitted come into use. The aim is to receive the signal from the
from the patient. RF excitation of the tissue volume of examination region with the highest possible SNR to
interest should be as homogeneous as possible. Here it achieve high image quality with good spatial resolution.
is important that the RF excitation flip angle for signal Surface RF coils are placed directly over the examination
excitation remain as stable as feasible, as the image region and receive signals from the immediate vicinity
contrast is fundamentally influenced by this parameter. of the examination volume. At the same time, these coils
Large cylindrical volume transmit RF coils fitted in the detect the unavoidable noise from a relatively limited
cylindrical magnet tunnel are used here, as they fulfill region so that the potentially achievable SNR with these
this requirement ( Fig. 2.1). coils is relatively high. At the same time, the disadvantage
On the signal receive side, homogeneous RF signal of surface coils lies in their severely restricted sensitivity
reception over as large an examination volume as possi- range. Several coil elements of this type are required to
ble is on the list of requirements. However, homogeneity cover larger anatomic regions; they can be combined as
plays a lesser role here compared with signal sensitivity, phased-array surface coils. Phased-array coils allow opti-
as inhomogeneous reception intensity affects only the mization of the SNR while extending the region for signal
brightness distribution over the volume, but not the reception ( Fig. 2.2). For imaging of large regions up
underlying image contrast. With regard to the homo- to whole-body MRA applications a full set of dedicated
geneous volume, the whole-body RF coil permanently local phased-array RF surface coils can be combined on
fitted in the magnet tunnel also offers an advantage here. state-of-the-art 1.5T and 3.0T MRI systems. A prerequi-
For signal reception and the maximum SNR achievable, site for exploiting the advantages inherent in this type of
such a large-volume coil is extremely limited, however. coil technology is a large number of RF receivers to which
20 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

tain best results, it is essential that the order of the phase-

encoding data acquisition scheme is known. Sampling of
the central lines of k-space should coincide with arrival
of the contrast agent in the target vessels, since central k-
2 space encodes contrast information in the reconstructed
MR images while the peripheral parts of k-space encode
spatial resolution ( Fig. 2.4).
Contrast-enhanced MRA is typically based on 3D
gradient-echo sequences with radiofrequency spoiling
[7]. The acquisition time for a 3D MRA data set is given
by: TAc = TR x Ny x Nz where TR is the repetition time,
Ny is the number of phase-encoding steps in in-plane di-
Fig. 2.2. Modern radiofrequency (RF) surface coil technology (Sie-
rection and Nz is the number of phase-encoding steps in
mens Medical Solutions, Erlangen, Germany). While the built-in RF slice direction [7]. From this it can be seen that reducing
body coil of the MR scanner transmits RF energy into the patients the acquisition time for a given resolution can simply be
body to excite the spins, MR signal is received with dedicated local RF achieved by reducing the TR to minimal achievable va-
surface coils that cover the region of interest. The patient can be cov-
lues. This already has been discussed for the specifications
ered with a full complement of dedicated RF surface coils. Whole-body
coverage of the patient with separate RF surface coils in combination
of the gradient system. Modern MR systems typically pro-
with motorized table movement permits whole-body MRA data ac- vide minimal TRs in the range of about 25 ms. Shorter
quisition with a relatively high SNR and the combination with parallel TRs might be limited technically by the gradient system
imaging techniques or even physiologically by the onset of peripheral nerve
stimulation of the patient.
In an effort to reduce the acquisition time with con-
ventional Cartesian Fourier encoding even further, while
the coil elements can be connected, either individually or maintaining the spatial resolution, one has to think about
in groups. This is also a basic prerequisite for the use of different strategies involving the reduction of phase-en-
parallel imaging techniques. coding steps. This will be discussed in the following

2.3 Data Acquisition: from k-Space

to MRA Images 2.3.1 Parallel Imaging Strategies

The MR signal that is detected by the RF receiver coils in Parallel imaging strategies in MRI are a technique to
the form of image raw data first is collected and stored increase the speed of the MRI acquisition by skipping a
in a two-dimensional (2D) or a three-dimensional (3D) number of phase-encoding lines in k-space during the
matrix also termed k-space. Following a mathematical process of MRI data acquisition. The basic idea behind
operation called fast Fourier-transformation (FFT), these this technique is to employ additional information in-
raw data are transformed into 2D or 3D MR images. This herent to the spatial geometry of the signal-receiving
context is illustrated in Fig. 2.3. For 2D data acquisition RF surface coil elements. Phased-array surface RF coils
a 2D matrix of complex data points with the size Nx x Ny is employing multiple individual coil elements are used
sampled. Nx here is the number of samples along the fre- to spatially encode the received MR signals. Thus, the
quency encoding direction. During each TR interval, an time-consuming acquisition of phase-encoding steps,
echo is formed and sampled with Nx points and filled into Ny and Nz, can be reduced by a specific reduction factor.
the matrix. Ny TR intervals (phase-encoding steps) are Missing information can be reconstructed by the known
necessary to complete the full 2D matrix. For the 3D case, spatial information provided by the individual signal-
Nz matrices have to be completed, where Nz is the number receiving coil elements. Two basic parallel imaging tech-
of phase-encoding steps in slice direction. Following data niques exist: SMASH (SiMultaneous Acquisition of Spa-
acquisition and the completion of the full 2D matrix, the tial Harmonics) [8] and SENSE (SENSitivity Encoding)
resulting 2D MR image can be reconstructed with a 2D [9]. All major vendors of MRI systems to date provide
FFT. In the 3D case, k-space is reconstructed with a 3D parallel imaging capabilities on their systems that are, in
FFT to a volume stack of 2D MR images. general, based on the generic techniques SMASH and/or
For contrast-enhanced MRA it is important to syn- SENSE and that differ in their specific implementations.
chronize the data acquisition sequence with injection of Parallel imaging techniques are also known under the
the contrast agent. For correct contrast timing and to ob- vendor-specific acronyms: SENSE (SENSivity Encoding,
2.3 Hardware
21 2

Fig. 2.3A,B. Raw data forming an MR image. The raw data for a 2D A mathematical operation, fast 2D Fourier transformation (2D FFT), is
image are scanned in the form of an Nx x Ny dimensioned matrix (A). used to obtain the final MRI image (B). In the 3D case, Nz 2D matrices
During each repetition time (TR) Nx complex data points are scanned are acquired and the raw data are stored in the form of a 3D-matrix, also
in the frequency-encoding direction and thereby fill a line of the matrix denoted as k-space. Accordingly, a 3D FFT transforms 3D k-space into a
with data. Ny TR times are required to fill the complete raw data matrix. stack of 2D images covering a 3D imaging volume

Fig. 2.4AF. Assignment of different regions of the raw data matrix. matrix (E) is reconstructed, the edges and fine detail information
The complete raw data matrix (A) is required to calculate a cor- are obtained (F); the resulting image, however, does not contain
responding image (B). If just the very central part of the raw data contrast information and only little signal. The high-signal center of
matrix (C) is reconstructed using 2D Fourier transformation to form the raw data matrix is therefore of key importance for MRA. This basic
an image, it becomes apparent that the contrast for the final image knowledge is, for instance, essential for the correct synchronization of
is contained within this part of the raw data matrix (D). However, the contrast agent arrival and the start of data acquisition using contrast
image (D) is still blurred. Conversely, if the outer part of the raw data agent-enhanced MRA
22 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

Philips Medical Systems, Best, The Netherlands); AS- or calibration scan that is independent of the subsequent
SET (Array Spatial Sensitivity Encoding Technique, GE imaging sequences realized by Philips Medical Systems.
Medical Systems, Milwaukee, USA); SPEEDER (Toshiba Alternatively, this calibration scan can be integrated into
Medical Systems, Tokyo, Japan); iPAT (integrated Par- the imaging sequences by acquisition of a few additional
2 allel Acquisition Techniques, Siemens Medical Solu- calibration k-space data lines into the imaging sequence
tions, Erlangen, Germany) featuring two reconstruction as realized by Siemens Medical Solutions. Depending on
algorithms mSENSE (modified SENSE) and GRAPPA the number of individual RF receiving coil elements that
(Generalized Autocalibrating Partially Parallel Acquisi- are available for spatial encoding in a specific orienta-
tions) [10,11]. tion, a specific reduction or acceleration factor (also
With SMASH this reconstruction is performed on the denominated as R-factor, PAT-factor, SENSE-factor) can
raw data in k-space, while with SENSE the individual im- be achieved to reduce the acquisition time. When 3D
ages of the individual coil elements are first reconstructed sequences are used, parallel imaging can be performed
( Fig. 2.5). These partial images are then added using a simultaneously in two orthogonal orientations of the two
mathematical unfolding algorithm. In both basic parallel phase-encoding directions. The acceleration factors in
imaging techniques, SMASH and SENSE, information of this case can simply be multiplied. In a clinical setting,
the signal sensitivity and of the individual coil element po- however, the maximum practical acceleration factor is
sition is used for image reconstruction. This information limited by the available SNR. If the imaging application is
is obtained by performing an additional short reference already limited by SNR, parallel imaging might degrade

Fig. 2.5AF. Basic parallel MRI techniques SMASH and SENSE. In par- direction to a fully reconstructed k-space (E). Thus a full MR image (F)
allel MRI, data acquisition of the object (A) is performed using phased- can be acquired and reconstructed in reduced time. In parallel MRI
array surface RF coils here with eight independent coil elements with SENSE the reduced k-space (B) is used to reconstruct a number of
connected to individual RF receivers. Depending on the number of RF partial images that show image folding (wrap around, C). The knowl-
coil elements and coil geometry, data acquisition of phase-encoding edge of the spatial coil sensitivities (D) is used to unfold these folded
steps can be reduced (B) by a specific reduction or acceleration factor. partial images and to reconstruct a full image (F). Note that while
In parallel MRI using SMASH, the a priori knowledge of spatial coil sen- SMASH reconstruction is performed in the k-space domain, SENSE
sitivities (D) is used to reconstruct missing raw data in phase-encoding reconstruction is performed in the image domain
2.4 Data Acquisition and Contrast Kinetics
23 2

image quality even further by adding additional noise. acquisition during the first pass of the contrast kinetics
In cases where sufficient SNR is available, as in contrast- only, MRA with Gadofosveset offers the additional pos-
enhanced MRA applications, parallel imaging can be an sibility of acquiring MRA data during the steady state of
attractive and powerful means to increase the efficiency the contrast kinetics. This opens up an additional time
of the administered contrast bolus [12-15]. Here paral- window in MRA for increasing the spatial resolution and/
lel imaging can be used to increase volume coverage per or the spatial coverage, especially in the above-mentioned
unit time for large FOV imaging of large vessel segments, applications [16].
to increase the spatial resolution within a given image
volume, or even to combine both effects [12-15]. Since
contrast-enhanced MRA in general is performed using 2.4.1 First pass MRA
3D sequences, simultaneous parallel imaging in two spa-
tial dimensions can be used for significantly increased First pass MRA using Gadofosveset can be performed as
data acquisition volume and spatial resolution in the with conventional extracellular contrast agents [17]. In
time window available by the administered contrast bolus order to optimize the contrast-to-tissue ratio, the flip an-
[12-15]. gle has to be slightly adapted to the increased relaxivity
of Gadofosveset. In general, a flip angle of approximately
35, which is higher than the Ernst angle, is optimal for
2.4 Data Acquisition and Contrast Kinetics a TR of approximately 25 ms and an assumed blood T1
relaxation time of about 50 ms [16]. As with conventio-
With conventional contrast agents, the image acquisition nal contrast agents, the scan should be timed such that
window for MRA of the arteries is limited, as the concen- acquisition of the central k-space coincides with the
tration of contrast rapidly declines due to fast clearance of peak concentration of Gadofosveset in the arterial target
the contrast agent. This ultimately limits the achievable vessels ( Fig. 2.6). Fluoroscopic triggering or automatic
spatial resolution for contrast-enhanced MRA. Imaging bolus tracking should be used to determine the con-
applications that suffer the most from this time constraint trast arrival in the target vessel. Use of a small-volume
are multi-station MRA of the peripheral arteries due to test bolus is also an option, although the intravascular
the bolus-chase nature of such applications, as well as nature of the contrast agent may lead to minor back-
MRA in regions where respiratory and/or cardiac mo- ground and venous enhancement in the actual first pass
tion compensation is necessary. Beyond performing data MRA scan [16].

MRA sequence acquisition time

center of k-space

arterial enhancement
venous enhancement
contrast injection

scan delay time-to-center

0 sec 4 sec 8 sec 12 sec 16 sec 20 sec time

Fig. 2.6. First pass MRA data acquisition as a function of the contrast agent kinetics. The individual circulation time was previously calculated
by administration of a test bolus (scan delay). Timing and initialization of the MRA sequence is adjusted such that acquisition of the center of
k-space falls into the arterial window of the contrast kinetics
24 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

MRA sequence acquisition time

arterial enhancement
2 venous enhancement
contrast injection

steady-state conditions a

0 sec 20 sec 40 sec 60 sec 3 min 5 min 7 min 9 min time

Fig. 2.7. Steady state MRA data acquisition as a function of the contrast agent kinetics. Initialization of the MRA sequence starts after the
contrast agent is evenly distributed in the arterial and venous vascular system (typically 35 min after contrast agent injection. Data acquisition
timing under steady state contrast conditions allows for a significantly prolonged data acquisition window for high-resolution MRA

2.4.2 Steady state MRA this allows for multi-planar reformation and viewing of
the results.
Following contrast administration, a steady state in the The larger contrast window in steady state MRA can
contrast kinetics is reached after about 3 min or later, be used to increase the signal-to-noise ratio in high-
depending on individual physiology ( Fig. 2.7). Due to resolution data sets by increasing the number of mea-
the slow clearance from blood, imaging is possible up to surements and/or by deactivation of partial k-space ac-
1 h after administration of Gadofosveset [16]. This signi- quisition schemes such as half Fourier imaging [16].
ficantly increases the data acquisition window for MRA The optimum flip angle for maximizing blood-to-muscle
applications and thus offers the possibility to acquire contrast-to-noise ratio is approximately 68 greater than
high-resolution MRA data sets or to use prepulses and the Ernst angle [19].
motion-compensation schemes.
Compared with first pass imaging, where a rela-
tively high concentration of contrast agent is present 2.4.3 Suppression of Venous Signal
in the arterial vasculature, blood T1 shortening is less
pronounced in the steady state phase of the contrast During the first pass the contrast agent is in the arteries
kinetics. This can be explained by the dilution effects, and is then also increasingly disseminated in the veins in
leading to a decreased local contrast agent concentration the steady state of the contrast kinetics. The arteries and
being distributed homogeneously in the entire blood veins are shown with a comparably high level of contrast
volume. In the steady state the T1 of blood at 1.5 Tesla in the steady state before the contrast agent is expelled over
is about 150 50 ms, which is substantially shorter than time via the kidneys. While MRA data acquisition in the
the T1 of fat and vessel-surrounding tissues [6]. In the first pass allows for pure arterial image contrast, imaging
case of Gadofosveset the intravascular concentration in the steady state consequently displays arteries and veins
remains relatively stable in the steady state, allowing for with comparable signal intensity. This can be problematic
lengthy MRA data acquisition schemes featuring increa- for image interpretation, since arteries and veins may be
sed spatial resolution through use of larger matrices, and difficult to separate under certain conditions. Here new
increasing the number of slices per volume, which ulti- strategies for data acquisition, data reconstruction, and
mately results in smaller voxels ( Fig. 2.8) [18]. Steady for data post-processing are advantageous. These will be
state MRA with Gadofosveset can be performed with discussed in the following chapter.
standard T1-weighted fast gradient echo MRA sequen- One approach for an alternative data acquisition re-
ces. Acquisition of isotropic voxels is recommended, as gime is to acquire the central, contrast-encoding parts of
2.4 Data Acquisition and Contrast Kinetics
25 2


first pass MRA steady state MRA

Fig. 2.8A,B. First pass (A) and steady state MRA (B) of the peripheral trast at this time point is equally distributed in the arterial and venous
run-off arteries in the calf of a patient with severe intermittent clau- vessel system. Consequently, the MIP of the steady state MRA shows
dication following the injection of Vasovist. For first pass MRA, data arteries and veins in the target region (B). Since the intravascular con-
acquisition has been timed such that the acquisition of central k-space trast agent remains in the vasculature for several more minutes, timing
of the MRA sequence (typically several seconds) falls into the arterial of steady state MRA is less critical and the increased data acquisition
contrast window of the contrast agent in target vessels. Consequently, window (typically several minutes) can be used for acquisition of high-
the maximum-intensity projection (MIP) of the first pass MRA shows resolution MRA here with an in-plane resolution of 200 m x 200 m.
only the arteries in the target region (A). For steady state MRA, data (Image examples (A, B) courtesy of Dr. Tim Leiner, MD, Department of
acquisition was started several minutes after contrast injection. Con- Radiology, University Hospital Maastricht, The Netherlands)

k-space during first pass of the contrast agent in the arte- central part of the first pass raw data matrix is now re-
ries [20, 21] while extending data acquisition of the peri- constructed together with the peripheral parts of k-space
pheral parts of k-space into the steady state of the contrast of the steady state raw data, an arterial MR angiogram
agent. This results in a high-resolution MRA of arteries with very high spatial resolution is obtained while venous
while suppressing venous overlay to a certain extent. Seg- superposition is avoided ( Fig. 2.9). This data acquisi-
mented k-space techniques such as CENTRA (Contrast- tion concept has been described as a proof of principle
ENhanced Time-Robust Angiography) have shown this by Foo et al. [23] for multi-station peripheral MRA using
potential for high-resolution first pass CE-MRA [22]. a conventional contrast agent. While the technique de-
Another approach combining alternative data acquisi- scribed still has to be considered as a research application,
tion and reconstruction regimes for CE-MRA using blood the combination of such segmented volume acquisition
pool contrast agents is displayed in Fig. 2.9. During the with the use of intravascular contrast agents holds high
first pass of the contrast agent through the arteries in potential for high-resolution MRA with venous signal
the target region, only the central part of the raw data suppression.
matrix is acquired. These data can be reconstructed to a Beyond applying data acquisition and data reconstruc-
high-contrast MR angiogram of the arteries however, tion schemes to provide venous signal suppression in
with reduced spatial resolution. During the steady state steady state MRA, this aim can also be achieved by a
of the contrast agent, a complete high-resolution data set number of post-processing techniques for artery/vein se-
is acquired. The corresponding MR angiogram contains paration [24, 25]. These techniques are examined in more
both arteries and veins with high spatial resolution. If the detail in the chapter on post processing.
26 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

Fig. 2.9AF. Possible raw data acquisition patterns for high-resolu- arterial MR angiogram with very high spatial resolution is obtained
tion MRA using a blood pool contrast agent. During the first pass (F), however, avoiding venous superposition. As opposed to MRA with
of the contrast agent through the arteries in the target region, only conventional extracellular contrast agent, this technique potentially
the central part of the raw data matrix is acquired, which contains the can be used together with an intravascular contrast agent to increase
contrast information (A). A high-contrast MR angiogram of the arteries the spatial resolution, because the contrast agent remains in the ves-
is reconstructed, however, with reduced spatial resolution (B). During sels during the time-consuming acquisition of the high-resolution
the steady state of the contrast agent, a complete high-resolution data set. Such segmented volume data acquisition today still has to
data set is acquired (C). The corresponding MR angiogram contains be considered as a research application. (Image examples (B, D, F)
both arteries and veins with high spatial resolution (D). If the central courtesy of Dr. Tim Leiner, MD, Department of Radiology, University
part of the first pass raw data matrix is now reconstructed (E) from Hospital Maastricht, The Netherlands)
(A) together with the remaining steady state raw data from (C), an

2.5 Extending the Field-of-View examinations, whereby the patient table, together with
the patient, is moved relative to the imaging volume in
MRI soon reaches its limits for comprehensive angiogra- between the partial measurements [2630]. These are
phic depiction of vessels with its limited FOV of appro- known as so-called multi-station techniques. They exploit
ximately 400500 mm. A comprehensive examination the fact that different body regions lie sequentially in the
of the peripheral arteries, however, typically requires an imaging region defined and limited by magnet homo-
FOV that covers the abdominal aorta, including the renal geneity, gradient linearity, and RF signal excitation and
arteries down to the pedal arteries a distance that can reception ( Fig. 2.1). These limitations remain applicable
be covered with an FOV of about 140 cm. One possibility for each partial imaging FOV with this method; however,
of extending the effective FOV for such an examination in this way the effective FOV may be expanded stepwise
is to divide the examination into several discrete partial over several stations.
2.5 Extending the Field-of-View
27 2

Fig. 2.10. Acquisition scheme for large-FOV MRA of the peripheral moved back to the starting position and contrast agent is injected. 3D
vasculature using multi-station moving table techniques. First over- contrast-enhanced acquisition of the first station is started according
view images or scouts are acquired at multiple stations covering to the previously determined individual bolus transit time. The 3D
the region of interest. One table movement takes place between the acquisitions of the second and third stations follow, after table move-
stations each time to reposition the respective anatomic region to the ment in each case. For additional acquisition of steady state MRA data
isocenter of the scanner. Following acquisition of the localizer images, following this first pass multi-station angiogram, the patient table
the coronal 3D MRA examination is planned and positioned on each can simply be stepped back through the isocenter of the magnet in
individual station. Prior to the injection of contrast agent, native 3D reversed order
data sets are acquired consecutively in all stations. The table is then

2.5.1 Multi-station Peripheral MRA The complex interaction between contrast agent injec-
tion and stepwise data acquisition with multiple table mo-
The basic requirement for the implementation of multi- vement requires careful planning of first pass bolus-chase
station techniques is the technical capability to move multi-station MRA examinations. Figure 2.10 illustrates
the patient table relative to the imaging field of the MR planning and acquisition of a three-station peripheral
scanner. Here, the maximum range of movement of the MRA examination ( Fig. 2.10). First, all three stations are
table directly determines the effective maximum FOV, consecutively positioned and overview images (locali-
which can be covered stepwise with an examination of zers or scouts) are produced. After the overview images
this type. Older-generation MR systems typically have for a station have been acquired, the table moves on to the
a range of movement limited to less than 150 cm. To next station and the next localizers are acquired. This is
examine a patient from head to toe it is therefore unavo- repeated until all anatomic regions of interest have been
idable to reposition the patient at least once during the covered stationwise with overview images ( Fig. 2.10).
examination. For example, in the first phase the patient The overview images serve for further planning of the 3D
is positioned head-first, in the second phase feet-first. MRA imaging volumes, for example in coronal slab ori-
The latest generation of MRI scanners now provide entation. After carefully planning the 3D MRA imaging
the option to move the table along a distance of about volumes at all stations, the examination begins without
200 cm, enough distance to provide whole-body MRA contrast agent administration starting at the abdomen
imaging capabilities. and moving towards the feet ( Fig. 2.10). Following data
28 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

Fig. 2.11AC. Peripheral multi-station 3D CE-MRA. The sketches the feet (B-C). Given the appropriate synchronization between the
show a longitudinal cross-section of a scanner. The patient with RF contrast agent bolus and the imaging time per station, the contrast
surface coils on top is first positioned with the thorax in the isocenter agent bolus is followed along its path from the aorta via the pelvis-
of the scanner (A). The coronal 3D acquisition volume (3D slab) is posi- leg vessels to the feet. This also allows vessels to be depicted whose
tioned within the isocenter. During the course of the examination, the dimensions exceed the size of a conventional FOV (MRA in A-C)
patient table, together with the patient, is advanced stepwise towards

acquisition and reconstruction, unenhanced native 3D station MRA imaging protocol by another one or two sta-
data sets are available from all three stations, which can tions even allows for whole-body MRA coverage [36-39].
later be subtracted from the 3D MRA contrast agent data Combining whole-body MRA with parallel imaging [40]
sets to improve image quality. now can be considered state-of-the-art on modern MRI
Taking into account the contrast agent bolus transit systems [41].
time previously determined, MRA data acquisition is
started with the appropriate delay after contrast agent
injection so that the data acquisition correlates with the 2.5.2 Continuously Moving Table Peripheral
arrival of the first pass contrast agent bolus in the vascular MRA
region to be imaged.
On completion of data acquisition at the first station, The multi-station bolus-chase MRA techniques pre-
the table is moved and acquisition follows at the second viously described help to overcome the constraints of
and the third station ( Figs. 2.10, 2.11). In this way, the conventional FOVs when larger vascular territories are
first pass contrast agent bolus is followed in its transit imaged. The stationwise examination of comprehensive
from the aorta, through the pelvis-leg arteries, towards vascular anatomy with a limited FOV, however, has other
the feet. 3D MRA data sets of the individual stations are inherent technical limitations: Data acquisition has to be
available at the end of a successful multi-station exami- interrupted for typically 34 s while the table is moving
nation ( Fig. 2.11AC). Multi-station MRA examinations between successive stations; the partial FOVs for each
are widespread in routine clinical CE-MRA of the peri- station typically overlap slightly by several centimeters,
pheral arteries nowadays [3135]. Extending the multi- requiring additional time for redundant data acquisition;
2.5 Extending the Field-of-View
29 2

time is necessary at each station for the re-establishment

of the equilibrium magnetization for the image sequence;
signal attenuation often arises on the margins of the
FOVs, which can lead to inhomogeneities in the resul-
ting combined overview images; and non-linearities in
the gradient system lead to geometric distortions on the
margins of the partial FOVs, which in turn can cause
artifacts on the boundaries between adjacent stations.
These limitations inherently reduce the time efficiency of
multi-station techniques and/or hamper the evaluation
of the images, especially on the boundaries of the partial
FOVs [42, 43]. The ideal situation would be a continuous
large-FOV 3D MRA data set with uninterrupted and
unchanging quality, which can be processed as a whole
during image viewing and assessment.
Various techniques have been investigated recently
as alternatives to the multi-station techniques [4247].
These techniques for continuous table movement during
data acquisition, also termed move during scan tech-
niques, place special requirements on data acquisition,
hardware, and image reconstruction: Inhomogeneity of
the main magnetic field and non-linearity of the gradients
generate image distortions at the margins of the FOV,
which are now a function of time, because the table is
moved through the scanner during acquisition [48]. In
addition, a number of adjustment parameters, which pre-
viously were modified once per station for the stationwise
examination of the patient, must now be automatically
adapted during table movement. These parameters in- Fig. 2.12A,B. Contrast-enhanced peripheral MRA acquired with the
clude the tuning of the RF transmitter coil, the transmis- multi-station technique (A) compared with the continuously moving
sion and receiver frequency, the RF transmission power, table technique (B), with which continuous data acquisition takes
place during patient table movement. The multi-station technique
the RF receiver gain, the shimming of the magnet, the
requires acquisition from several (here three) discrete, slightly over-
activation of the RF surface coil elements located in the lapping FOVs (A), whereas the continuously moving table technique
imaging region and other parameters, all of which are delivers a large, seamless FOV. With the multi-station technique, gradi-
needed to ensure optimal, homogeneous image quality ent distortions on the margins of the individual FOVs potentially lead
and the safety of the patient. Otherwise, the excitation flip to discontinuity artifacts. Both technologies have the potential to be
expanded to whole-body MRA
angle can, for instance, have a different value in the pelvic
region than in the thoracic region and the image contrast
varies unpredictably.
Initial clinical studies of continuously moving table well as adaptation of the continuously moving table speed
MRA in direct comparison with multi-station MRA with to the individual travel velocity of the contrast bolus [52].
volunteers and patients show promising results [49, 50] These steps will further increase the efficient use of the
( Fig. 2.12). Parallel imaging techniques can also be ap- contrast bolus. Intravascular contrast agents will funda-
plied here through the use of surface coils combined with mentally support the use of continuously moving table
independent RF receivers [4951] and the potential for techniques by increasing the flexibility of the potential
faster imaging, larger anatomic coverage in the same time, data acquisition schemes. Beyond combining moving ta-
and/or improved spatial resolution can be exploited. The ble techniques with conventional first pass bolus chase,
application of such continuously moving table MRA tech- the addition of steady state imaging opens up a whole new
niques is rewarded by the depiction of extended anatomic time window for significantly increased data acquisition
structures in large seamless images and has the potential times. Consequently, the data acquisition schemes pro-
of achieving a paradigm shift in MRA data acquisition, posed above and illustrated in Fig. 2.9 that make use
reconstruction, post-processing and evaluation [49, 50]. of the steady state to increase spatial resolution, can be
Further technical developments will feature increased combined with the continuously moving table techniques
spatial resolution of the large FOV 3D MRA data sets as and thus provide significantly increased spatial resolution
30 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

in large-FOV MRA. Initial experiments combining these numerous hardware and software developments, and,
data acquisition schemes with multi-station MRA and as a result, has replaced X-ray-based invasive catheter
Vasovist have shown promising results [53]. The combi- angiography for many diagnostic investigations. In the
nation of such data acquisition and reconstruction tech- process, MRA itself has initiated numerous technical
2 niques with continuously moving tables and blood pool developments in imaging from which modern MRI as a
contrast agents is envisaged. whole has profited enormously. The clinical advance of
new intravascular contrast agents such as Gadofosveset
allows both for first pass and steady state imaging, thus
2.6 Conclusion offering an additional window for high-resolution MRA
imaging. This impact of intravascular contrast agents is
Magnetic resonance angiography has achieved a high set to bring attractive applications and to further expand
diagnostic status in a short time, not least due to the the diagnostic spectrum of MRA.

Take home messages

The three basic hardware components of an MRI When injecting blood pool contrast agents for MRA,
system are: 1) The main magnet providing high the arterial first pass contrast phase known from
magnetic field strength [B0] for magnetization of the conventional extracellular contrast agents is fol-
spins in the tissue; 2) the gradient system providing lowed by an additional prolonged steady state of
spatial and time varying magnetic fields in three di- signal enhancement in the arteries and veins. This
mensions [Gx, Gy, Gz] for spatial encoding of the MRI provides enhanced flexibility for MRA data acquisi-
signal; 3) The radiofrequency (RF) system that excites tion, e.g. for increasing spatial resolution.
the spins by sending RF energy [B1] with the built-in Multi-station techniques can be used for extending
body RF coil and that receives weak RF signals being the limited field-of-view (FOV) of an MRI system in
emitted from the tissue. order to perform MRA of the peripheral arteries or
MR signals are collected and stored as raw data in even whole-body MRA. For MRA data acquisition
a two-dimensional (2D) or three-dimensional (3D) with multi-station techniques the patient table
matrix also termed k-space. Following a math- is moved through the isocenter of the magnet in
ematical operation called fast Fourier-transformation multiple successive stations while the first pass
(FFT), these raw data are transformed into 2D or 3D of contrast agent is traveling towards the patients
MR images. feet.
Parallel imaging strategies (e.g. SMASH, SENSE) in Continuously moving table techniques are currently
MRI are a technique to increase the speed of the being developed to replace multi-station MRA ex-
MRI acquisition by skipping a number of phase- aminations. Continuous techniques instead of multi-
encoding lines in k-space during data acquisition station techniques provide one seamless large-FOV
an otherwise time-consuming process. In MR MRA data set that is virtually free of discontinuity
angiography (MRA) parallel imaging techniques artifacts. In combination with continuous moving
can be used to reduce data acquisition time, to table techniques blood pool contrast agents provide
increase the spatial resolution per unit acquisition enhanced flexibility and efficiency for large-FOV
time or even to combine both effects. MRA data acquisition.

References 5. Caravan P, Cloutier NJ, Greenfield MT, McDermid SA, Dunham SU,
Bulte JW, Amedio JC Jr, Looby RJ, Supkowski RM, Horrocks WD
Jr, McMurry TJ, Lauffer RB (2002) The interaction of MS-325 with
1. Prince MR (1994) Gadolinium-enhanced MR aortography. Radiol- human serum albumin and its effect on proton relaxation rates. J
ogy 191:155164 Am Chem Soc 124:31523162
2. Yucel EK, Anderson CM, Edelman RR, Grist TM, Baum RA, Man- 6. Hartmann M, Wiethoff AJ, Hentrich HR, Rohrer M (2006) Initial
ning WJ, Culebras A, Pearce W (1999) AHA scientific statement. imaging recommendations for Vasovist angiography. Eur Radiol
Magnetic resonance angiography: update on applications for 16 [Suppl] 2:B1523
extracranial arteries. Circulation 100:22842301 7. Laub G (1999) Principles of contrast-enhanced MR angiography.
3. Lauffer RB (1991) Targeted relaxation enhancement agents for Basic and clinical applications. Magn Reson Imaging Clin N Am
MRI. Magn Reson Med 22:339342 7:783795
4. Lauffer RB, Parmelee DJ, Dunham SU, Ouellet HS, Dolan RP, Witte 8. Sodickson DK, Manning WJ (1997) Simultaneous acquisition of
S, McMurry TJ, Walovitch RC (1998) MS-325: albumin-targeted spatial harmonics (SMASH): fast imaging with radiofrequency coil
contrast agent for MR angiography. Radiology 207:529538 arrays. Magn Reson Med. 38:591603
31 2

9. Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P (1999) 27. Wang Y, Lee HM, Khilnani NM, Jagust MB, Winchester PA, Bush
SENSE: sensitivity encoding for fast MRI. Magn Reson Med HL, Sos TA, Sostman HD (1998) Bolus-chase MR digital sub-
42:952962 traction angiography in the lower extremity. Radiology 207:
10. Griswold MA, Jakob PM, Heidemann RM, Nittka M, Jellus V, Wang 263269
J, Kiefer B, Haase A (2002) Generalized autocalibrating partially 28. Meaney JF, Ridgway JP, Chakraverty S, Robertson I, Kessel D, Rad-
parallel acquisitions (GRAPPA). Magn Reson Med 47:12021210 jenovic A, Kouwenhoven M, Kassner A, Smith MA (1998) Stepping-
11. Blaimer M, Breuer F, Mueller M, Heidemann RM, Griswold MA, table gadolinium-enhanced digital subtraction MR angiography
Jakob PM (2004) SMASH, SENSE, PILS, GRAPPA: how to choose the of the aorta and lower extremity arteries: preliminary experience.
optimal method. Top Magn Reson Imaging.15:223236 Radiology 211:5967
12. Weiger M, Pruessmann KP, Kassner A, Roditi G, Lawton T, Reid 29. Ruehm SG, Hany TF, Pfammatter T, Schneider E, Ladd ME, Debatin
A, Boesiger P (2000): Contrast-enhanced 3D MRA using SENSE. J JF (2000) Pelvic and lower extremity arterial imaging: diagnostic
Magn Reson Imaging 12:671677 performance of three-dimensional contrast-enhanced MR angi-
13. Sodickson DK, McKenzie CA, Li W, Wolff S, Manning WJ, Edelman ography. AJR Am J Roentgenol 174:11271135
RR (2000): Contrast-enhanced 3D MR angiography with simulta- 30. Leiner T, Ho KY, Nelemans PJ, de Haan MW, van Engelshoven JM
neous acquisition of spatial harmonics: A pilot study. Radiology (2000) Three-dimensional contrast-enhanced moving-bed infu-
217:284289 sion-tracking (MoBI-track) peripheral MR angiography with flex-
14. Wilson GJ, Hoogeveen RM, Willinek WA, Muthupillai R, Maki JH ible choice of imaging parameters for each field of view. J Magn
(2004) Parallel imaging in MR angiography. Top Magn Reson Imag- Reson Imaging 11:368377
ing 15:169185 31. Goyen M, Ruehm SG, Barkhausen J, Kroger K, Ladd ME, Truemmler
15. Glockner JF, Hu HH, Stanley DW, Angelos L, King K (2005) Parallel KH, Bosk S, Requardt M, Reykowski A, Debatin JF (2001) Improved
MR imaging: a users guide. Radiographics 25:12791297 multi-station peripheral MR angiography with a dedicated vascu-
16. Rohrer M, Geerts-Ossevoort L, Laub G (2007) Technical require- lar coil. J Magn Reson Imaging 13:475480
ments, biophysical considerations and protocol optimization with 32. Ho KY, Leiner T, de Haan MW, van Engelshoven JM (1999) Periphe-
magnetic resonance angiography using blood-pool agents. Eur ral MR angiography. Eur Radiol 9:17651774
Radiol 17 [Suppl 2]:B7B12 33. Busch HP, Hoffmann HG, Rock J, Schneider C (2001) MR angiog-
17. Prince M, Grist TM, Debatin JF (2003) 3-D contrast MR angiogra- raphy of pelvic and leg vessels with automatic table movement
phy, 3rd edn. Springer, Berlin, Heidelberg, New York technique (MobiTrak): Clinical experience with 450 studies.
18. Nikolaou K, Kramer H, Grosse C, Clevert D, Dietrich O, Hartmann RoFo Fortschr Rntgenstr 173:405409
M, Chamberlin P, Assmann S, Reiser MF, Schoenberg SO (2006) 34. Huber A, Scheidler J, Wintersperger B, Baur A, Schmidt M, Re-
High-spatial-resolution multistation MR angiography with parallel quardt M, Holzknecht N, Helmberger T, Billing A, Reiser M (2003)
imaging and blood pool contrast agent: initial experience. Radiol- Moving-table MR angiography of the peripheral runoff vessels:
ogy 241:861872 comparison of body coil and dedicated phased array coil systems.
19. Wang MS, Haynor DR, Wilson GJ, Leiner T, Maki JH (2007) Maximiz- AJR Am J Roentgenol 180:13651373
ing contrast-to-noise ratio in ultra-high resolution peripheral MR 35. Leiner T, Nijenhuis RJ, Maki JH, Lemaire E, Hoogeveen R, van
angiography using a blood pool agent and parallel imaging. J Engelshoven JM (2004) Use of a three-station phased array coil
Magn Reson Imaging 26:580588 to improve peripheral contrast-enhanced magnetic resonance
20. Huston J 3rd, Fain SB, Riederer SJ, Wilman AH, Bernstein MA, Busse angiography. J Magn Reson Imaging 20:417425
RF (1999) Carotid arteries: maximizing arterial to venous contrast 36. Ruehm SG, Goyen M, Barkhausen J, Kroger K, Bosk S, Ladd ME,
in fluoroscopically triggered contrast-enhanced MR angiography Debatin JF (2001) Rapid magnetic resonance angiography for
with elliptic centric view ordering. Radiology 211:265273 detection of atherosclerosis. Lancet 357:10861089
21. Huston J 3rd, Fain SB, Wald JT, Luetmer PH, Rydberg CH, Covarru- 37. Ruehm SG, Goyen M, Quick HH, Schleputz M, Schleputz H, Bosk S,
bias DJ, Riederer SJ, Bernstein MA, Brown RD, Meyer FB, Bower TC, Barkhausen J, Ladd ME, Debatin JF (2000) [Whole-body MRA on
Schleck CD (2001) Carotid artery: elliptic centric contrast-enhan- a rolling table platform (AngioSURF)] Rofo Fortschr Rntgenstr
ced MR angiography compared with conventional angiography. 172:670674
Radiology 218:138143 38. Goyen M, Quick HH, Debatin JF, Ladd ME, Barkhausen J, Herborn
22. Willinek WA, Gieseke J, Conrad R, Strunk H, Hoogeveen R, von CU, Bosk S, Kuehl H, Schleputz M, Ruehm SG (2002) Whole-body
Falkenhausen M, Keller E, Urbach H, Kuhl CK, Schild HH (2002) three-dimensional MR angiography with a rolling table platform:
Randomly segmented central k-space ordering in high-spatial- initial clinical experience. Radiology 224:270277
resolution contrast-enhanced MR angiography of the supraaortic 39. Herborn CU, Goyen M, Quick HH, Bosk S, Massing S, Kroeger K,
arteries: initial experience. Radiology 225:583588 Stoesser D, Ruehm SG, Debatin JF (2004). Whole-body 3D MR
23. Foo TK, Ho VB, Hood MN, Hess SL, Choyke PL (2001) High-spatial- angiography of patients with peripheral arterial occlusive disease.
resolution multistation MR imaging of lower-extremity peripheral AJR Am J Roentgenol 182:14271434
vasculature with segmented volume acquisition: feasibility study. 40. Quick HH, Vogt FM, Maderwald S, Herborn CU, Bosk S, Gohde S,
Radiology 219:835841 Debatin JF, Ladd ME (2004) High spatial resolution whole-body
24. van Bemmel CM, Spreeuwers LJ, Viergever MA, Niessen WJ (2003) MR angiography featuring parallel imaging: initial experience.
Level-set-based artery-vein separation in blood pool agent CE-MR RoFo Fortschr Rntgenstr 176:163169
angiograms. IEEE Trans Med Imaging 22:12241234 41. Kramer H, Schoenberg SO, Nikolaou K, Huber A, Struwe A, Winnik
25. Lei T, Udupa JK, Odhner D, Nyl LG, Saha PK (2003) 3DVIEWNIX- E, Wintersperger BJ, Dietrich O, Kiefer B, Reiser MF (2005) Cardio-
AVS: a software package for the separate visualization of arte- vascular screening with parallel imaging techniques and a whole-
ries and veins in CE-MRA images. Comput Med Imaging Graph body MR imager. Radiology 236:300310
27:351362 42. Dietrich O, Hajnal JV (1999) Extending the coverage of true vol-
26. Ho KY, Leiner T, de Haan MW, Kessels AG, Kitslaar PJ, van Engels- ume scans by continuous movement of the subject. ISMRM, 7th
hoven JM (1998) Peripheral vascular tree stenoses: evaluation Scientific Meeting and Exhibition, Philadelphia, p 1653
with moving-bed infusion-tracking MR angiography. Radiology 43. Kruger DG, Riederer SJ, Grimm RC, Rossman PJ (2002) Continu-
206:683692 ously moving table data acquisition method for long FOV con-
32 Chapter 2 Technical Aspects of Contrast Enhanced MRA First Pass and Steady State

trast-enhanced MRA and whole-body MRI. Magn Reson Med

44. Zhu Y, Dumoulin CL (2003) Extended field-of-view imaging with
table translation and frequency sweeping. Magn Reson Med
2 45.
Fain SB, Browning FJ, Polzin JA, Du J, Zhou Y, Block WF, Grist TM,
Mistretta CA (2004) Floating table isotropic projection (FLIPR)
acquisition: a time-resolved 3D method for extended field-of-
view MRI during continuous table motion. Magn Reson Med
46. Madhuranthakam AJ, Kruger DG, Riederer SJ, Glockner JF, Hu HH
(2004) Time-resolved 3D contrast-enhanced MRA of an extended
FOV using continuous table motion. Magn Reson Med 51:568
47. Zenge MO, Ladd ME, Vogt FM, Brauck K, Barkhausen J, Quick HH
(2005) Whole-body magnetic resonance imaging featuring mov-
ing table continuous data acquisition with high-precision position
feedback. Magn Reson Med 54:707711
48. Polzin JA, Kruger DG, Gurr DH, Brittain JH, Riederer SJ (2004) Cor-
rection for gradient nonlinearity in continuously moving table MR
imaging. Magn Reson Med 52:181187
49. Zenge MO, Vogt FM, Brauck K, Jkel M, Barkhausen J, Kannengie-
sser S, Ladd ME, Quick HH (2006) High-resolution continuously
acquired peripheral MR angiography featuring partial parallel
imaging GRAPPA. Magn Reson Med 2006 56:859865
50. Vogt FM, Zenge MO, Ladd ME, Herborn CU, Brauck K, Luboldt
W, Barkhausen J, Quick HH (2007) Peripheral vascular disease:
comparison of continuous MR angiography and conventional MR
angiography--pilot study. Radiology 243:229238
51. Keupp J, Aldefeld B, Bornert P (2005) Continuously moving table
SENSE imaging. Magn Reson Med 53:217220
52. Kruger DG, Riederer SJ, Polzin JA, Madhuranthakam AJ, Hu HH,
Glockner JF (2005) Dual-velocity continuously moving table ac-
quisition for contrast-enhanced peripheral magnetic resonance
angiography. Magn Reson Med 53:110117
53. Kinner S, Zenge MO, Vogt FM, Ucan B, Ladd ME, Barkhausen J,
Quick HH (2007) Ultra-high resolution peripheral MRA with k-
space segmentation featuring a blood-pool contrast agent and
venous suppression. Proc ISMRM 19-25 May, Berlin

Part II Safety and Efficacy

Chapter 3 Risks and Safety Issues Related to Radiological Imaging,

in Particular MRI 35
Gunnar Brix

Chapter 4 Summary of Safety of Vasovist at 0.03 mmol/kg Body Weight

Dose Clinical Data 43
Matthias Voth and Andrea Lwe

Chapter 5 Efficacy of Vasovist: Overview of the Clinical Development

Program 51
Mathias Goyen

Risks and Safety Issues Related

to Radiological Imaging, in Particular MRI
Gunnar Brix

3.1 Introduction 36

3.2 Risks Related to Imaging Procedures Using Ionizing Radiation 36

3.3 Risks Related to MRI Procedures 36

3.3.1 Safety Regulations and Operating Modes 36
3.3.2 Static Magnetic Fields 37
3.3.3 Time-varying Magnetic Gradient Fields 37
3.3.4 Radiofrequency Electromagnetic Fields 38
3.3.5 Special Safety Issues and Contraindications 40

References 40
36 Chapter 3 Risks and Safety Issues Related to Radiological Imaging, in Particular MRI

3.1 Introduction In line with this philosophy, the Directive on Health

Protection of Individuals Against the Dangers of Ionizing
Radiological imaging techniques always pose some risks Radiation in Relation to Medical Exposure [7], issued
of adverse health effects to both patients and medical per- by the Council of the European Communities in 1997,
sonnel. Therefore, this issue has to be thoroughly evalu- requires that the following basic principles of radiation
ated in particular when introducing new imaging strate- protection be applied in clinical practice:
3 gies into clinical routine. As no ionizing radiation is used Justification: Medical exposure ... shall show a suf-
in MRI, it is generally deemed safer than conventional ficient net benefit, weighting the total potential diag-
X-ray or nuclear medicine procedures in terms of health nostic or therapeutic benefits it produces ... against the
risks. Nevertheless, there are possible risks and health ef- individual detriment that the exposure might cause,
fects associated with the use of diagnostic MR devices that taking into account the efficacy, benefits and risks
have to be considered carefully [13]. of available alternative techniques having the same
In this context, a fundamental difference between ion- objective but involving no or less exposure to ionizing
izing and non-ionizing radiation has to be noted: radiation radiation.
exposures related to diagnostic X-ray procedures result in Optimization: All doses due to medical exposure for
stochastic effects, whereas biological effects of magnetic radiological purposes ... shall be kept as low as reason-
fields used in MRI are deterministic. A stochastic proc- ably achievable (ALARA principle) consistent with
ess is one where the exposure determines the probability obtaining the required diagnostic information, taking
of the occurrence of an event but not the severity of the into account social and economic factors.
effect. In contrast, the severity of a deterministic effect is
related to the level of exposure, and a threshold may be This means, for example, that digital subtraction angiog-
defined [4]. As a consequence, the probability of detri- raphy (DSA) should be replaced by contrast-enhanced
mental effects caused, for example, by X-ray examinations MR angiography (CE-MRA) in clinical routine as soon as
performed over many years accumulates, whereas physi- there is sufficient evidence from clinical studies that both
ological stress induced by MR procedures is related to the angiographic techniques (a) are at least equivalent from a
acute exposure levels of a particular examination and does diagnostic point of view and (b) that the overall health risk
not, to our present knowledge, accumulate over years. (taking into account risks from the imaging technique,
the administration of the contrast agent, and the invasive-
ness of the procedure) related to a CE-MRA is lower than
3.2 Risks Related to Imaging Procedures that of a conventional DSA.
Using Ionizing Radiation

Detrimental effects of ionizing radiation, the most sig- 3.3 Risks Related to MRI Procedures
nificant being induction of cancer, have been identified
in human beings through various epidemiological studies In MR imaging three variants of magnetic fields are em-
at intermediate and high doses, i.e., organ or whole-body ployed to form cross-sectional images of the human body:
doses exceeding 50100 mGy, delivered acutely or over (a) a high static magnetic field generating a macroscopic
prolonged periods. There is, however, considerable con- nuclear magnetization, (b) rapidly alternating magnetic gra-
troversy regarding the radiological risk of low-level radia- dient fields for spatial encoding of the MR signal, and (c)
tion, typical for diagnostic radiation exposures, since the radio-frequency (RF) electromagnetic fields for excitation
stochastic risks evaluated at these dose levels are not based and preparation of the spin system. In the following, the
on experimental evidence but are rather extrapolated biophysical interaction mechanisms and biological effects of
down from risks observed at high doses [5, 6]. The ex- these fields will be briefly summarized and exposure limits
trapolation is based on the so-called linear non-threshold and precautions to be taken to minimize health hazards and
(LNT) hypothesis, which assumes that (a) any radiation risks to patients undergoing MR procedures will be noted1.
dose no matter how small may cause detrimental This is of special importance for CE-MRA examinations
health effects and that (b) the probability of these effects that are performed with extremely fast sequences relying on
is directly proportional to the dose absorbed in the tissue. the use of high-performance gradient and RF systems.
Although the risks evaluated at low dose levels are thus
hypothetical, it is prudent to presume that these risks ex-
ist and that the LNT model represents an upper boundary 3.3.1 Safety Regulations and Operating Modes
for them. It is for this reason that current radiation pro-
tection standards as well as risk assessments are generally In the recent past, regulations concerning the MR safety
based on the LNT hypothesis. of patients have been largely harmonized. There are two
3.3 Risks Related to MRI Procedures
37 3

comprehensive reviews by international commissions that [11]. A large number of studies have been conducted to
form the basis for both national safety standards and the detect biological responses to static magnetic fields rang-
implementation of monitor systems by the manufacturers ing in flux density from the milli-Tesla range to several
of MR devices: Teslas. These studies have been reviewed comprehen-
The technical product standard IEC 60601-2-33 is- sively among others by the ICNIRP [12] and WHO
sued by the International Electrotechnical Commis- [13]. Overall, there is little convincing evidence from cel-
sion (IEC) in 2002 [9] lular, animal, and epidemiological studies for biologically
The safety recommendation issued by the Interna- harmful effects of short-term exposure resulting from
tional Commission on Non-ionizing Radiation Pro- static magnetic fields with flux densities up to several
tection (ICNRIP) in 2004 [10]. Teslas.
Although there are initial experiences concerning the
In order to reflect the still existing uncertainty about examination of volunteers and patients using ultra-high
deleterious effects of (electro)magnetic fields and to offer MR systems with magnetic flux densities of up to 8 T,
the necessary flexibility for the development and clinical most clinical MR procedures have been performed so far
evaluation of new MR technologies, both safety guidelines at static magnetic fields below 3 T. The literature does not
give exposure limits for three different modes of opera- indicate any serious adverse health effects from the expo-
tion: sure of healthy human subjects up to a flux density of 8 T
Normal operating mode: Routine MR examinations [10, 13]. However, because movements in static magnetic
that do not cause physiological stress to patients. fields above 2 T can produce nausea and vertigo, both the
Controlled operating mode: Specific MR examinations IEC standard and the ICNIRP recommendation stipulate
outside the normal operating range where discomfort that MR examinations above this static magnetic flux
and/or physiological stress to some patients may occur. density should be performed in the controlled operat-
Therefore, a clinical decision must be taken to balance ing mode under medical supervision. The recommended
such effects against foreseen benefits and exposure upper limit for the controlled operating mode is 4 T, due
must be carried out under medical supervision. to the limited information concerning possible effects
Experimental operating mode: Experimental MR pro- above this magnetic flux density. For MR examinations
cedures with exposure levels outside the controlled performed in the experimental operating mode, there is
operating range. In view of the potential risks for no defined upper limit for the magnetic flux density. In
patients and volunteers, special ethical approval and a safety document issued in 2003, the Food and Drug
adequate medical supervision are required. Administration (FDA) deems MR devices a significant
risk only when a static magnetic field of more than 8 T
is used [14].
3.3.2 Static Magnetic Fields

The basic actions of static magnetic fields are (a) mag- 3.3.3 Time-varying Magnetic Gradient Fields
neto-mechanical effects, including the orientation of
magnetically anisotropic structures in uniform fields and The magnetic flux density of gradient fields used in MRI
the translation of paramagnetic or ferromagnetic materi- is about two orders of magnitude lower than that of the
als in magnetic field gradients, (b) electrodynamic inter- static magnetic field. Therefore, time-varying magnetic
actions with ionic conduction currents, and (c) effects on fields produced by gradient coils in MRI can be neglected
electron spin states of chemical reaction intermediates compared with the strong static magnetic field as far as
interactions of magnetic fields with biological tissues and
organisms are concerned (cf. Sect. 3.3.2). In contrast,
Limit and action values for personnel exposed to time-varying however, biophysical effects related to the electric fields
electric, magnetic, and electromagnetic fields were established by and currents induced by the magnetic fields have to be
the European Union (EU Physical Agents Directive (EMF) 2004/40/
considered carefully. According to Faradays law, the
EC) in 2004 on the basis of guidelines published by the International
Commission on Non-ionizing Radiation Protection [8]. The regulations
field strength of the induced electric field is proportional
of the EU directive, which should be transposed into national law by to the time rate of change of the magnetic flux density,
April 2008, may seriously compromise occupational handling in dB(t)/dt.
the periphery of an MR system, for example, the preparation or In general, rise times of magnetic gradients in MRI
positioning of the patient and administration of a contrast agent
are longer than 100 s, resulting in time-varying electric
during the examination. When this text was written, it was not clear
whether special arrangements would be established for medical
fields and currents with frequencies below 100 kHz. In
workplaces by the EU or the member states to avoid the expected this frequency range, the conductivity of cell membranes
negative consequences for the clinical use of MRI. is several orders of magnitude lower than that of the ex-
38 Chapter 3 Risks and Safety Issues Related to Radiological Imaging, in Particular MRI

tra- and intracellular fluid [15]. This has two important

consequences: First, the cell membrane tends to shield the
interior of cells very effectively from current flow, which
is thus restricted to the extracellular fluid. Second, volt-
ages are induced across the membrane of cells. When the
electric voltages are above a tissue-specific threshold level,
3 they can stimulate nerve and muscle cells.
The primary concern with regard to time-varying
magnetic fields is cardiac fibrillation, because it is a
life-threatening condition. In contrast, peripheral nerve
stimulation is of practical concern because uncomfort-
able or intolerable stimulations would interfere with the
examination (e.g., due to patient movements) or even
result in a termination of the examination. As determined
in animal experiments, excitation thresholds for the heart Fig. 3.1. Threshold for cardiac stimulation [16] and limits for normal
and controlled operation of an MR device. Data are expressed as dB/
are substantially greater than those for nerves as long as
dt as a function of the effective stimulus duration teff. The limit for the
the pulse duration is sufficiently less than about 3 ms [16]. controlled operation mode is given by the median threshold for the
Therefore, the avoidance of peripheral sensations in a perception of peripheral nerve stimulation
patient provides a conservative safety margin with respect
to cardiac stimulation.
Peripheral nerve stimulation has been investigated
in various volunteer studies. A systematic evaluation of relevant for MR procedures. Instead of using the param-
the available data was presented by Schaefer, Bourland eterization given by Eq. (1), mean perception threshold
and Nyenhuis in 2000 [17]. They recalculated published levels can be determined by the manufacturers for a given
threshold levels often reported for different gradient type of gradient system by means of experimental studies
coils and shapes in different terms to the maximum on human volunteers.
dB/dt occurring during the maximum switch rate of
the gradient coil at a radius of 20 cm from the central
axis of the MR system, i.e., at the border of the volume 3.3.4 Radiofrequency Electromagnetic Fields
normally accessible to patients. Bourland et al. also
analyzed their stimulation data in the form of cumulative Time-varying magnetic fields used for the excitation and
frequency distributions that relate a dB/dt level to the preparation of the spin system in MRI typically have fre-
number of subjects that had already reported on per- quencies above 10 MHz. In this RF range, the conductiv-
ceptible, uncomfortable, or even intolerable sensations ity of cell membranes is comparable to that of the extra-
[18]. The results indicate that the lowest percentile for and intracellular fluid, which means that no substantial
intolerable stimulation is approximately 20% above the voltages are induced across the membranes [15]. For this
median threshold for the perception of peripheral nerve reason, stimulation of nerve and muscle cells is no longer
stimulation, which can be parameterized by the following a matter of concern. Instead, thermal effects due to tis-
empirical relation: sue heating by the induced electric tissue currents are of
importance. The parameter relevant for the evaluation
= 20 1 +
dB 0.36
(inT/s) . Eq. (3) of biological effects of RF fields is the increase in tissue
temperature, which is dependent not only on localized
In this equation, teff is the effective stimulus duration (in power absorption and the duration of RF exposure, but
ms) defined as the duration of the period of monotonic also on heat transfer and the activation of thermoregula-
increasing or decreasing gradient. tory mechanisms leading to thermal equalization within
In the current safety recommendations issued by the the body.
IEC and ICNRIP, the maximum recommended exposure Absorption of energy in the human body strongly
level for time-varying magnetic fields is set equal to a dB/ depends on the size and orientation of the body with
dt of 80% of the median perception threshold given by respect to the RF field as well as on the frequency of the
Eq. (3) for normal operation and 100% of the median for field. Theoretical and experimental considerations reveal
controlled operation. As shown in Fig. 3.1, the threshold that RF absorption in the body approaches a maximum
for cardiac stimulation [16] is well above the median per- when the wavelength of the field is on the order of the
ception threshold for peripheral nerve stimulation, except body size. Unfortunately, the wavelength of the RF fields
at very long pulse durations, which are, however, not used in MRI falls into this resonance range.
3.3 Risks Related to MRI Procedures
39 3

Table 3.1. Basic restrictions for body temperature rise and partial-body temperatures for patients undergoing MRI procedures [9, 10]

Operating mode Rise of body-core temperature Spatially localized temperature limits

Head Trunk Extremities
(C) (C) (C)

Normal 0.5 38 39 40

Controlled 1 38 39 40

Experimental >1 > 38 > 39 > 40

Table 3.2. SAR limits (in W/kg) for patients when a volume coil is used to excite a greater field-of-view [9, 10]. The values hold at environ-
mental temperatures below 24C

Averaging time 6 min

Body region Whole-body Partial-body

Operating mode
any region, except head headb

Normal 2 210a 3.2

Controlled 4 410a 3.2

Experimental >4 > (410) > 3.2

Short-term SAR: The SAR limit over any 10-s period shall not exceed three times the corresponding average SAR limit.
Partial-body SARs scale dynamically with the ratio r between the patient mass exposed and the total patient mass:
normal operating mode: SAR = (10 - 8 . r) W/kg
controlled operating mode: SAR = (10 - 6 . r) W/kg
Partial-volume SARs given by IEC; ICNIRP limits SAR exposure to the head to 3 W/kg.

The time rate of electromagnetic energy absorption by cal temperatures under exposure to the head, trunk, and
a unit of tissue mass is defined as specific absorption rate extremities are limited for each of the two operating
(SAR in W/kg). In good approximation, the time-averaged modes to the values given in Table 3.1.
SAR of an RF pulse is proportional (a) to the square of the However, temperature changes in the different parts
static magnetic field, B0, which means that energy absorp- of the body are difficult to control during an MR proce-
tion is markedly higher with high-field than with low-field dure in clinical routine. Therefore, SAR limits have been
MR systems, and (b) to the square of the pulse angle, so that derived on the basis of experimental and theoretical stud-
MRA sequences with large-angle pulses will result in rela- ies which should not be exceeded, in order to limit the
tively high SAR levels. From the point of view of safety, CE- temperature rise to the values given in Table 3.1. As only
MRA at 3 T is thus much more challenging than at 1.5 T parts of the body at least in the case of adult patients
and makes it necessary to use optimized coil designs as well are exposed simultaneously during an MR procedure,
as sophisticated pulse sequences utilizing SAR reduction not only the whole-body SAR but also partial-body SAR
techniques such as parallel imaging or hyperechos. values for the head, the trunk, and the extremities have to
According to published studies, no adverse health be estimated by means of suitable patient models [19] and
effects are expected if the RF-induced increase in body- limited to the values given in Table 3.2 for the normal
core temperature does not exceed 1C. In case of infants, and controlled operating mode. With respect to the ap-
pregnant women, or persons with cardiocirculatory im- plication of the SAR limits defined in Table 3.2, it has to
pairment, it is desirable to limit body-core temperature be taken into account that they do not relate to an indi-
increases to 0.5C [10]. As indicated in Table 3.1, these vidual MR sequence, but rather to running SAR averages
values have been laid down in the current safety recom- computed over each 6-min period, which is assumed to
mendations to limit the body-core temperature for the be a typical thermal equilibration time of smaller masses
normal and controlled operating mode. Additionally, lo- of tissue [20].
40 Chapter 3 Risks and Safety Issues Related to Radiological Imaging, in Particular MRI

3.3.5 Special Safety Issues and carefully interviewing the patient, evaluating the patients
Contraindications file, and contacting the implanting clinician and/or the
manufacturer for advice on MR safety and compatibility
To protect volunteers, patients, accompanying persons, of the implant. MR examinations of patients with active
and uninformed health-care workers from possible haz- implants are strictly contraindicated, provided that the
ards and accidents associated with the MR environment, patient implant card does not explicitly state their safety in
3 it is indispensable to perform an appropriate control of the MR environment. Comprehensive information on the
access to the MR environment. The greatest potential MR compatibility of more than 1200 implants and other
hazard comes from metallic, in particular ferromagnetic metallic objects is available in a reference manual published
materials (such as coins, pins, hair clips, pocket knives, by Shellock [21] and online at
scissors, nail clippers, etc.) that are accelerated in the Pregnant women undergoing MR examinations are
inhomogeneous magnetic field in the periphery of an MR exposed to the combined magnetic and electromagnetic
system and quickly become dangerous projectiles (mis- fields used in MRI. The few studies on pregnancy out-
sile effect). This risk can be minimized only by a strict come in human beings following MR examinations have
and careful screening of all individuals entering the MR not revealed any adverse effects, but they are very limited
environment for metallic objects. because of the small numbers of patients involved and dif-
MR examinations of patients with passive implants ficulties in the interpretation of the results [10]. It is thus
(e.g., vascular clips and clamps, intravascular stents and advisable that MR procedures be performed in pregnant
filters, vascular access ports and catheters, heart valve pros- patients, in particular in the first trimester, only after
theses, orthopedic prostheses, sheets and screws, intrau- critical risk/benefit analysis and with verbal and written
terine contraceptive devices), active implants (e.g., cardiac informed consent of the mother or parents [22]. The
pacemakers and defibrillators, cochlear implants, electronic standard of care is that MRI may be used in pregnant
drug-infusion pumps), or other objects of ferromagnetic or women, if other non-ionizing forms of diagnostic imaging
unknown material (pellets, bullets) are always associated (e.g., sonography) are inadequate or if the examination
with a serious risk, even if all procedures are performed provides important information that would otherwise re-
within the established exposure limits summarized in the quire exposure to ionizing radiation (e.g., when perform-
previous sections. This risk can be minimized only by ing conventional DSA instead of MRA) [22, 23].

Take home messages

There are possible risks and health effects associ- Exposure limits for the different magnetic fields
ated with the use of diagnostic MR devices that used in MRI are specified in recent safety regula-
have to be considered, such as nerve stimulation tions and have to be considered in clinical routine.
caused by time-varying magnetic gradient fields MR examinations of patients with implants are
or tissue heating induced by electromagnetic always associated with a serious risk, even if all
radiofrequency fields. procedures are performed within the established
Physiological stress induced by MR procedures is exposure limits.
related to the acute exposure levels of a particular From the perspective of radiation hygiene, digital
examination and does not in contrast to stochas- subtraction angiography should be replaced by
tic radiation effects caused by X-ray procedures contrast-enhanced MR angiography as far as pos-
accumulate over years. sible.

References 5. United Nations Scientific Committee on the Effects of Atomic Ra-

diation (2000) Sources and effects of ionizing radiation. Report to
1. Brix G (2007) Risks and safety issues related to MR examinations. the General Assembly, with scientific annexes. Vol II: Effects, Annex
In: Reiser M, Semmler W, Hricak H (eds) Magnetic resonance tom- G. United Nations, New York
ography, 1st English edn. Springer-Verlag, Berlin Heidelberg New 6. Committee to Assess Health Risks from Exposure to Low Levels of
York, Chap. 2.9 pp 153167 Ionizing Radiation, National Research Council (2006) Health risks
2. Ordidge R, Shellock FG, Kanal E (2000) Special issue: MR safety. J from exposure to low levels of ionizing radiation: BEIR VII Phase 2.
Mag Reson Imaging 12 National Academies Press, Washington, DC
3. Shellock FG (ed) (2001) Magnetic resonance procedures: health 7. Council of the European Union (1997) Council directive 97/43/
effects and safety. CRC Press, Boca Raton Euratom of 30 June 1997 on Health Protection Against the Dan-
4. International Commission on Non-ionizing Radiation Protection gers of Ionizing Radiation in Relation to Medical Exposure, and
(ICNIRP) (2002) General approach to protection against non-ioniz- Repealing Directive 84/466/Euratom. Document 397L0043. Of-
ing radiation. Health Physics 82:540548 ficial Journal NO. L 180, 09/07/1997 pp 00220027
41 3

8. International Commission on Non-ionizing Radiation Protection

(ICNIRP) (1998) Guidelines for limiting exposure to time-varying
electric, magnetic and electromagnetic fields. Health Physics
9. International Electrotechnical Commission (2002) IEC 60601-2-33
(2nd edn) Particular requirements for the safety of magnetic reso-
nance equipment for medical diagnosis.
10. International Commission on Non-ionizing Radiation Protection
(2004) Medical magnetic resonance (MR) procedures: protection
of patients. Health Physics 87:197216
11. Schenck JF (2005) Physical interactions of static magnetic fields
with living tissues. Prog Biophys Molec Biol 87:185204
12. International Commission on Non-Ionizing Radiation Protection
(ICNIRP) (2003) Exposure to static and low frequency electromag-
netic fields, biological effects and health consequences (0-100
kHz). Matthes R., McKinlay AF, Bernhardt JH, Vecchia P, Veyret, B,
(eds) Oberschleissheim, Germany: publication ICNIRP 13/2003
13. World Health Organization (WHO) (2006) Environmental health
criteria 232. Static fields. World Health Organization, Geneva
14. Food and Drug Administration. Center for Devices and Radio-
logical Health (2003) Criteria for significant risk investigations
of magnetic resonance diagnostic devices.
15. Foster KR, Schwan HP (1996) Dielectrical properties of tissues. In:
Polk C, Postow E (eds) Handbook of biological effects of electro-
magnetic fields. CRC Press, Boca Raton, pp 25102
16. Reilly JP (1998) Applied bioelectricity: from electrical stimulation
to electro-pathology. Springer, New York
17. Schaefer DJ, Bourland JD, Nyenhuis JA (2000) Review of patient
safety in time-varying gradient fields. J Magn Reson Imaging
18. Bourland JD, Nyenhuis JA, Schaefer DJ (1999). Physiologic effects
of intense MR imaging gradient fields. Neuroimaging Clin N Am
19. Brix G, Reinl M, Brinker G (2001) Sampling and evaluation of spe-
cific absorption rates during patient examinations performed on
1.5-Tesla MR systems. Magn Reson Imaging 19:769779
20. Brix G, Seebass M, Hellwig G, Griebel J (2002). Estimation of heat
transfer and temperature rise in partial-body regions during MR
procedures: an analytical approach with respect to safety consid-
erations. Magn Reson Imaging 20:6576
21. Shellock FG (2008) Reference manual for magnetic resonance
safety, implants, and devices: 2008 edition. Biomedical Research
Publishing Company, Los Angeles
22. Colletti PM (2001) Magnetic resonance procedures and preg-
nancy. In: Shellock FG (ed) Magnetic resonance procedures: health
effects and safety. CRC Press, Boca Raton, pp 149182
23. Shellock FG, Crues JV (2004). MR procedures: biologic effects,
safety, and patient care. Radiology 232:635652

Summary of Safety of Vasovist

at 0.03 mmol/kg Body Weight Dose
Clinical Data
Matthias Voth and Andrea Lwe

4.1 Introduction 44

4.2 Vasovist 45
4.2.1 Clinical studies 45
4.2.2 Recommended clinical dose (0.03 mmol Gd/kg) 46
4.2.3 Safety in postmarketing surveillance 48
4.2.4 Special populations 48
4.2.5 Discussion 48

References 49
44 Chapter 4 Summary of Safety of Vasovist at 0.03 mmol/kg Body Weight Dose Clinical Data

4.1 Introduction cellular contrast agents for imaging of all body regions
and/or organs. GBCAs of the next generation yielded
The first gadolinium-containing contrast-enhancing better images through higher relaxivity and higher con-
agent for magnetic resonance imaging (MRI) Magne- centration [e.g., Gadobutrol (Gadovist, Bayer Schering
vist (Gd-DTPA, Bayer Schering Pharma AG, Berlin, Pharma AG, Berlin, Germany)] [9]. Still more recently,
Germany), was introduced in 1988 [1, 2]. Today a va- GBCAs such as Gadoxetic acid (Gd-EOB-DTPA, Primov-
riety of different Gd-based contrast agents (GBCAs) are ist, Bayer Schering Pharma AG, Berlin, Germany) have
available for clinical use, and they have become a stand- been introduced that are tailored for use in specific organ
ard part of the radiologists diagnostic repertoire. systems or tissues. Most recently, a Gd-chelate containing
4 All of these agents have two structural features in blood pool contrast agent allowing for an extended imag-
common: a central gadolinium (Gd3+) ion, which pos- ing window, i.e., Vasovist (Gadofosveset, Bayer Schering
sesses the paramagnetic properties required for contrast Pharma AG, Berlin, Germany), has been introduced for
enhancement, and a surrounding organic molecule (lig- use in contrast-enhanced MR-Angiography (CE-MRA).
and) that binds the gadolinium ion. The purpose of the Despite the differences in imaging preconditions and
ligand is threefold: It allows contact between the gado- the range of approved indications between the available
linium ion and the surrounding water molecules while GBCAs in use today, the common structures and basic
at the same time tightly binding the ion and thus pre- similarities of these substances have so far resulted in
venting exposure of the patient to toxic free Gd3+. Based largely similar safety profiles. In nearly all patients these
on their chemical structures GBCAs may be divided in agents are well-tolerated, and side effects, where present,
general into two groups: linear (open-chain) Gd-chelates are most often mild and transient [10, 11]. Nevertheless,
and macrocyclic Gd-chelates [3]. The likelihood that a some safety-related questions have been raised recently,
particular Gd-chelate will release Gd3+ ions depends on which indicate that the existing differences in physico-
that particular chelates chemical structure and respec- chemical and pharmacokinetic properties of the available
tive in vivo stability. Finally, its properties determine the GBCAs may result in substantial differences in some
distribution of the contrast agent to the various body aspects of the safety profile of these agents, in particular
compartments to which the contrast agent is exposed for special populations under particular circumstances.
a factor that, in turn, determines the mechanism and First of all, a possible interference of gadoliniuim-based
kinetic (t1/2) by which the contrast agent is excreted. contrast agents with laboratory measurements is de-
The contrast-enhancing power and pharmacokinetic scribed. The ability of some gadolinium contrast agents
properties, the safety and range of indications of a Gd- to falsify calcium measurements, resulting in apparent
chelate-containing contrast agent thus all depend to a hyper- or hypocalcemia, has been documented for some
large extent upon the choice of organic ligand (chelate) GBCA, such as Gadodiamide (Omniscan) and Gadover-
( Table 4.1). setamide (OptiMARK) [1217], in which the thermody-
While basic structural similarities exist, the available namic stability of the metal-ion binding is known to be
GBCAs differ in their pharmacokinetic and pharma- less strong [18]. However, this effect has not been seen
codynamic properties and thus in their major imaging with Gd-DTPA (Magnevist), Gadobutrol (Gadovist), or
efficacy profiles resulting in their use for a variety of Gadofosveset (Vasovist) [17, 19].
diagnostic questions: GBCAs of the first generation [e.g., Very recently a very rare disease has been described
Gd-DTPA (Magnevist)] are used as non-specific extra- and termed nephrogenic systemic fibrosis (NSF; formerly

Table 4.1. Physiochemical properties of linear chelates [48] related to stability

Contrast agent Charge Thermodynamic stability Conditional stability Excess-free ligand

(log Ktherm, pH 14) (log Kcond, pH 7.4) in formulation

Omniscan (Gadodiamide) Non-ionic 16.9 14.9 12 mg/ml (5%)

OptiMark (Gadoversetamide) Non-ionic 16.8 15.0 28.4 mg/ml (10%)

Magnevist (Gadopentetate dimeglumine) Ionic 22.5 18.4 0.4 mg/ml (~0.1%)

MultiHance (Gadobenate dimeglumine) Ionic 22.6 18.4 0 mg/ml (0%)

Primovist (Gadoxetate disodium) Ionic 23.5 18.7 1.0 mg/ml (~0.6%)

Vasovist (Gadofosveset trisodium) Ionic 22.1 18.9 0.27 mg/ml (~0.1%)

4.2 Vasovist
45 4

known as nephrogenic fibrosing dermopathy, NFD); it space. Vasovist is eliminated from the body not only via
has been observed only in patients with severe renal im- the kidneys, but to a small percentage also via the hepato-
pairment. A possible association of gadolinium as a con- biliary route. While the amount of the administered dose
tributing factor in was first suggested in early 2006 [20]. excreted via the bile in healthy volunteers ranges up to 9%,
Since then, numerous reports of NSF in association with this amount increases up to 22% of the administered dose
the administration of gadolinium-based contrast agents in patients with severe renal impairment [31].There is no
have been published in the scientific literature, the major- evidence of any metabolism [30].
ity of them associated with Gadodiamide (Omniscan).
[2023, 2527].
The purpose of this chapter is to provide an up-to- 4.2.1 Clinical studies
date summary of knowledge on the overall safety profile
of Gadofosveset (Vasovist), a blood pool contrast agent. All doses (0.01 0.15 mmol Gd/kg)
Since the first clinical trial with Vasovist was conducted
in healthy volunteers, more than 20 clinical studies have
4.2 Vasovist been performed to investigate the safety and tolerability
as well as the efficacy of Vasovist-enhanced MRI in vari-
Vasovist is a gadolinium-based intravascular blood pool ous indications. Of this clinical development program,
contrast agent which received the first marketing au- safety data of a subset of studies have thus far been pub-
thorization in the EU in 2005. It can be used for first pass lished in one detailed overview. In this overview a total
or for steady state MRA or both following a single i.v. of 1438 subjects, including 117 healthy volunteers, 1203
bolus administration. Vasovist binds reversibly to hu- patients with known or suspected vascular disease, and
man serum albumin, the bound fraction predominating 118 other patients were included in a total of 18 clinical
(~85%), and is eliminated predominantly renally. The studies: seven phase I/II pharmacological studies, two
albumin-binding fraction of the contrast agent shows a dose-ranging phase II studies, four pivotal phase III stud-
slow spin, spinning at the same rate as albumin, thereby ies in various vascular territories, and five additional
increasing relaxivity and thus generating a brighter signal feasibility studies. Doses administered ranged between
than would be possible with unbound circulating Gd- 0.01 and 0.15 mmol Gd/kg with the majority of patients
chelate. Furthermore the prolonged blood pool retention receiving the clinically recommended dose of 0.03 mmol
limits the amount of background signal and prolongs Gd/kg [30, 32].
blood half-life, which allows steady state blood pool ac- For the 1321 Vasovist-treated patients in these trials,
quisition of MR angiograms within a time window of up 415 adverse events were reported which were considered
to ~60 min following administration [2830]. A 0.25-M to be at least possibly treatment-related; of the latter, 81%
formulation was developed to obtain an injection volume were mild, 17% were moderate, and 2% were severe. The
that is suitable for bolus injection and permits a clinically most common treatment-related events are shown in Ta-
reasonable injection speed. Its use is indicated in adult ble 4.3 and included pruritus (7%), paresthesia (6%), nau-
patients with suspected or known vascular disease.1 sea, vasodilatation, burning sensation (all 5%), dysgeusia,
Vasovist belongs to the group of ionic linear contrast headache (both 3%), and feeling cold (1%). The incidence
agents. It is provided as a 0.25-M solution and is approved was determined to be dose dependent ( Table 4.2).
for use as a single i.v. injection at a dose of 0.03 mmol Twelve patients experienced a total of 15 serious ad-
Gd/kg (=0.12 ml/kg). The plasma concentration-time verse events. Two of these were experienced by patients
profile of i.v. administered Vasovist conforms to a two- who received <0.03 mmol Gd/kg, five by patients who
compartment open model. Following i.v. administration received 0.03 mmol Gd/kg, four by patients who received
of the clinically recommended dose of 0.03 mmol Gd/ 0.05 mmol Gd/kg, and two by patients who received
kg in healthy volunteers, the mean half-life of the distri- >0.05 mmol Gd/kg. Four serious adverse events (SAEs),
bution phase (t1/2) is 0.48 0.11 h. The renal clearance including one death, were considered by the investiga-
(5.5 ml/h/kg) and total clearance (6.6 ml/h/kg) are similar, tor to be possibly or probably related to Vasovist: These
and mean terminal elimination half-life is ~18 h [30]. adverse events were allergic reaction, chest pain, and
The volume of distribution at steady state is 148 16 ml/ anaphylactoid reaction. The fourth case was reported with
kg, roughly equivalent to that of the extracellular fluid an SAE of arteriosclerotic cardiovascular disease and had
a fatal outcome. This event was considered by the inves-
tigator to be possibly related to Vasovist with a received
Approval in Switzerland and Turkey has been granted for visualization
dose of 0.07 mmol Gd/kg. There was no evidence of dose
of vascular disease and in Australia, Canada and the EU for abdominal dependency regarding frequency of SAEs overall or for
or limb vascular disease individual events [31].
46 Chapter 4 Summary of Safety of Vasovist at 0.03 mmol/kg Body Weight Dose Clinical Data

4.2.2 Recommended clinical dose the 767 patients (0.5%) experienced five serious adverse
(0.03 mmol Gd/kg) events, which included aggravated coronary artery disease,
hypoglycemia (both of these reported by one patient),
Patients from this subgroup received a dose of 0.03 mmol/ gangrene, chest pain, and anaphylactoid reaction. The
kg Vasovist across all clinical studies using Vasovist instances of gangrene and chest pain were deemed un-
and part of the overall assessment of all dose ranges likely to be related to the study drug. Overall, no clinically
(0.010.15 mmol Gd/kg) as described above. For the significant changes in any laboratory parameters were seen
recommended dose of 0.03 mmol Gd/kg, safety results after injection of Vasovist. Hematology, coagulation, and
in phase II and III studies have been described in detail urine values were normal, with sporadic deviations only.
4 for a pooled collective of 767 patients [64]. In summary, Specifically, there was no evidence of a depletion of zinc,
the incidence rates and the profile of adverse events in calcium, or iron from serum following administration of
patients who had received Vasovist at the clinically rec- Vasovist. Individual changes in immunological variables
ommended dose were comparable with those in patients (complement C3, complement C4 plasma histamine, IgE,
who had received placebo, and in fact the proportion of tryptase-CAPS), ECG parameters and the results of physi-
patients with adverse events was lower in the Vasovist cal examination all raised no safety concerns.
group than in the placebo group. Most (~82%) of the ad- One prospectively planned phase IIIb study was con-
verse events in the Vasovist group were assessed as mild ducted between 2006 and 2007 [77] for which a total of
in intensity. Seventeen severe adverse events were reported 264 patients with aortoiliac disease Fontaine stage IIbIV
for 15 patients (2%) in the Vasovist group. Six events were were enrolled; 261 patients received Vasovist. The age of
considered to be at least possibly related to the treatment the patients ranged from 32 to 86 years (mean 65 years);
with Vasovist and included burning sensation, dysgeusia, 171 patients (65.5%) were male and 90 (34.5%) were
headache, pruritus, rash, and pain. Individual adverse female, and they received a dose of 0.03 mmol Gd/kg
events are shown in Table 4.3 including the subgroup Vasovist. A total of 39 patients (14.9%) experienced 51
for the clinically recommended dose of 0.03 mmol Gd/kg. adverse events; 33 adverse events were assessed as being
The most common adverse events ( Table 4.2) following (probably: 16, possibly: 17) related to the study drug. The
treatment with Vasovist (possibly or probably related to most common events were grouped to gastrointestinal
treatment) were pruritus (4%), nausea (4%), paresthesia disorders (nausea), followed by nervous system disor-
(3%), and vasodilatation (3%). Onset was generally (70% ders (burning sensations) and reproductive system/breast
of all treatment-related events) within 2 h of dosing, disorders (genital burning). The intensities of all drug-
and symptoms usually resolved in most cases (75% of related adverse events were mild or moderate. No serious
treatment-related events) within 2 h after onset. Four of adverse drug reactions were reported.

Table 4.2. Number and percentage of patients who experienced the most common Gadofosveset-related adverse events [30]

Placebo Dose (mmol Gd/kg) All doses

<0.03 0.03 0.05 >0.05

n=49 n=95 n=767 n=348 n=111 n=1321

n % n % n % n % n % n %

Any Gadofosveset-related AE 16 33 14 15 176 23 150 43 75 67 415 31

Pruritus * 1 2 1 1 34 4 39 11 20 18 94 7

Paresthesia 1 2 20 3 39 11 19 17 78 6

Headache * 2 4 1 1 17 2 13 4 3 3 34 3

Nausea 1 1 29 4 25 7 5 5 60 5

Vasodilatation 1 1 22 3 19 5 22 20 64 5

Burning sensation* 15 2 28 8 17 15 60 5

Dysgeusia 6 12 2 2 17 2 20 6 5 5 44 3

Feeling cold 2 2 5 1 10 3 17 1

*Not otherwise specified

Cut-off: 1% of all patients. n is the total number of patients in the dose group; % is based on n
4.2 Vasovist
47 4
Summary conventions. A total of five serious adverse events were ex-
Adverse events which were recorded as related to the intra- perienced by patients who received 0.03 mmol/kg Vasovist
venous injection of Vasovist in all company-sponsored clin- across all clinical studies using Vasovist. These events were
ical studies to date are available for a total of 1821 subjects. coronary artery disease, hyperglycemia, gangrene, chest
They are grouped in Table 4.3 according to MedDRA [35] pain, and anaphylactoid reaction. Only the anaphylactoid
system organ classes and adapted to the current version of reaction was considered to be probably related to Vasovist
MedDRA terminology, as well as to applicable frequency and was resolved within 5 min.

Table 4.3. Frequency of adverse reactions from clinical trial data. Based on experience of more than 1800 patients, the following un-
desirable effects have been observed and classified by investigators as drug-related (possibly, probably, or definitely). The Table below
reports adverse reactions by MedDRA system organ classes (MedDRA SOCs)

System Organ Class Common Uncommon Rare

(MedDra) (1:100) (1:1000, <1:100) (<1:1000)

Infections and infestations nasopharyngitis cellulites

Immune system disorders hypersensitivity

Metabolism and nutrition hyperglycemia, electrolyte imbalance (incl.

disorders hypocalcemia)

Psychiatric disorders anxiety, confusion

Nervous system disorders headache, par- ageusia, dizziness (excl. vertigo), tremor, hy-
esthesia, dysgeusia, poesthesia, parosmia, muscle contractions
burning sensation involuntary

Eye disorders lacrimation increased, vision abnormal

Ear and labyrinth disorders ear pain

Cardiac disorders atrioventricular block first degree, tachy- cardiac flutter, bradycardia, pal-
cardia pitations

Vascular disorders vasodilatation (incl. hypertension, phlebitis, peripheral coldness hypotension, anaphylactoid
flushing) reaction

Respiratory, thoracic and dyspnea, cough

mediastinal disorders

Gastrointestinal disorders nausea Vomiting, diarrhea, abdominal discomfort,

abdominal pain, dry mouth, flatulence, hy-
poesthesia lips, salivary hypersecretion, dys-
pepsia, pharyngolaryngeal pain, pruritus ani

Skin and subcutaneous tis- pruritus urticaria, erythema, rash, sweating in- swelling face
sue disorders creased

Musculoskeletal, connective muscle cramps, muscle spasms, neck pain,

tissue and bone disorders pain in limb

Renal and urinary disorders hematuria, microalbuminuria, glycosuria micturition urgency

Reproductive system and genital burning sensation, genital pruritus pelvic pain
breast disorders

General disorders and admi- feeling cold pain, chest pain, fatigue, feeling abnormal, pyrexia, rigors
nistration site conditions groin pain, feeling hot, injection site reac-

Investigations electrocardiogram QT prolonged , electro- electrocardiogram ST segment

cardiogram abnormal depression, electrocardiogram
T-wave amplitude decreased,
eosinophil count increased
48 Chapter 4 Summary of Safety of Vasovist at 0.03 mmol/kg Body Weight Dose Clinical Data

4.2.3 Safety in postmarketing surveillance extracellular GBCAs ( Table 4.1). The profile of adverse
events is comparable to that of the established extracellu-
Post-marketing reports of adverse drug reactions from lar Gd-containing contrast agents and these are generally
health-care professionals as well as patients are recorded characterized by mild severity, early onset, and a transient
in a data base at the Pharmacovigilance department of nature with usually rapid resolution.
Bayer Schering Pharma AG (Berlin, Germany). Within One of the advantages of Gd-chelate-based contrast
the first 18 months since the first marketing authoriza- agents over iodinated contrast agents is the lower inci-
tion of Vasovist in October 2005 spontaneous reports dence of anaphylactoid reactions [36, 37].
of adverse drug reactions have been received with a total While Gd-chelates have generally been considered as
4 reporting rate of 0.1%, only three classified as serious. having an excellent safety record in patients with renal
No outcome was fatal, and no unlisted symptoms were impairment, the use of any GBCA today is seen more
reported. The data are consistent with and confirm the critically in patients with severe renal impairment due to
known good safety profile of Vasovist. the recently described disease NSF. NSF is a rare disease
which was first described in the medical literature in
2000, with the first case report going back to 1997 [38];
4.2.4 Special populations a possible association between NSF and GBCAs was first
described in early 2006 [20]. It is primarily characterized
Use in pediatric patients by thickening, induration, and hardening of the skin.
No published data on the safety and efficacy of Vasovist Systemic involvement of organs such as the lung and the
in pediatric patients younger than 18 years are available. heart may also occur. In approximately 5% of patients
the course of the disease is rapidly progressive and it may
potentially lead to a fatal outcome. This disease has been
Use in renally impaired patients reported to date only in patients with severe renal insuf-
The safety and tolerability as well as the pharmokinetic ficiency. The etiology of NSF is likely to be multifactorial;
properties of Vasovist were investigated in patients with various other concomitant symptoms such as metabolic
different degrees of renal impairment (i.e., mild, moder- acidosis or vascular injury [25, 39] are also considered
ate, severe renal impairment, dialysis-dependent end- to be possible co-factors in the pathogenesis of NSF.
stage renal disease) [30]. Histopathology on a skin biopsy specimen is necessary to
The pharmacokinetics of Vasovist were not signifi- make a definitive diagnosis of NSF.
cantly influenced by mild renal impairment. In moderate The prevailing theory regarding gadolinium and
and severe renal impairment the serum elimination half- NSF is that gadolinium (Gd3+) ions are released from
life increased and plasma clearance decreased; the AUC the Gd-chelate complex of MRI contrast agents and
increased for the moderate and severe impairment group accumulate in tissues such as skin, thereby initiating a
by a factor of 2 and 3, respectively. High-flux filters are toxic reaction for which the precise pathomechanism
recommended for the effective elimination of Vasovist in is not yet known. The likelihood that a particular Gd-
hemodialysis patients. Plasma pharmacokinetics and pro- chelate will release Gd3+ ions in vivo depends strongly
tein binding of Vasovist were not significantly influenced on the chelates physicochemical properties (in particu-
by moderate hepatic impairment [30]. lar its chemical structure and complex stability, which
To date there are no reports of NSF in association with may be characterized by distinct parameters including
the sole administration of Vasovist. thermodynamic and conditional stability, excess ligand
in the formulation, dissociation half-life) and is in-
creased in case of reduced elimination of the Gd-chelate
4.2.5 Discussion from the body. Based on their chemical structures,
Gd-containing contrast agents may be divided into two
The data on safety of the blood pool contrast agent Vaso- groups: linear (open-chain) chelates and macrocyclic
vist have been reviewed. chelates. For further details on kinetic stability see else-
Regarding the physicochemical and pharmacokinetic where [31, 40].
properties for Vasovist, clear differences exist from the Vasovist belongs to the group of linear (open-chain)
other currently available GBCAs. Vasovist binds revers- chelates, which are all based on a DTPA backbone. Ionic
ibly to serum albumin, resulting in a prolonged intravas- linear chelates such as Vasovist are generally character-
cular retention time. ized by a much higher complex stability compared with
Despite these characteristic differences of Vasovist, non-ionic linear chelates, which have a higher concentra-
it is tolerated very well by patients overall and exhibits tion of excess ligand included in the formulation in an
a safety profile similar to those of the other marketed effort to rescue unchelated gadolinium (Gd3+) ions in
49 4

1. Claussen C, Laniado M, Schrner W, et al (1985) Gadolinium-DTPA

in MR imaging of glioblastomas and intracranial metastases. Am J
Neuroradiol 6:669674
2. Niendorf HP, Laniado M, Semmler W, Schorner W, Felix R (1987)
Dose administration of gadolinium-DTPA in MR imaging of intrac-
ranial tumors. Am J Neuroradiol 8:803815
3. Ide JM, Port M, Raynal I, Schaefer M, Le Greneur S, Corot C (2006)
Clinical and biological consequences of transmetallation induced
by contrast agents for magnetic resonance imaging: a review.
Fundam Clin Pharmacol 20:563576. Review. Erratum in (2007)
Fundam Clin Pharmacol 21:335
4. Cacheris WP et al (1990) The relationship between thermodynam-
ics and the toxicity of gadolinium complexes. Magn Reson Imag-
ing 8:467481
5. White DH et al (1991) The thermodynamics of complexation
of lanthanide (III) DTPA-bisamide complexes and their implica-
tion for stability and solution structure. Invest Radiol 126 [Suppl
Fig. 4.1. Used amount of contrast media and Gd dose for MR angi- 1]:S226S228
ography of the renal arteries based on a 75-kg patient. Similar first pass 6. Schmitt-Willich H, Brehm M, Ewers CL, et al (1999) Synthesis and
enhancement was considered.*For Multihance MR angiography is an physicochemical characterization of a new gadolinium chelate:
off-label-use. (From Michaely et al. [39]) the liver-specific magnetic resonance imaging contrast agent Gd-
EOB-DTPA. Inorg Chem 38:11341144
7. Uggeri F et al (1995) Novel contrast agents for magnetic resonance
imaging. Synthesis and characterization of the ligand BOPTA and
the solution. For the respective values of charge, ther- its LN(III) complexes (LnGd, La, Lu). X-ray structure of disodium
modynamic stability, conditional stability, and amount (TPS-9-145337286-C-S)-(4-carboxy-5,8,11-tris(carboxamethyl)-l-
of excess-free ligand in the formulation of various linear phenyl-2-oxa-5,8,11-triazatridecan-13-oato(5-)) gadolinite(2-) in a
mixture with its enantiomer. Inorg Chem 34:633642
Gd-chelates including Vasovist, Table 4.1.
8. Caravan P et al (2001) Thermodynamic stability and kinetic inert-
Furthermore, a much lower dose of Gd, 0.03 mmol/ ness of MS-325, a new blood-pool agent for magnetic resonance
kg, is administered for contrast-enhanced magnetic reso- imaging. Inorg Chem 40:21702176
nance angiography (CE-MRA) using Vasovist in compari- 9. Rohrer M, Bauer H, Mintorovitch J, Requardt M, Weinmann H-J
son to CE-MRA with any of the other marketed GBCAs, (2005) Comparison of magnetic properties of MRI contrast me-
dia solutions at different magnetic field strengths. Invest Radiol
for which doses between 0.1 and 0.3 mmol Gd/kg are
administered. This factor needs to be taken into account 10. Shellock FG, Kanal E (1999) Safety of magnetic resonance imaging
when comparing the various GBCAs in terms of safety contrast agents. J Magn Reson Imaging 10:477484
and tolerability ( Fig. 4.1). 11. Knopp MV, Balzer T, Esser M, Kashanian F, Paul P, Niendorf HP
Even though the excretion half life of Vasovist is (2006) Assessment of utilization and pharmacovigilance based
on spontaneous adverse event reporting of gadopentetate
higher than with conventional extracellular MRCM due
dimeglumine as a magnetic resonance contrast agent after 45
to the albumin-binding properties, it is currently unclear million administrations and 15 years of clinical use. Invest Radiol
if and how this could be related to the pathophysiology 41:491499
of NSF. Overall, Vasovist should be used only after care- 12. Prince MR, Erel HE, Lent RW, et al (2003) Gadodiamide administra-
ful risk/benefit assessment in patients with severe renal tion causes spurious hypocalcemia. Radiology 227:639646
13. Choyke PL, Knopp MV (2003) Pseudohypocalcemia with MR imag-
impairment (GFR <30 ml/min/1.73m2), including those
ing contrast agents: a cautionary tale. Radiology 227:627628
with an acute renal insufficiency of any severity due to 14. Emerson J, Kost G (2004) Spurious hypocalcemia after Omniscan-
the hepatorenal syndrome or in the perioperative liver or Optimark-enhanced magnetic resonance imaging. Arch Pathol
transplantation period. Lab Med 128:11511156

Take home messages

Vasovist has been demonstrated to be well toler- Vasovist shows a profile of adverse events com-
ated by all patients investigated during clinical parable to those of other extracellular Gd-based
studies. MR contrast media (GBCAs), characterized overall
Vasovist can effectively be removed from the by events of mild severity, early onset, and a tran-
body with high flux dialysis if necessary. sient nature with rapid resolution.
50 Chapter 4 Summary of Safety of Vasovist at 0.03 mmol/kg Body Weight Dose Clinical Data

15. Moore CD, Newman RC, Caridi JG(2006) Spurious hypocalcemia cal Manufacturers and Associations (IFPMA) (2007). Available at:
after gadodiamide-enhanced magnetic resonance imaging: A
case report and review of the literature. Rev Urol 8:165168 36. Nelson KL, Gifford LM, Lauber-Huber C, et al (1995) Clinical safety
16. Zhang HL, Ersoy H, Prince MR (2006) Effects of gadopentetate of gadopentetate dimeglumine. Radiology 196:439443
dimeglumine and gadodiamide on serum calcium, magnesium 37. Hagan JB (2004) Anaphylactoid and adverse reactions to radio-
and creatinine measurements. J Magn Res Imaging 23:383387 contrast agents. Immunol Allergy Clin N Am 24 507519
17. Lwe A, Balzer T, Hirt U (2005) Interference of gadolinium-contain- 38. Cowper SE, Robins HS, Steinberg HM, Su LhD, Gupta S, Leboit PE
ing contrast-enhancing agents with colorimetric calcium labora- (2000) Scleromyxedema-like cutaneous diseases in renal-dialysis
tory testing. Invest Radiol 40:521525 patients. Lancet 356:10001001
18. Caravan P, Ellison JJ, McMurry TJ, Lauffer RB (1999) Gadolinium(III) 39. Michaely HJ, Thomsen HS, Reiser MF, Schoenberg SO (2007) Ne-
chelates as MRI contrast agents: structure, dynamics, and applica- phrogenic systemic fibrosis (NSF) implications for radiology.
4 tions. Chem Rev 99:22932352
19. Kang HP, Scott MG, Joe BN, Narra V, Heiken J, Parvin CA (2004) 40.
Radiologe 47:785793
Schmitt-Willich H (2007) Stability of linear and macrocyclic gado-
Model for predicting the impact of gadolinium on plasma cal- linium based contrast agents. Br J Radiol 80:581582; author reply
cium measured by the o-cresolphthalein method. Clin Chem 584585
20. Grobner T (2006) Gadolinium a specific trigger for the develop-
ment of nephrogenic fibrosing dermopathy and nephrogenic sys-
temic fibrosis? Nephrol Dial Transplant 11 (epub ahead of print)
21. Marckmann P, Skov L, Rossen K, et al (2006) Nephrogenic systemic
fibrosis: suspected causative role of gadodiamide used for con-
trast-enhanced magnetic resonance imaging. J Am Soc Nephrol
22. Maloo M, Abt P, Kashyap R, et al (2006) Nephrogenic systemic
fibrosis among liver transplant recipients: a single institution ex-
perience and topic update. Am J Transplan 6:2212
23. Thomsen HS, Morcos SK, Dawson P (2006) Is there a causal relation
between the administration of gadolinium-based contrast media
and the development of nephrogenic systemic fibrosis (NSF)? Clin
Radiol 61:905906
24. Deo A, Fogel M, Cowper SE (2007) Nephrogenic systemic fibrosis:
a population study examining the relationship of disease develop-
ment to gadolinium exposure. Clin J Am Soc Nephrol 2:264267
25. Broome DR, Girguis MS, Baron PW, Cottrell AC, Kjellin I, Kirk GA
(2007) Gadodiamide-associated nephrogenic systemic fibrosis:
why radiologists should be concerned. AJR Am J Roentgenol
26. Khurana A, Runge VM, Narayanan M, Greene JF Jr, Nickel AE
(2007) Nephrogenic systemic fibrosis: a review of 6 cases tempo-
rally related to gadodiamide injection (omniscan). Invest Radiol
27. Sadowski EA, et al (2007) Nephrogenic systemic fibrosis : risk fac-
tors and incidence estimation. Radiology 243:148157
28. Goyen M, Edelman M, Perreault P, et al (2005) MR angiography
of aortoiliac occlusive disease: a phase III study of the safety and
effectiveness of the blood-pool contrast agent MS-325. Radiology
29. Rapp JH, Wolff SD, Quinn SF, et al (2005) Aortoiliac occlusive disease
in patients with known or suspected peripheral vascular disease:
safety and efficacy of Gadofosveset-enhanced MR angiography
multicenter comparative phase III study. Radiology 236:7178
30. European Medicines Agency (2006) Vasovist: summary of product
characteristics. [online]. Available from URL: http://www.emea.
31. European Medicines Agency. Vasovist: Scientific discussion. [on-
line] (2007). Available from URL:
32. Henness S, Keating GM. (2006) Gadofosveset. Drugs 66:851-857
33. Shamsi K, Yucel EK, Chamberlin P (2006) A summary of safety of
Gadofosveset (MS-325) at 0.03 mmol Gd/kg body weight dose.
Invest Radiol 41:822830
34. [online] (2007). Available from URL http://
35. The Medical Dictionary for Regulatory Activities; MedDRA. Reg-
istered trademark of the International Federation of Pharmaceuti-

Efficacy of Vasovist: Overview of the Clinical

Development Program
Mathias Goyen

5.1 Introduction 52

5.2 Phase II Dose-Finding Studies 53

5.3 Phase III Trials 53

5.3.1 Iliac Arteries 54
5.3.2 Leg Arteries 54
5.3.3 Renal Arteries 54
5.3.4 Foot Arteries 54
5.3.5 Stenosis Quantification 56

5.4 Efficacy of Vasovist at Low Albumin Levels 56

References 56
52 Chapter 5 Efficacy of Vasovist: Overview of the Clinical Development Program

5.1 Introduction not of diagnostic or therapeutic benefit to the participants

and thus financially compensated most of the time, are
Prior to approval Vasovist (Gadofosveset, development primarily for collecting tolerance and pharmacology data
name MS-325) was subjected to extensive testing to evalu- in healthy volunteers.
ate the safety and efficacy of the drug. Since the majority Once the initial safety of the study drug has been con-
of the clinical trials were carried out in the USA, the firmed in phase I trials, phase II trials are performed on
particular circumstances of the US-approval process and larger groups (20300 persons) and are designed to assess
the discussions with the health authorities, the Food and how well the drug works, as well as to continue phase I
Drug Administration (FDA), are reflected in the study safety assessments in a larger group of volunteers and pa-
design and results. tients. When the development process for a new drug fails,
Vasovist was developed primarily for use with MR this usually occurs during phase II trials, when the drug is
angiography. Unlike in Europe, where, for example, the discovered not to work as planned or to have toxic effects.
5 unspecific extracellular MR contrast agent Gadovist (Bayer Phase III studies are randomized, controlled, multi-
Schering Pharma AG, Berlin, Germany) is specifically ap- center trials with large patient groups (3003000 or more,
proved for MR angiography, there is presently no MR con- depending upon the disease/medical condition studied)
trast agent on the US market which is approved for MR an- and are aimed at being the definitive assessment of how
giography. Therefore, the trial design of all the clinical trials effective the drug is, in comparison to current gold stand-
conducted does not include a comparison to other MR con- ard treatment. In order to be considered for approval, all
trast agents. Scientifically validated from the perspective of trials must be carried out in exact accordance with good
the regulatory authorities are only non-contrast-enhanced clinical practice (GCP) regulations.
MRA methods [for example, time-of-flight (TOF) MRA]. As part of the clinical development and approval of
Therefore, all approval trials are based on the comparison Vasovist a total of 18 trials were carried out. A total of
between Gadofosveset-enhanced MRA and non-contrast- 1524 patients and volunteers and patients participated in
enhanced MRA. Digital subtraction angiography (DSA), these trials on a voluntary basis.
a methodology viewed as providing either true or 100% Specifically, the following trials were performed:
correct results, serves as the gold standard for all trials per- Eight clinical pharmacological phase I and phase II
formed. All MRI images and DSA results were analyzed by trials
standardized, centralized, independent, blinded readings. Two phase II dose-finding studies on patients with
The clinical development of Vasovist can be divided vascular disease
into three phases. In phase I, the substance is adminis- Four phase III trials ( Table 5.1) and
tered to healthy volunteers or patients for whom the Four additional trials whose results only were used for
diagnostic agent is not indicated. These trials, normally assessing the tolerance to Vasovist

Table 5.1. Phase III trials

Study no. Study title MS-325 dose No. of patients

(mmol/kgBW) (ITT population)

MS-325-12 Multicenter, comparative, phase III study 0.03 268

Safety and Efficacy of Gadofosveset-Enhanced MRA for Evalu-
ation of Aortoiliac Occlusive Disease in Patients with Known or
Suspected Peripheral Vascular Disease
Turbulent flow

MS-325-13 Multicenter, phase III study 0.03 174

Safety and Efficacy of Gadofosveset-Enhanced MRA in Patients
with Suspected Peripheral Vascular Disease
Turbulent flow

MS-325-14 Phase III multicenter study 0.03 136

Safety and Efficacy of MS-325-enhanced MRA in Patients with
Known or Suspected Renal Arterial Disease
Flow to an organ

MS-325-15 Multicenter, randomized, open-label, two-dose, comparative, 0.03 93

phase III study 0.05 87
Safety and Efficacy of MS-325-enhanced MRA in Patients with
Known or Suspected Pedal Arterial Disease
Slow flow
5.3 Phase III Trials
53 5

Of these volunteers and patients, 1445 received Vasovist unchanged, and patients with identified arterial occlusive
as an injection. The remaining 79 study participants disease of the pelvic and lower extremity vessels received
received instead an injection of isotonic saline solution doses of 0.005, 0.01, 0.03, 0.05, and 0.07 mmol/kg body
(NaCl). Administration of NaCl (as a so-called placebo) weight. This study also included comparing the variables
is an integral and legally mandated part of early (phase with the DSA results. A ROC (receiver operating curve)
I and II) clinical development and is intended to reduce analysis, which takes into consideration the examiner
methodical errors during evaluation of results concerning confidence of stenosis diagnosis, was used as a statisti-
tolerance and efficacy (e.g., improvement of results after cal test. The study revealed that (a) all doses were clearly
contrast agent administration is not due to the substance diagnostically superior to administered placebo, and (b) a
itself but due to the post-contrast sequence and therefore dose of 0.03 mmol/kg body weight is required for mini-
would also be observed after NaCl has been given). mizing the risk of non-diagnostic MRA examination.
In the following the two dose-finding studies as well A consideration of all results pertaining to Vasovist
as the four phase III trials (Controlled Clinical Studies tolerability and efficacy in patients with clinically sus-
Pertinent to Claimed Indications) are discussed. pected arterial occlusive disease of the pelvic and lower
extremity vessels showed that a dosage of 0.03 mmol/
kg body weight was the most suitable. Further dose
5.2 Phase II Dose-Finding Studies increases did not coincide with statistically significant
increases in diagnostic accuracy, and therefore would
The aim of the study of patients suffering from a disease expose the patient to unnecessarily high contrast agent
for which the substance is planned to be indicated in the concentrations.
future is to collect additional data regarding substance tol- In addition, an evaluation of both the clinical out-
erability, to present the clinical efficacy of the substance, come parameters and the technical parameters (contrast-
and to identify the optimal dose. noise-ratio; signal-noise-ratio) for the area of a small
As part of the clinical development of the contrast distal artery (the ascending branch of the lateral cir-
agent Vasovist, two independent dose-finding studies cumflex femoral artery, average diameter ~2 mm) was
of 346 patients were carried out. 309 of these patients performed. In this vessel, a dose of 0.05 mmol/kg body
received the contrast agent at various doses and 37 pa- weight showed improved efficacy compared with a dose
tients a volume-equivalent injection of normal saline as of 0.03 mmol/kg body weight. Thus, a phase III trial
placebo. to examine the efficacy and tolerability of Vasovist in
In order to minimize the influences of a preselected patients with vascular pathology and reduced flow in
patient population and of the dependence on specific peripheral foot arteries also included a second dose of
instrument types or settings, each trial was conducted as 0.05 mmol/kg body weight.
a multicenter study. As clinically dependent variables, the
sensitivity, the specificity, and the overall diagnostic ac-
curacy of Vasovist at doses of 0.01, 0.03, and 0.05 mmol/ 5.3 Phase III Trials
kg body weight for detecting arterial vascular stenosis or
occlusion were investigated. The MR angiographies were After consultation with the US Food and Drug Admin-
performed in patients with evidence of corresponding istration (FDA), the indication was expanded in phase
vascular pathologies in the area of the carotid artery [1,2] III of clinical development to now include patients with
or the pelvic and leg arteries. DSA gold standard results vascular pathologies involving the pelvic, leg, renal, and
were available for these patients. foot arteries in the studies ( Table 5.1).
In order to perform the analysis of the diagnostic Analogous to the previously performed phase II stud-
validity of Vasovist in MR angiography of carotid, pelvic, ies, a DSA carried out 30 days before or after MRA was re-
and leg arteries as objectively as possible, the images were quired for all phase III trials and was used as the standard
evaluated both in the study centers themselves and by a of reference. Primary inclusion criterion was that patients
centralized, standardized, blinded reading. For all three had clinically significant arterial stenosis that reduced the
administered doses, the images of patients with vascular vessel diameter by 50%. To minimize the risk of a poten-
pathologies involving the carotid (n=26), iliac (n=31), tial bias on study analysis, MRA and DSA gold standard
and femoral arteries (n=16) showed a good correlation assessment, each of the two trials were evaluated sepa-
with DSA. However, only for patients with pathologies rately via central blinded reading by three (MRA) or two
in the iliac arteries was a dependency between diagnostic (DSA) experienced radiologists. Evaluation of DSA was
accuracy and administered dose observed. performed according to the consensus method. If there
Thus, a second study was carried out using a larger was a discrepancy in the assessment between the two ra-
number of patients [3]. The dependent variables remained diologists, the diagnosis of a third experienced radiologist
54 Chapter 5 Efficacy of Vasovist: Overview of the Clinical Development Program

was obtained. The MRA findings were presented both as 5.3.2 Leg Arteries
first pass and steady state data sets.
The four phase III studies corresponded to the vari- The study design for leg arteries was nearly identical and
ous vascular regions. The dependent variables, namely included 174 patients in 17 study centers [5]. In this study
sensitivity and specificity, as well as the overall diagnostic as well, all three radiologists found a significant increase
accuracy of Vasovist, were investigated in the following in the sensitivity (between 21.9% and 30.8%), specificity
patient group: (between 8.5% and 11.9%), and diagnostic accuracy (be-
Vascular pathology of the large iliac vessels, typically tween 10.5% and 13.1%) ( Table 5.2). Also, the number
representing vessels with turbulent flow of inaccessible vessel segments after Vasovist application
Vascular pathology of the leg arteries was clearly lower compared with TOF MRA ( Table 5.3).
Vascular pathology of the renal arteries (representing In addition, the study showed that after Gadofosveset
organ flow) application the evaluators are more confident in making
5 Vascular pathology of the foot arteries, typically repre- the diagnosis.
senting vessels with slow flow This study included not only patients with diagnosed
stenosis (>50) but also 14 patients with diagnosed ab-
dominal aortic aneurysm as well as pelvic and femoral
5.3.1 Iliac Arteries aneurysm. Nearly all aneurysms were accurately detected
after Vasovist administration ( Table 5.2), correspond-
A total of 274 patients from 37 study centers were in- ing to a specificity of >93%.
cluded in this study [4]. The objective of the study was
to determine the statistical superiority of Gadofosveset-
enhanced MRA at a dose of 0.03 mmol/kg body weight 5.3.3 Renal Arteries
compared with non-contrast-enhanced 2D TOF MRA
for detecting clinically significant stenosis (>50%) in the Also in this study, which included 145 patients with sus-
pelvic area. pected significant stenosis of the proximal/distal renal ar-
The blinded reading revealed that the Gadofosveset- tery or an accessory renal artery, Gadofosveset-enhanced
enhanced images correlated significantly better with DSA MRA showed a significant superiority in diagnostic ac-
findings than did the TOF sequences. On average, the curacy.
evaluating radiologists achieved an increase in sensitivity In this study as well, all three radiologists found a
of 14.5%, in specificity of 12.6%, and in diagnostic accu- significant increase in the sensitivity (between 24.5% and
racy of 12.8%. For all three radiologists both the specificity 41.5%), specificity (between 22.7% and 29.3%), and diag-
and the diagnostic accuracy were significantly superior. nostic accuracy (between 23.0% and 28.7%) ( Table 5.2).
All three evaluators also witnessed significantly improved
sensitivity; however, only two of the three radiologists
found these increases to be significant ( Table 5.2). 5.3.4 Foot Arteries
A secondary approach also included evaluating the
non-diagnostic data sets. Here, Gadofosveset-enhanced This phase III trial of 185 patients with suspected stenosis
MRA showed significant gains: While with TOF MRA or occlusion of foot arteries revealed several differences
4.7%, 17.6%, and 21.9% of examinations were classified as compared with the studies on renal, pelvic, or leg arteries.
inconclusive, the rates after Gadofosveset application were Analysis of quantitative signal parameters, in particular
only 0.4%, 0.4%, and 1.2%, respectively. Since all phase the signal-to-noise ratio (SNR), which were representa-
III approval studies consisted of showing first pass and tively assessed in the ascending branch of the lateral
steady state images after Vasovist injection, always at the femoral circumflex artery compared with thigh muscles,
same time, it is not measurable which image in particular revealed a statistically significant increase when a higher
was responsible for the reduction in this unsatisfactory dose of 0.05 mmol/kg body weight was used compared
examination outcome. However, it can be assumed that with 0.03 mmol/ kg body weight. However, the primary
the very fact of Vasovists expanded possibilities, namely, outcome parameters of the study, namely sensitivity, spe-
the acquisition of first pass and steady state data sets, has cificity, and diagnostic accuracy, showed no difference
made this significant improvement possible. When con- between the two doses.
ventional, unspecific, extracellular MRI contrast agents In addition, the study confirmed the known limita-
are used, only first pass MRA images can be acquired, tions [6] of TOF angiography in vessels with slow flow.
consequently reducing the robustness of the examination The very slow flow velocity in the range of a few centim-
results if timing problems, injection problems, or move- eters per second, leads to signal losses that can contribute
ments during first pass data acquisition occur. to misdiagnosis of vascular occlusion and stenosis. For ex-
5.3 Phase III Trials
55 5

Table 5.2. Sensitivity, specificity and diagnostic accuracy as the primary variable of all phase III trials

Trial (vessel type) TOF MRA (in %) Vasovist-enhanced MRA Difference (in %)
(in %)

Sensiti- Specifi- Diagnostic Sensi- Speci- Diagnostic Sensiti- Specificity Diagnostic

vity city accuracy tivity ficity accuracy vity accuracy

Iliac artery
(n=268 patients)

Reader A 62.0 75.1 73.2 80.2 84.5 83.8 18.1*** 9.4*** 10.6***

Reader B 66.7 84.8 82.2 73.0 93.2 90.3 6.3 n.s. 8.4*** 8.1***

Reader C 41.8 75.4 70.6 60.8 95.3 90.3 19.0*** 19.9*** 19.7***

Femoral artery
(n=175 patients)

Reader A 52.1 70.7 68.4 82.9 80.0 80.3 30.8*** 9.2*** 11.9***

Reader B 60.3 74.5 72.7 84.2 83.0 83.2 24.0*** 8.5*** 10.5***

Reader C 48.6 78.2 74.5 70.5 90.1 87.6 21.9*** 11.9*** 13.1***

Renal artery
(n=136 patients)

Reader A 30.2 48.0 44.7 56.6 77.3 73.4 26.4** 29.3*** 28.7***

Reader B 41.5 59.0 55.7 66.0 81.7 78.7 24.5** 22.7*** 23.0***

Reader C 22.6 57.2 50.7 64.2 82.5 79.1 41.5*** 25.3*** 28.4***

Foot arteries
(n=93 patients****)

Reader A 77.0 38.3 63.0 93.0 59.5 80.7 16.0*** 20.7* 17.7***

Reader B 86.5 31.9 66.5 77.5 66.4 73.4 (-9.0) n.s. 34.5*** 7.0 n.s.

Reader C 78.0 28.4 59.8 78.5 62.9 72.8 0.5 n.s. 34.5*** 13.0**

**** Number of patients evaluated for efficiency who received a dose of 0.03 mmol/kg body weight
n.s. Not significant

Table 5.3. Number of inaccessible vessel segments / patients with inaccessible vessel segments

Trial Blinded reader TOF MRA Contrast-enhanced MRA after

injection of Vasovist

Pelvic arteries [1], n=274 patients A 45/256 (17.6%) 3/256 (1.2%)

B 12/256 (4.7%) 1/256 (0.4%)

C 56/256 (21.9%) 1/256 (0.4%)

Leg arteries [2], n=174 patients A 235/1206 (19.5%) 30/1206 (2.5%)

B 150/1206 (12.4%) 21/1206 (1.7%)

C 197/1206 (16.3%) 31/1206 (2.6%)

56 Chapter 5 Efficacy of Vasovist: Overview of the Clinical Development Program

ample, the average specificity across the precontrast TOF Take home messages
sequence was only around 31% and could be increased Data from all the phase III studies showed an ex-
after Gadofosveset application to between 59.4 and 66.4%. cellent overall efficacy compared with DSA as a
Sensitivity was also increased by Vasovist application clinical standard.
( Table 5.3). However, clinically this value is of only mi- Due to significantly higher relaxivity compared
nor significance, since for patients with an indication for with extracellular agents, Vasovist provides much
foot artery angiography the question of the possibility of stronger (brighter) signal, thereby improving the
extremity-saving bypass surgery, and therefore primarily signal to-noise and contrast-to-noise ratios.
a high degree of accuracy with unequivocal exclusion Individual MRA readers had a mean accuracy for
(true negatives, which are factored into the calculation of Gadofosveset-enhanced MRA of 88%, while the
specificity, Table 5.2) of significant vascular stenosis is mean accuracy for non-contrast MRA was 75%.
the primary focus. Only vessels appearing non-stenosed In addition, it was shown that the proportion of
5 in angiography, with findings reflecting reality, are suit- non-diagnostic angiographies after Vasovist ad-
able as connecting vessels for bypass. ministration falls to below 1%; this is clearly lower
Contrary to the other phase III trials, the study than in TOF MRA (~14%) and even below DSA
performed on 156 patients with suspected stenosis/oc- (~8%).
clusion of the foot arteries showed for both non-contrast-
enhanced TOF MRA and Gadofosveset-enhanced MRA
unusually high sensitivity values in the detection of References
vascular stenosis. A statistically significant sensitivity
increase due to Vasovist administration was shown for 1. Grist TM, Korosec FR, Peters DC, Witte S, Walovitch RC, Dolan RP,
only one reader. However, for all three blinded readers, Bridson WE, Yucel EK, Mistretta CA (1998) Steady state and dy-
namic MR angiography with MS-325: initial experience in humans.
the specificity was statistically significantly superior after Radiology 207:539544
Vasovist administration. The diagnostic accuracy in- 2. Bluemke DA, Stillman AE, Bis KG, Grist TM, Baum RA, DAgostino
creased with the administration of Vasovist for all three R, Malden ES, Pierro JA, Yucel EK (2001) Carotid MR angiogra-
readers and was statistically significant for two of the phy: phase II study of safety and efficacy for MS-325. Radiology
three ( Table 5.3). 219:114122
3. Perreault P, Edelman MA, Baum RA, Yucel EK, Weisskoff RM, Shamsi
K, Mohler ER 3rd (2003) MR angiography with gadofosveset triso-
dium for peripheral vascular disease: phase II trial. Radiology
5.3.5 Stenosis Quantification 229:811820
4. Rapp JH, Wolff SD, Quinn SF, Soto JA, Meranze SG, Muluk S, Ble-
All four phase III trials not only examined the accuracy of bea J, Johnson SP, Rofsky NM, Duerinckx A, Foster GS, Kent KC,
Moneta G, Middlebrook MR, Narra VR, Toombs BD, Pollak J, Yucel
detection of significant vascular stenosis compared with EK, Shamsi K, Weisskoff RM (2005) Aortoiliac occlusive disease in
DSA, but also included quantification (in % stenosis) by patients with known or suspected peripheral vascular disease:
measurement. These studies revealed that Gadofosveset- safety and efficacy of gadofosveset-enhanced MR angiography
enhanced MRA not only far surpasses the accuracy of multicenter comparative phase III study. Radiology 236:7178
TOF angiography but also delivers results comparable to 5. Goyen M, Edelman M, Perreault P, ORiordan E, Bertoni H, Taylor
J, Siragusa D, Sharafuddin M, Mohler ER 3rd, Breger R, Yucel EK,
those of DSA. Shamsi K, Weisskoff RM (2005) MR angiography of aortoiliac oc-
clusive disease: a phase III study of the safety and effectiveness of
the blood-pool contrast agent MS-325. Radiology 236:825833
5.4 Efficacy of Vasovist at Low Albumin 6. Pipe J (2001) Limits of time-of-flight magnetic resonance angiog-
Levels raphy. Top Magn Reson Imaging 2:163174

The contrast behavior of Vasovist is critically dependent

on its interaction with albumin. Thus, an analysis across
several studies examined whether the diagnostic accuracy
of MRA after administration of Vasovist decreases in
patients with low serum albumin levels. The unchanged
values show that even in patients with pathologically low
albumin levels a sufficient number of albumin molecules
are still available for ensuring satisfactory efficacy of

Part III Blood Pool Agents in MRA:

Indications, Clinical Applications
and Benefits

Chapter 6 Head and Neck MRA 59

Marco Essig and Frederik Giesel

Chapter 7 Pulmonary MRA 69

Christian Fink, Ulrike Attenberger, and Konstantin Nikolaou

Chapter 8 Abdominal MRA 81

Henrik J. Michaely

Chapter 9 MRA of the Peripheral Vasculature 93

Tim Leiner and Jeffrey H. Maki

Chapter 10 Magnetic Resonance Venography 115

Giles Roditi

Chapter 11 Whole-body MRA 131

Harald Kramer, Maximilian F. Reiser, and Konstantin Nikolaou

Chapter 12 Endoleak Imaging 139

Sandra A.P. Cornelissen, Mathias Prokop, Hence J.M. Verhagen,
and Lambertus W. Bartels

Chapter 13 Gastrointestinal Bleeding 147

Joachim Lotz

Head and Neck MRA

Marco Essig and Frederik Giesel

6.1 Introduction 60

6.2 Traditional MRA Techniques for Imaging the Head and Neck Vasculature:
TOF and PC MRA 61

6.3 CE-MRA: Benefit over non-contrast techniques 61

6.4 Technical Challenges Facing CE-MRA 61

6.5 Accuracy of Contrast-enhanced MRA in the Assessment

of the Supra-aortic Vasculature 61

6.6 Protocols for Supra-aortic Vessel Visualization 62

6.7 Benefits of the Blood pool MR Contrast Agent Vasovist 62

6.8 Conclusions 63

References 67
60 Chapter 6 Head and Neck MRA

6.1 Introduction angiography using the DSA technique and over computed
tomography (CT) angiography.
Vascular disease, mainly atherosclerosis, is a major cause First, the method does not require ionizing radia-
of death and disability in the Western world. As a gen- tion and is much less invasive, especially compared with
eralized process it involves nearly all vascular territories, conventional angiography. As a result, the level of patient
and the incidence increases with age. Responsible for safety is higher. MRA can be performed on an outpatient
different severe clinical findings, it is also the major basis and requires little preparation. It offers a more fa-
cause of cerebrovascular disease. In the CNS it is the vorable cost-benefit ratio, especially with regard to the
main cause of stroke, which is fatal in about 40% of cases much lower complication rates. The latter particularly
[1]. However, every year approximately 20 million peo- applies to both brain-supplying vessel territories the ca-
ple worldwide survive stroke, making it a major cause of rotid artery and the vertebral arteries because compli-
disability as well. Many of the stroke survivors are left cations with these vessels can frequently lead to cerebral
with significant permanent disability [2] that requires symptoms, often with extensive subsequent neurological
costly clinical care [3]. Although ischemia accounts for deficits [14]. The diagnostic performance of MRA of
80% of all strokes, in some instances it is caused by in- the cervical vessels is comparable to that of CTA and
6 tracranial hemorrhage, usually secondary to rupture of conventional techniques, which is why MR angiography
an intracranial aneurysm or vascular malformations [4]. should be selected as the primary diagnostic modality for
Atheromatous debris from focal atherosclerotic plaque, diagnostic workup prior to treatment [15] or for staging
usually at the carotid bifurcation, is the most common atherosclerosis.
identifiable cause of carotid artery stenosis [5], present Most of the MR angiographic techniques available
in about 25% of all strokes [6]. In patients with sig- can also be used to depict brain-supplying vessels. These
nificant stenoses (i.e., >70% luminal narrowing) there include 2D and 3D time-of-flight (TOF) techniques,
are a variety of therapeutic options, including carotid 2D and 3D phase-contrast techniques, and contrast-en-
endarterectomy or carotid stenting [79]. Both proce- hanced MR angiography. By comparison with the two
dures have a risk of 510%, and in order to stratify pa- former techniques, contrast -enhanced MR angiography
tients into those requiring surgery, a precise diagnostic has considerable advantages, especially with regard to the
workup before initiating treatment is essential to clearly speed of examination, anatomical coverage, and vascular
define the degree of luminal compromise as, for mild-to- contrast.
moderate stenoses, the risks associated with intervention Magnetic resonance imaging in combination with
outweigh the possible benefits [10]. MRA also offers the possibility of visualizing not only the
Because of the clearly defined threshold (70%) for vascular lumen alone, but also the vascular wall, which
surgical versus non-surgical treatment and the fact that provides information about the pathophysiology of the
the internal carotid artery is relatively small (typically disease [16].
58 mm), identification of a clinically significant inter- Imaging of the intracerebral vasculature has long
nal carotid artery stenosis requires high spatial resolu- been a domain of MR angiography [17, 18]. Due to its
tion imaging, of both the vessel lumen and the vessel non-invasiveness and excellent ease of use, the compet-
wall. ing procedures such as conventional digital subtraction
Until recently, conventional digital subtraction X-ray angiography (DSA) and CTA are indicated only for
angiography (XRA) was the standard method for depict- certain medical questions and acute diagnosis. Another
ing the location and extent of vascular disease. However, advantage of MRI over both procedures is the limited
the method requires arterial catheterization and the need for the use of contrast agents. Another key advan-
use of substantial volumes of iodinated contrast agent. tage of MR angiography is the ability to integrate the
Today, less invasive imaging techniques using computed technique into a conventional MRI examination of the
tomography (CT) or magnetic resonance imaging (MRI) brain and combine it with other, so-called functional
have been developed. These imaging methods, either MRI methods such as MR perfusion or MR diffusion.
with the use of contrast agents (iodine-based for CT Especially the combination with perfusion measure-
and gadolinium-based for MRI) or without exogenous ments and diffusion-weighted MRI makes it possible to
contrast agents (MRI only) are increasingly being used arrive at a more precise diagnosis of the various vascular
and their sensitivity and accuracy have reached the level lesions.
of XRA. In this chapter, we discuss recent advances in MRA
Because of its non-invasiveness and the lack of radia- for visualization of the head and neck vasculature, includ-
tion, MRA has become the method of choice for diagnos- ing use of the new intravascular contrast agent Vasovist
ing diseases of the supra-aortic vascular system [1113]. (Gadofosveset, Bayer Schering Pharma AG, Berlin, Ger-
MR angiography has many advantages over conventional many).
6.5 Accuracy of Contrast-enhanced MRA in the Assessment of the Supra-aortic Vasculature
61 6
6.2 Traditional MRA Techniques for to the patient, thus replacing the previous gold standard
Imaging the Head and Neck Vasculature: of invasive angiography for evaluation of the head and
TOF and PC MRA neck vessels.

The main benefit of time-of-flight (TOF) MRA is that it

exploits the well-understood phenomenon of inflow as 6.4 Technical Challenges Facing CE-MRA
the basis of vessel contrast. As it does not require intra-
venous contrast agents, it truly represents a non-invasive Despite numerous advances, such as parallel imaging
technique. However, disadvantages include long acquisi- and higher field strengths [24], conventional contrast-
tion time, flow-related artifacts beyond stenoses, false- enhanced MRA is still beset with some technical prob-
positive occlusions, susceptibility-induced signal loss due lems that restrict its effectiveness. For example, as the
to the presence of dense bone at the skull base, and arteriovenous contrast agent transit time for the carotid
misregistration arteiacts secondary to patient motion and vessels is short, making bolus timing more critical com-
swallowing [19]. Despite these limitations, many studies pared with other areas. Time-resolved MRA can reduce
have documented the excellent sensitivity and specificity the need for precise timing of the bolus, but this is
of this technique for demonstrating carotid artery stenosis possible only with a short acquisition time, which com-
at the 70% threshold [20, 21]; however, identification of promises spatial resolution and causes blurring in the
ulceration is poor [20]. slice-encoding direction [25]. Although a test bolus can
Vessel contrast for phase-contrast MRA relies on de- be used to determine the time delay between injection
tection of changes (phase shifts) in the phase of blood as it and arrival of the contrast agent bolus in the target vessel
moves along a magnetic field gradient. The bipolar gradi- [26], automatic contrast bolus arrival can be detected us-
ent employed gives a zero phase shift for stationary spins ing dedicated software (e.g., SmartPrep; General Electric
(background tissues), which are therefore completely sup- Healthcare, Milwaukee, WI, USA) [27], or preferably
pressed, but a non zero phase shift for moving (blood) with magnetic resonance fluoroscopy (e.g., Bolus Trak;
protons, proportional to velocity, resulting in bright ves- Philips Medical Systems, Best, The Netherlands) [28].
sels. Arteries or veins can be selectively highlighted by tai- Image resolution may also be reduced if the carotid and
loring the velocity-encoding gradient (Venc) to the range intracranial vessels are to be viewed in a single proce-
of velocities expected within the vessel, therefore selective dure because of the larger field-of-view required in the
arteriography and venography is possible. Additionally, slice-select direction.
blood velocity and flow rates can be calculated using a 2D
cardiac-gated approach. Although PCA is highly effective
for evaluation of intracranial venous sinus patency, it is 6.5 Accuracy of Contrast-enhanced MRA
no longer routinely used for evaluation of intracranial in the Assessment of the Supra-aortic
arteries [1718, 22]. Vasculature

A large number of studies have examined the accuracy

6.3 CE-MRA: Benefit over non-contrast of CE-MRA for the identification of carotid artery steno-
techniques sis, reporting variable results. However, there have been
few direct comparisons with DSA. In a meta-analysis of
Since its introduction in the late 1990s, contrast-enhanced studies reporting on non-enhanced MRA techniques an
(CE) MRA has replaced non-contrast MRA because it increasing role was described for MRA on the basis of its
is more time efficient, yields better vessel contrast, and high sensitivity and specificity in the detection of high-
higher resolution, and is associated with substantially grade stenoses (>70%). [29, 30]. For CE-MRA different
fewer artifacts [23]. The basis of CE-MRA is shortening of authors found a better discriminatory power compared
the blood T1 by injecting a sufficient volume of contrast with duplex ultrasonography in diagnosing 7099% ste-
agent to lower the blood T1 to less than that of fat (usually nosis (i.e., differentiating critical stenosis from occlusion)
the brightest structure on the image), so that the contrast- and reported that MRA had a pooled sensitivity of 95%
enriched vessels can be extracted by the MIP algorithm and a pooled specificity of 90%, compared with respective
and reprojected onto the final angiographic type image. values of 86 and 87% for duplex ultrasonography [31].
Advances in hardware, software, and contrast agents for For detecting carotid artery occlusion, MRA was found to
contrast-enhanced MRA have ensured that it now gener- have the same sensitivity and specificity as conventional
ates images of the head and neck vessels with high spatial X-ray angiography.
resolution, high temporal resolution if necessary, and Although diagnostic workup algorithms for carotid
excellent vascular contrast with essentially minimal risk stenosis clearly focus on atherosclerotic occlusive disease
62 Chapter 6 Head and Neck MRA

of the carotid bifurcation, CE-MRA also yields useful 6.7 Benefits of the Blood pool MR Contrast
supplemental information about such conditions as coex- Agent Vasovist
istent infarction, ischemia, intracranial aneurysms [32],
and arteriovenous malformations [33], as well as arterio- Vasovist is the first intravascular contrast agent approved
venous fistulae. Crucially, and in addition, although CE- for use with MRA in the European Union, Switzerland,
MRA depicts only the lumen accurately, non-enhanced, Turkey, Canada, and Australia. The agent reversibly binds
mainly fat-suppressed T1- and T2-weighted imaging to albumin, providing extended intravascular enhance-
proved to be effective for detection of vessel dissection, ment [37], which should overcome the limitations of con-
vascular wall and parenchymal hematoma, and subarach- ventional contrast agents in MRA. The long intravascular
noid bleeding. These diagnoses are not achievable with residence time, combined with the highest available T1
other imaging modalities [34]. relaxivity of all approved agents, enables high-resolution
imaging of the carotid vessel walls and yields morphologi-
cal and functional information with a single injection.
6.6 Protocols for Supra-aortic Vessel The optimum dose, clinical efficacy, and safety of
Visualization Vasovist have been evaluated in two phase II and four
6 multicentre, phase III clinical trials. The optimum dose
MRA provides the opportunity for both morphologi- was found to be 0.03 mmol/kg and an injection time of
cal and functional imaging. If multiple 3D data sets are 23 ml/s is recommended for first pass imaging [38, 39].
acquired during the passage of the contrast agent bolus, The phase III trials showed that the overall accuracy of
the dynamics of blood flow can be monitored, provid- Vasovist-enhanced MRA was similar to that of catheter-
ing a functional dimension to the technique that allows based DSA, as determined by blinded readings [39]. In
detection of blood-flow abnormalities. With extracellular the four studies, nine of 12 readers reported statistically
gadolinium-based contrast agents and a 1.5-T scanner, a significant improvement in sensitivity, 12 of 12 readers
series of MRA data sets can be obtained with a 3D fast greater specificity, and 11 of 12 readers greater accuracy
low-angle shot (FLASH) pulse sequence [35]. However, in both vascular beds representative of areas of turbulent
for most extracellular contrast agents, a conflict arises blood flow and of slow flow.
between the spatial resolution required for diagnostic ac- The excellent performance of the agent in MRA is
curacy of stenosis and the temporal resolution required based on two mechanisms: First, the agent binds revers-
for functional imaging. Therefore, for a combination ibly and non-covalently to albumin, allowing an elimina-
of morphological and functional imaging, two separate tion half-life of approximately 15 h. The protein binding
injections are usually necessary when an extracellular also reduces the tumbling rate of the molecule, resulting
contrast agent is used. in a significantly higher relaxivity and an extended imag-
Technical innovations, including the introduction ing time compared with other contrast agents [40]. The
of parallel imaging allow reduction in acquisition time relaxivity (at 20 mHz) of Vasovist measured in human
or increase the spatial resolution, or both [36]. The plasma and ex vivo samples from rabbits and monkeys is
result is that, by virtue of improved sub-millimeter approximately six to ten times greater that that of gado-
resolution, increased accuracy of contrast-enhanced linium diethylenetriaminepentaacetic acid (Magnevist,
MRA may be anticipated, thus meeting the require- Gd-DTPA, BayerSchering Pharma AG, Berlin, Germany)
ments for detection and accurate grading of clinically [41]. Because Vasovist remains in the blood vessels in
significant carotid stenosis, notwithstanding the fact the steady state longer than extravascular agents, an
that comparison with catheter angiography (DSA) is imaging window of at least 3060 min is provided which
unlikely now that it is no longer a widely employed makes it possible to obtain not only a first pass MRA
diagnostic tool. ( Fig. 6.1), but also high-resolution steady state imaging
Another robust technique, the time-resolved echo- with no time restrictions from the contrast media side
shared angiography technique (TREAT), provides ad- [4244] ( Fig. 6.2). Compared with first pass imaging,
ditional functional information in areas or lesions with steady state imaging sequences yield more morphologi-
rapid blood flow. TREAT MRA combines morphological cal information than conventional MRA type sequences.
and functional information of the cerebral vasculature This makes it possible to get additional information
and addresses the specific physiological demands of about parenchymal structures such as the vessel wall
some vascular lesions, e.g., short arteriovenous transit or adjacent pathologies ( Fig. 6.3). Thus, there is the
time. The advantage of using TREAT in dynamic MRA is potential to integrate the different diagnostic require-
that it provides images with reasonable resolution (voxel ments of morphological and functional imaging in one
size 1.5 1.3 3 mm) that approach the accuracy of comprehensive scan, with only one injection of contrast
DSA. agent [39] ( Fig. 6.3).
6.7 Benefits of the Blood pool MR Contrast Agent Vasovist
63 6

Fig. 6.1A,B. Supra-aortic 1.5-Tesla (Magnetom

Avanto, Siemens Medical Solutions, Erlangen,
Germany) first pass MRA with 0.03 mmol/kg of
Vasovist in a healthy volunteer (A) and in a pa-
tient with cerebrovascular disease (B). The con-
trast media allows a homogeneous enhance-
ment of all displayed vessel segments including
the intracerebral vasculature. MIP projection (A)
and volume-rendered display (B) both allow a
high-quality diagnosis of the complete supra-
aortic vasculature with visualization of even the
small peripheral branches. In the patient study
(B) the pathology, a tandem stenosis of the left
carotid artery, was clearly visualized using a
standard dose. Due to the high vessel contrast
and the possibility to achieve high-resolution
A B data even a subtotal stenosis can be diagnosed
with high accuracy


Fig. 6.2A,B. High-resolution 1.5-T (Magnetom Avanto, Siemens Medi- 0.03 mmol/kg of Vasovist. Both the MIP projection (A) and the volume-
cal Solutions, Erlangen, Germany) steady state MRA in a patient with a rendered display (B) enable a detailed visualization of the complex
large cerebral arteriovenous malformation. The spatial resolution was angioarchitecture, including the venous drainage patterns. This infor-
0.5 x 0.5 x 0.5 mm using a fast GRE sequence 5 min after injection of mation is very important for treatment decision and planning

The reversible binding of Vasovist to albumin en- With exact timing, first pass imaging shows excellent im-
hances its efficiency and, for diagnostic purposes, allows age quality, which is comparable to that obtained with ex-
the administration of a smaller dose compared with exist- tracellular contrast agents. Processing of the images using
ing extracellular contrast agents. The result is a reduced thin maximum intensity projection or volume rendering
injection volume and flow rate, which avoids the need techniques reduced venous overlay in the reconstructed
to change sequence parameters, as in conventional MRA image.
[45]. After reaching the equilibrium phase, Vasovist ena-
Correct bolus timing for the first pass imaging is along bles repeated imaging at a high or ultra high spatial
routine lines with either a test bolus of Vasovist of 1 ml resolution. Even using low-level systems, steady state
injected at 1 ml/s followed by a 20-ml saline bolus injected acquisition of isotropic voxels of <1 mm3 is possible, with
at the same rate or, preferably, using MR fluoroscopy [44]. acquisition times 1 min. As well as determining high
64 Chapter 6 Head and Neck MRA

resolution, this allows for reformatting of the data into resulted in high sensitivity and specificity (100% and
any projection without compromising image quality. A 96%, respectively) for detection of significant stenosis
recent study found that, for steady state imaging of the [44]. In cases with suboptimal bolus timing during first
carotid arteries, an isotropic voxel dimension of 0.80 mm pass imaging, steady state imaging enables several re-
appeared to be the ideal compromise between resolution peated acquisitions.
and acquisition time (40 s) to avoid movement-based Because of the intravascular persistence and the pos-
artifacts and allowed the differentiation of the venous sibility to achieve high-resolution steady state imaging,
structures from the carotid arteries [44]. Vasovist has the potential to be used for an improved
When Vasovist is used for steady state imaging of diagnosis of intracranial vascular disease, such as aneu-
the carotid arteries, it provides a high diagnostic image rysms and vascular malformations ( Fig.6.4), and can
quality without the need for special post-processing. also demonstrate tumor vascular supply ( Fig. 6.5).
Steady state imaging provides all diagnostic information The protein binding increases both the T1 and the T2
in contrast to first pass data. Vasovist-enhanced MRA relaxivity. The improved T1 relaxivity is responsible for

Fig. 6.3AD. High-resolution 3T

(Magnetom Trio, Siemens Medi-
cal Solutions, Erlangen, Germany)
steady state Vasovist (0.03 mmol/
kg)-enhanced MRA in a patient with
squamous cell carcinoma. The GRE
VIBE sequence used allows for ultra
high spatial resolution imaging while
still obtaining anatomical information
(AC). This enables a clear description
of the relation of the tumor (arrow in
B) to the right carotid artery. As an
additional finding a stenosis of the left
carotid artery was diagnosed. Steady
state imaging also allows for better
depiction of the vessel wall, as can be
seen in a different patient with carotid
artery disease panel D (arrow)
6.7 Benefits of the Blood pool MR Contrast Agent Vasovist
65 6


Fig. 6.4AC. Vasovist-enhanced MRA

(0.03 mmol/kg) in a patient with small residual
arteriovenous malformation after bleeding who
was referred for treatment decision (A) While
TOF-MRA was not able to diagnose residual AVM
components, the contrast-enhanced 3-T MRA
(Magnetom Trio, Siemens Medical Solutions,
Erlangen, Germany) using TRICKS (B) showed a
small residual AVM nidus at the lateral border
of the cerebellum, fed by an early-appearing
branch from the left PICA. The high-resolution
steady state MRA that followed (C) allowed a
detailed 3D view of the nidus in relation to the
feeding arteries and draining vein. Based on the
MRA, the patient was identified as a candidate
for high-precision radiosurgical treatment

Fig. 6.5. Visualization of a radiotherapeutic induced blood-brain to illustrate residual components. However, the enhancement in the
barrier breakdown in a patient with arteriovenous malformation. Fol- AVM bed was well-displayed due to an extravasation of Vasovist via
lowing injection of a standard dose of Vasovist both first pass mul- the disrupted blood-brain barrier
tiphase TRICKS MRA and high-resolution steady state were not able
66 Chapter 6 Head and Neck MRA


Fig. 6.6A,B. 3-Tesla (Magnetom Trio, Siemens Medical Solutions, substantially increased, with enhanced display of small venous vas-
Erlangen, Germany) susceptibility-weighted MRI prior to (A) and after cular structures. On the displayed identical slices the venous vascular
(B) injection of 0.03 mmol/kg of Gadofosveset in a cavernoma patient. malformation appears to be larger due to the stronger susceptibility
Following contrast medium injection the quality of the SWI MRI effect, induced by the higher relaxivity of the agent

the substantially better MRA performance, and because 6.8 Conclusions

of the improved T2 relaxivity it is possible to enhance the
image quality of new alternative angiographic techniques, Over the past decade, a series of technological advances,
e.g., susceptibility-weighted MR imaging. such as fast gradient echo imaging and better understand-
High-resolution susceptibility-weighted imaging (SWI) ing of the contrast dynamics of extravascular contrast
(initially called high-resolution magnetic resonance blood agents, has resulted in the development of highly ac-
oxygenation level-dependent venography) is a three-di- curate, safe imaging of the head and neck arteries. The
mensional MRA method that has been developed within development of Vasovist the first approved blood pool
the past few years [46, 47]. contrast agent combined with advances in acquisition
The BOLD (blood oxygenation level-dependent) sen- strategies and more advanced post-processing techniques
sitive method combines magnitude and phase image infor- has resulted in sub-millimeter resolution of carotid steno-
mation from a high-resolution 3D T2*-weighted gradient sis. This degree of resolution and accuracy of CE-MRA
echo sequence to visualize small cerebral vessels, mainly has already resulted in MRA replacing DSA as the gold
veins, in high spatial resolution and detail. The technique standard for assessing carotid artery disease and identify-
has so far been applied for imaging vascular malforma- ing patients who are likely to benefit most from surgical
tions [48], brain tumors [49], trauma, stroke, microhemor- intervention. Future studies will have to address the opti-
rhages, and as a functional imaging method [49]. mization of imaging time, reconstruction time, and inter-
Methods to improve this interesting new angiographic pretation strategies as well as the further optimization of
technique are the use of 3T and the use of a T1-shorten- image quality for steady state imaging data. Although it
ing paramagnetic contrast agent [50]. First experience has yet to be evaluated extensively in the clinical setting,
with the use of Vasovist has shown that the quality of initial studies suggest that Vasovist gives important ad-
susceptibility-weighted MRA is superior with respect to ditional information in the evaluation of both carotid and
vessel visualization and description of extent of disease intracranial disease, not only at the structural but also the
( Fig. 6.6). functional level.
67 6
Take home messages
Vasovist as a strongly protein-binding contrast agent Initial studies proved the superiority of the agent in
has the potential for high quality contrast enhanced MR the assessment of the vessel wall.
imaging of the carotids and intracerebral vasculature. Additional controlled comparative studies are on-
The agent allows a parallel assessment of both the going to better assess the clinical potential of these
arterial and venous section of the vascular bed, both agents, the optimal imaging time and parameters as
from the morphological as well as from the functional well as the advantage over the currently established
perspective. MR contrast agents.

References 16. Nieman K, van der Lugt A, Pattynama PM, de Feyter PJ (2003)
Noninvasive visualization of atherosclerotic plaque with electron
1. Feigin VL, Lawes CM, Bennett DA, Anderson CS (2003) Stroke beam and multislice spial computed tomography. J Interv Cardiol
epidemiology: a review of population-based studies of incidence, 16:123128
prevalence, and case-fatality in the late 20th century. Lancet Neu- 17. Adams WM, Laitt RD, Thorne J, Jackson A (1999) MRA visualization
rol 2:4353 of cerebral aneurysms. Medica Mundi 43:29
2. Beers MHH, Porter RS (2003) The Merck manual of diagnosis and 18. Ozsarlak O, Van Goethem JW, Maes M, Parizel PM (2004) MR angi-
therapy, 18th edn. Merck Research Laboratories, Whitehouse Sta- ography of the intracranial vessels: technical aspects and clinical
tion applications. Neuroradiology 46:955972
3. World Health Organization (2003) Cardiovascular disease fact file. 19. Wilkerson DK, Keller I, Mezrich R, Schroder WB, Sebok D, Gronlund-
World Health Organization, Geneva Jacobs J, Conway R, Zatina MA (1991) The comparative evaluation
4. Chang HS (2006) Simulation of the natural history of cerebral of three-dimensional magnetic resonance for carotid artery dis-
aneurysms based on data from the International Study of Unrup- ease. J Vasc Surg 14:803-809
tured Intracranial Aneurysms. J Neurosurg 104:188194 20. Chiesa R, Melissano G, Castellano R, Triulzi F, Anzalone N, Veg-
5. Bots ML, Hoes AW, Hofman A, Witteman JCM, Grobbee DE (1999) lia F, Scotti G, Grossi A (1993) Three dimensional time-of-flight
Cross-sectionally assessed carotid intima-media thickness relates magnetic resonance angiography in carotid artery surgery: a
to long-term risk of stroke, coronary heart disease and death as comparison with digital subtraction angiography. Eur J Vasc Surg
estimated by available risk functions. J Intern Med 245:269276 7:171176
6. Hademenos GJ, Massoud TF (1997) Biophysical mechanisms of 21. DeMarco JK, Huston J, 3rd, Bernstein MA (2004) Evaluation of clas-
stroke. Stroke 28:20672077 sic 2D time-of-flight MR angiography in the depiction of severe
7. European Carotid Surgery Trialists Collaborative Group (1998) carotid stenosis. AJR Am J Roentgenol 183:787793
Randomised trial of endarterectomy for recently symptomatic 22. Carr JC, Shaibani A, Russell E, Finn JP (2001) Contrast-enhanced
carotid stenosis: final results of the MRC European Carotid Surgery magnetic resonance angiography of the carotid circulation. Top
Trial (ECST). Lancet 351:13791387 Magn Reson Imaging 12:349357
8. Barnett HJ, Taylor DW, Eliasziw M, Fox AJ, Ferguson GG, Haynes 23. U-King-Im JM, Trivedi RA, Graves MJ, Higgins NJ, Cross JJ, Tom BD,
RB, Rankin RN, Clagett GP, Hachinski VC, Sackett DL, Thorpe KE, Hollingworth W, Eales H, Warburton EA, Kirkpatrick PJ, Antoun
Meldrum HE, Spence JD (1998) Benefit of carotid endarterectomy NM, Gillard JH (2004) Contrast-enhanced MR angiography for ca-
in patients with symptomatic moderate or severe stenosis. North rotid disease: diagnostic and potential clinical impact. Neurology
American Symptomatic Carotid Endarterectomy Trial Collabora- 62:12821290
tors. N Engl J Med 339:14151425 24. Nael K, Ruehm SG, Michaely HJ, Pope W, Laub G, Finn JP, Villablanca
9. Mayberg MR, Wilson SE, Yatsu F, Weiss DG, Messina L, Hershey LA, JP (2006) High spatial-resolution CE-MRA of the carotid circulation
Colling C, Eskridge J, Deykin D, Winn HR (1991) Carotid endar- with parallel imaging: comparison of image quality between 2 dif-
terectomy and prevention of cerebral ischemia in symptomatic ferent acceleration factors at 3.0 Tesla. Invest Radiol 41:391399
carotid stenosis. JAMA 266:32893294 25. Naganawa S, Koshikawa T, Fukatsu H, Sakurai Y, Ichinose N,
10. Barth A, Arnold M, Mattle HP, Schroth G, Remonda L (2006) Ishiguchi T, Ishigaki T (2001) Contrast-enhanced MR angiography
Contrast-enhanced 3-D MRA in decision making for carotid endar- of the carotid artery using 3D time-resolved imaging of contrast
terectomy: a 6-year experience. Cerebrovasc Dis 21:393400 kinetics: comparison with real-time fluoroscopic triggered 3D-
11. Kaufmann TJ, Kallmes DF (2005) Utility of MRA and CTA in the eval- elliptical centric view ordering. Radiat Med 19:185192
uation of carotid occlusive disease. Semin Vasc Surg 18:7582 26. Earls JP, Rofsky NM, DeCorato DR, Krinsky GA, Weinreb JC (1996)
12. U-Kim-Im JM, Trivedi Ra, Graves MJ et al (2004) Contrast-enhanced Breath-hold single-dose gadolinium-enhanced three-dimen-
MR angiography for carotid disease: diagnostic and potential clini- sional MR aortography: usefulness of a timing examination and
cal impact. Neurology 27:12821290 MR power injector. Radiology 201:705-710
13. Jewells V, Castillo M (2003) MR angiography of the extracranial 27. Foo TK, Saranathan M, Prince MR, Chenevert TL (1997) Automated
circulation. Magn Reson Imaging Clin N Am 11:585597 detection of bolus arrival and initiation of data acquisition in fast,
14. Beltramello A, Piovan E, Rosta L (1994) Double blind comparison three-dimensional, gadolinium-enhanced MR angiography. Radi-
of safety and efficaxo of iomeprol and iopamidol in carotid digital ology 203:275280
subtraction angiography. Eur J Radiol 18:S6772 28. Wilman AH, Riederer SJ, King BF, Debbins JP, Rossman PJ, Ehman
15. Barth A, Arnold M, Mattle HP, Schroth G, Remonda L (2006) RL (1997) Fluoroscopically triggered contrast-enhanced three-
Contrast-enhance 3-D MRA in decision making for carotid endar- dimensional MR angiography with elliptical centric view order:
terectomy: a 6-year experience. Cerebrovasc Dis 21:393400 application to the renal arteries. Radiology 205:137146
68 Chapter 6 Head and Neck MRA

29. Blakeley DD, Oddone EZ, Hasselblad V, Simel DL, Matchar DB 48. Essig M, Reichenbach JR, Schad LR, et al (1999) High-resolution
(1995) Noninvasive carotid artery testing. A meta-analytic review. MR venography of cerebral arteriovenous malformations. Mag Res
Ann Intern Med 122:360367 Imaging 17:14171425
30. Nederkoorn PJ, van der Graaf Y, Hunink MG (2003) Duplex ul- 49. Sehal V, Delproposto Z, Haacke EM, et al (2005) Clinical applica-
trasound and magnetic resonance angiography compared with tions of neuroimaging with susceptibility-weighted imaging. J
digital subtraction angiography in carotid artery stenosis: a sys- Magn Reson Imaging 22:439450
tematic review. Stroke 34:13241332 50. Barth M, Noebauer-Huhmann I-M, Reichenbach JR, et al (2003)
31. Atlas SW, Sheppard L, Goldberg HI, Hurst RW, Listerud J, Flamm High resolution, three dimensional contrast-enhanced blood oxy-
E (1997) Intracranial aneurysms: detection and characterization genation level-dependent magnetic resonance venography of
with MR angiography with use of an advanced postprocessing brain tumors at 3 Tesla: first experience and comparison with 1.5
technique in a blinded-reader study. Radiology 203:807814 Tesla. Invest Radiol 38:409414
32. Duran M, Schoenberg SO, Yuh WT, Knopp MV, van Kaick G, Essig M
(2002) Cerebral arteriovenous malformations: morphologic evalu-
ation by ultrashort 3D gadolinium-enhanced MR angiography. Eur
Radiol 12:29572964
33. Essig M, Reichenbach JR, Schad LR, Schoenberg SO, Debus J,
Kaiser WA (1999) High-resolution MR venography of cerebral arte-

6 riovenous malformations. Magn Reson Imaging 17:14171425

34. Schellinger PD, Fiebach JB (2004) Intracranial hemorrhage: the
role of magnetic resonance imaging. Neurocrit Care 1:31-45
35. Fellner C, Strotzer M, Fraunhofer S, Held P, Spies V, Seitz J, Fellner
F (1997) MR angiography of the supra-aortic arteries using a
dedicated head and neck coil: image quality and assessment of
stenoses. Neuroradiology 39:763771
36. Weiger M, Pruessmann KP, Kassner A, Roditi G, Lawton T, Reid
A, Boesiger P (2000) Contrast-enhanced 3D MRA using SENSE. J
Magn Reson Imaging 12:671677
37. Caravan P, Cloutier NJ, Greenfield MT, McDermid SA, Dunham SU,
Bulte JW, Amedio JC Jr, Looby RJ, Supkowski RM, Horrocks WD Jr,
McMurry TJ, Lauffer RB (2002) The interaction of MS-325 with hu-
man serum albumin and its effect on proton relaxation rates. J Am
Chem Soc 124:31523162
38. Bluemke DA, Stillman AE, Bis KG, Grist TM, Baum RA, DAgostino
R, Malden ES, Pierro JA, Yucel EK (2001) Carotid MR angiogra-
phy: phase II study of safety and efficacy for MS-325. Radiology
39. Perreault P, Edelman MA, Baum RA, Yucel EK, Weisskoff RM, Shamsi
K, Mohler ER 3rd (2003) MR angiography with gadofosveset triso-
dium for peripheral vascular disease: phase II trial. Radiology
40. Goyen M, Shamsi K, Schoenberg SO (2006) Vasovist-enhanced MR
angiography. Eur Radiol 16 [Suppl 2]:B9B14
41. Lauffer RB, Parmelee DJ, Dunham SU, Ouellet HS, Dolan RP, Witte
S, McMurry TJ, Walovitch RC (1998) MS-325: albumin-targeted
contrast agent for MR angiography. Radiology 207:529538
42. Rohrer M, Bauer H, Mintorovitch J, Requardt M, Weinmann HJ (2005)
Comparison of magnetic properties of MRI contrast media solutions
at different magnetic field strengths. Invest Radiol 40:715724
43. Grist TM, Korosec FR, Peters DC, Witte S, Walovitch RC, Dolan RP,
Bridson WE, Yucel EK, Mistretta CA (1998) Steady state and dy-
namic MR angiography with MS-325: initial experience in humans.
Radiology 207:539544
44. Nikolaou K, Kramer H, Grosse C, Clevert D, Dietrich O, Hartmann
M, Chamberlin P, Assmann S, Reiser MF, Schoenberg SO (2006)
High-spatial-resolution multistation MR angiography with parallel
imaging and blood pool contrast agent: initial experience. Radiol-
ogy 241:861872
45. Hartmann M, Wiethoff AJ, Hentrich HR, Rohrer M (2006) Initial
imaging recommendations for Vasovist angiography. Eur Radiol
16 [Suppl 2]:B15B23
46. Haacke EM, Xu Y, Cheng YC, Reichenbach JR (2004) Susceptibility
weighted imaging (SWI). Magn Reson Med 52:612618
47. Reichenbach JR, Venkatesan R, Schillinger et al (1997) Small ves-
sels in the human brain: MR venography with deoxyhemoglobin
as an intrinsic contrast agent. Radiology 204:272277

Pulmonary MRA
Christian Fink, Ulrike Attenberger, and Konstantin Nikolaou

7.1 Introduction 70

7.2 Technical Considerations 70

7.2.1 General Considerations 70
7.2.2 Introduction of Intravascular Contrast Agents 70
7.2.3 Perfusion MRI of the Lungs: Qualitative and Quantitative Approach 71
7.2.4 Imaging the Pulmonary Vasculature at 3 Tesla 73

7.3 Clinical Applications 73

7.3.1 Pulmonary MRA in Acute Pulmonary Embolism 74
7.3.2 Venous Imaging 74
7.3.3 Imaging of Patients with Pulmonary Arterial Hypertension 75

7.4 Conclusions 76

References 77
70 Chapter 7 Pulmonary MRA

7.1 Introduction of the pulmonary vasculature may be performed. Along

with improved arteriovenous separation, time-resolved
Computed tomography angiography (CTA) of the lung pulmonary MRA is less sensitive for incorrect bolus timing
is nowadays considered the clinical gold standard for the and less sensitive for motion artifacts. The latter may be rel-
assessment of pulmonary vascular disease [1, 2]. Different evant in dyspneic patients. Also, and probably most impor-
to other vascular territories (e.g., the peripheral arteries, tant, time-resolved MRA also makes it possible to obtain
renal arteries), contrast-enhanced magnetic resonance functional information of the pulmonary circulation, such
angiography (MRA) is usually not considered a first-line as the characterization of shunts or the assessment of capil-
imaging tool for the assessment of the pulmonary circula- lary perfusion of the lung parenchyma. Several studies have
tion. When compared with CT, the inferior spatial resolu- shown the feasibility of pulmonary perfusion MRI using a
tion and long breath-hold times of pulmonary MRA are time-resolved MRA technique [711].
regarded as major drawbacks. The anatomy of the pulmonary circulation has some
On the other hand, there are potential advantages of specific implications for pulmonary MR imaging. The
MRA for the evaluation of lung disease. Above all, this multiple air-tissue interfaces of the lungs result in a very
includes the lack of ionizing radiation, which is of major high susceptibility. This substantially affects the achiev-
importance in congenital disease and chronic diseases re- able signal intensity of small peripheral lung vessels in
quiring frequent follow-up examinations. Moreover, MRI MRI. On the other hand, the low signal of the lung pa-
7 is considered the standard of reference for the assessment renchyma usually results in a high vessel-to-background
of cardiac function; thus MRI can provide a comprehen- contrast of pulmonary MRA. In general, short echo times,
sive assessment of pulmonary vascular disease, includ- i.e., potentially lower than 23 ms, should be used to elim-
ing measurement of the right heart function pulmonary inate susceptibility effects [12]. Moreover, the lungs have
artery pressure [3, 4]. Ongoing technical developments, a very short transit time, i.e., in the range of 35 s [13]. As
such as parallel imaging techniques, have further im- a consequence, first pass pulmonary MRA often has sub-
proved the feasibility of pulmonary MRA and moved stantial venous contamination, potentially affecting the
the application from a pure research tool to a clinical diagnostic accuracy. To reduce venous contamination, the
practice. The recent introduction of the blood pool MR aim should be a very compact bolus profile. Using ECCM,
contrast agent Vasovist (Gadofosveset, Bayer Schering this can be achieved by using a low contrast agent volume
Pharma AG, Berlin, Germany) has further increased the and high injection rates (e.g. 5 ml/s).
options of pulmonary MRA [5]. The albumin-binding
characteristic of the contrast agent extends the vascular
lifetime and thus allows for longer vascular imaging 7.2.2 Introduction of Intravascular Contrast
time, potentially higher spatial resolution, and greater Agents
anatomic coverage. During the first pass of the contrast
agent, pulmonary perfusion can be assessed. During the Compared with standard ECCM, the intravascular contrast
equilibrium or steady state phase of the contrast, the agent Vasovist provides a much smaller and therefore
pulmonary vasculature as well as other parts of the body very compact injection volume of approximately 710 ml
vasculature such as the veins of the abdomen and lower for a standard dose (0.12 ml/kg body weight). Similar
extremities can be scanned at high spatial resolution, to standard ECCM, Vasovist can also be used for first
without additional injections of contrast agent required. pass high-resolution MRA or first pass perfusion imag-
With the introduction of such contrast agents, limitations ing [14]. Because the relaxivity of Vasovist is five to
of todays thoracic MRA might be overcome, and imaging seven times higher than that of conventional 0.5 mol/l
strategies could further be developed and improved [6]. extracellular contrast agents (used at 1.5 T), the injection
of Vasovist at 1 ml/s translates into a conventional con-
trast agent injection rate of about 57 ml/s, resulting in a
7.2 Technical Considerations high vascular signal intensity (SI) and high image quality
of the Vasovist-enhanced first pass imaging data. This
7.2.1 General Considerations benefit of high signal-to-noise ratio during the first pass
of Vasovist might be of particular advantage for time-
Three-dimensional gradient echo MRA after injection of resolved imaging. For first pass MRA, regardless of using
an extracellular MR contrast agent (ECCM) has been es- ECCM or Vasovist, the shortest achievable TE and TR
tablished as the method of choice for pulmonary MRA. should be used. However, in contrast to ECCM, Vasovist
In general, there are two different approaches to first pass can also be used for an additional high-resolution MRA
contrast-enhanced pulmonary MRA. In addition to a single during steady state distribution of the contrast agent
high-resolution 3D MRA, time-resolved multiphase MRA ( Fig. 7.1). If possible, the sequence parameters for steady
7.2 Technical Considerations
71 7

state MRA with Vasovist should be adapted using longer 7.2.3 Perfusion MRI of the Lungs: Qualitative
TR (720 ms) and lower flip angles (1525) [15]. How- and Quantitative Approach
ever, using a longer TR might increase the scan time, so
that a breath-hold exam is no longer feasible. Also, car- For a number of clinical indications, knowledge of the
diac motion artifacts can pose considerable problems in regional pulmonary microcirculation is essential, such
acquisition optimization. One solution to maintain a high as in patients suffering from acute or chronic pulmonary
spatial resolution of the steady state MRA datasets could embolism or various other forms of pulmonary hyperten-
be the acquisition of two separate sagittal slabs covering sion. Presurgical information on the most poorly per-
the pulmonary arteries and both lungs in two subse- fused lung area before lung reduction is as important as
quent breath-holds [16]. In addition, refined acquisition postoperative quantification of lung perfusion after lung
techniques involving free-breathing navigator approaches transplantation or in postoperative complications, such as
possibly combined with electrocardiographic gating obliterative bronchiolitis. The desire for both qualitative
could be effective to optimize image quality. Navigator- and quantitative information on pulmonary microcircula-
gated MRA has already been evaluated for free-breathing tion in a single study without exposure to ionizing radia-
pulmonary MRA in various animal and volunteer studies tion drives the investigation of magnetic MRI techniques
[1720]. A potential drawback of navigator-gated MRA for the assessment of pulmonary perfusion.
is the increased scan time, which might limit the value of More than 10 years ago it was reported that qualita-
this imaging technique in clinical practice. tive assessment of pulmonary microcirculation in human
Several studies have already evaluated the use of vari- subjects was feasible using a contrast-enhanced inversion
ous blood pool MR contrast agents for pulmonary MRA. recovery fast gradient-echo technique with ultra-short
In a study by Nolte-Ernsting et al., pulmonary MRA was echo times [25]. Subsequent clinical studies have shown
performed in pigs using a superparamagnetic iron oxide the clinical utility of dynamic MRI for the evaluation of
blood pool agent. In correlation to conventional X-ray lung diseases such as pulmonary embolism by detecting
angiography, the pulmonary vasculature was visualized the first pass of a paramagnetic contrast medium (CM)
down to the first order sub-segmental branches, includ- through lung tissue [26, 27]. The clinical introduction of
ing vessel diameters of approximately 1.5 mm [21]. A parallel imaging techniques has brought along significant
subsequent feasibility study by Weishaupt et al. evalu- advantages, primarily in terms of acquisition time and/
ated ultra-small superparamagnetic iron oxide-enhanced or an increase of spatial resolution [10, 28]. These advan-
pulmonary MRA in two volunteers and a pig model of tages are particularly attractive for time-resolved, con-
pulmonary hemorrhage [22]. The use of the blood pool trast-enhanced MRI of tissue microcirculation, because
agent allowed excellent delineation of central, segmental, a high temporal resolution is necessary for acquisition of
and sub-segmental pulmonary arteries in the volunteer sufficient data points during the first passage of a contrast
studies. In the pig model, pulmonary hemorrhage was bolus. With the application of these techniques, MRI has
visualized. In a volunteer study by Ahlstrom et al., pulmo- been shown to be sensitive to gravity-dependent differ-
nary MRA was performed with a superparamagnetic iron ences of pulmonary perfusion and to agree well with con-
oxide using navigator respiratory gating. With increasing ventional radionuclide perfusion scintigraphy in healthy
doses of the contrast agent, higher signal intensities and volunteers and in patients with lung cancer [7].
vessel branch order visualization were achieved. A subse- In addition to qualitative imaging of lung perfusion,
quent animal study by Abolmaali et al. demonstrated the the quantitative assessment of the pulmonary microcircu-
feasibility of navigator-gated pulmonary MRA using the lation would be desirable for several reasons: It would en-
gadolinium-based blood pool contrast agent Gadomer-17 able interindividual comparison of patients with various
[20]. Zheng et al. evaluated the protein-binding blood pathologies, and an intraindividual assessment after sur-
pool contrast agent B-22956/1 for the assessment of pul- gical or other medical treatment, e.g., longitudinal assess-
monary embolism by perfusion MRI and high-resolution ment of drug treatment of pulmonary hypertension using
MRA in a pig model [23]. They found that whole- prostaglandin derivates [29]. However, several technical
lung-coverage perfusion MRI and high-resolution target difficulties in absolute quantification of lung perfusion
MRA could be performed after a single contrast bolus have to be overcome: susceptibility artifacts reducing the
injection. Motivated by these studies, we assessed the obtainable signal, necessary optimization of the dosage
feasibility of contrast-enhanced 3D perfusion MRI and of contrast agent to avoid systematic errors in perfusion
MRA of pulmonary embolism using a single injection of quantification, and choosing the correct theoretical or
Gadomer-17. Perfusion MRI allowed the visualization of mathematical model to calculate lung perfusion. Dif-
typical wedge-shaped perfusion defects. Compared with a ficulties in obtaining MR signal from lung parenchyma
standard ECCM, the blood pool contrast agent achieved can be decreased by using an ultra-short echo time (TE)
a higher SNR [24]. of <1 ms, which reduces the effect of local magnetic field
72 Chapter 7 Pulmonary MRA

Fig. 7.1AC. High-resolution steady state MRA of the lung acquired isotropic spatial resolution of about 1.01.5 mm3 can be achieved in a
with Vasovist in a healthy volunteer. After first pass imaging (dur- scan time of approximately 20 s. The nearly isotropic spatial resolution
ing which either perfusion imaging or first pass high-resolution MR allows for a true 3D volumetric assessment of the pulmonary arterial
angiography can be performed), a long acquisition window of up to tree [A coronal maximum intensity projection (MIP); B sagittal MIPs of
60 min can be used for high-resolution imaging of various vascular the right lung; C 3D volume rendering of the complete MRA dataset].
7 beds. Using parallel imaging for high-resolution steady state MRA, (Modified from [50], with permission)

inhomogeneities caused by the multiple air/soft tissue for quantification of pulmonary blood flow and volume
interfaces of pulmonary alveolar architecture. Promising must be markedly lower than the dose typically used for
results concerning quantification of pulmonary perfusion qualitative, visual assessment of pulmonary perfusion.
in an animal model using a time-resolved inversion re- Furthermore, for a quantitative estimation of pulmonary
covery technique with ultra-short echo times have been blood flow and volume, typically simplified models such
published [30]. Initial studies have presented data on as an open one-compartment model are being used.
absolute quantification of pulmonary perfusion in hu- However, this simple model reflects the presumption that
man subjects, including a limited number of patients and CM extravasation is negligible during first pass through
reporting a broad interindividual variation of results [1, the lungs (or, least probable, that CM exchange between
10, 32]. Also as mentioned above - the effect of the CM the capillaries and the interstitial space is extremely
dosage and CM type on the correctness of quantification fast). The use of blood pool agents such as Vasovist may
of pulmonary microcirculation in human subjects has resolve this dilemma because their prolonged retention
been brought up and seems to be crucial for the reliability time circumvents the possibility of extravasation during
of quantitative results [10, 33]. CM dosages appropriate the first pass [36]. Several initial studies have employed
for absolute quantification are small, resulting in only a blood pool contrast agent for qualitative evaluation of
small signal increases. So far, only a few studies have pulmonary perfusion, with promising results ( Fig. 7.1
systematically investigated the effect of the amount and and 7.2) [23, 24, 37]. One study has also demonstrated
type of the paramagnetic contrast agent applied in time- the successful quantification of perfusion indices on a 3T
resolved pulmonary MRA [10]. It has been reported that system, using an intravascular contrast agent [38]. The
administration of a CM with a higher concentration of prolonged intravascular time of an intravascular contrast
gadolinium chelates (e.g., 1.0 M Gadobutrol, Gadovist) agent ensures the linearity of the signal response to tracer
but maintaining the absolute amount of chelates does concentration, which is necessary for application of the
not offer significant advantages over standard 0.5 M Gd- indicator dilution theory [39]. This results in an improved
DTPA for contrast-enhanced 3D MRI of the lung [33]. It fitting, and accurate quantification of perfusion indices
seems reasonable that the use of a higher total CM dose such as time-to-peak (TTP), mean-transit-time (MTT),
tended to provide a higher degree of visual parenchymal maximal-signal-intensity (MSI) and others becomes fea-
enhancement [34, 35]. However, using these high CM sible. Overall, in terms of qualitative and quantitative as-
concentrations, absolute quantification of microcircula- sessment of pulmonary perfusion, intravascular contrast
tory parameters is not practicable because of saturation agents such as Vasovist have two major advantages: First,
effects affecting the shape of the arterial input func- after perfusion imaging during first pass of the contrast,
tion. An approximately linear correlation between sig- it is still possible to obtain high-resolution steady state
nal intensity and CM concentration is an indispensable MRA data. Second, due to the lack of contrast extravasa-
requirement for reliable assessment of the arterial input tion during first pass, absolute quantification should be
function. For this reason, the CM dose administered more reliable and reproducible.
7.3 Clinical Applications
73 7

Fig. 7.2A,B. Time-resolved MRA (A, B) of the lung acquired with fusion MRI). Using parallel imaging techniques and a state-of-the-art
Vasovist in a healthy volunteer. A: Selected frames from a time-re- MR scanner with strong gradients, single 3D slabs can be acquired
solved MRA acquired during the first pass of the contrast agent bolus in about 1.01.2 s, acquiring a dynamic dataset of 20 slabs in an ac-
showing a homogeneous lung perfusion B: Color-coded perfusion ceptable breath-hold time of less than 25 s during the first pass of
map computed from the time-resolved MRA data (pulmonary per- Vasovist. (Modified from [50], with permission)

7.2.4 Imaging the Pulmonary Vasculature to 1.5 T [46]. Dielectric resonances and radiofrequency
at 3 Tesla (RF) eddy currents can also be potentially troublesome at
3 T [47]. The use of an intravascular contrast agent with
The implementation of parallel acquisition techniques shorter T1 may be advantageous to trade off the signal loss
generally results in a reduction of the minimum scan associated with magnetic susceptibility effects and further
time, which then can be used to increase the speed and/ support highly accelerated parallel acquisition. Also, since
or the spatial resolution. However, the signal-to-noise the longitudinal relaxation time (T1) of unenhanced blood
ratio (SNR) penalty of parallel imaging can be limiting increases with field strength, the sensitivity to injected
in clinical practice, depending on several factors, such as gadolinium (Gd) agents is heightened, and thus less con-
the degree of K-space undersampling [40]. Strategies to trast material is required [48]. However, initial studies
preserve SNR while maximizing the efficiency of parallel have shown a relatively lower parenchymal enhancement
acquisition include raising the baseline signal by increas- at 3 T in comparison to 1.5 T. One possible explanation
ing field strength [41, 42], minimizing noise amplification is that signal loss occurs due to magnetic susceptibility
by using new array coils [43], and using contrast agents gradients of the lungs resulting from multiple air/soft-
with higher concentration or T1 relaxivity [44, 45]. In a tissue interfaces. Since the susceptibility effect increases
recently published study, Nael et al. evaluated the feasibil- with magnetic field [32], the T2* of the lung is expected to
ity of contrast-enhanced pulmonary perfusion MRI at be smaller at 3.0 T than at 1.5 T [49]. This signal loss has
3 T using Gadomer-17 in a pig model [38]. The high T1 constrained the application of parallel acquisition to an ac-
relaxivity of the blood pool contrast agent facilitated the celeration factor of 2 for pulmonary imaging at both 1.5 T
acquisition of perfusion data sets with high image quality and 3.0 T. An intravascular contrast agent with shorter T1,
and allowed for the calculation of quantitative perfusion such as Vasovist, may be used to enhance the perform-
parameters. ance of pulmonary perfusion protocols at 3.0 T and sup-
In general, the introduction of multichannel 3 T MR port parallel acquisition with a higher acceleration factor.
systems into clinical practice has the potential to signifi-
cantly improve the performance of pulmonary MR per-
fusion and high-resolution MRA in terms of spatial and 7.3 Clinical Applications
temporal resolution [41]. However, in the lungs the higher
magnetic susceptibility gradients associated with high Pulmonary MRA has been proposed for several pulmo-
magnetic field can be limiting, resulting in signal loss and nary vascular disorders. The imaging technique can be
relatively lower parenchymal enhancement in comparison tailored to the clinical problem, ranging from detailed
74 Chapter 7 Pulmonary MRA

morphological evaluation using high spatial resolution in 62 patients with suspected PE. Using parallel imag-
MRA to functional time-resolved MRA studies character- ing their MRA technique achieved a substantially higher
izing shunts or lung perfusion. In the following, the most spatial resolution (0.7 1.2 1.5 mm) than the previous
important applications are reviewed in detail. studies in a short acquisition time of 15 s. With this tech-
nique pulmonary contrast-enhanced MRA achieved a
sensitivity of 81% and a specificity of 100% in comparison
7.3.1 Pulmonary MRA in Acute Pulmonary to 16-slice CTA [56].
Embolism In addition to these results of high-resolution MRA
of the pulmonary vasculature, time-resolved contrast-
Several studies have evaluated the performance of pulmo- enhanced MRA has been evaluated for the assessment of
nary contrast-enhanced MRA for the diagnosis of pulmo- patients with suspected PE. In a feasibility study, Goyen
nary embolism (PE) [50]. In a pivotal study, Meaney et al. et al. examined eight dyspneic patients with known or
examined 30 patients with suspected PE using contrast- suspected PE using a time-resolved contrast-enhanced
enhanced pulmonary MRA and DSA. Contrast-enhanced MRA with a scan time of less than 4 s. Pulmonary
MRA yielded a sensitivity and specificity for the detection contrast-enhanced MRA allowed the assessment of the
of PE between 75100% and 95100%, respectively. In pulmonary arterial tree up to a sub-segmental level and
addition, a good interobserver agreement (kappa values identified PE in all four subsequently confirmed cases.
7 of 0.570.83) was demonstrated [51]. A subsequent study All patients were able to hold their breath for at least
by Gupta et al. evaluated the accuracy of pulmonary 8 s, during which a dataset with an angiogram of the
contrast-enhanced MRA in 36 patients with intermedi- pulmonary arteries was obtained [57]. In a more recent
ate or low probability lung scintigraphy. In this study study Ohno et al. compared the diagnostic accuracy
as well, all patients underwent pulmonary DSA, which of time-resolved contrast-enhanced MRA with parallel
demonstrated PE in 13 patients. Pulmonary contrast- imaging (SENSE) versus CTA and ventilation-perfusion
enhanced MRA diagnosed 12 patients as having PE but scintigraphy (VQ scan) in 48 patients with suspected
missed two cases. This resulted in a sensitivity of 85% and PE. Conventional pulmonary DSA served as the gold
specificity of 96%. Both missed pulmonary emboli were standard. In this study time-resolved MRA had a higher
isolated and sub-segmental in location [52]. In the larg- sensitivity [92% vs. 83% (CTA) and 67% (VQ scan)] and
est study so far, Oudkerk et al. assessed the accuracy of specificity [94% vs. 94% (CTA) and 78% (VQ scan)] for
pulmonary contrast-enhanced MRA in 141 patients with the detection of PE than CTA and VQ scan [58].
an abnormal perfusion lung scintigraphy and compared
the findings with those of pulmonary DSA. Contrast-
enhanced MRA detected 27 of 35 cases with confirmed 7.3.2 Venous Imaging
PE, resulting in an overall sensitivity of 77%. The sensi-
tivity for isolated sub-segmental, segmental, and central/ It has been estimated that in the United States alone, as
lobar PE was 40%, 84%, and 100%, respectively. Contrast- many as 5 million episodes of pulmonary embolism, a
enhanced MRA demonstrated emboli in two patients with dreaded complication of deep venous thrombosis, occur
a normal angiogram, i.e., the specificity was 98% [53]. In a every year. Radiographic venography or venous ultra-
more recent study, Blum et al. examined 89 patients with sonography is still considered the gold standard for the di-
suspected PE using coronal, axial, and sagittal-orientated agnosis of deep venous thrombosis; however, venography
contrast-enhanced MRA. The images were interpreted in- is an invasive X-ray based procedure requiring iodinated
dependently by two teams of radiologists. Different to the contrast agent, and ultrasonography suffers from diag-
previous studies, a heterogeneous combination of clinical nostic limitations, e.g., of the pelvic veins. Recently, MR
probability, D-dimer testing, spiral CT, compression ve- venography has also been proposed for the assessment
nous ultrasound, and pulmonary DSA served as the gold of the venous system as a potential source of PE. Non-
standard in this study [54]. In addition, the study cohort enhanced as well as direct and indirect contrast-enhanced
had a much higher prevalence of PE (i.e. 71%). Depend- MR venography techniques have been described. Accord-
ing on the team of readers, the sensitivity and specificity ing to a recently published meta-analysis [59], the sensi-
of contrast-enhanced MRA ranged between 3171% and tivity of MR venography ranges between 0% and 100%,
8592%. In a study of 48 patients with suspected PE, while the specificity ranges between 43% and 100%. The
Pleszewski reported a sensitivity and specificity of 82% pooled estimate of the sensitivity and specificity of the
and 100%, respectively [55]; similar to the study by Blum included studies was 91.5% and 94.8%, respectively. In
et al., a combination of different imaging methods served a recent study, 207 patients underwent a comprehensive
as the gold standard in this study. In a study by Kluge et MR protocol of non-enhanced and contrast-enhanced
al., pulmonary contrast-enhanced MRA was performed pulmonary MRA, pulmonary perfusion MRI and con-
7.3 Clinical Applications
75 7

trast-enhanced MR venography. The imaging time for the speed up the postprocessing phase. To date no clinical
entire protocol (including the patient positioning) was studies on Vasovist-enhanced pulmonary MRA of pa-
less than 20 min. However, agreement with venous ultra- tients with PE have been published. However, based on
sound was overall only moderate [60]. Also, for contrast- our own experience from an ongoing study, Vasovist can
enhanced MR venous angiography, timing of the contrast be used for a comprehensive evaluation of PE and deep
material injection is crucial, and contrast administration venous thrombosis using first pass pulmonary perfusion
has to be repeated to image more than one anatomic level. MRI, high-resolution steady state pulmonary MRA, and
Intravascular contrast agents such as Vasovist have sev- steady state whole-body MR venography ( Fig. 7.3, 7.4).
eral advantages. For example, the bolus transit time does
not need to be determined, which simplifies the injection
technique and avoids timing errors. Furthermore, insuf- 7.3.3 Imaging of Patients with Pulmonary
ficient spatial resolution is one of the major problems with Arterial Hypertension
contrast-enhanced MR angiography using extracellular
contrast agents. Intravascular contrast agents allow for Pulmonary arterial hypertension (PAH) is a progressive
longer scanning times than extracellular contrast agents, disease that leads to substantial mortality and eventually
thus potentially improving the spatial resolution. A third death. Elevated pulmonary arterial pressure and pulmo-
advantage of contrast-enhanced MR angiography using nary vascular resistance with right ventricular failure of
intravascular contrast agents is the possibility of scanning varying degree are the main hemodynamic features of this
as many vascular regions as needed after administra- disease. Many different disorders have been identified as
tion of a single contrast bolus. An obvious gain is the causing pulmonary hypertension, and medical manage-
visualization of the pelvic and abdominal vessels and the ment of hypertension is based on these underlying con-
deep femoral vein, compared with what can be depicted ditions [62]. Recurrent pulmonary thromboembolism is
on radiographic venography. Furthermore, scanning for one of the causes of PAH. Hypertension that is caused by
deep venous thrombosis and pulmonary embolism can be recurrent pulmonary thromboembolism is classified as
performed during a single MRA session [61]. A potential chronic thromboembolic pulmonary arterial hyperten-
drawback of contrast-enhanced MR venography using sion (CTEPH) and may be treated with surgery [63];
intravascular contrast agents is the need for postprocess- other forms of pulmonary hypertension can be treated
ing, because the contrast agent is present in arteries as with vasodilating drugs. Comprehensive workup to dif-
well as in veins. In some anatomic regions, such as below ferentiate the various causes of pulmonary hypertension,
the knee, this limitation could constitute a problem. Here, especially CTEPH, typically requires a series of clinical
multiplanar reconstructions and maximum intensity pro- examinations and imaging procedures. Because treat-
jections are crucial to differentiate venous from arterial ment of CTEPH differs considerably from that of other
structures. Improvements of these tools (e.g., in vessel causes of pulmonary hypertension, an accurate diagnosis
separation) are being developed and will most certainly is essential. A prerequisite for the correct and reliable

Fig. 7.3AC. Contrast-enhanced MR venography using the intravas- lower extremities. Coronal maximum-intensity-projection images of
cular contrast agent Vasovist in a healthy volunteer. After injection abdomen and pelvis (A), the upper leg vasculature (B), and the knee/
of the contrast agent, an acquisition window of up to 60 min can be proximal lower leg (C) show normal anatomy of arterial and venous
used not only for high resolution of the thoracic vasculature, but also structures with no evidence of venous thrombosis
for imaging the complete venous anatomy of the abdomen and the
76 Chapter 7 Pulmonary MRA

Fig. 7.4AC. Images of a 53-year-old man with dyspnea. A 64-slice tal perfusion defect in the right lower lobe owing to the thromboem-
CT angiography demonstrates a large thrombus in the right lower lobe bolic occlusion (arrow). C The high-resolution MR angiogram of same
pulmonary artery (arrow). B The maximum-intensity-projection (MIP) patient obtained during the steady state window after injection of
image of a single perfusion phase obtained during contrast-enhanced, Vasovist demonstrates the dark thromboembolic material in the pul-
first pass MR perfusion imaging with Vasovist clearly shows a segmen- monary artery of the lower right lobe (arrow), similar to the CTA image

diagnosis of CTEPH is the depiction of occluding throm- searching for sometimes subtle perfusion defects in PAH
botic material and concomitant perfusion defects. As dis- patients typically high injection rate of about 5 ml/s
cussed above, pulmonary magnetic resonance perfusion have to be used when applying extracellular MR contrast
imaging and high-resolution angiography are rather new agents. Here, contrast agents with a higher relaxivity (such
and promising non-invasive imaging techniques for as- as Vasovist provides) could be beneficial. In a recent
sessing the pulmonary circulation. Detection of chronic study, a combined approach for the correct differentiation
occlusive and non-occlusive changes in the pulmonary of CTEPH from other forms of PAH was applied by per-
arteries at the segmental or sub-segmental level requires forming time-resolved dynamic MR perfusion imaging
high spatial resolution, which is limited by the imaging and MRA with parallel imaging acquisition techniques
time of a single breath hold. Recently, parallel imaging at the same time [11]. The combination of first pass
techniques have become widely available. With these and high-resolution MRA data significantly enhanced
techniques, the temporal and spatial resolution can be diagnostic accuracy, which reached 90%. Thus, the ap-
improved substantially [9, 10]. In combination with in- plication of an intravascular contrast agent could enable a
travascular contrast agents such as Vasovist, an opti- comprehensive protocol in patients with PAH, combining
mized, fully comprehensive protocol for the assessment MR first pass perfusion imaging with high-resolution pul-
and differential diagnosis of patients with pulmonary monary MR angiography and MR venography in steady
hypertension could be implemented. Such a protocol state, providing complementary diagnostic information,
could include first pass perfusion imaging (potentially and increasing the diagnostic value of pulmonary MR
with quantification of results see above), steady state imaging in these patients.
high-resolution MRA, assessment of venous structures as
a potential source of recurrent emboli, and an examina-
tion of the right heart to look for changes caused by the 7.4 Conclusions
increased pulmonary arterial pressure. This comprehen-
sive protocol would be feasible after a single injection of Recent technical developments in the field of MR hard-
intravascular contrast. ware and pulse sequence design have substantially im-
MR perfusion imaging adds valuable information in proved the potential of MRI for the assessment of the pul-
patients with PAH; it can demonstrate the exact extent of monary circulation. With the clinical introduction of the
perfusion defects and has increased diagnostic accuracy first intravascular MR contrast agent Vasovist, new op-
in the differentiation of various causes of PAH, depending tions arise. The main advantages of this contrast agent for
on the specific patterns of perfusion defects, differentiat- imaging the pulmonary vasculature include the prolonged
ing occlusive (segmental) from non-occlusive (patchy) acquisition time during the equilibrium phase of the
forms of defects. contrast agent and its higher relaxivity as compared with
To maintain a sufficient SNR during first pass per- most extracellular contrast agents. This enables the setup
fusion imaging which is a crucial prerequisite when of a comprehensive MR protocol for various vascular dis-
77 7

eases of the thorax, including first pass perfusion imaging, additional functional measurements, such as assessments
high-resolution steady state MRA, and MR venography. of cardiac function in the right side of the heart and MR
Due to its intravascular character, Vasovist could also venography to detect the potential source of emboli. A
be ultimately implemented for the calculation of true similar protocol would be ideally suited for patients with
quantitative parameter maps of parenchymal perfusion. pulmonary arterial hypertension, reliably differentiating
The higher available SNR at 3 T and the high T1 relaxiv- chronic thrombembolic forms of the disease from other
ity of Vasovist could efficiently support highly accelerated causes of PAH or including an assessment of the treat-
parallel acquisition and generate an optimized combina- ment effects of vasodilative drug therapy in such patients.
tion of temporal and spatial resolution. In certain patient Acquisition techniques will have to be further adapted in
groups, such as subjects with contraindications to iodi- future studies. Perhaps the use of electrocardiographic
nated contrast media, or younger women with a low clini- or respiratory gating techniques can further improve the
cal probability for PE or during pregnancy, Vasovist-MRA image quality that is achievable during steady state MRA
might soon be considered as a radiation-free first-line of the thorax. Future studies will also have to address the
imaging test for acute PE. Furthermore, Vasovist MRA optimization of reconstruction times and interpretation
of the pulmonary vasculature could be complemented by strategies.

Take home messages

Recent technical developments in the field of MR of true quantitative parameter maps of parenchy-
hardware and pulse sequence design have sub- mal pulmonary perfusion.
stantially improved the potential of MRI for the The higher available SNR at 3 T and the high T1
assessment of the pulmonary circulation. relaxivity of Vasovist could efficiently support
The introduction of an intravascular contrast highly accelerated parallel acquisition and generate
agent such as Vasovist brings along a number of an optimized combination of temporal and spatial
potential advantages for imaging the pulmonary resolution.
vasculature, including the prolonged acquisition In certain patient groups, such as subjects with
time during the equilibrium phase of the contrast contraindications to iodinated contrast media, or
agent and its higher relaxivity as compared with younger women with a low clinical probability for
most extracellular contrast agents, enabling the pulmonary embolism or during pregnancy, Vaso-
setup of a comprehensive MR protocol for various vist-MRA might soon be considered as a radiation-
vascular diseases of the thorax, including first pass free first-line imaging test for acute pulmonary
perfusion imaging, high-resolution steady state embolism.
MRA, and MR venography. In patients with pulmonary embolism, Vasovist MR
Due to its intravascular character, Vasovist could venography could be routinely implemented during
also be ultimately implemented for the calculation steady state to detect the potential source of emboli.

References 6. Perrault LP, Edelman RR, Baum RA, Yucel EK, Weisskoff RM, Shamsi
K et al (2003) MR angiography with gadofosveset trisodium for
1. Schoepf UJ, Costello P (2004) CT angiography for diagnosis of peripheral vascular disease: phase II trial. Radiology 229:811820
pulmonary embolism: state of the art. Radiology 230:329337 7. Fink C, Puderbach M, Bock M, Lodemann KP, Zuna I, Schmahl A et
2. Remy-Jardin M, Pistolesi M, Goodman LR, Gefter WB, Gottschalk A, al (2004) Regional lung perfusion: assessment with partially paral-
Mayo JR et al (2007) Management of suspected acute pulmonary lel three-dimensional MR imaging. Radiology 231:175184
embolism in the era of CT angiography: a statement from the 8. Fink C, Bock M, Puderbach M, Schmahl A, Delorme S (2003) Partially
Fleischner Society. Radiology 245:315329 parallel three-dimensional magnetic resonance imaging for the as-
3. Kreitner KF, Ley S, Kauczor HU, Mayer E, Kramm T, Pitton MB sessment of lung perfusion initial results. Invest Radiol 38:482488
et al (2004) Chronic thromboembolic pulmonary hypertension: 9. Ohno Y, Kawamitsu H, Higashino T, Takenaka D, Watanabe H, Van
pre- and postoperative assessment with breath-hold MR imaging Cauteren M et al (2003) Time-resolved contrast-enhanced pulmo-
techniques. Radiology 232:535543 nary MR angiography using sensitivity encoding (SENSE). J Magn
4. Sanz J, Kuschnir P, Rius T, Salguero R, Sulica R, Einstein AJ et al Reson Imaging 17:330336
(2007) Pulmonary arterial hypertension: noninvasive detection 10. Nikolaou K, Schoenberg SO, Brix G, Goldman JP, Attenberger U,
with phase-contrast MR imaging. Radiology 243:7079 Kuehn B et al (2004) Quantification of pulmonary blood flow
5. Fink C, Goyen M, Lotz J (2007) Magnetic resonance angiography and volume in healthy volunteers by dynamic contrast-enhanced
with blood-pool contrast agents: future applications. Eur Radiol magnetic resonance imaging using a parallel imaging technique.
17 [Suppl 2]:B3844.:B38B44 Invest Radiol 39:537545
78 Chapter 7 Pulmonary MRA

11. Nikolaou K, Schoenberg SO, Attenberger U, Scheidler J, Dietrich O, 28. Sodickson DK, McKenzie CA (2001) A generalized approach to
Kuehn B et al (2005) Pulmonary arterial hypertension: diagnosis parallel magnetic resonance imaging. Med Phys 28:16291643
with fast perfusion MR imaging and high-spatial-resolution MR 29. Sitbon O, Humbert M, Simonneau G (2002) Primary pulmonary
angiography preliminary experience. Radiology 236:694703 hypertension: Current therapy. Prog Cardiovasc Dis 45:115128
12. Prince MR, Grist TM, Debatin JF (2003) 3D Contrast MR angiogra- 30. Rizi RR, Saha PK, Wang B, Ferrante MA, Lipson D, Baumgardner J
phy, 3rd edn. Springer Verlag, Berlin Heidelberg New York et al (2003) Co-registration of acquired MR ventilation and per-
13. Fishman AP (1963) Dynamic of the pulmonary circulation. In: fusion images validation in a porcine model. Magn Reson Med
Hamilton WF, editor. Handbook of physiology, Sect 2: Circulation. 49:1318
Washington, DC: American Physiological Society, p 1708 31. Fink C, Risse F, Buhmann R, Ley S, Meyer FJ, Plathow C et al (2004)
14. Rohrer M, Geerts-Ossevoort L, Laub G (2007) Technical require- Quantitative analysis of pulmonary perfusion using time-resolved
ments, biophysical considerations and protocol optimization with parallel 3D MRI initial results. Rofo Fortschr Geb Rontgenstr
magnetic resonance angiography using blood-pool agents. Eur Neuen Bildgeb Verfahr 176:170174
Radiol 17 [Suppl 2]:B712.:B712 32. Ley S, Fink C, Puderbach M, Plathow C, Risse F, Kreitner KF et al
15. Hartmann M, Wiethoff AJ, Hentrich HR, Rohrer M (2006) Initial (2004) [Contrast-enhanced 3D MR perfusion of the lung: appli-
imaging recommendations for Vasovist angiography. Eur Radiol cation of parallel imaging technique in healthy subjects]. Rofo
16 [Suppl 2]:B1523.:B15B23 Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 176:330334
16. Nikolaou K, Kramer H, Grosse C, Clevert D, Dietrich O, Hartmann M 33. Fink C, Puderbach M, Ley S, Plathow C, Bock M, Zuna I et al (2004)
et al (2006) High-spatial-resolution multistation MR angiography Contrast-enhanced three-dimensional pulmonary perfusion mag-
with parallel imaging and blood pool contrast agent: initial experi- netic resonance imaging: intraindividual comparison of 1.0 M
ence. Radiology 241:861872 gadobutrol and 0.5 M Gd-DTPA at three dose levels. Invest Radiol
17. Hui BK, Noga ML, Gan KD, Wilman AH (2005) Navigator-gated three- 39:143148
7 dimensional MR angiography of the pulmonary arteries using 34. Halliburton SS, Paschal CB, Rothpletz JD, Loyd JE (2001) Estima-
steady state free precession. J Magn Reson Imaging 21:831835 tion and visualization of regional and global pulmonary perfusion
18. Wang Y, Rossman PJ, Grimm RC, Riederer SJ, Ehman RL (1996) with 3D magnetic resonance angiography. J Magn Reson Imaging
Navigator-echo-based real-time respiratory gating and triggering 14:734740
for reduction of respiration effects in three-dimensional coronary 35. Hatabu H, Gaa J, Kim D, Li W, Prasad PV, Edelman RR (1996) Pul-
MR angiography. Radiology 198:5560 monary perfusion: qualitative assessment with dynamic contrast-
19. Ahlstrom KH, Johansson LO, Rodenburg JB, Ragnarsson AS, Ake- enhanced MRI using ultra-short TE and inversion recovery turbo
son P, Borseth A (1999) Pulmonary MR angiography with ultras- FLASH. Magn Reson Med 36:503508
mall superparamagnetic iron oxide particles as a blood pool agent 36. Dong Q, Hurst DR, Weinmann HJ, Chenevert TL, Londy FJ, Prince
and a navigator echo for respiratory gating: pilot study. Radiology MR (1998) Magnetic resonance angiography with gadomer-17.
211:865869 An animal study original investigation. Invest Radiol 33:699
20. Abolmaali ND, Hietschold V, Appold S, Ebert W, Vogl TJ (2002) 708
Gadomer-17-enhanced 3D navigator-echo MR angiography of 37. Berthezene Y, Vexler V, Price DC, Wisner-Dupon J, Moseley ME,
the pulmonary arteries in pigs. Eur Radiol 12:692697 Aicher KP et al (1992) Magnetic resonance imaging detection of
21. Nolte-Ernsting C, Adam G, Bucker A, Berges S, Bjornerud A, Gunther an experimental pulmonary perfusion deficit using a macromo-
RW (1998) Experimental evaluation of superparamagnetic iron lecular contrast agent. Polylysine-gadolinium-DTPA40 [published
oxide particles in pulmonary MR angiography (in German). Rofo erratum appears in Invest Radiol (1992) 27:582]. Invest Radiol
Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 168:508513 27:346351
22. Weishaupt D, Hilfiker PR, Schmidt M, Debatin JF (1999) Pulmonary 38. Nael K, Saleh R, Nyborg GK, Fonseca CG, Weinmann HJ, Laub G
hemorrhage: imaging with a new magnetic resonance blood et al (2007) Pulmonary MR perfusion at 3.0 Tesla using a blood
pool agent in conjunction with breath-held three-dimensional pool contrast agent: Initial results in a swine model. J Magn Reson
magnetic resonance angiography. Cardiovasc Intervent Radiol Imaging 25:6672
22:321325 39. Atkinson DJ, Burstein D, Edelman RR (1990) First pass cardiac per-
23. Zheng J, Carr J, Harris K, Saker MB, Cavagna FM, Maggioni F et al fusion: evaluation with ultrafast MR imaging. Radiology 174 (3 Pt
(2001) Three-dimensional MR pulmonary perfusion imaging and 1):757762
angiography with an injection of a new blood pool contrast agent 40. Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P (1999)
B-22956/1. J Magn Reson Imaging 14:425432 SENSE: sensitivity encoding for fast MRI. Magn Reson Med 42:952
24. Fink C, Ley S, Puderbach M, Plathow C, Bock M, Kauczor HU (2004) 962
3D pulmonary perfusion MRI and MR angiography of pulmonary 41. Nael K, Michaely HJ, Kramer U, Lee MH, Goldin J, Laub G et al
embolism in pigs after a single injection of a blood pool MR con- (2006) Pulmonary circulation: contrast-enhanced 3.0-T MR angi-
trast agent. Eur Radiol 14:1291-1296 ographyinitial results. Radiology 240:858868
25. Hatabu H, Gaa J, Kim D, Li W, Prasad PV, Edelman RR (1996) Pul- 42. Michaely HJ, Kramer H, Dietrich O, Nael K, Lodemann KP, Reiser MF
monary perfusion and angiography: evaluation with breath-hold et al (2007) Intraindividual comparison of high-spatial-resolution
enhanced three-dimensional fast imaging steady state precession abdominal MR angiography at 1.5 T and 3.0 T: initial experience.
MR imaging with short TR and TE. AJR Am J Roentgenol 167:653 Radiology 244:907913
655 43. Weiger M, Pruessmann KP, Leussler C, Roschmann P, Boesiger P
26. Berthezene Y, Croisille P, Wiart M, Howarth N, Houzard C, Faure (2001) Specific coil design for SENSE: a six-element cardiac array.
O et al (1999) Prospective comparison of MR lung perfusion and Magn Reson Med 45:495504
lung scintigraphy. J Magn Reson Imaging 9:6168 44. Meaney JF, Goyen M (2007) Recent advances in contrast-enhanced
27. Amundsen T, Torheim G, Kvistad KA, Waage A, Bjermer L, Nordlid magnetic resonance angiography. Eur Radiol 17 [Suppl 2]:B2B6
KK et al (2002) Perfusion abnormalities in pulmonary embolism 45. Rohrer M, Geerts-Ossevoort L, Laub G (2007) Technical require-
studied with perfusion MRI and ventilation-perfusion scintigra- ments, biophysical considerations and protocol optimization with
phy: an intra-modality and inter-modality agreement study. J magnetic resonance angiography using blood-pool agents. Eur
Magn Reson Imaging 15:386394 Radiol 17 [Suppl 2]:B7B12
79 7
46. Nael K, Michaely HJ, Lee M, Goldin J, Laub G, Finn JP (2006) Dy-
namic pulmonary perfusion and flow quantification with MR imag-
ing, 3.0T vs. 1.5T: initial results. J Magn Reson Imaging 24:333-339
47. Kangarlu A, Baertlein BA, Lee R, Ibrahim T, Yang L, Abduljalil AM et
al (1999) Dielectric resonance phenomena in ultra high field MRI.
J Comput Assist Tomogr 23:821-831
48. Rinck PA, Muller RN (1999) Field strength and dose dependence of
contrast enhancement by gadolinium-based MR contrast agents.
Eur Radiol 9:9981004
49. Marzola P, Osculati F, Sbarbati A (2003) High field MRI in preclinical
research. Eur J Radiol 48:165170
50. Fink C, Ley S, Schoenberg SO, Reiser MF, Kauczor HU (2007)
Magnetic resonance imaging of acute pulmonary embolism. Eur
Radiol 17:25462553
51. Meaney JF, Weg JG, Chenevert TL, Stafford-Johnson D, Hamilton
BH, Prince MR (1997) Diagnosis of pulmonary embolism with
magnetic resonance angiography. N Engl J Med 336:14221427
52. Gupta A, Frazer CK, Ferguson JM, Kumar AB, Davis SJ, Fallon MJ et
al (1999) Acute pulmonary embolism: diagnosis with MR angiog-
raphy. Radiology 210:353359
53. Oudkerk M, van Beek EJ, Wielopolski P, van Ooijen PM, Brouwers-
Kuyper EM, Bongaerts AH et al (2002) Comparison of contrast-
enhanced magnetic resonance angiography and conventional
pulmonary angiography for the diagnosis of pulmonary embo-
lism: a prospective study. Lancet 359:16431647
54. Blum A, Bellou A, Guillemin F, Douek P, Laprevote-Heully MC,
Wahl D (2005) Performance of magnetic resonance angiogra-
phy in suspected acute pulmonary embolism. Thromb Haemost
55. Pleszewski B, Chartrand-Lefebvre C, Qanadli SD, Dery R, Perreault
P, Oliva VL et al (2006) Gadolinium-enhanced pulmonary mag-
netic resonance angiography in the diagnosis of acute pulmo-
nary embolism: a prospective study on 48 patients. Clin Imaging
56. Kluge A, Luboldt W, Bachmann G (2006) Acute pulmonary em-
bolism to the subsegmental level: diagnostic accuracy of three
MRI techniques compared with 16-MDCT. AJR Am J Roentgenol
57. Goyen M, Laub G, Ladd ME, Debatin JF, Barkhausen J, Truemmler
KH et al (2001) Dynamic 3D MR angiography of the pulmonary ar-
teries in under four seconds. J Magn Reson Imaging 13:372377
58. Ohno Y, Higashino T, Takenaka D, Sugimoto K, Yoshikawa T, Kawai H
et al (2004) MR angiography with sensitivity encoding (SENSE) for
suspected pulmonary embolism: comparison with MDCT and ven-
tilation-perfusion scintigraphy. AJR Am J Roentgenol 183:9198
59. Sampson FC, Goodacre SW, Thomas SM, van Beek EJ (2007) The
accuracy of MRI in diagnosis of suspected deep vein thrombosis:
systematic review and meta-analysis. Eur Radiol 17:175181
60. Kluge A, Mueller C, Strunk J, Lange U, Bachmann G (2006) Ex-
perience in 207 combined MRI examinations for acute pulmo-
nary embolism and deep vein thrombosis. AJR Am J Roentgenol
61. Sandstede JJ, Krause U, Pabst T, Hoffmann V, Braun H, Kenn W et
al (2000) Deep venous thrombosis and consecutive pulmonary
embolism as the first sign of an ovarian cancer: MR angiography
using an intravascular contrast agent (CLARISCAN). J Magn Reson
Imaging 12:497500
62. Rich S (2000) Primary pulmonary hypertension. Curr Treat Options
Cardiovasc Med 2:135140
63. Fedullo PF, Auger WR, Kerr KM, Rubin LJ (2001) Chronic throm-
boembolic pulmonary hypertension. N Engl J Med 345:1465

Abdominal MRA
Henrik J. Michaely

8.1 Introduction 82

8.2 Technical Considerations for Imaging with Vasovist 82

8.3 Contrast Agent Administration 83

8.4 Clinical Indications 84

8.5 New Diagnostic Possibilities 89

References 91
82 Chapter 8 Abdominal MRA

8.1 Introduction

More than 10 years after the groundbreaking publica-

tion on Gd-enhanced abdominal magnetic resonance
angiography (MRA) by Prince and colleagues, MRA has
established itself as the problem-solving clinical vascular
imaging modality [1, 2]. The previously used MRA tech-
niques, time-of-flight (TOF) MRA and phase-contrast
(PC) MRA [3], have major limitations for imaging of
the abdominal vessels. Today, virtually every scanner in
use meets the hardware requirements for 3D CE-MRA.
A complete examination of the abdominal organs and
vessels can be easily performed in 20 min and yields
high-resolution, high-contrast images of the abdominal
arteries [2]. Especially in patients with suspected renal ar-
tery disease, the widespread use of MRA is a success story
which has been fostered by several factors [4]. In contrast
to the main competitors Doppler-ultrasound (DUS) and
computed tomography angiography (CTA), MRA has
striking advantages. MRA is operator independent, and Fig. 8.1. Oblique coronal 30-mm thin MIP of a first pass MRA in a
8 the resulting three-dimensional images can be post-proc- patient with suspected renal artery stenosis. The spatial resolution
was 1.2 x 1.0 x 1.0 mm3 acquired at 1.5 T. The segmental renal arteries
essed to allow for better visualization of the structures of
can be clearly seen due to the good vascular enhancement and the
interest. The three-dimensional character of the MRA relatively low enhancement of the kidney
and CTA data makes it possible to choose any desired
imaging plane for image interpretation ( Fig. 8.1). Unlike
with CTA, no ionizing radiation has to be employed for from involuntary diaphragmatic random motion [7]. As
image acquisition. Therefore, MRA can also be acquired with spatial resolution, the resulting SNR is proportional
with multiple phases (non-enhanced, arterial-dominant, to the square root of the acquisition time, so that minimal
venous-dominant) without any restrictions. More recent acquisition times also lead to a decrease in SNR.
sequencing techniques allow for time-resolved imaging of 3D CE-MRA is typically performed using a spoiled
contrast-kinetics over a large field-of-view which cannot gradient echo (GRE) sequence acquiring the entire 3D
be achieved with the small field-of-view of DUS or with data set in 2030 s. In order to reduce acquisition times,
CTA due to the tremendous dose that would be adminis- partial Fourier and asymmetric echo techniques are com-
tered to the patient. Other technical advances of the past monly used. Typically, a spatial resolution of at least 1.5
decade, particularly multi-channel MR technology with x 1.5 x 1.5 mm3 is achieved. Yet the spatial resolution
parallel imaging, as well as ultra-high field imaging at of MRA is still significantly inferior to that of DSA by
3.0 Tesla, have substantially expanded the limits of spatial a factor of 35. The current standard 1.5 T MR systems
resolution of MRA [5]. typically allow acceleration factors of 2 to 4. All vendors
offer parallel imaging options for their scanners marketed
as ASSET (GE Healthcare, Little Chalfont, UK), iPAT (Sie-
8.2 Technical Considerations for Imaging mens Medical Solutions, Erlangen, Germany), or SENSE
with Vasovist (Philips Medical Systems, Best, The Netherlands). Using
an acceleration factor of 2 or 3 on a high-end MR system,
Abdominal MRA has to fulfill three fundamental proper- a 3D CE-MRA with a sub-millimeter isotropic resolution
ties: high spatial resolution with isotropic voxels, mini- of 0.9 mm can be acquired within a single breath-hold
mized acquisition times, and good image contrast. Iso- period; this is particularly true at 3.0 T, where the higher
tropic, high spatial resolution is a prerequisite for exact SNR countervails the possible SNR drop [5, 8].
detection and grading of vascular lesions [6]. As the result- All sequences that have been developed for fast ab-
ing signal-to-noise (SNR) ratio is proportional to the voxel dominal MRA during the first pass with minimum TR and
size, small voxels as with high spatial resolution yield TE times can be equally applied to the first pass of Vasovist
a lower SNR. High spatial resolution is combined ideally (Gadofosveset, Bayer Schering Pharma AG, Berlin, Ger-
with isotropic voxels to foster the application lossless refor- many), as the high relaxivity of Vasovist leads to significant
mats. Finally, a short image acquisition is desirable to min- T1 shortening. The T1 of non-enhanced blood is about
imize motion artifacts from bulk patient motion but also 1250 ms, Vasovists R1 at 1.5 T is 19 s1(mmol/l)1 [9].
8.3 Contrast Agent Administration
83 8

Using an amount of 10 ml Vasovist (equalling 2.5 mmol)

Table 8.1. Sequence parameters for a steady state abdomi-
at a flow rate of 2 ml/s will yield a bolus length in the aorta nal VIBE
after the pulmonary passage of roughly 12 s. In a stand-
ard patient with a cardiac output of 6 l/min and hence of Parameter Value
1.2 l in 12 s, this will lead to a Vasovist concentration of TR / TE (ms) 3.8/1.5
2 mmol/l in the abdominal aorta, resulting in a T1 time of Flip angle () 15
blood of 25 ms. As the blood will show a high longitudi-
FOV (mm2) 420 x 90%
nal relaxation, fast imaging approaches with maximal flip
angles to suppress the background tissue effectively and Matrix 320 x 320

higher factors of parallel imaging can be chosen. However, Acquisition time (s) 29
a completely different situation is encountered when in Parallel imaging 2
the steady state. The initial bolus of Vasovist has now Phase Partial Fourier 6/8
been dispersed completely and the Vasovist molecules
are distributed over the entire blood volume of 6 l to yield
an initial steady state concentration of 0.4 mmol/l with given in Table 8.1: This sequence combines prolonged
a resulting T1 of blood of 120 s. Due to the decreased TR times and low flip angles with a high acquired spatial
concentration of Vasovist during the steady state the resolution which is achieved by using parallel imaging
maximally achievable SNR is decreased as well. As the factor 2. The longer TR times allow for decreasing the
longitudinal relaxation of blood is decreased by a factor of readout bandwidth to gain back some signal.
5 compared with the first pass, different imaging strategies These theoretical considerations also hold true for
have to be implemented in the abdomen. The decrease in 3.0 T imaging. The higher baseline SNR at 3.0 T theo-
blood T1 leads to a decrease in the maximally achievable retically allows for increasing the spatial resolution. As
SNR. With regard to MR physics, the signal intensity of a the acquisition time is the limiting factor for abdominal
T1-weighted GRE sequence is dependent on T2*, the flip imaging, the matrix can be increased only if parallel imag-
angle , and the ratio TR/T1. As T2* cannot be altered, the ing with higher factors is utilized. Initial publications on
user has to optimize TR/T1 and in order to optimize the first pass imaging with dedicated 32-element coils report
signal intensity. From model simulations an increase in TR the feasibility of parallel imaging with acceleration factors
and a decrease in can be deduced as a recommendation of up to 6 at 3.0 T during the first pass of extracellular
for steady state imaging [9]. TR values for the steady state contrast agent [11]. At 3.0 T, particular attention must be
ought to be increased to 7 ms; this has been successfully given to the TE, as the in-phase and opposed-phase times
shown to yield very good image quality in the lower legs are halved compared with 1.5 T.
[10]. However, these recommended changes in sequence New sequence techniques such as the Dixon water-fat
parameters cannot easily be adopted for abdominal imag- saturation, which yields water-only and fat-only images,
ing. In abdominal imaging longer image acquisition times may play an important role in the future. While the basic
result in blurring of the distal parts of the renal arteries idea was already presented in 1984 [12], the technique has
and probably also of the mesenteric arteries [7]. Acquir- not become commercially available until now. Apart from
ing a higher spatial resolution, as is potentially feasible homogeneous background suppression, the inherent dif-
during the steady state, lowers the SNR. Overall, the rec- ferentiation of fat and water may be useful for vessel wall
ommendations for steady state imaging cannot be applied imaging. Dixon imaging requires longer TR times, as at
to abdominal imaging. The main hurdle is the limitation least two echoes are acquired. While in the steady state af-
of the image acquisition to a single breath-hold period. ter Vasovist injection the TR times ought to be increased,
So far, no respiratory-triggered T1-weighted sequences Dixon imaging techniques can be applied without further
are commercially available that would allow the acquisi- time penalties ( Fig. 8.2).
tion of high-spatial resolution images of the abdomen.
At his time, there are no solutions to this dilemma. One
potential improvement for abdominal imaging during the 8.3 Contrast Agent Administration
steady state is to employ fat-saturated 3D volume-inter-
polated sequences (LAVA, THRIVE, VIBE) instead of the Vasovist can be injected as a bolus intravenously followed
typically employed sequences FLASH or SPGR. Volume- by a saline chaser. Initial recommendations include a slow
interpolated sequences provide slightly higher SNR due injection rate of 12 ml/s. This relatively slow rate was cho-
to their specific K-space sampling properties and because sen for two reasons. First, the small administered amount
they use fewer asymmetric echoes. Typical sequence pa- of contrast agent is taken into account. With a slow flow
rameters for a steady state sequence (Siemens Avanto rate for injection, the bolus length can be stretched to per-
1.5 T, Siemens Medical Solutions, Erlangen, Germany) are fectly match the duration of central K-space acquisition. It
84 Chapter 8 Abdominal MRA

Fig. 8.2AC. Source images (1.2-mm thin) of 3.0 T steady state VIBE- water-only reconstruction where the fat signal is completely eliminated
MRA with Dixon fat suppression demonstrating the in-phase recon- (star). In the fat-only image the vessels cannot be evaluated. Due to the
struction (A), the water-only reconstruction (B), and the fat-only recon- recommended longer TR-times during the steady state Dixon fat sup-
struction (C). The right renal artery (arrow) can be best visualized on the pression can be applied without additional time penalty

must be kept in mind, however, that the bolus length is pro- As most abdominal MRA-indications focus on the
longed by 68 s during the pulmonary passage. Second, as renal vasculature, this topic is discussed here in detail.
Vasovists relaxivity depends on the albumin interaction, a Renal MRA should be performed with isotropic voxels
slow injection rate is supposed to allow for a higher amount for lossless post-processing including the calculation of
of albumin-bound Vasovist and hence a higher signal dur- curved multiplanar reformats and cross-sectional views.
ing first pass. This thesis has not yet been proven. This allows determination of the area stenosis in patients
From our own experience we can state that there is with RAS. The area stenosis correlates with DSA better
no visual difference in the perceived signal intensity be- than the diameter stenosis [6] and leads to a better inter-
tween first pass images with high flow rate (34 ml/s) of observer agreement ( Figs. 8.4, 8.5). This is particularly
Vasovist and those with a lower flow rate (12 ml/s). But true for intermediate-grade RAS, where the discordance
a slower flow rate yields more images with purely arterial between two readers is greater than with high-grade ste-
signal, while venous return can be seen earlier with high nosis. Fast image acquisition is another mandatory factor
flow rates ( Fig. 8.3). for a successful CE-MRA examination, as there is also
random diaphragmatic motion even during breath hold
[7]. Therefore, the faster the acquisition, the fewer motion
8.4 Clinical Indications artifacts will be introduced by involuntary motion. Fast
image acquisition is also advantageous if the distal renal
All MRA-indications for which standard extracellular con- arteries are the focus of the examination. With longer ac-
trast agents have been used so far can be equally well exam- quisition times and extracellular contrast renal parenchy-
ined with Vasovist. mal enhancement often obscures the distal renal arteries
Due to the smaller amount of contrast agent adminis- [8]. In contrast, Vasovist remains mainly intravascular
tered when using Vasovist the injection rates have to be and hence leads to less parenchymal enhancement. This
adapted. With a typical acquisition time of 25 s with Car- allows for very good depiction of the distal parts of the
tesian K-space sampling a bolus length of 12 s is required. renal arteries during the first pass and considerably good
Assuming an injected volume of 8 ml and a bolus pro- depiction during the steady state. 3.0 T MRA provides
longation of 7 s after the pulmonary passage, an injection higher image contrast and higher spatial resolution [5],
rate of 1.5 ml/s should be chosen to yield a high SNR. An which will probably also increase the sensitivity and spe-
initial study on whole-body MRA, Vasovist was found to cificity of renal MRA.
be equally as good as conventional extracellular contrast Seventeen to 24% of the kidneys are supplied by an ac-
agents [13]. In this study the image quality was rated bet- cessory renal artery [14, 15]. Accessory renal arteries are
ter with Vasovist than with extracellular contrast agents, aberrant branches originating most of the time directly
but the difference was not statistically significant. from the aorta and serving only a small portion of the kid-
8.4 Clinical Indications
85 8

Fig. 8.3. Time-resolved (TREAT) Vasovist-enhanced MRA at 1.5 T tion small vessels (fourth-order pulmonary vessels, segmental renal
(2 x 2 x 5 mm2 with a temporal resolution of 1.6 s/3D volume) after arteries, branches of the liver artery) can be seen. In the late phases a
contrast injection at 2 ml/s. The high relaxivity of Vasovist leads to complete depiction of the portal venous system is obtained
a strong vascular signal. Despite the relatively coarse spatial resolu-

Fig. 8.4AC. A Volume-rendered first pass MRA taken from a time- isotropic resolution a small plaque with low-grade lumen narrowing of
resolved MRA study. A proximal renal artery narrowing on the left side the right renal artery can be seen. Note that the venous enhancement
can be seen. B and C In the 10-mm thin steady state MRA with 1.3-mm does not interfere with diagnostic image reading
86 Chapter 8 Abdominal MRA

Fig. 8.5. A 10-mm thin coronal maximum intensity projection (left) of the slice planes perpendicular to the vessel axis to yield the vessel
of a steady state MRA 10 min after the injection of intravascular area in the stenosis (upper row) and after the stenosis (lower row). From
contrast agent (Gadofosveset) demonstrates a proximal renal artery the vessel areas the area stenosis can be calculated. (Reprinted with
stenosis on the right side. This figure demonstrates the placement permission from [32])

ney. However, in rare cases they can also originate from stenosis. By this means the diameter stenosis can be meas-
other arteries. They are often longer and narrower than ured, which is more robust than the typically used inplane
regular renal arteries, with a lower perfusion pressure stenosis [19] ( Figs. 8.4, 8.5).
and higher vascular resistance [14]. With respect to the Atherosclerotic RAS is only one manifestation of
detection of accessory renal arteries, several studies report systemic atherosclerotic disease. In patients undergoing
accuracies between 80% and 100%. Vasovist-enhanced cardiac catheterization a prevalence of RAS of 39% was
steady state imaging is an ideal means of detecting acces- found, with 14.3% of these RAS being high-grade stenoses
sory renal arteries, as it allows for imaging after a homo- [20]. Also in patients with lower extremity occlusive dis-
geneous distribution of the contrast agent. The possibility ease or with aortic aneurysms a prevalence of RAS of 39%
of covering a large volume of interest is another advantage was reported [21]. Steady state imaging allows extending
of steady state imaging ( Fig. 8.6). the field of view into the pelvis or below to detect other
Renal artery stenosis is the most common cause of manifestations of atherosclerotic disease.
secondary hypertension and is caused by atherosclerotic Fibromuscular Dysplasia (FMD) is a rare cause of RAS
vessel disease in 90% of cases [16]. Most patients are accounting for fewer than 10% of cases [16]. The most
in or beyond their fifth decade and usually also suffer common type of FMD, medial fibroplasia, is characterized
from other manifestations of atherosclerotic disease and by the classical string-of-beads appearance of the affected
hypertension such as CAD, peripheral artery disease, vessel. Detection of FMD requires high spatial resolution
and diabetes. Atherosclerotic disease also affects the ab- and fast acquisition [22]. Due to the rare occurrence of this
dominal aorta and the other main branches that need to disease, no data exist on the value of Vasovist so far.
be incorporated in the diagnostic reading. Renal artery Renal vein thrombosis (RVT) is most common in
stenosis is located mainly at the ostium and can occur in patients with membranous glomerulonephritis, but it also
3278% of patients bilaterally [17]. Most stenoses show occurs in patients with malignant disease. Radiology is
post-stenotic dilatation and they are often marked by an essential to establish the diagnosis. Typical radiographic
excentric position [18]. This is important, as excentric findings of RVT include swelling of the affected kidney,
stenoses may be underestimated depending on the angle low signal intensity in both T1- and T2-weighted se-
of vessel projection used in the analysis. We therefore quences, and compression of the collecting system [23].
strongly recommend obtaining reformats perpendicular Venous collateral vessels as well as dilatation of the left
to the vessel axis at the site of stenosis and distal to the gonadal vein may be seen in chronic RVT [24]. The gold
8.4 Clinical Indications
87 8

Fig. 8.6A,B. A A 40-mm thin coronal MIP in a patient with complex patient in the steady state. The vessel originating from the iliac artery
horseshoe kidney. In the Vasovist-enhanced MRA at least four renal can now be visualized much better (arrow). In this phase the venous
arteries (arrows) can be seen, originating from the aorta as well as from return from the renal pelvis can be clearly seen (star)
the right common iliac artery. B A 40-mm thin coronal MIP of the same

Fig. 8.7A,B. Nutcracker syndrome. A The venous VIBE MRA (20-mm The thin axial MIP (10 mm) demonstrates a filiform left renal vein (ar-
thin MIP, 1.3 mm isotropic resolution) demonstrates a dilated left row) compressed by the superior mesenteric artery. The venous return
testicular vein (arrow) in a patient with suspected renal artery stenosis from the kidney is maintained over the dilated testicular vein (arrow-
which could not be confirmed. The left kidney and ureter reveal initial head). This finding is typical of renal vein compression syndrome, also
contrast enhancement, indicating normal renal excretory function. B known as nutcracker syndrome, which is clinically often not apparent

standard for diagnosis of RVT, conventional renal venog- propriate. Like conventional RVT, tumor-caused renal
raphy, is not popular and is rarely used in clinical routine vein thrombosis can be imaged. Other pathologies of the
[25]. MRA has been shown to be a useful and accurate renal vein include the so-called nutcracker syndrome, a
test for the detection of RVT [24]. It can be assumed that compression of the renal vein by the superior mesenteric
with the use of Vasovist better venous opacification and artery ( Fig. 8.7). In this case the venous return from the
hence a higher diagnostic confidence will be reached. left kidney flows over the dilated testicular/ovarian vein
To detect RVT in the steady state, Vasovist is most ap- into the venous plexus in the pelvis.
88 Chapter 8 Abdominal MRA

Table 8.2. Added value of Vasovist as shown for typical abdominal MRA indications

MRA application Standard approach Added value of Vasovist

Grading of stenotic Determine vessel diameter Characterize vessel wall


Renal donor evaluation Multiple acquisitions necessary: Single acquisition approach feasible, high signal from arteries and
arterial phase, venous phase, veins
excretory phase with decreas-
Robust assessment of renal vein anatomy
ing SNR
Combine dynamic and high-resolution MRA to quantify perfusion after
single contrast agent injection

Dissecting aneurysms Dynamic and high-resolution Dynamic and high-resolution MRA with injection, high relaxivity
MRA with two separate injec- may allow to increase spatial and/or temporal resolution of the
tions dynamic series

Large FoV (e.g. thoracic and abdominal aorta) feasible

EKG-gating feasible

Follow-up of aortic Multi-phase MRA (34 phases) Delayed imaging feasible which seems to be more sensitive to slow
stents of the abdominal aorta flow often seen in endoleaks

8 Detection of venous
Acquisition of limited (n=12)
number of three-dimensional
Multiple volumes with high SNR can be acquired

Imaging can be delayed to allow for sufficient venous return

volumes with decreasing SNR
FoV can be extend when required

Functional vascular Multiple injections required Repeated MRA exams feasible ( Fig. 8.8)

Fig. 8.8AC. MRA in a young patient with abdominal pain. A In stenosis of the celiac axis can be seen. C In the 10-mm thin sagittal
the 30-mm thin coronal MIP of a first pass Vasovist-enhanced MRA MIP of a steady state Vasovist-enhanced MRA during inspiration the
a strong collateral vessel from the superior mesenteric artery to the stenosis resolves completely, confirming the diagnosis of a Dunbar
common liver artery can be seen. B In the 10-mm thin sagittal MIP of syndrome an impingement of the celiac axis caused by the crus of
a first pass Vasovist-enhanced MRA during expiration a high-grade the diaphragm
8.5 New Diagnostic Possibilities
89 8
8.5 New Diagnostic Possibilities Living-donor kidney transplantation is increasingly
used when HLA-compatible relatives are available. Poten-
Vasovist is not only as good as standard extracellular con- tial donors undergo a thorough radiological examination
trast agents it broadens the spectrum of MRA applica- to answer four fundamental questions: Are there any re-
tions and the imaging strategy significantly. nal masses preventing transplantation? How many renal
As high spatial resolution imaging can be performed arteries and veins are there? What is the course of these
in the steady state, a missed bolus still allows for diag- vessels? How many ureters originating from one or two
nostic image reading. This implies that the first pass, the renal pelvises can be seen?
inherent dynamic information of the bolus injection, can Potential renal donor examination is done today with
be used to acquire time-resolved functional MRA data CT-angiography (CTA) [29] or MRA [30]. CT has the
( Fig. 8.3). Two approaches should be considered. One advantages of being fast, robust, and cheap. However, it
is to use the time-resolved MRA to seek target lesions. is associated with a considerable dose of ionizing radia-
These target lesions can then be exactly characterized in tion, because three phases normally have to be acquired:
steady state measurements. Another approach is to use an arterial-dominant phase, a portal-venous phase and a
the time-resolved data to quantify the renal or pulmonary delayed phase. MR imaging, including MRA, is typically
perfusion parameters [26, 27]. Either way, new, previously performed in a similar way with acquisition of multiple
not-available information can be gathered after a single phases. Steady state imaging with Vasovist facilitates in
injection of Vasovist. these patients the detection and the description of the
Looking back at the RADISH study, disappointing course of the renal arteries and the renal veins. As the
results were reported for the specificity and sensitivity non-albumin-bound fraction of Vasovist is excreted as
of renal MRA regarding the detection and grading of quickly as conventional extracellular contrast agents and
renovascular disease [28]. Among the potential sources of Vasovist has a high relaxivity, sufficient enhancement
error were poor patient selection and a wide range of par- from the renal pelvis and the ureters can be achieved
ticipating institutions, from university hospitals to rural after 5 min. Therefore, with Vasovist theoretically a
hospitals. It seems likely that if Vasovist had been used, a single MRA acquisition can answer all relevant questions
smaller number of MRA examinations would have been ( Fig. 8.9).
non-diagnostic for two reasons. First, a non-diagnostic In patients who have undergone endovascular ther-
MRA examination virtually can no longer occur. Second, apy of an abdominal aortic aneurysm with stent place-
the concept of identifying a target lesion and closely ment, the long-term success of the stent depends on the
examining the target lesions in the steady state makes it presence of endoleaks. Endoleaks are typically located
possible to fit any MRA exam closely to the clinical re- at the proximal (type Ia) or distal end (type Ib) of the
quirements individually. stent and caused by a non-perfect fit of the stent to the

Fig. 8.9AC. (A) arterial, (B) venous, and (C) late-phase MRA of a aneurysm (arrow) are also seen with good signal. Ten minutes after
young patient who was evaluated as a potential renal donor. Arterial the initial contrast injection a strong enhancement of the renal pelvis
and venous vessels can be clearly delineated. Due to the high relax- and of the ureters is present, allowing for an exact assessment of the
ivity of Vasovist, the distal renal arteries and a small splenic artery excretion
90 Chapter 8 Abdominal MRA

8 Fig. 8.10. On the coronal 30-mm thin MIP of a venous-phase MRA (13) a filiform lumen of the inferior vena cava (arrow in 1), occlusion
multiple venous collaterals in a patient with Ormonds disease caused of the right common iliac vein (arrow in 2), and distal reconstitution of
by occlusion of the inferior vena cava can be seen in the retroperito- the external iliac vein (arrow in 3) can be seen. Note the large number
neum, particularly on the right side. On three axial thin-section cuts of small retroperitoneal collateral vessels in all three axial images

abdominal vessel wall. Retrograde flow over the lumbar hancement properties of different kinds of plaques (fatty
arteries or over the inferior mesenteric artery (type II) is vs. fibroproliferative). If this information were clinically
another form of endoleak. Endoleaks type III and IV are reliable Vasovist would make a perfect match for whole-
caused by a defect in the stent fabric or by stent fabric body MRA to detect and characterize atherosclerotic
porosity. CTA and MRA are most often used for fol- vessel disease.
low-up of patients post aortic stent repair. CTA has the Venous imaging reaches a new level of quality with
disadvantage of being non-dynamic, and hence yielding Vasovist. With conventional extracellular contrast
only a single phase. MRA can be acquired with multiple agents the imaging window for venous imaging stud-
phases, but the successful acquisition of delayed images ies is very short and the resulting image contrast often
is not feasible, as conventional extracellular contrast poor. Also the timing for venous studies is very difficult,
agents are excreted very quickly. Using Vasovist delayed so that the overall results are poor. The weak protein-
images with high vascular signal and also high signal binding contrast agent Gd-BOPTA allows for slightly
from endoleaks can be expected. In an initial study early better venous imaging than the other extracellular con-
steady state images were subtracted from delayed steady trast agents, but dedicated venous studies with high
state images acquired after 1 h [31]. As the vascular SNR are not feasible with Gd-BOPTA either. Due to the
contrast remains almost unchanged over the course of strong albumin interaction, Vasovist is perfectly suited
1 h, the vessel signal can be completely removed by the for venous imaging.
subtraction. The remaining paravascular signal then can Conventional MRA sequences or steady state opti-
be clearly attributed to slow-flowing blood in endoleaks. mized sequences can be applied with high venous SNR.
Even though there is no large amount of clinical data Indications for venous MRA are suspected thrombi or
available for this approach it seems very attractive, in vein occlusion with collateral flow ( Fig. 8.10), which
that late image acquisition clearly also demonstrates often occurs post radiation therapy, in patients with Or-
slow flow. monds disease, or those with retroperitoneal tumors.
Initial reports suggest that Vasovist leads to a good
enhancement of the vessel wall during the steady state.
Thus atherosclerotic plaque can be easier marked off
the vessel wall. No published data exist so far on the en-
91 8
Take home messages
Vasovist MRA of the abdominal vessels can Dynamic first pass MRA can be combined with
be performed with good results during the first high-spatial resolution steady state imaging over an
pass. extended field of view.
Image quality will be typically at least equal to or The optimal sequence for abdominal steady state
better than that with conventional extracellular imaging has not yet been determined.
contrast agents. As a good compromise between spatial resolution,
In addition, the unique pharmacodynamic proper- signal-to-noise ratio, and image acquisition time,
ties of Vasovist foster the acquisition of compre- volume-interpolated 3D gradient-echo sequences
hensive MR examinations. can be applied.

15. Bude RO, Forauer AR, Caoili EM, et al (2003) Is it necessary to study
References accessory arteries when screening the renal arteries for renovas-
cular hypertension? Radiology 226:411416
1. Prince MR, Narasimham DL, Stanley JC, et al (1995) Breath-hold 16. Safian RD, Textor SC (2001) Renal-artery stenosis. N Engl J Med
gadolinium-enhanced MR angiography of the abdominal aorta 344:431442
and its major branches. Radiology 197:785792 17. Klatte EC, Worrell JA, Forster JH, et al (1971) Diagnostic criteria of
2. Michaely HJ, Dietrich O, Nael K, Weckbach S, Reiser MF, Schoen- bilateral renovascular hypertension. Radiology 101:301304
berg SO (2006) MRA of abdominal vessels: technical advances. Eur 18. Prince MR, Schoenberg SO, Ward JS, et al (1997) Hemodynamically
Radiol 16:16371650 significant atherosclerotic renal artery stenosis: MR angiographic
3. Debatin JF, Spritzer CE, Grist TM, et al (1991) Imaging of the renal features. Radiology 205:128136
arteries: value of MR angiography. AJR Am J Roentgenol 157:981 19. Schoenberg SO, Rieger J, Weber C, et al (2004) High-resolution
990 MRA of the renal arteries using parallel acquisition techniques:
4. Michaely HJ, Schoenberg SO, Rieger JR, Reiser MF (2005) MR angi- value of isotropic cross-sectional reformats compared to digital
ography in patients with renal disease. Magn Reson Imaging Clin subtraction angiography and intravascular ultrasound. In: Proc
N Am 13:131151 of the International Society for Magnetic Resonance in Medicine.
5. Michaely HJ, Kramer H, Nael K, et al (2007) Intraindividual com- Kyoto: International Society for Magnetic Resonance in Medicine,
parison of high-spatial resolution abdominal MRA at 1.5T and p 235
3.0T. Radiology 244:907913 20. Buller CE, Nogareda JG, Ramanathan K, et al. (2004) The profile
6. Schoenberg SO, Rieger J, Weber CH, et al (2005) High-spatial- of cardiac patients with renal artery stenosis. J Am Coll Cardiol
resolution MR angiography of renal arteries with integrated paral- 43:1606-1613
lel acquisitions: comparison with digital subtraction angiography 21. Olin JW, Melia M, Young JR, et al (1990) Prevalence of atheroscle-
and US. Radiology 235:687698 rotic renal artery stenosis in patients with atherosclerosis else-
7. Vasbinder GBC, Maki JH, Nijenhuis RJ, et al (2002) Motion of the where. Am J Med 88:46N-51N
distal renal artery during three-dimensional contrast-enhanced 22. Michaely HJ, Nael K, Schoenberg SO, et al. (2005) The feasibility of
breath-hold MRA. J Magn Reson Imaging 16:685696 spatial high-resolution magnetic resonance angiography (MRA) of
8. Michaely HJ, Herrmann KA, Kramer H, et al (2006) High-resolution the renal arteries at 3.0 T. Rofo 177:800804
renal MRA: comparison of image quality and vessel depiction 23. Kim SH, Byun HS, Park JH, et al (1994) Renal parenchymal ab-
with different parallel imaging acceleration factors. J Magn Reson normalities associated with renal vein thrombosis: correlation
Imaging 24:95100 between MR imaging and pathologic findings in rabbits. AJR Am
9. Hartmann M, Wiethoff AJ, Hentrich HR, et al (2006) Initial imag- J Roentgenol 162:13611365
ing recommendations for Vasovist angiography. Eur Radiol 16 24. Tempany CM, Morton RA, Marshall FF (1992) MRI of the renal
[Suppl 2]:B1523 veins: assessment of nonneoplastic venous thrombosis. J Comput
10. Nikolaou K, Kramer H, Grosse C, et al (2006) High-spatial-resolution Assist Tomogr 16:929-934
multistation MR angiography with parallel imaging and blood 25. Kanagasundaram NS, Bandyopadhyay D, Brownjohn AM, et al
pool contrast agent: initial experience. Radiology 241:861872 (1998) The diagnosis of renal vein thrombosis by magnetic reso-
11. Fenchel M, Nael K, Deshpande VS, et al (2006) Renal magnetic nance angiography. Nephrol Dial Transplant 13:200202
resonance angiography at 3.0 Tesla using a 32-element phased- 26. Michaely HJ, Nael K, Schoenberg SO, et al (2006) Renal perfusion:
array coil system and parallel imaging in 2 directions. Invest Radiol comparison of saturation-recovery TurboFLASH measurements
41:697703 at 1.5T with saturation-recovery TurboFLASH and time-resolved
12. Dixon WT (1984) Simple proton spectroscopic imaging. Radiology echo-shared angiographic technique (TREAT) at 3.0T. J Magn
153:189194 Reson Imaging 2006;24:1413-1419
13. Klessen C, Hein PA, Huppertz A, et al (2007) First pass whole-body 27. Nael K, Michaely HJ, Lee M, et al (2006) Dynamic pulmonary
magnetic resonance angiography (MRA) using the blood-pool perfusion and flow quantification with MR imaging, 3.0T vs. 1.5T:
contrast medium gadofosveset trisodium: Comparison to gado- initial results. J Magn Reson Imaging 24:333339
pentetate dimeglumine. Invest Radiol 42:659664 28. Vasbinder GB, Nelemans PJ, Kessels AG, et al (2004) Accuracy of
14. Gupta A, Tello R (2004) Accessory renal arteries are not related computed tomographic angiography and magnetic resonance
to hypertension risk: a review of MR angiography data. AJR Am J angiography for diagnosing renal artery stenosis. Ann Intern Med
Roentgenol 182:15211524 141:674682
92 Chapter 8 Abdominal MRA

29. Raman SS, Pojchamarnwiputh S, Muangsomboon K, et al (2007)

Surgically relevant normal and variant renal parenchymal and vas-
cular anatomy in preoperative 16-MDCT evaluation of potential
laparoscopic renal donors. AJR Am J Roentgenol 188:105114
30. Kock MC, Ijzermans JN, Visser K, et al (2005) Contrast-enhanced
MR angiography and digital subtraction angiography in living
renal donors: diagnostic agreement, impact on decision making,
and costs. AJR Am J Roentgenol 185:448456
31. Cornelissen SA, Prokop M, Adriaensen ME (2007) Visualizing slow-
flow endoleak after endovascular abdominal aortic aneurysm
repair with the new blood pool agent Vasovist. In: Proc of the
International Society for Magnetic Resonance in Medicine. Berlin:
International Society for Magnetic Resonance in Medicine
32. Michaely HJ, Kramer H, Attenberger U, et al (2007) Renal magnetic
resonance angiography at 3.0 T: technical feasibility and clinical
perspectives. Top Magn Reson Imaging 18:117125


MRA of the Peripheral Vasculature

Tim Leiner and Jeffrey H. Maki

9.1 Introduction 94

9.2 Technical Considerations and Imaging Protocols 94

9.2.1 Vasovist Versus Conventional Contrast Media 94
9.2.2 Considerations with Regard to Vasovist-enhanced Peripheral MRA Imaging Protocols 95
9.2.3 Vasovist Injection Protocol 96

9.3 Clinical Indications 96

9.3.1 Upper Extremities 96
9.3.2 Lower Extremity CE-MRA 102

9.4 Diagnostic Accuracy of MRA for Evaluation of Peripheral Arterial

Occlusive Disease 112

9.5 Vasovist-enhanced Imaging 112

9.6 Conclusions 112

94 Chapter 9 MRA of the Peripheral Vasculature

9.1 Introduction of how Vasovist can be used to improve imaging of the

peripheral vasculature.
Since the first description of moving-bed contrast-en-
hanced magnetic resonance angiography (CE-MRA) in
1998 [1], the technique has seen numerous refinements 9.2 Technical Considerations and Imaging
and is now widely applied in clinical practice throughout Protocols
the world. The most widely used technique is to obtain a
luminogram during first arterial passage of an extracellular 9.2.1 Vasovist Versus Conventional Contrast
MR contrast agent, often in combination with background Media
subtraction to improve vessel-to-background contrast.
There are several important boundary considerations Vasovist can be used exactly the same way as extracel-
when performing peripheral MRA. Because of the inher- lular agents with regard to first pass imaging, as has been
ently slow MR image acquisition process it is very important demonstrated by Klessen et al. [3]. The advantage of using
to carefully synchronize first arterial passage of the contrast this agent for first pass imaging lies in its much higher
medium with acquisition of the center K-space lines. Also, relaxivity (see also the chapter by Rohrer). This means
one always has to make a trade-off between acquisition that a higher signal-to-noise ratio can be obtained when
duration, spatial resolution, and the capability of the patient parameters are kept identical or, conversely, that spatial
to remain motionless and to sustain a breath hold. Further- resolution can be increased while the same signal-to-
more, manual table movement was needed in the earliest noise ratio is maintained. The truly interesting property of
iterations of the technique, whereas today all major MR blood pool agents in general, however, is the much longer
vendors market hard- and software that allows for easy and intravascular residence time. Equilibrium phase imaging
near full-automated workflow once the 3D volumes are in is possible because, despite the fact that dilution of the
the desired position in relation to the vasculature. injected contrast medium after first arterial passage leads
9 A major improvement that challenges these traditional to a T1 increase of the blood pool compared with the first
conceptions about CE-MRA is the recent clinical intro- pass, the value is still much lower than that of fat. Hart-
duction of the first blood pool agent, Vasovist (Gadofos- mann et al. estimated that T1 of blood in the equilibrium
veset, Bayer Schering Pharma AG, Berlin, Germany). First phase, 35 min after injection of 0.03 mmol/kg Vasovist is
described in 1998 [2], Vasovist came on the market in about 130 ms, increasing to about 150 ms after 1015 min
the European Union in 2005 and has since expanded the [4]. This prolonged T1 reduction offers the opportunity to
diagnostic armamentarium of the radiologist, not only by obtain images of the peripheral vascular tree up to about
providing superior first pass imaging, but also by enabling 4560 min after injection. The extended imaging window
ultra-high spatial resolution equilibrium-phase imaging of can be used to acquire images with much higher spatial
both arteries and veins at previously unseen spatial resolu- resolution without a significant loss of vessel-to-back-
tions that come very close to that of the gold standard in ground contrast ( Fig. 9.1). In clinical practice this means
vascular imaging, conventional intra-arterial X-ray-based that scan duration is no longer determined by the transient
digital subtraction angiography (IA-DSA). T1 shortening, but by the capacity of the patient to sustain
This chapter discusses some of the theoretical consid- a breath hold or to remain motionless. An exciting work-
erations and provides practical advice about and examples in-progress that is likely to further improve image quality

Fig. 9.1. Source images of equilibrium-

phase acquisitions at three different iso-
tropic spatial resolutions demonstrating
progressively improving visualization of
the right anterior tibial artery surrounded
by accompanying veins (arrowheads). Left
panel: 1.0 x 1.0 x 1.0 mm3 (1000 m3); mid-
dle panel: 0.75 x 0.75 x 0.75 mm3 (422 m3);
right panel: 0.5 x 0.5 x 0.5 mm3 (125 m3)
resolution. Note relative preservation of
vessel-to-background enhancement de-
spite the eightfold decrease in voxel size
9.2 Technical Considerations and Imaging Protocols
95 9

is the combination of navigator gating with equilibrium- 9.2.2 Considerations with Regard to
phase imaging of the abdominal aorta and its branches. By Vasovist-enhanced Peripheral MRA
synchronizing the acquisition to end-expiration by means Imaging Protocols
of tracking diaphragmatic motion it will be possible to ex-
tend imaging to several minutes as well ( Fig. 9.2). The MR sequence parameters used in conventional first
The apparent drawback of using a blood pool agent is pass CE-MRA are not suitable for equilibrium phase im-
the simultaneous enhancement of venous structures close aging, as short repetition (TR) and echo times (TE), high
to arteries. This phenomenon is a well-known problem at bandwidths, high flip angles, and relatively low spatial
first pass imaging, often resulting in images that cannot resolutions are used to achieve short scan times aimed at
be used for clinical decision-making. However, because obtaining a selective arterial phase within a breath hold.
equilibrium phase images can be acquired at much higher Imaging in the equilibrium-phase places other de-
spatial resolution, often with a 5- to 15-fold decrease in mands on the imaging protocol. For instance, much
voxel size compared with first pass protocols, arteries can higher spatial resolution is needed in order to reliably
be readily separated from accompanying veins. Empirically, separate arteries from veins. However, higher resolution
the authors recommend that the acquisition for the iliac ar- comes at the expense of signal-to-noise ratio (SNR),
teries last no longer than 20 s, because usually a breath hold which must remain sufficient to provide adequate vessel-
is desired. When imaging the vascular system in the pelvis to-background contrast. Since there is no requirement to
only, no breath hold is needed and the acquisition can be complete the acquisition in the period of first arterial pas-
extended to a few minutes. For the upper and lower legs we sage of contrast material, significantly longer acquisition
recommend that acquisitions last no longer than 5 min. times can be used in the equilibrium phase to compensate


Fig. 9.2AC. Whole-volume MIP of first pass acquisition (A) in a

35-year-old patient suspected of having renovascular hypertension.
There is slight venous enhancement of the left renal vein and inferior
vena cava (resolution was 1.2 x 1.0 x 2.2 mm3). Cardiac-triggered and
respiratory navigator-gated equilibrium-phase acquisition (B) demon-
strates improved sharpness of the vessel borders compared with the
first pass images. Resolution was 1.0 x 1.0 x 1.0 mm3. Zoomed detail of
the right renal artery confirms high-fidelity depiction of the renal ar-
tery in the equilibrium-phase (C, right panel) compared with first pass
C (c, left panel). Again, note sharper delineation of the right renal artery
(arrow) and aorta (arrowheads)
96 Chapter 9 MRA of the Peripheral Vasculature

for the loss in SNR due to increased spatial resolution. 9.3 Clinical Indications
Both in vitro and in vivo work by our groups with regard
to equilibrium-phase sequence parameter optimization 9.3.1 Upper Extremities
has demonstrated that optimal image quality demands an
increase of TR from to 24 to about 1012 ms, a TE below Anatomical and Physiological Considerations
5 ms, and a decrease in flip angle from about 3540 to With the elbow extended, the upper extremity of a 170-cm-
2029 degrees, depending on the type of coil and parallel tall adult is about 75 cm long. This implies that at least two
imaging factor used [5]. These parameters are optimized fields-of-view (FOV) of about 40 cm in the craniocaudal
for an equilibrium-phase voxel size of 0.5 x 0.5 x 0.5 mm3 direction are needed to depict all the relevant arteries. A
(125 m), which provides sufficiently high spatial resolu- FOV of 2025 cm suffices if the main area of interest is the
tion to clearly distinguish arteries from veins at still rea- distal forearm and hand. The diameter of the subclavian
sonable imaging times. artery is about 1.01.3 cm, and the diameter of the axillary
Practically, we recommend an imaging protocol that artery is about 68 mm. Distally, the arteries progressively
consists of time-of-flight localizer images, followed by taper to a diameter of about 12 mm at the level of the
non-enhanced mask images. Subsequently, we use real- palmar arch in the hand. It is of paramount importance
time bolus tracking of the injected contrast material, that depiction of the arteries in the distal forearm and
which is immediately followed by high-resolution first hand requires higher spatial resolution compared with
pass imaging of the pelvic, upper leg, and lower leg sta- more proximal vessels. The preferred voxel size is on the
tions using moving-table software. After the first pass order of 1 mm3 or even smaller. While this is not easily
acquisitions are completed, equilibrium-phase images can achieved in combination with first pass imaging, acquisi-
be obtained using the relatively minor modifications in tion in the equilibrium-phase facilitates ultra-high spatial
imaging parameters as described above, yielding high- resolution images.
quality ultra-high spatial resolution equilibrium-phase The mean time of selective arterial opacification in the
9 images of the entire peripheral vascular tree. We recom- upper extremity is relatively short. Winterer et al. found
mend limiting the 125-m equilibrium-phase acquisition a mean bolus arrival time of 27 s measured at the level of
to the upper and lower leg stations. For the aortoiliac the terminal radial and ulnar arteries and a mean time
station we use the same imaging parameters as in the first window of selective arterial enhancement of 1114 s [6].
pass acquisition in order to combine the acquisition with However, large interindividual differences in contrast me-
breath holding. dium circulation times were found. The period between
When imaging the upper extremity a very similar arterial and venous enhancement can be substantially
imaging protocol in terms of sequences and spatial resolu- lengthened by applying supra- or sub-systolic compres-
tion can be used. Voxel sizes down to 0.4 x 0.4 x 0.4 mm3 sion (>200 mmHg) just proximal to the wrist with an MR-
are feasible. compatible blood pressure cuff. This effectively increases
the period of selective arterial enhancement and hence
the imaging window needed to achieve the ultra-high spa-
9.2.3 Vasovist Injection Protocol tial resolution required for optimal evaluation of arterial
occlusive disease in the small hand vessels [7, 8].
The approved clinical dose of Vasovist is 0.03 mmol/kg
body weight. Considering that the drug comes in 10-ml
vials at a concentration of 0.25 mol/l, the typical 75-kg Imaging Protocol and Parameters
adult would receive a dose of 9 ml. Conversely, a single High-quality first pass upper extremity CE-MRA de-
vial can be used for patients up to 83.3 kg. Considering mands large FOV imaging with high spatial resolution,
the rather small volume of the contrast injection com- short imaging times, and dedicated multi-element paral-
pared with regular extracellular chelates, we recommend lel imaging capable surface coils. In combination with a
an injection rate of 1.0 ml/s at the highest in order to multiphase acquisition a timing bolus is not needed. This
avoid having too short a bolus. This is especially perti- is advantageous because of the small volume of contrast
nent when imaging the lower extremity vasculature in a agent that is available.
multiple station imaging protocol because of the desire A multiphase acquisition ensures adequate synchroni-
to achieve uniform vascular opacification over the en- zation of peak arterial contrast enhancement with acquisi-
tire duration of the first pass acquisition. The contrast tion of the center of K-space. Alternatively, real-time bo-
injection should be followed by at least 2530 ml of lus visualization software (e.g. BolusTrak [Philips Medical
saline chaser injected at the same rate, in order to avoid Systems]; SmartPrep [GE Health] or CareBolus [Siemens
pooling of the contrast medium in upper extremity Medical Solutions]) can be used [911]. In Tables 9.1
veins. and 9.2, recommended sequence parameters are listed for
Table 9.1. Recommended MR sequence parameters for first pass and equilibrium-phase MRA of the aortic arch and proximal upper extremity

Sequence Orientation Coverage TR/TE FA FOV Matrix Slice thickness* No. of Duration (s) Comments
(ms) (degrees) (mm) (mm) partitions

bSSFP Coronal Aortic arch to distal bra- 6.0 / 3.0 90 430 x 430 400 x 256 510 3060 3060
chial artery
9.3 Clinical Indications

2D PCA Transverse Neck / carotid and verte- 20 / 5 15 260 x 260 256 x 256 10 1 3060 To detect subclavian ste-
bral arteries al set VENC to 120 cm/s

3D GRE (first Coronal Aortic arch to distal bra- <4.0 / <2.0 3040 430 x 300 430 x 300 1.02.0 Tailor to 4050 (2 dyn) Use parallel imaging
pass) chial artery anatomy

3D GRE (equili- Sagittal Upper extremity with <12.0 / 20 385 x 270 550 x 386 0.7 Tailor to Several min Use parallel imaging
brium-phase) suspected abnormality. <3.0 anatomy
Include aortic arch

bSSFP balanced steady state free precession; TR repetition time; TE echo time; FA flip angle; FOV field-of-view (frequency x phase); BW band width; NSA number of signals averaged (number below 1 indicates
partial or fractional echo); dyn dynamic scans; * non-interpolated, truly acquired slice thickness

Table 9.2. Recommended MR sequence parameters for first pass and equilibrium-phase MRA of the forearm and hand

Sequence Orientation Coverage TR/TE FA FOV Matrix Slice thickness* No. of Duration (s) Comments
(ms) (degrees) (mm) (mm) partitions

bSSFP Coronal Distal brachial artery to 6.0 / 3.0 90 430 x 430 400 x 256 510 3060 3060

3D GRE (first Coronal Distal brachial artery to <4.0 / <2.0 3040 430 x 300 430 x 300 1.0 Tailor to 4050 Use parallel imaging
pass) digits anatomy (2 dyn)

3D GRE (equili- Sagittal Upper extremity with <12.0 / 20 385 x 270 770 x 540 0.5 Tailor to Several min Use parallel imaging
brium-phase) suspected abnormality. <3.0 anatomy
Include aortic arch

bSSFP balanced steady state free precession; TR repetition time; TE echo time; FA flip angle; FOV field-of-view (frequency x phase); BW band width; NSA number of signals averaged (number below 1 indicates

partial or fractional echo); dyn dynamic scans; * non-interpolated, truly acquired slice thickness
98 Chapter 9 MRA of the Peripheral Vasculature

localizer and timing scans as well as high spatial resolu- with that of the lower extremities. The most frequent
tion 3D CE-MRA and equilibrium-phase imaging of the site of occlusion is the brachiocephalic trunk or the
upper extremity arteries. proximal subclavian artery. In case of occlusion collat-
When information is needed about both the proximal eral blood supply is established via the side branches of
and distal upper extremity arteries, two FOVs need to be the subclavian artery and the axillary artery. Subclavian
imaged. This can be done either with a single injection of artery stenosis or occlusion is not infrequently accompa-
contrast medium in combination with table movement nied by retrograde collateral blood flow in the ipsilateral
in between acquisitions, or by performing two separate vertebral artery. This is also known as subclavian steal.
acquisitions with separate injections of half the total Retrograde flow can be depicted on the 2D time-resolved
amount of contrast medium each. Owing to the short cir- test-bolus scan or with non-enhanced time-of-flight or
culation times in the upper extremity, the latter approach phase-contrast MRA acquisitions [1315]. Although in
is recommended because it usually leads to better results, general equilibrium-phase acquisitions are not needed to
especially when the forearm and hand are imaged first. establish the diagnosis in the proximal arterial tree, they
Another important caveat of upper extremity CE-MRA are very useful for the depiction of abnormalities in the
is to inject contrast medium on the asymptomatic side venous system ( Fig. 9.3).
in order to avoid susceptibility-induced intra-arterial
signal loss [12.]. When both upper extremities need to Vasculitis
be imaged it is advisable to inject contrast medium on The vasculitides regularly affect upper extremity arter-
the right because of the greater distance of the right ies. Wall thickening and enhancement on post-contrast
brachiocephalic vein and the superior vena cava to the T1-weighted imaging as well as central aneurysms with-
large arteries. out a history of trauma or infection suggest a vasculitis
[16, 17]. Vasculitides affecting the subclavian artery are
Takayasus disease and Behets syndrome. The axillary
9 Clinical Applications of Upper Extremity MRA artery is involved in giant cell arteritis and systemic lupus
CE-MRA is a powerful technique for the diagnosis of, erythematosus (SLE). Thromboangiitis obliterans (TAO;
pre-interventional treatment planning and post-interven- Buergers disease) often specifically affects the forearm
tional follow-up of a broad spectrum of congenital and ac- arteries in patients with this disease. The angiographic
quired upper extremity occlusive disease ( Table 9.3). It is hallmark of TAO is widespread arterial occlusion with
important to keep in mind that sub-millimeter isotropic numerous corkscrew collaterals [18, 19]. Arteries of the
resolutions are necessary to reliably depict distal upper hand are discussed below.
extremity arterial disease. The use of Vasovist facilitates
both high-quality first pass imaging as well as exquisitely Aneurysmal disease
detailed images of the small vessels in the hand in the Aneurysms of upper extremity arteries are frequently
equilibrium phase. secondary to iatrogenic arterial trauma (e.g., arterial cath-
eterizations or dialysis access fistula creation) or as the
Atherosclerotic occlusive disease result of blunt or penetrating trauma to the chest. Less
Although atherosclerosis is the most common cause of frequent causes include vasculitis and hematogenous in-
upper extremity arterial stenosis in the large arteries fection (mycotic aneurysm). Because the first pass con-
down to the wrist, its incidence is much lower compared trast-enhanced portion of the exam provides only a lu-

Table 9.3. Indications for Vasovist-enhanced MRA of the upper extremity

Indication Advantages of using Vasovist

Atherosclerotic occlusive disease of aortic High relaxivity in F/P; Ultra-high spatial resolution in E/P
arch and branches

Aneurysmal disease High relaxivity in F/P; Definition of mural thrombus in E/P

Vasculitis High relaxivity in F/P; Ability to depict vessel wall in E/P

Suspected thoracic outlet syndrome Ability to depict both arteries and veins

Hand imaging High relaxivity in F/P; Ultra-high spatial resolution and definition of thrombus in E/P

F/P first pass; E/P equilibrium phase

9.3 Clinical Indications
99 9

Fig. 9.3AC. Whole-volume MIPs of dy-

namic acquisition in the arterial (A, left panel)
and early venous phases (A, right panel) in
a 54-year-old patient suspected of having
thoracic outlet syndrome. The subclavian
artery shows no abnormalities. Color volume
renderings (cVR) at different thresholds of
equilibrium-phase acquisitions (0.7 x 0.7 x
0.7 mm3) demonstrate a large superficial
venous collateral in the left anterior oblique
view (B, left panel). When the cVR threshold
is increased, an interruption is suggested in
the subclavian vein (demarcated by arrow-
heads in B, right panel). Review of equilib-
rium-phase source images confirms normal
subclavian artery (C, left panel) and near oc-
clusion of the subclavian vein (demarcated
by arrowheads in C, right panel)
100 Chapter 9 MRA of the Peripheral Vasculature

Fig. 9.4A,B. A 67-year-old patient post kidney transplant with a

history of dialysis access surgery presented with swelling in the left
upper arm. Dynamic MRA shows normal filling of the aortic arch and
branches (A, left two panels), but no filling yet of the distal brachial
artery (asterisk, left panel). The source images demonstrate a cylindri-
cal, partially opacifying mass medial to the humerus (A, arrowheads in
right panel). Equilibrium-phase images (0.7 x 0.7 x 0.7 mm3) in coronal
(B, upper left panel), sagittal (B, upper right panel), and transverse
9 (B, lower panel) orientations confirm this is a partially thrombosed
former dialysis access shunt (arrowheads), incompletely excluded from
the circulation. *, brachial artery; h, humerus

ultra-high spatial resolution (on the order of 0.5 x 0.5

x 0.5 mm3) ensures that arteries can be easily be distin-
guished from veins when thin-slab maximum intensity
projections (MIP) and curved multiplanar reformations
(MPR) are used.
Indications for imaging of the hand are suspected
emboli in patients with atrial fibrillation or atherosclerotic
disease in proximal upper extremity arteries, preopera-
tive imaging in patients undergoing plastic surgery of the
hand, vibration tool users with symptoms of arterial oc-
B clusion in the hand, hypothenar hammer syndrome, and
suspected vasculitis. Connective tissue disorders such as
scleroderma, rheumatoid arthritis, SLE, and antiphos-
pholipid syndrome are associated with multiple occlu-
sions of the small arteries of the hand palm and digits
minogram, an equilibrium-phase T1-weighted acquisition ( Fig. 9.5). Small distal aneurysms are associated with
should always be performed to assess the true size of the polyarteritis nodosa.
aneurysm including any mural thrombus ( Fig. 9.4).
Dialysis access fistulae
CE-MRA of hand arteries Arteriovenous fistulae (AVF) for hemodialysis vascular
Contrast-enhanced MR angiography of peripheral arte- access have a high risk of thrombosis due to stenoses
rial disease in palmar and digital arteries is challenging that jeopardize flow and patency, and early detection and
but possible. Best results are achieved with a dedicated treatment of these stenoses can prevent thrombosis [20,
small FOV coil in the equilibrium phase and the use of 21]. In current clinical practice, stenosis detection and
high parallel imaging factors. Although there usually grading in AVF and the venous outflow segments are
is bright enhancement of venous structures, the use of performed using digital subtraction angiography (DSA).
9.3 Clinical Indications
101 9

Fig. 9.5AC. Comprehensive vascular hand imaging in 43-year-old

female patient with known anti-phospholipid syndrome now present-
ing with digital gangrene. Dynamic acquisitions (A) in the first pass
of a 10-ml bolus of Vasovist demonstrate severely delayed filling of
the radial artery and a complete lack of opacification of digital arter-
ies. Equilibrium-phase acquisitions (0.5 x 0.5 x 0.5 mm3) show both
arteries and veins, but these can be separated visually because of the
ultra-high spatial resolution (B). Whole volume MIP (B, left panel) and
angulated thin-slab MIPs of hand palm (B, middle panel) and fingers
(B, right panel) correspond well with first pass images. Note depiction
of patent proper digital arteries (arrowheads) in the second and third
fingers, not seen on first pass imaging, as well as intra-arterial DSA with
vasodilatation (C). r, radial artery; u, ulnar artery; *, palmar arch
102 Chapter 9 MRA of the Peripheral Vasculature

The arterial part of the AVF is visualized according to the around 1.02.0 mm. In most patients, peripheral arteries
method of Staple, using a proximal cuff to interrupt flow are slightly curved in the anteroposterior direction and
in order to achieve retrograde filling of the arterial part of the coverage needed is usually less than 10 cm. In the
the AVF [22]. Analysis of DSA images acquired with this presence of an aortic aneurysm, iliac arterial elongation,
method can be difficult due to vessel overlap, especially at collaterals bridging iliac or superficial femoral arterial ob-
the level of the anastomoses. Owing to the temporary flow structions, or a femorofemoral crossover bypass graft the
interruption, the hemodynamic situation is altered, which AP coverage needed to depict these vessels may be mark-
potentially limits the value of DSA. Furthermore, because edly increased (up to 1520 cm). Review of the transverse
of incomplete retrograde filling, the feeding artery and localizer images ensures that these structures are not
arterial part of the AVF are not always depicted in their excluded from the 3D CE-MRA imaging volume. This is
entirety [23, 24]. particularly important if a patient has a femorofemoral
Because of the reasons outlined above, the pre-inter- crossover bypass graft because these grafts are usually not
ventional workup in patients with hemodialysis access seen on TOF MIPs due to in-plane saturation artifacts
failure requires imaging from the aortic valve and supe- [28, 29]. Other patients who require special attention are
rior vena cava down to the distal arteriovenous anasto- those with (thoraco-) abdominal aortic aneurysms where
mosis in the elbow or forearm region. The demands of a flow may be markedly slower than in patients without
large FOV combined with the extremely high flow rates aneurysms [30]. If insufficient delay time is observed
in these shunts make parallel imaging particularly useful between injection of contrast and imaging, this will result
for imaging dialysis access fistulae and supplying arteries in incomplete opacification of the aneurysm at the time
and draining veins. Furthermore, 3D CE-MRA provides of imaging. To avoid this problem, either a longer delay
unlimited viewing angles that are very helpful in plan- between injection and start of acquisition, or a multiphase
ning the proper intervention. For instance, Planken et acquisition should be used [31]. In addition, the use of
al. demonstrated that multiphase 3D CE-MRA detected Vasovist enables an equilibrium-phase acquisition, which
9 flow-limiting stenoses in two patients (13%) with failing should always be performed when possible.
hemodialysis access fistulae that were not seen with con- The mean time of selective arterial opacification at
ventional DSA. Only after these stenoses were removed different levels in the lower extremity varies widely. Pa-
did flow improve [25]. tients with a history of heart failure usually exhibit slower
Despite these obvious advantages of MRA, patients on venous return, while patients with ulcers often have an
dialysis are not good candidates to receive gadolinium- extremely short arteriovenous window. Prince et al. dem-
based contrast agents because of the probability of devel- onstrated that after arrival in the common femoral artery
oping nephrogenic systemic fibrosis (NSF). Although NSF (CFA), intravenously injected contrast material travels
has been described mainly in relation to administration of down the peripheral arteries at about 6 s per station. In
Gadodiamide (Omniscan, GE Health, Chalfont St-Giles, 87 patients undergoing time-resolved 2D CE-MRA, the
United Kingdom), this is certainly not the only compound mean travel time of contrast material to the CFA was 24 s,
that has been associated with the disease [26]. with an additional 5 s to reach the popliteal artery and 7 s
to reach the ankle arteries. In the same study they found
that the mean time window of arterial enhancement was
9.3.2 Lower Extremity CE-MRA 49 s in the pelvis, 45 s in the thigh, and 35 s in the calf
Anatomical and Physiological Considerations It was recently found that applying venous compres-
The lower extremity peripheral arterial tree extends from sion could substantially lengthen the period between
the infrarenal aorta down to the feet and has a total arterial and venous enhancement. Suppression of ve-
length of about 90120 cm. To depict all relevant arter- nous enhancement in the distal lower extremity is al-
ies, at least three separate acquisitions are needed: one ready achieved with a blood pressure cuff inflated to
acquisition to cover the aortoiliac arteries, one to cover 5060 mmHg either just proximal to or just distal to the
the superficial femoral and proximal popliteal arteries, knee joint [3335].
and an additional acquisition to cover the distal popliteal,
lower leg, and pedal arteries. As outlined in the section on
upper extremity CE-MRA above, best results are obtained Imaging Protocol and Parameters
when the acquisition is tailored to the arterial dimensions Because of the lower injection rate and hence the lower
in the specific FOV [27]. The diameter of the infrarenal vessel-to-background contrast, non-enhanced mask ac-
aorta is usually around 2.02.5 cm, the diameter of the quisitions are usually acquired prior to the contrast-
superficial femoral artery is about 710 mm, and the ar- enhanced acquisitions in patients undergoing lower
teries at the level of the ankle and foot have diameters of extremity MRA. After the acquisition has ended, the
9.3 Clinical Indications
103 9

non-enhanced images are subtracted from the enhanced first of all, by lowering TR and TE to the shortest possible
images to suppress non-vascular background signal. Mask value without excessively increasing bandwidth. In addi-
scans are not needed for equilibrium-phase acquisitions, tion, partial or fractional echo should be used. When the
as MIP is not the preferred mode of evaluation. 3-consecutive station approach is used it is particularly
Currently, the most commonly used and easiest ap- important to image the first two stations (i.e., aortoiliac
proach to lower extremity CE-MRA is acquisition of three and upper legs) as fast as possible. Once the acquisition
consecutive anatomically tailored imaging volumes dur- of the lower legs is started, this will allow a relatively
ing injection of a fixed dose (0.20.3 mmol/kg) or volume long (and thus high-resolution) scan, given that none
(3045 ml) of 0.5 M gadolinium chelate [1, 36, 37]. or minimal venous enhancement is already present and
Although the injected volume is smaller when using centric K-space filling is used in this station. With the in-
Vasovist, an identical approach can be used for the first troduction of multi-element (peripheral) surface coils and
pass acquisition [3]. The first pass technique is well suited parallel imaging acquisition speed can be increased fur-
for the diagnostic imaging workup of the vast majority ther [3941]. Another way to obtain venous free images
of patients with intermittent claudication. However, it is of the lower legs is to switch from a 3D high-resolution
of paramount importance that inferior image quality due acquisition to a 2D projectional acquisition, analogous to
to insufficient spatial resolution as well as disturbing ve- IA-DSA [42). A disadvantage of this latter method is that
nous enhancement becomes clinically relevant in patients additional views demand separate injections of contrast
with severely diseased lower leg arteries. Specific groups medium.
in whom first pass imaging with a 3-consecutive station To avoid the limitations of imaging three consecu-
approach does not work well are patients with diabetes tive stations, an alternative approach is a dual-injection
mellitus, especially with concomitant arteriovenous mal- protocol in which the lower legs are imaged first and the
formations, who are known to primarily have very distal aortoiliac arteries and upper legs are imaged afterwards
disease [38], patients with cellulitis, and patients with in a separate acquisition. The rationale for this hybrid
chronic critical ischemia who have severe stenoses and approach [43] is that it is easier in this way to obtain
occlusions from the aorta down to the feet (multilevel high-resolution 3D images of the lower leg station free of
disease). disturbing venous enhancement. The initial acquisition
There are several general acquisition strategies that of the lower legs is typically done using up to 1520 ml
can be used to decrease the chance for disturbing venous 0.5 M Gd-DTPA and can be either mono- or multiphasic.
enhancement ( Table 9.4). These are: (a) lowering TR Because synchronization of central K-space lines is opti-
and TE; (b) use of a separate acquisition for the lower mized with regards to contrast enhancement in the lower
leg station; (c) use of centric K-space filling; (d) use of a leg arteries, venous enhancement is virtually eliminated.
time-resolved acquisition strategy (keyhole or TRICKS); After imaging of the lower legs is completed, a moving-
(e) use of infrasystolic venous compression, and (f) use of table acquisition is performed to image the aortoliac and
parallel imaging to amplify the other strategies. The most upper leg arteries using the remaining volume of contrast
straightforward way of preventing venous enhancement is agent. When Vasovist is used, the entire dose of contrast
by shortening acquisition duration. This should be done, agent can be used for the multiphase lower leg station

Table 9.4. Strategies to reduce venous enhancement in first pass lower extremity MRA

Strategy Advantages Drawbacks

Separate acquisition for lower legs Multiphase DSA-like images Aortoiliac and upper leg vessels need to be
acquired separately

Infrasystolic venous compression Delays venous filling Can be perceived as uncomfortable by pati-
ents; takes time to apply

Lowering TR/TE Faster acquisition Lowers SNR

Centric K-space filling Suppression of venous enhancement None, even when contrast present in veins

Time-resolved acquisition Multiphase DSA-like images Still requires compromise between spatial and
temporal resolution

Parallel imaging Can substantially speed up acquisition Lowers SNR; can lead to disturbing artifacts

SNR signal-to-noise ratio

104 Chapter 9 MRA of the Peripheral Vasculature

acquisition and the proximal stations are imaged in the Several authors have shown that the use of such coils is
equilibrium phase. The greatest benefit from the hybrid beneficial in terms of both increased signal-to-noise as
approach can obviously be expected in patients with well as increased anatomical coverage [4446]. Recom-
chronic critical ischemia, patients with AVF (such as in mended sequence parameters for 3-consecutive station
diabetes mellitus), and patients with cellulitis [42]. high spatial resolution 3D CE-MRA of the lower extrem-
As mentioned above, one of the most important prob- ity are listed in Table 9.5. In Table 9.6 recommended
lems with first pass imaging of the vessels in the lower parameters are listed for lower leg and pedal CE-MRA.
extremity is disturbing venous enhancement or overlay.
Although simultaneous opacification of venous structures
per se does not hinder interpretation of the images, it is Clinical Applications
the generally poor spatial resolution of first pass acquisi- As is true for the upper extremity, CE-MRA is a powerful
tions that makes reliable differentiation of arteries from technique for diagnostic and preinterventional workup of
veins difficult or sometimes even impossible. The pos- lower extremity occlusive disease ( Table 9.7). The pos-
sibility of adding a higher resolution equilibrium-phase sibility to add on ultra-high spatial resolution equilibrum-
acquisition is especially fortuitous in this respect, as it phase acquisitions has really been quite a revolution
allows for easy distinction of arteries and veins while si- in terms of image quality, confidence of diagnosis, and
multaneously enabling assessment of both systems, even comprehensive evaluation of both the arterial and venous
in the lower legs ( Fig. 9.6). The chapter by Willinek and systems.
Hadizadeh also provides a perspective on the benefits of
this approach, as well as many practical examples. Atherosclerotic occlusive disease
Independent of which approach one chooses to pe- The vast majority (>80%) of cases of lower extremity arte-
ripheral MRA, an absolute requirement is the use of rial disease are due to atherosclerotic peripheral arterial
surface coils, at least for the lower legs. The use of such occlusive disease (APAOD). APAOD is a major health
9 surface coils increases SNR, which allows for an increase care problem in Western society, with an estimated preva-
in spatial resolution; this in turn increases arterial con- lence in the general population of 2.5% in the age-group
spicuity and improves disease characterization in small, over 50 years of age. In the age-group over 70 years old,
distal, diseased arteries. Conversely, the increase in SNR the prevalence is estimated at about 7% [47, 48]. Patients
can be used to increase acquisition speed in order to avoid who have complaints of APAOD usually present with a
venous enhancement. Almost all vendors have phased- history of intermittent claudication. Intermittent clau-
array surface body or torso coils that can be used to image dication is caused when the blood flow to exercising leg
the lower leg. In addition, all major MRI vendors are now and calf muscles is insufficient compared with metabolic
selling parallel imaging-capable dedicated multistation demands. Upon cessation of exercise, complaints usually
peripheral vascular coils with coverage exceeding 100 cm. disappear rapidly. The clinical course of intermittent clau-

Fig. 9.6. Transverse reconstructions of equilibrium phase acquisi- Left: 1.0 x 1.0 x 1.0 mm3 (1000 m3); middle: 0.75 x 0.75 x 0.75 mm3
tions at three different isotropic spatial resolutions. Especially at higher (412 m3); right: 0.5 x 0.5 x 0.5 mm3 (125 m3) resolution. ata, anterior
spatial resolutions, arteries can easily be visually separated from veins. tibial artery; fib, fibular artery; pta, posterior tibial artery
Table 9.5. Recommended 3D CE-MRA sequence parameters for first pass and equilibrium-phase 3-station peripheral CE-MRA

Sequence Orientation Coverage Coil TR/TE FA FOV Matrix Slice thick- No. of Duration Comments
(ms) (degrees) (mm) ness* (mm) partitions (s)

3D GRE (first Coronal Abd aorta to prox Body/ sur- 5 / 2.5 3040 430 x 300 400 x200 2.5 Tailor to 1015 Use parallel imaging
pass) SFA face PA anatomy when possible
9.3 Clinical Indications

3D GRE (first Coronal CF to trifurcation Body/ sur- 4/1 3040 430 x 350 400 x 256 2.02.5 Tailor to 712 Use parallel imaging
pass) face PA anatomy when possible

3D GRE (first Coronal P3 to pedal arch Surface PA < 4 / <2 3040 430 x 300 430 x 300 <1.01.5 Tailor to <30 Use parallel imaging
pass) anatomy when possible

3D GRE (equili- Coronal or Area of interest Body/ sur- <12 / 20 Tailor to Tailor to isotropic 0.7 Tailor to Several Use parallel imaging
brium-phase) Sagittal face PA <3.0 anatomy voxel size of 0.7 x anatomy min when possible
0.7 x 0.7 mm3

T1w T1-weighted; GRE gradient recalled echo; SE spin-echo; TR repetition time; TE echo time; FA flip angle; FOV field-of-view (frequency x phase); BW band width; NSA number of signals averaged (number below
1 indicates partial or fractional echo); * non-interpolated, truly acquired slice thickness; Abd abdominal; prox proximal; SFA superficial femoral artery; CF common femoral artery; P3 popliteal artery below knee
joint; PA phased array

Table 9.6. Recommended imaging parameters for first pass and equilibrium-phase lower leg and pedal CE-MRA

Sequence Orientation Coverage Coil TR/TE FA FOV Matrix Slice thick- No. of Duration Comments
(ms) (degrees) (mm) ness* (mm) partitions (s)

3D CE- T1w Coronal or P3 to pedal arch Body/ <5 / <2 3040 430 x 300 Same value as 1.01.5 Tailor to <10, MP Use parallel imaging
GRE sagittal Surface PA FOV anatomy <30 when possible
< 45

3D CE- T1w Sagittal Ankle to pedal arch Head / Knee <5 / <2 3040 430 x 300 Same value as 1.0 Tailor to <10, MP Use parallel imaging
GRE (single foot only) FOV anatomy when possible

2D CE-T1w GRE Sagittal P3 to pedal arch Body/ <10 / 6080 Tailor to 256 Up to 100 Single thick 35 (per
(single foot only) Surface <2 anatomy slice, tailor phase)
PA to anatomy

3D GRE (equili- Coronal or Area of interest Body/ <12 / 20 Tailor to Tailor to isotropic 0.7 Tailor to Several Use parallel imaging

brium-phase) sagittal Surface <3.0 anatomy voxel size of 0.7 x anatomy min when possible
PA 0.7 x 0.7 mm3

TR repetition time; TE echo time; FA flip angle; FOV field-of-view (frequency x phase); NSA number of signals averaged (number below 1 indicates partial or fractional echo); * non-interpolated, truly acquired
slice thickness; P3 popliteal artery below knee joint; PA phased array; MP multiphasic acquisition. when acquired as last station in 3-station runoff protocol; when acquired as dedicated lower leg examination
with infrasystolic venous compression
106 Chapter 9 MRA of the Peripheral Vasculature

Table 9.7. Indications for Vasovist-enhanced MRA of the lower extremity

Indication Advantages of using Vasovist

Atherosclerotic occlusive disease of High relaxivity in F/P; ultra-high spatial resolution in E/P with ability to assess plaque calci-
infrarenal aorta and lower extremities fications

Aneurysmal disease High relaxivity in F/P; definition of mural thrombus in E/P

Vasculitis High relaxivity in F/P; ability to depict vessel wall in E/P

Foot imaging High relaxivity in F/P; ultra-high spatial resolution and definition of thrombus in E/P

F/P first pass; E/P equilibrium phase

dication in terms of progression of disease in the sympto- tage of ultra-high spatial resolution equilibrium-phase
matic limb is usually benign, since only about a quarter of imaging is that it also facilitates assessment of the degree
patients will ever significantly deteriorate [47]. With an of calcification of the atherosclerotic vascular changes
intervention rate of approximately 5%, only 13% of all ( Fig. 9.7).
patients with intermittent claudication will ever undergo If desired, the aortoiliac arteries can also be imaged
a major amputation [49]. Angiographically, intermittent in the equilibrium phase ,albeit at first pass resolution so
claudication is characterized by either a short, isolated as not to compromise image quality because of breath-
aortoiliac or superficial femoral lesion or a longer iliac or ing artifacts. In cases where the primary interest is the
superficial femoral occlusion. In more severe cases com- iliac arteries or the arteries in the pelvis we found spatial
9 bined lesions are seen. resolutions of 0.7 x 0.7 x 0.7 mm3 to be feasible as well
About 5% of patients with intermittent claudication ( Fig. 9.8).
progress to chronic critical ischemia; that is, the oxygen The angiographic hallmark of chronic critical ischemia
and nutrient supply of the distal lower extremity fall is bilateral, multiple severe stenoses and occlusions at dif-
below the level for maintenance of normal cellular proc- ferent levels in the peripheral arterial tree ( Fig. 9.9).
esses in resting conditions. Clinically, this is manifested Patients with diabetes are a well-recognized subgroup
by rest pain and tissue loss (i.e., non-healing ulcers and with primarily distal atherosclerotic occlusive disease and
gangrene). The incidence of chronic critical ischemia is preservation of normal inflow ( Fig. 9.10). Obtaining a
estimated to be between 300 and 1000 per million per high-quality full anatomic study from the infrarenal aorta
year in the general population [47]. Because of the sever- down to the lower leg and pedal arteries is essential in
ity of complaints an invasive intervention is attempted in the pre-interventional workup of distal peripheral arterial
all but the worst cases. Despite advances in endovascular disease. The hybrid approach as described above is best
and vascular surgical therapy, however, the rate of major suited to achieve this aim.
amputation in the general population remains in excess
of 0.03% in industrialized countries [50]. In contrast to Aneurysmal disease
patients with intermittent claudication, imaging outflow An aneurysm is defined as a focal enlargement of an ar-
arteries is an essential part of the imaging workup and tery to more than 1.5 times its normal diameter. For the
potentially limb saving in patients with chronic critical aorta the general cut-off value is 3 cm, for the iliac arteries
ischemia, particularly when it is taken into account that 1.8 cm. Aneurysms are either true when intima, media,
IA-DSA is known to fail in visualizing patent crural and adventitia are involved, or false if fewer than three
and foot arteries that are demonstrated with other mo- layers are involved. False aneurysms are frequently asso-
dalities [5153]. In order to achieve the best possible ciated with confined rupture of arterial vessel wall, either
image quality, the hybrid approach as described above spontaneous, or due to trauma.
is highly recommended when dealing with patients with Patients with gross aneurysmal dilatation of the tho-
chronic critical ischemia, followed by ultra-high spatial racic and abdominal aorta need to be imaged with a
resolution equilibrium-phase acquisitions of the upper multiphase protocol to ensure optimal opacification of all
and lower leg stations. The authors believe that a spa- the arteries in the FOV. Equilibrium-phase acquisitions
tial resolution ranging between 0.7 x 0.7 x 0.7 mm3, or in combination with parallel imaging are very helpful to
343 m (upper legs), and 0.5 x 0.5 x 0.5 mm3, or 125 m precisely outline disease. Such acquisitions also allow for
(lower legs), represents a good compromise between reliable detection of mural thrombus, and hence the true
spatial resolution and imaging time. An added advan- diameter of the artery in question. In addition, they are
9.3 Clinical Indications
107 9



Fig. 9.7AD. A 65-year-old male subject presenting with intermittent equilibrium-phase acquisition (B), on the other hand, clearly demon-
claudication of the left lower extremity after having undergone PTA strates the highly calcified nature of the plaque, as evidenced by the
of the right popliteal artery. MIP of first pass acquisition demonstrates uniformly low signal intensity, similar to that of cortical bone (arrow in
a moderate stenosis in the left popliteal artery (arrow in A). Note that B). Additional reformations in the transverse plane at the level of the
first pass acquisition is a pure luminogram, providing no information distal femur (C) and femoral condyles (D) also confirm this (arrows in C
about the underlying vascular pathology. Coronal source image of the and B). fh, fibular head; pv, popliteal vein; f, femur; fc, femoral condyle

essential in outlining arterial as well as venous anatomy in in these patients for two main reasons: (a) steady state
case of incomplete opacification of these structures in the imaging is better suited to characterize the exact degree
first pass ( Figs. 9.11 and 9.12). of stenosis at the sites of proximal and distal anastomosis
( Fig. 9.13A, B); (b) because of the ability to perform mul-
Bypass graft imaging tiple subsequent ultra-high spatial resolution acquisitions,
Equilibrium-phase imaging confers substantial added the high-resolution FOV becomes virtually unlimited,
value over first pass imaging in patients with bypass thus in effect greatly extending anatomic coverage com-
grafts. Ultra-high spatial resolution imaging has actually pared with what can be imaged during first arterial pas-
evolved into an important adjunct to first pass imaging sage of contrast medium ( Fig. 9.13C).
108 Chapter 9 MRA of the Peripheral Vasculature

Fig. 9.8A,B. Feasibility and

added value of equilibrium-
phase imaging of the pelvic
arteries in a 65-year-old patient
presenting with intermittent
claudication of the right leg.
First pass imaging of the runoff
arteries demonstrates total oc-
clusion of the right common il-
iac artery over its entire length
(A, arrowhead). Note excellent
9 image quality of the distal
runoff vessels despite injec-
tion of just 10 ml of Vasovist.
Fig. 9.9. CE-MRA appearance of different types of peripheral arte-
Equilibrium-phase acquisition
rial occlusive disease. Type I (left panel): involvement limited to aorta
(0.7 x 0.7 x 0.7 mm3) source
and common iliac arteries; Type II (middle panel): involvement limited
image, however, shows a much
to suprainguinal peripheral arteries; Type III (right panel): central and
shorter occlusion length com-
peripheral arterial occlusive disease (i.e. multilevel) disease
pared with first pass imaging
(B, arrowhead). This finding
allows for a more confident
recanalization attempt by the
interventional radiologist. Also
note small aortic aneurysm
with excellent definition of the
outer wall and mural thrombus
(B, asterisk)

Fig. 9.10. A 45-year old female patient with

diabetes mellitus type 2. There is extensive
occlusive disease in the right lower leg arter-
ies. The proximal arteries are spared. This
pattern is typical for patients with diabetes
9.3 Clinical Indications
109 9

Fig. 9.11AC. A 71-year-old male patient with aneurysmal dilation is susceptibility-induced signal loss due to bilateral knee prostheses.
from the infrarenal abdominal aorta down to the proximal lower-leg Equilibrium-phase acquisition (0.7 x 0.7 x 0.7 mm3) shows the femo-
arteries. The patient presented with a rapidly enlarging mass in the left ropopliteal bypass (B, left panel) except for a small area of signal loss,
popliteal fossa and reported having undergone bilateral knee replace- as well as a partially thrombosed huge aneurysm (B, right panel). Note
ments as well as bypass surgery to exclude a popliteal aneurysm on that the aneurysm is connected to the deep venous circulation (B, long
the left side. First pass imaging (A) shows gross aneurysmal dilatation arrow in left panel). Additional reformations (C) demonstrate the an-
from the aorta downwards. Due to slow flow of contrast material, there eurysm (arrowheads) as well as the bypass and connections to the ve-
is incomplete filling of the distal common iliac and common femoral nous system in exquisite detail. a, aneurysm; s, susceptibility-induced
arteries as well as of the popliteal arteries (* in left two panels). There signal loss; b, femoropopliteal bypass; p, popliteal vein
110 Chapter 9 MRA of the Peripheral Vasculature

Fig. 9.12AC. A 70-year-old male patient who presented with a femoral arteries (B, left panel) and a clover-leaf-shaped aneurysm
pulsatile mass of the medial aspect of the ankle. First pass imaging of the posterior tibial artery (B, arrowhead in right panel). A more
(A) shows a mistimed acquisition with no filling in the initial phase detailed examination of the aneurysm reveals connections with
(A, left panel) and incomplete filling of the distal lower leg arteries posterior tibial artery (C, left panel) and that the aneurysm is partially
in the second phase (A, right panel). Subsequent equilibrium-phase thrombosed (C, right panel). a, aneurysm; pt, posterior tibial artery
acquisitions (0.5 x 0.5 x 0.5 mm3) demonstrate normal superficial
9.3 Clinical Indications
111 9


Fig. 9.13AC. Multi-station maximum-intensity projection of first (216 m3) corroborates first pass findings and shows normal caliber
pass CE-MRA in a 55-year-old patient with extra-anatomic bypass of distal bypass and runoff into the distal popliteal artery (arrow in
suffering from renewed intermittent claudication in the left leg. The B). Because of the extended intravascular retention of Vasovist it is
right lower leg had been amputed previously. The patient was re- possible to image additional arterial territories as well as the venous
ferred for CE-MRA because routine follow-up of the femoropopliteal compartment. C Curved multiplanar reformation (left panel) and
bypass with duplex ultrasonography in the left leg showed slow flow subvolume maximum-intensity projection (right panel) of the origin
and questionable stenosis of the distal anastomosis. First pass imag- and proximal part of the extra-anatomic bypass. Despite the fact
ing after injection of 0.03 mmol/kg Vasovist shows no major stenoses that vessel-to-background is lower compared with first pass imag-
in the distal part of the graft (boxed area in A). Subsequent higher ing, a flow-limiting, significant stenosis can clearly be appreciated
resolution imaging in the equilibrium phase at 0.6 x 0.6 x 0.6 mm3 (arrowheads)
112 Chapter 9 MRA of the Peripheral Vasculature

Non-atherosclerotic causes of peripheral arterial CE-MRA improves diagnostic accuracy compared with
disease 2D TOF MRA [58]. These meta-analyses included only
Approximately 20% of chronic arterial occlusions are studies published up to 2000, but more recent studies
based on diseases other than atherosclerosis that cause also report high sensitivity and specificity figures ranging
luminal narrowing and/or occlusion. The presence, lo- between 85 and 100% [55].
cation, and progression of arterial stenoses and occlu-
sions may reflect an underlying systemic disease or
may be an expression of regional inflammatory or de- 9.5 Vasovist-enhanced Imaging
generative processes. Lower extremity arterial diseases
other than APAOD are popliteal entrapment syndrome, Limited data exist on the efficacy of Vasovist-enhanced
cystic adventitial disease, arterial fibrodysplasia, non- MRA. It is important to note that no reports have been
specific aortoarteritis (Takayasus disease), and a host of published that specifically focused on the first pass ac-
uncommon vasculitides, the most important of which is curacy. Only studies reporting on the combined first
thromboangiitis obliterans (Buergers disease) [54, 55]. pass and steady state evaluation were reported in the
In addition, congenital connective tissue diseases such as literature, four of which focused on peripheral arterial
Marfans syndrome, Ehlers-Danlos syndrome, pseudox- disease [5962]. All four Vasovist publications reported
anthoma elasticum, homocystinurea, and neurofibroma- sensitivity and specificity values in comparison to X-ray
tosis are well-known for their peripheral arterial involve- angiography as the standard of reference. Sensitivity and
ment [56]. Congenital and acquired clotting disorders specificity values as reported in these trials are lower
may also manifest with symptoms of peripheral arterial compared with extracellular agent trials performed in
disease. Rare causes of peripheral arterial disease are pri- the same vascular territory. The main explanations for
mary tumors, radiation therapy, and the iliac syndrome this finding are probably that trials not performed un-
in cyclists. An acute arterial occlusion always necessitates der regulatory oversight are not truly blinded, and that,
9 a prompt search for atrial fibrillation, ulcerative endocar- on average, much lower numbers of subjects were in-
ditis, and mural thrombi overlying the site of myocardial cluded.
infarction. Finally, as mentioned above, Klessen et al. confirmed
that Vasovist-enhanced first pass MRA was of equivalent
quality to extracellular-enhanced first pass MRA in a
9.4 Diagnostic Accuracy of MRA for comparison of 40 patients who underwent whole-body
Evaluation of Peripheral Arterial MR angiography [3].
Occlusive Disease

Due to tremendous improvements in MR hard- and soft- 9.6 Conclusions

ware over the past decade, CE-MRA is now the preferred
MRA technique for evaluation of the full spectrum of Contrast-enhanced MRA has emerged as the MRA tech-
patients with symptomatic peripheral arterial disease in nique of choice for vascular imaging of both the upper
both the upper and lower extremities. Sensitivities and and lower extremities. The higher relaxivity and pro-
specificities for the detection of flow-limiting stenoses longed intravascular residence time of blood pool agents
are, with few exceptions, very high. Two relatively recent yield subjectively better first pass image quality as well
meta-analyses, by Nelemans et al. [57] and Koelemay et as the possibility to obtain additional information about
al. [58], confirm that 3D CE-MRA is significantly more both the arterial and the venous system in the delayed
accurate than 2D TOF MRA. Nelemans et al. found that phase. The latter property has led to a fundamental para-
the relative diagnostic odds ratio (a measure of the dis- digm shift in MR imaging of the vasculature, enabling the
criminatory power of an examination which represents migration to equilibrium-phase ultra-high spatial resolu-
the odds of a positive examination result among diseased tion imaging sequences. Ongoing technical improvements
persons relative to the odds of a positive examination have significantly improved not only image quality but
result among non-diseased persons) of 3D CE-MRA com- also speed, reliability, and ease of use, and will decrease
pared with 2D TOF MRA was 7.46 [57]. This finding in- imaging times greatly in the near future. Without excep-
dicates that 3D CE-MRA is superior to 2D TOF MRA for tion, published studies that have compared CE-MRA with
the detection and grading of peripheral arterial disease. IA-DSA demonstrate high accuracy for the detection and
In addition, the same study demonstrated that review of characterization of vascular pathology and confirm the
source images or multiplanar reformations led to an al- clinical value of the technique.
most fivefold increase in the relative diagnostic odds ratio
[57]. Koelemay et al. also came to the conclusion that 3D
113 9
Take home messages
Vasovist can be used for first pass imaging of the Equilibrium-phase Vasovist-enhanced MRA can
peripheral vasculature in the exact same way as reveal details about the vasculature that are not
conventional extracellular contrast agents. In cases well seen with first pass MRA, such as very small
where the contrast bolus is missed the acquisition arterial branches. In addition, equilibrium-phase
can be repeated in the equilibrium phase with MRA enables a better assessment of the true degree
little or no loss of image quality. of stenosis due to the superior resolution.
Equilibrium-phase Vasovist-enhanced MRA im- Vasovist is the imaging agent of first choice for
ages can be acquired at vastly improved spatial depiction of both the central and the peripheral
resolution compared with first pass imaging. The venous system due to its safety-profile, its non-inva-
achievable level of detail is now approaching that sive nature, and its extended intravascular half-life.
of conventional intra-arterial DSA, and in some Vasovist-enhanced MRA can be used for the
cases it even exceeds the latter technique in diag- workup of the full spectrum of upper and lower
nostic capabilities. extremity peripheral arterial disease.

References 12. Siegelman ES, Charafeddine R, Stolpen AH, Axel L (2000) Suppres-
1. Ho KY, Leiner T, de Haan MW, Kessels AG, Kitslaar PJ, van En- sion of intravascular signal on fat-saturated contrast-enhanced
gelshoven JM (1998) Peripheral vascular tree stenoses: evaluation thoracic MR arteriograms. Radiology 217:115118
with moving-bed infusion-tracking MR angiography. Radiology 13. Tay KY, JM UK-I, Trivedi RA, et al (2005) Imaging the vertebral ar-
206:683692 tery. Eur Radiol 15:13291343
2. Lauffer RB, Parmelee DJ, Dunham SU, et al (1998) MS-325: al- 14. Wu C, Zhang J, Ladner CJ, Babb JS, Lamparello PJ, Krinsky GA
bumin-targeted contrast agent for MR angiography. Radiology (2005) Subclavian steal syndrome: diagnosis with perfusion met-
207:529538 rics from contrast-enhanced MR angiographic bolus-timing exam-
3. Klessen C, Hein PA, Huppertz A, et al (2007) First pass whole-body ination--initial experience. Radiology 235:927933
magnetic resonance angiography (MRA) using the blood-pool 15. Bitar R, Gladstone D, Sahlas D, Moody A (2004) MR angiography of
contrast medium gadofosveset trisodium: comparison to gado- subclavian steal syndrome: pitfalls and solutions. AJR Am J Roent-
pentetate dimeglumine. Invest Radiol 42:659664 genol 183:18401841
4. Hartmann M, Wiethoff AJ, Hentrich HR, Rohrer M (2006) Initial 16. Bley TA, Wieben O, Vaith P, Schmidt D, Ghanem NA, Langer M
imaging recommendations for Vasovist angiography. Eur Radiol (2004) Magnetic resonance imaging depicts mural inflammation
16 [Suppl 2]:B1523 of the temporal artery in giant cell arteritis. Arthritis Rheum
5. Wang MS, Haynor DR, Wilson GJ, Leiner T, Maki JH (2007) Maximiz- 51:10621063; author reply 1064
ing contrast-to-noise ratio in ultra-high resolution peripheral MR 17. Bley TA, Wieben O, Uhl M, Thiel J, Schmidt D, Langer M (2005)
angiography using a blood pool agent and parallel imaging. J High-resolution MRI in giant cell arteritis: imaging of the
Magn Reson Imaging 26:580588 wall of the superficial temporal artery. AJR Am J Roentgenol
6. Winterer JT, Scheffler K, Paul G, et al (2000) Optimization of 184:283287
contrast-enhanced MR angiography of the hands with a timing 18. Olin JW, Young JR, Graor RA, Ruschhaupt WF, Bartholomew JR
bolus and elliptically reordered 3D pulse sequence. J Comput As- (1990) The changing clinical spectrum of thromboangiitis obliter-
sist Tomogr 24:903908 ans (Buergers disease). Circulation 82:IV38
7. Wentz KU, Frohlich JM, von Weymarn C, Patak MA, Jenelten R, 19. Olin JW (2000) Thromboangiitis obliterans (Buergers disease). N
Zollikofer CL (2003) High-resolution magnetic resonance angiog- Engl J Med 343:864869
raphy of hands with timed arterial compression (tac-MRA). Lancet 20. Murphy GJ, White SA, Nicholson ML (2000) Vascular access for
361:4950 haemodialysis. Br J Surg 87:13001315
8. Bilecen D, Aschwanden M, Heidecker HG, Bongartz G (2004) Opti- 21. NKF-DOQI clinical practice guidelines for vascular access. (1997)
mized assessment of hand vascularization on contrast-enhanced National Kidney Foundation-Dialysis Outcomes Quality Initiative.
MR angiography with a subsystolic continuous compression tech- Am J Kidney Dis 30:S150191
nique. AJR Am J Roentgenol 182:180182 22. Staple TW (1973) Retrograde venography of subcutaneous arte-
9. Wikstrom J, Johansson L, Karacagil S, Ahlstrom H (2003) The im- riovenous fistulas created surgically for hemodialysis. Radiology
portance of adjusting for differences in proximal and distal con- 106:223224
trast bolus arrival times in contrast-enhanced iliac artery magnetic 23. Waldman GJ, Pattynama PM, Chang PC, Verburgh C, Reiber JH, de
resonance angiography. Eur Radiol 13:957963 Roos A (1996) Magnetic resonance angiography of dialysis access
10. Foo TK, Saranathan M, Prince MR, Chenevert TL (1997) Automated shunts: initial results. Magn Reson Imaging 14:197200
detection of bolus arrival and initiation of data acquisition in fast, 24. Bos C, Smits JH, Zijlstra JJ, et al (2001) MRA of hemodialysis ac-
three-dimensional, gadolinium-enhanced MR angiography. Radi- cess grafts and fistulae using selective contrast injection and flow
ology 203:275280 interruption. Magn Reson Med 45:557561
11. Luccichenti G, Cademartiri F, Ugolotti U, Marchesi G, Pavone P 25. Planken RN, Tordoir JH, Dammers R, et al (2003) Stenosis detec-
(2003) Magnetic resonance angiography with elliptical order- tion in forearm hemodialysis arteriovenous fistulae by multiphase
ing and fluoroscopic triggering of the renal arteries. Radiol Med contrast-enhanced magnetic resonance angiography: preliminary
(Torino) 105:4247 experience. J Magn Reson Imaging 17:5464
114 Chapter 9 MRA of the Peripheral Vasculature

26. Leiner T, Herborn CU, Goyen M (2007) Nephrogenic systemic 44. Goyen M, Ruehm SG, Barkhausen J, et al (2001) Improved multi-
fibrosis is not exclusively associated with gadodiamide. Eur Radiol station peripheral MR angiography with a dedicated vascular coil.
17:19211923 J Magn Reson Imaging 13:475480
27. Leiner T, Ho KY, Nelemans PJ, de Haan MW, van Engelshoven JM 45. Fellner FA, Requardt M, Lang W, Fellner C, Bautz W, Cavallaro A
(2000) Three-dimensional contrast-enhanced moving-bed infu- (2003) Peripheral vessels: MR angiography with dedicated phased-
sion-tracking (MoBI-track) peripheral MR angiography with flex- array coil with large-field-of-view adapter feasibility study. Radiol-
ible choice of imaging parameters for each field of view. J Magn ogy 228:284289
Reson Imaging 11:368377 46. Leiner T, Nijenhuis RJ, Maki JH, Lemaire E, Hoogeveen R, van
28. Ho KY, Leiner T, de Haan MW, van Engelshoven JM (1999) Periph- Engelshoven JM (2004) Use of a three-station phased array coil
eral MR angiography. Eur Radiol 9:17651774 to improve peripheral contrast-enhanced magnetic resonance
29. Ho KY, Leiner T, van Engelshoven JM (1999) MR angiography of angiography. J Magn Reson Imaging 20:417425
run-off vessels. Eur Radiol 9:12851289 47. Dormandy JA, Rutherford RB Management of peripheral arterial
30. Nijenhuis RJ, Leiner T, Cornips EM, et al (2004) Spinal cord feed- disease (PAD). TASC Working Group. TransAtlantic Inter-Society
ing arteries at MR angiography for thoracoscopic spinal surgery: Concensus (TASC). J Vasc Surg 2000 31:S1-S296
feasibility study and implications for surgical approach. Radiology 48. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes
233:541547 FG Inter-Society Consensus for the Management of Peripheral
31. Schoenberg SO, Londy FJ, Licato P, Williams DM, Wakefield T, Arterial Disease (TASC II). J Vasc Surg 2007 45 Suppl S:S5-67
Chenevert TL (2001) Multiphase-multistep gadolinium-enhanced 49. Suggested standards for reports dealing with lower extrem-
MR angiography of the abdominal aorta and runoff vessels. Invest ity ischemia. Prepared by the Ad Hoc Committee on Reporting
Radiol 36:283291 Standards, Society for Vascular Surgery/North American Chapter,
32. Prince MR, Chabra SG, Watts R, et al (2002) Contrast material travel International Society for Cardiovascular Surgery. J Vasc Surg 1986
times in patients undergoing peripheral MR angiography. Radiol- 4:80-94
ogy 224:5561 50. Dormandy JA, Ray S (1996) The natural history of peripheral arte-
33. Herborn CU, Ajaj W, Goyen M, Massing S, Ruehm SG, Debatin JF rial disease. In: Tooke JE, Lowe GD (eds) A textbook of vascular
(2004) Peripheral vasculature: whole-body MR angiography with medicine. Arnold, London, pp 162175
midfemoral venous compression--initial experience. Radiology 51. Beard JD, Scott DJ, Evans JM, Skidmore R, Horrocks M (1988)
230:872878 Pulse-generated runoff: a new method of determining calf vessel
9 34. Bilecen D, Schulte AC, Aschwanden M, et al (2004) MR angiog-
raphy with venous compression. Radiology 233:617618; author
patency. Br J Surg 75:361363
52. Owen RS, Carpenter JP, Baum RA, Perloff LJ, Cope C (1992) Mag-
reply 618619 netic resonance imaging of angiographically occult runoff vessels
35. Bilecen D, Schulte AC, Bongartz G, Heidecker HG, Aschwanden in peripheral arterial occlusive disease. N Engl J Med 326:1577
M, Jager KA (2004) Infragenual cuff-compression reduces venous 1581
contamination in contrast-enhanced MR angiography of the calf. 53. Wilson YG, George JK, Wilkins DC, Ashley S (1997) Duplex assess-
J Magn Reson Imaging 20:347351 ment of run-off before femorocrural reconstruction. Br J Surg
36. Meaney JF, Ridgway JP, Chakraverty S, et al (1999) Stepping-table 84:13601363
gadolinium-enhanced digital subtraction MR angiography of the 54. Rudofsky G (2003) Peripheral arterial disease: chronic ischemic
aorta and lower extremity arteries: preliminary experience. Radi- syndromes. In: Lanzer P, Topol EJ (eds) PanVascular Medicine. Ber-
ology 211:5967 lin, Springer, pp 13631422
37. Ruehm SG, Hany TF, Pfammatter T, Schneider E, Ladd M, Debatin 55. Leiner T (2005) Magnetic resonance angiography of abdominal and
JF (2000) Pelvic and lower extremity arterial imaging: diagnostic lower extremity vasculature. Top Magn Reson Imaging 16:2166
performance of three-dimensional contrast-enhanced MR angi- 56. Rooke TW, Joyce JW (2000) Uncommon arteriopathies. In: Ruther-
ography. AJR Am J Roentgenol 174:11271135 ford RB (ed) Vascular surgery. W.B. Saunders, Philadelphia, pp 418-
38. Menzoian JO, LaMorte WW, Paniszyn CC, et al S(1989) ymptoma- 434.
tology and anatomic patterns of peripheral vascular disease: dif- 57. Nelemans PJ, Leiner T, de Vet HC, van Engelshoven JM (2000) Pe-
fering impact of smoking and diabetes. Ann Vasc Surg 3:224228 ripheral arterial disease: meta-analysis of the diagnostic perform-
39. Maki JH, Wilson GJ, Eubank WB, Hoogeveen RM (2002) Utilizing ance of MR angiography. Radiology 217:105114
SENSE to achieve lower station sub-millimeter isotropic resolution 58. Koelemay MJ, Lijmer JG, Stoker J, Legemate DA, Bossuyt PM (2001)
and minimal venous enhancement in peripheral MR angiography. Magnetic resonance angiography for the evaluation of lower ex-
J Magn Reson Imaging 15:484491 tremity arterial disease: a meta-analysis. JAMA 285:13381345
40. Bezooijen R, van den Bosch HC, Tielbeek AV, et al (2004) Peripheral 59. Perreault P, Edelman MA, Baum RA, et al (2003) MR angiogra-
arterial disease: sensitivity-encoded multiposition MR angiogra- phy with gadofosveset trisodium for peripheral vascular disease:
phy compared with intraarterial angiography and conventional phase II trial. Radiology 229:811820
multiposition MR angiography. Radiology 231:263271 60. Goyen M, Edelman M, Perreault P, et al (2005) MR angiography
41. de Vries M, Nijenhuis RJ, Hoogeveen RM, de Haan MW, van En- of aortoiliac occlusive disease: a phase III study of the safety and
gelshoven JM, Leiner T (2005) Contrast-enhanced peripheral MR effectiveness of the blood-pool contrast agent MS-325. Radiology
angiography using SENSE in multiple stations: feasibility study. J 236:825833
Magn Reson Imaging 21:3745 61. Rapp JH, Wolff SD, Quinn SF, et al (2005) Aortoiliac occlusive
42. Wang Y, Winchester PA, Khilnani NM, et al (2001) Contrast-en- disease in patients with known or suspected peripheral vascu-
hanced peripheral MR angiography from the abdominal aorta to lar disease: safety and efficacy of gadofosveset-enhanced MR
the pedal arteries: combined dynamic two-dimensional and bolus- angiography--multicenter comparative phase III study. Radiology
chase three-dimensional acquisitions. Invest Radiol 36:170177 236:7178
43. Morasch MD, Collins J, Pereles FS, et al (2003) Lower extremity 62. Bosch E, Kreitner KF, Peirano MF, Thurner S, Shamsi K, Parsons
stepping-table magnetic resonance angiography with multilevel EC, jr (2008) Safety and efficacy of gadofosveset-enhanced MR
contrast timing and segmented contrast infusion. J Vasc Surg angiography for evaluation of pedal arterial disease: multicenter
37:6271 comparative phase 3 study. AJR Am J Roentgenol 190:179186

Magnetic Resonance Venography

Giles Roditi

10.1 Introduction 116

10.2 Magnetic Resonance Venography 116

10.3 Contrast Administration for MRV 117

10.4 Technique for MRV with Blood pool Contrast Agent 118

10.5 Lower Limbs 118

10.5.1 Lower Limb Thrombosis 119
10.5.2 Venous Mapping and Arteriovenous Malformations 121

10.6 Abdominal and Pelvic Veins 123

10.6.1 Inferior Vena Cava and Pelvic Veins 123
10.6.2 Gonadal Veins 123
10.6.3 Renal Veins 124
10.6.4 Portal Venous Circulation 126

10.7 Upper Limbs and Thorax 126

10.8 Conclusion 128

References 129
116 Chapter 10 Magnetic Resonance Venography

10.1 Introduction

The literature of conventional contrast angiography has

traditionally focused on arteriography and in fact, many
texts make scant reference to venography. Indeed, good-
quality conventional venography has long been a chal-
lenging procedure for several reasons. First, unlike arte-
riography, where smaller distal branches are visualized
from a central injection in a larger artery, the vein in
question or one of its smaller tributaries needs to be di-
rectly cannulated for opacification as flow is away from
the point of injection towards larger vessels; e.g., a small
foot vein must be cannulated in order to study the veins
of a lower limb. Second, full evaluation of all the lower
limb veins for exclusion of thrombus with conventional
venography can actually be quite a complex and techni-
cally demanding undertaking, requiring large volumes of
Fig. 10.1. Conventional venography with DSA. Bilateral arm injec-
contrast medium, careful biplane filming of all veins, and tions to demonstrate the central veins in a patient with pacemaker in
intermittent application of tourniquets [1]. Furthermore, situ, unsuitable for MRI examination. Note that the axillary and jugular
the study of large central veins requires simultaneous bi- veins etc. are not opacified on this study with preferential cephalic
lateral limb injections in order to opacify these capacious venous inflow
low-pressure vessels, for example, bilateral arm injections
to show the SVC ( Fig. 10.1). a CT pulmonary angiography study is used to assess
However, even with this technique of bilateral simul- the large veins of the pelvis and thighs. However, this
taneous injection the full opacification of large central is not widely practiced and due to contrast extraction
10 veins can be problematic due to the influx of unopacified in the limbs, venous opacification is often poor while
blood from other tributaries, e.g., from the jugular veins small deep calf veins are not assessable [2]. In the thorax
to the brachiocephalic veins, from the renal veins to the central venous assessment by CT is often hampered by
IVC, etc. Meanwhile, some veins cannot be routinely beam hardening artifacts at the thoracic inlet, and inflow
evaluated even with a meticulous technique as their tribu- of unopacified blood may be problematic, similar to
taries are not accessible (e.g., the deep femoral vein), and conventional venography. The techniques of magnetic
diagnosis of deep venous thrombosis (DVT) in evaluable resonance venography (MRV) have largely overcome
veins often has to be inferred from the non-opacification many of the pitfalls of other modalities for diagnostic
of those that would be expected to be seen, rather than use [3], particularly the contrast-enhanced techniques,
from more direct visualization of thrombosis such as a though their clinical uptake has until now been slow out
contrast-outlining clot shown as a filling defect. Finally, with specialist centres. Now with the ease of perform-
particularly for lower limb venography, large volumes of ance of MRV using blood pool contrast agents reducing
iodinated contrast medium are required and even with timing errors etc. this is likely to change. The following
modern contrast media, the development of phlebitis or discussion examines the history of body MRV with a
deep venous thrombosis is a recognized complication of brief description of the applicable techniques, the use of
the procedure, while provocation of pulmonary throm- blood pool contrast agents, and how they enhance MRV.
boembolism has also been recorded. This is followed by an examination of the use of MRV in
Other non-invasive modalities have had success in various body territories, starting with the lower limbs as
venous imaging but have their shortcomings in compre- currently the most frequent site of venous assessment.
hensive assessment; for example, while ultrasound tech-
niques are widely and successfully employed in evalua-
tion of the superficial veins of the legs and in suspected 10.2 Magnetic Resonance Venography
lower-limb DVT, significant limitations persist for the
evaluation of deeply situated or inaccessible veins and Magnetic Resonance Angiography (MRA) was initially
delineation of the small deep calf veins remains chal- developed in the 1980s with the flow sensitive MRA tech-
lenging, particularly in the swollen leg. CT has been little niques (Time of Flight and Phase Contrast) relying on
used for primary diagnosis of DVT, although there are maximisation of signal difference between flowing blood
advocates of its use as an adjunct to the investigation of in the lumen versus the vessel wall and surrounding tis-
pulmonary embolism where a delayed scan 3 min after sues. This can be problematic for venous imaging where
10.3 Contrast Administration for MRV
117 10

the flow may be very slow, reducing available signal differ-

ences. Furthermore, the problems of bulk motion artifacts
in deep body parts, in-plane flow saturation effects, and
long imaging times have meant that, outside of intracra-
nial neuroradiolgical applications, these sequences have
been superseded in large part by contrast-enhanced MRA
(CE-MRA) techniques for body applications [4].
CE-MRA relies on the effect of gadolinium-based
paramagnetic contrast agents, increasing blood pool sig-
nal through a profound shortening of the T1 relaxation
time. For most CE-MRA imaging it is helpful to obtain a
3D mask image volume before contrast administration to
serve as a reference, enabling subsequent image subtrac-
tion to increase the contrast-to-noise ratio (CNR) just as
for conventional digital subtraction angiography. These
mask images should also be reviewed, as on occasion
local areas of high T1 signal pre-contrast (e.g. methemo-
globin in thrombus) may confound subsequent contrast-
enhanced image interpretation. This latter phenomenon
has been the basis for the elegant technique of direct
thrombus imaging, whereby T1w sequences tuned to
emphasize thrombus with appropriately short T1 due to
the presence of methemoglobin have been successfully
used to diagnose thrombosis [5, 6]. However, the T1 of
thrombus at a given field strength varies as a function
of its age over the first few days and weeks, reducing
the applicability of the technique, which has so far not
been widely clinically adopted in comparison to contrast-
Fig. 10.2. Direct contrast-enhanced MRV of upper thigh and pelvic
enhanced studies. veins following injection of dilute gadolinium contrast into right foot
vein in a patient with chronic right iliofemoral venous occlusion. Con-
trast flows via long saphenous vein to groin with tortuous cross-pelvic
10.3 Contrast Administration for MRV collateral flow shown to left iliac veins

A sufficient concentration of gadolinium contrast agent

must be achieved in the vein of interest to reduce blood tion can be remote from the actual veins to be studied,
T1 to less than that of fat (i.e. 270 ms for a 1.5-T system) and this is frequently performed in combination with
in order to produce effective contrast. In practice, this a first pass arterial CE-MRA of the region of interest.
requires an intravascular gadolinium concentration of Indeed, contrast arrival in the vascular bed to be studied
greater than 1 mM for conventional extracellular space is usually timed to the arteriographic phase, followed by
agents. However, while at low contrast agent concentra- several repeat measurements acquired as contrast moves
tions the T1 shortening effects predominate, at higher from the arterial to venous compartments. An advantage
concentrations both T2 and T2* shortening effects (which of the indirect method over direct injection is that all
also act over a wider region than the T1 shortening ef- the veins of the vascular bed of interest will be perfused
fects) assume more importance, explaining the drop in with contrast agent so that hitherto inaccessible veins can
signal on T1w sequences seen at high contrast concentra- potentially be visualized. A further potential advantage of
tions. This T1 signal abolition at high agent concentra- this approach is that the arteriographic-phase 3D datat-
tions means that gadolinium contrast for direct MR set can be subtracted from the later equilibrium phase
venography injected into the relevant tributary veins of where there is enhancement of both arteries and veins
the territory to be imaged (e.g., pedal injection for a lower to produce an image set with only venous enhancement.
limb study, Fig. 10.2) must be suitably diluted usually This technique is known as VESPA (venous-enhanced
a 1:20 dilution of a conventional 0.5-M ECS agent with subtracted peak arterial Fig. 10.3), but it does require
isotonic saline [4]. good registration of the 3D studies, and this may not al-
Indirect contrast-enhanced MR venography is more ways be successful in body regions requiring breath-hold
commonly and easily accomplished, as intravenous injec- acquisitions [79].
118 Chapter 10 Magnetic Resonance Venography

off or ideally removed before proceeding to venous imag-

ing. In the arms a neutral position with the arms by the
patients side is best, as with arm elevation the axillary/
subclavian veins may be pinched in the thoracic outlet in
the prescalene space, costoclavicular space, or subcoracoid
tunnel and this is not necessarily pathological, unlike arte-
rial compression [10]. For the pelvis and limbs the issue
of respiratory motion is not relevant and longer imaging
sequence acquisitions than for usual CE-MRA can be em-
ployed. Since the imaging time window for a blood pool
contrast agent is considerably extended as compared with
a conventional ECS gadolinium agent, this this time can be
expolited to improve spatial resolution with smaller voxel
sizes which can be made isovolumetric to allow multipla-
nar reformation in post-processing. With higher matrix
imaging for smaller voxels, longer TR/TE than for a first
pass CE-MRA acquisition will be mandated for imaging a
useful volume. With small voxels SNR will be lower, hence
sequences with reduced bandwidth without the use of par-
tial K-space or half-Fourier techniques are needed to boost
signal. The other consideration is that since the contrast
concentration is lower at this stage, the T1 of blood at 150
ms is not as short as during first pass; this, along with the
Fig. 10.3. Contrast-enhanced MRV of thoracic veins with VESPA increased TR, means that the Ernst angle is lower, and the
technique. Patient with normally patent central veins; portal, hepatic, imaging pulse sequence must be tailored to account for
and azygous systems also demonstrated to advantage
10 this with a reduction in flap angle for example to 2025
for a sequence with TR 10 ms [11]. For imaging the central
veins such as the SVC and IVC and their major tributar-
The limitations of indirect MRV are that with ECS ies in the thorax and upper abdomen, where respiratory
gadolinium contrast agents the first pass tissue extrac- motion can be problematic, standard breath-hold MRA
tion can limit venous contrast concentration and imaging sequences suffice since spatial resolution for these larger
of the central veins may suffer from low signal. This is veins is adequate. In the first few minutes following injec-
particularly the case for the iliac veins and lower IVC, as tion these breath-hold type sequences need little if any
contrast extraction in the legs is greater than in the arms, modification, except that for imaging later in the steady
resulting in lower central venous contrast concentrations state the flip angle should be dropped, as discussed above.
and hence reduced enhancement. For this reason a large
dose of contrast agent has typically been employed e.g.,
double or triple dose of a conventional 0.5-M ECS agent. 10.5 Lower Limbs
Another disadvantage is that while superficial veins are
often visualized early in the legs, the deep veins fill with The indications for venous imaging of the lower limbs
contrast rather more slowly and may indeed not be visible include the investigation of leg swelling (particularly the
during the limited blood pool residence time of conven- evaluation of suspected DVT), assessment of venous in-
tional ECS gadolinium contrast agents. competence and varicosity prior to surgical or endovascu-
lar intervention and evaluation of venous malformations
as part of either isolated lesions or more extensive dis-
10.4 Technique for MRV with Blood pool orders such as the Klippel-Trenaunay-Weber syndrome
Contrast Agent (KTWS). An increasing source of referral is for assess-
ment of the integrity, caliber, and quality of the superfi-
When imaging the limbs, care must be taken to ensure that cial veins of the legs in order to evaluate their suitability
the veins are not compressed at any point by coils, pads, for use as conduits, either in-situ for bypass grafting in
or supports. In the legs the popliteal vein is most prone occlusive peripheral vascular disease or indeed for coro-
to compression behind the knees to which particular at- nary artery bypass grafting particularly where there has
tention should be paid, and obviously any tourniquets or been previous surgery or a past history of thrombosis/
cuffs used to enhance arterial imaging must be slackened thrombophlebitis.
10.5 Lower Limbs
119 10
10.5.1 Lower Limb Thrombosis symptomatic DVT will very often exhibit this contrast en-
hancement of the adjacent venous wall around the filling
The MR imaging signs of deep venous thrombosis are defect, the bulls-eye sign [12]. This venous wall thicken-
those of a filling defect within the venous lumen, often ing and enhancement diminishes over time as the throm-
with expansion of the vein, and this filling defect may bus organizses and contracts; thus this phenomenon can
be completely or partially occlusive. Phlebitis may ac- help in gauging the age of thrombosis. One phenomenon
company thrombosis and is recognized by enhancement, that should be appreciated is the potential for hematocrit
thickening, and poor definition of the venous wall. Acute layering of gadolinium contrast agent in stagnant veins,
though this is obviously different from the intravascular
filling defect of thrombosis, with which it should not be
confused [13]. The advantages of MRI for evaluation of
lower limb DVT over ultrasonography are that it is not
limited for iliac and IVC assessment of thrombus exten-
sion as ultrasound can be, by depth and bowel gas etc., and
that with high spatial resolution sequences MRI is equally
accurate in assessing the small calf veins, where ultrasound
can be difficult in the swollen leg ( Figs. 10.410.9). MRI

Fig. 10.4A,B. Contrast-enhanced MRV of femoral veins in patient
with recurrent right lower limb swelling. A Overview coronal MIP Fig. 10.5A,B. Contrast-enhanced MRV of lower limb veins in steady
shows extensive venous occlusion. Note lack of long saphenous and state in a patient with new left calf swelling. A Overview coronal MIP
superficial femoral veins. Common and external iliac vein occlusion shows expansion and occlusion of left gastrocnemius veins while
results in collateral flow via obturator veins to the pelvis; compare deeper calf veins are patent. B Transverse thin-slab MIP shows ex-
with normal left side. B Transverse thin-slab MIP shows contracted, panded occlusive thrombus with surrounding enhancement typical of
non-occlusive thrombus in right femoral vein acute thrombosis; contrast with opposite limb
120 Chapter 10 Magnetic Resonance Venography

also maps collateral venous drainage pathways in an angi-

ographic form and this can be important in the post-phle-
bitic limb when potential interventions and being planned.
Thus MRI is a comprehensive tool for delineating throm-
bosis in the whole lower limb as an objective method that
can evaluate the full extent of DVT while gauging the age
of thrombosis and assess collateral pathways, allowing ac-
curate follow-up.


Fig. 10.7A,B. First pass arterial and steady state phase of contrast-
C enhanced MRA/V of pelvic and thigh veins, MIP projections. Patient
with critical ischemia of left leg showing occluded left common femo-
Fig. 10.6AC. Contrast-enhanced MRV of lower limb veins in steady ral and deep femoral arteries. The ex-intravenous drug abuser also has
state in an intravenous drug abuser with painful left leg swelling. occluded common femoral and iliac veins with collateralization via
Coronal thin-slab MIP shows expansion and occlusion of left thigh obturator veins to internal iliac veins (note similar venous problems in
veins by thrombus with extensive enhancing vein wall thickening plus right groin). Patient surprisingly still has patent long saphenous veins
extension of thrombus up into iliac veins as potential bypass conduit for arterial repair
10.5 Lower Limbs
121 10
10.5.2 Venous Mapping and Arteriovenous

While ultrasonography is the mainstay of assessment of

simple superficial venous varicosity, CE-MRV in the steady
state phase with blood pool contrast agent may be ex-
tremely helpful in challenging cases, as it provides a read-
ily understandable angiographic map, and workstation
analysis can rapidly evaluate vessel caliber and length to
aid the planning of surgery or endovenous techniques.
CE-MRV also clearly delineates the anatomical sites of
perforators that cross fascial planes to communicate be-
tween the veins of the superficial and deep compartments.
The technique also easily demonstrates variant anatomy
such as persistent sciatic veins, Giacomini veins, and other
aberrant connections while simultaneously evaluating for
potential complicating deep venous thrombosis. This is
particularly important in the KTWS which is a triad of
A soft tissue and/or bony hypertrophy along with cutane-
ous capillary malformation (port-wine type staining) and
venous varicosities/malformations ( Figs. 10.10, 10.11).

Fig. 10.8A,B. Steady state phase of contrast-enhanced MRA of right

calf in patient with critical lower limb ischemia and ulceration. In ad-
dition to severe occlusive arterial disease, the thin-slab coronal (a) and
transverse (b) MIP reconstructions revealed clinically unsuspected
calf-limited DVT

Fig. 10.9. Steady state phase of contrast-enhanced MRA of lower B

limbs in patient with exacerbation of longstanding brawny left lower
limb swelling. Ultrasound examination proved futile. MRV in steady Fig. 10.10A,B. Patient with extensive varicose veins of right leg.
state shows chronic edematous changes in calf but patent calf veins. Steady state phase of contrast-enhanced MRV of thigh veins thin
At the groin (not shown) the common femoral vein was found to be slab coronal & transverse MIP reconstructions. Note enlarged, persist-
chronically occluded along with flow across to the right iliac veins via ent sciatic vein connecting to deep femoral veins of thigh and large
tortuous collaterals Hunter type perforator compare to normal left side
122 Chapter 10 Magnetic Resonance Venography

In the management and diagnosis of KTWS it is particularly agents. They give excellent first pass arterial imaging to de-
important, prior to considering surgical intervention or lineate inflow anatomy, and also provide exquisite detail of
sclerotherapy, to eliminate the possibility of DVT and check the AVM nidus and any muscular or skeletal involvement
the integrity of the deep veins, which are often aberrant with as well as the venous drainage pathways ( Fig. 10.12).
enlarged lateral thigh veins or persistent sciatic veins [14]. In the assessment of lower limb superficial veins as po-
The assessment of arteriovenous malformations (AVM) tential conduits for bypass grafting, CE-MRV again provides
is a further example of the utility of blood pool contrast a rapidly interpretable visual map, and modern workstation

Fig. 10.11. Steady state phase of contrast-en-

hanced MRV of thigh veins. Volume-rendered
slab reconstruction (left) with accompanying
10 automated CPR unfolded vessel view (right)
and orthogonal reconstructions axial to vein
(middle). Patient with extensive varicose veins
of left leg and particularly dilated and tortuous
long saphenous vein


Fig. 10.12a,b. Patient with deep AVM, left anterior thigh. (A) First anterior to distal left femur, including small phleboliths manifest as
pass coronal MRA shows normal vasculature of thighs apart from small tiny foci of susceptibility artifact. There is no bony involvement, and
lateral superior geniculate artery branch feeding deep AVM on the left. no large draining veins are demonstrated
(B) Steady state phase shows deep nidus of AVM in vastus intermedius
10.6 Abdominal and Pelvic Veins
123 10

analysis software with automated center-line definition for 10.6 Abdominal and Pelvic Veins
curved MPR allows accurate quantification of vessel caliber
throughout the length of the vein. Clearly, this is a major 10.6.1 Inferior Vena Cava and Pelvic Veins
advantage of using a blood pool contrast agent such as
Vasovist when performing CE-MRA for peripheral arterial The inferior vena cava (IVC) is most commonly evalu-
occlusive disease, since the simple addition of steady state ated along with the pelvic veins when lower limb venous
high-resolution sequences at the end of the standard first thrombosis is assessed, since extension of DVT to the large
pass scan will provide all the venous information required iliac veins and IVC poses an increased potential threat of
to plan any potential surgical intervention, obviating the pulmonary thromboembolism. The use of MRI for IVC
scheduling of additional studies ( Figs. 10.13, 10.14). evaluation is advantageous as it may be poorly opacified
on CT, the interpretation of which can be challenging since
the IVC is prone to flow artifacts with complex admixture
of unopacified blood. IVC patency and caliber determina-
tion are important to determining the safety of potential
intervention in terms of caval filter placement; in particular,
the size of the IVC will determine the type of caval filter to
be used, as in cases of megacava a birds-nest type filter
is required. Determining the position of the main renal
veins and any accessory/aberrant renal veins can alter the
position of deployment, as these may necessitate suprarenal
placement, a priori knowledge of which is helpful prior to
obtaining the patients consent and planning the interven-
tion. Variants of the IVC itself such as duplicated infrarenal
IVC (which occurs in up to 3% of subjects; solitary left
infrarenal IVC is less common at ~1% [15]) will also affect
management. As in other areas- the use of blood pool con-
trast agent is beneficial and steady state high-resolution im-
aging is easily achieved for the fixed retroperitoneal IVC- as
it is not disturbed by respiratory motion artifacts.
An advantage of simultaneous venous and arterial
Fig. 10.13. Thin-slab MIPs of normal (4-mm caliber) long saphenous imaging in the steady state is ready depiction of the re-
veins suitable for use as bypass conduits lationships of the common iliac vessels for the diagnosis
of May-Thurner syndrome, where thrombosis of the left
iliac veins is precipitated through compression of the
proximal left common iliac vein by the right common
iliac artery crossing point. While May-Thurner syndrome
as a cause of left iliofemoral DVT is well recognized it has
been thought relatively uncommon; however, recent work
has suggested a more subtle and hitherto unrecognized
role of iliac vein compression in non-thrombotic chronic
venous insufficiency [16, 17].

10.6.2 Gonadal Veins

For male patients with simple scrotal varicocele the pre-

procedural evaluation of gonadal venous anatomy is sel-
dom required prior to embolotherapy. However, in female
patients with vulval varicosities or suspected pelvic con-
gestion syndrome it is helpful to define whether there is
abnormality of ovarian veins, particularly as their pelvic
Fig. 10.14. Coronal overview MIP of calves in patient with critical
connections may be complex. MRV for gonadal veins with
lower limb ischemia showing poor-quality long saphenous veins with blood pool contrast agent can be performed with stand-
nothing suitable for bypass conduit ard CE-MRA acquisitions or fat-suppressed T1w GRE se-
124 Chapter 10 Magnetic Resonance Venography

quences as per renal venous imaging, though the optimized

steady state high-resolution sequences work very well for
these retroperitoneal and pelvic structures that are little
affected by respiratory excursions. A dilated ovarian vein
alone does not denote pelvic congestion syndrome; this is a
common asymptomatic finding even when reflux is evident
and can be found in nearly 40% of asymptomatic parous
women, as shown by several MRI and CT studies. Hence
pelvic congestion syndrome should be diagnosed only in an
appropriate clinical context and with other supporting fea-
tures, such as evidence of severe ovarian vein dilation with
reflux plus pelvic varicosities around the uterus and ovaries
or vulval varicosities [1824]. Vulval varicosities are a not
uncommon feature of pregnancy that resolve post partum
( Figs. 10.15, 10.16); however, in the absence of pregnancy
they are much more unusual and associated not only with
pelvic congestion syndrome but also with ipsilateral lower
limb varicosites and even KTWS.
Disorders of the penile circulation are another area A

that blood pool contrast agent MRA is opening up to

investigation. The high resolution afforded by steady state
imaging allows delineation of the small bulbar and penile
arteries in the deep pelvis which otherwise, with conven-
tional angiography, require super-selective catheterization
for visualization ( Fig. 10.17). The use of blood pool agent
10 imaging additionally allows assessment of the integrity of
the penile venous drainage, i.e., the corpora, allowing a
comprehensive assessment.

10.6.3 Renal Veins

Specific MR imaging of the renal veins is most commonly

required today as part of the workup of potential live re-
nal transplant donors. Altruistic would-be donors require
careful assessment to ensure that their anatomy is suitable
for surgery, and this is not confined to the number of ure-
ters and arteries as both multiple and aberrant renal veins
can be just as problematic; this is particularly the case if
laparoscopic nephrectomy is being considered. MRA is
advantageous for potential renal donor assessment since
the least-harm approach is mandatory for these other-
wise healthy people. CE-MRA ensures no radiation dose
and has been proven accurate in assessment of arterial
anatomy. Congenital anomalies of the renal veins are not
Fig. 10.15A,B. Coronal MIP (A) and transverse MPRs (B) of same da-
uncommon with an incidence of 12% for a purely retro- taset of pelvis in steady state after Vasovist administration in patient
aortic left renal vein and approximately 8% for the more with vulval varicosities demonstrating ipsilateral deep venous vari-
common circumaortic renal venous collar, where the left cosities in right thigh drain to internal iliac pelvic venous varicosities
kidney is drained by both a normally situated pre-aortic (incidental uterine fibroid)
vein and a more caudally draining retro-aortic vein [24].
However, delineation of the renal venous vasculature With a blood pool contrast agent this is extended, and
with ECS agents may be more problematic than arterial using dynamic time-resolved techniques renal venous
imaging, particularly for the short right renal vein, and conspicuity is increased due to the relatively lower re-
there is only a limited time window for image acquisition. nal extraction in first pass. Renal MRV with blood pool
10.6 Abdominal and Pelvic Veins
125 10

Fig. 10.16. Dynamic multiphase first pass MIPs and subsequent early reflux in incompetent left ovarian vein. (Courtesy of Dr J Bottom-
steady state phase slab MIP of contrast-enhanced MRA/V of abdomi- ley, Sheffield Teaching Hospitals, England, UK)
nopelvic veins. Patient with vulval varicosities demonstrating rapid

Fig. 10.17AC. Sagittal (A) and transverse (B,C) thin-slab MIP images
of male pelvis in steady state in a patient with erectile dysfunction fol-
lowing injury. Images demonstrate clear patency of artery to the bulb
and penile arteries plus normal enhancement of corpus spongiosus
and proximal corpora but lack of enhancement of corpora cavernosa C
in the shaft, presumably from post-traumatic thrombotic injury
126 Chapter 10 Magnetic Resonance Venography

contrast agent is usually performed with standard breath- use of a low dose of a kinetically stable agent may be less
hold CE-MRA ( Fig. 10.18) or fat-suppressed T1w GRE risky than administration of iodinated contrast medium,
sequences (such as THRIVE, VIBE, and LAVA) to elimi- precipitating a worsening of renal function.
nate respiratory artifacts, though the optimized steady
state high-resolution non-breath-hold sequences may be
surprisingly successful for these retroperitoneal structures 10.6.4 Portal Venous Circulation
even without the addition of navigator techniques which
are being researched. Portal and mesenteric venous imaging with contrast-
Despite the excellent image quality now afforded by enhanced MRA has been very successful but with the use
modern multislice CT imaging there is still occasion to of ECS contrast agents this relies on accurate timing of
use MRI when investigating suspected renal vein throm- breath-hold acquisitions, which may be unreliable, and
bosis. MRI is particularly indicated in younger patients the ideal portal phase can be missed unless time-resolved
without malignant disease and in pregnancy, for example, MRA sequences are available. With a blood pool contrast
in nephrotic syndrome and other systemic conditions agent the imaging window is prolonged and hence more
such as the connective tissue disorders. Although renal easily captured and since there is little extraction to the
vein thrombus is often well-visualized when renal and ECS compared with conventional agents the venous con-
adrenal carcinomas are staged on multislice CT, the dif- spicuity in the first pass of the portal phase is increased.
ferentiation of bland from malignant thrombus may not Portal MRV with a blood pool contrast agent is currently
be easy, and studies have shown that gadolinium enhance- best performed with standard resolution breath-hold CE-
ment of the thrombus on delayed-phase MRI, indicating MRA acquisitions or fat-suppressed T1w GRE sequences
neovascularity, is predictive of malignancy. There are (such as THRIVE, VIBE and LAVA) since the optimized
of course concerns about the use of gadolinium-based steady state high-resolution sequences applicable in the
contrast agents in severe renal failure (i.e., patients with limbs are not suitable for the upper abdomen, where res-
chronic kidney disease stage 5 on dialysis), but in milder piratory movement artifacts are problematic. This limita-
degrees of renal impairment (CKD stages 3 and 4) the tion may in future be overcome with the use of navigator
10 techniques.

10.7 Upper Limbs and Thorax

The indications for venous imaging in the upper limbs

and thorax are similar to those for the lower limbs, except
that nowadays the most common request is to evaluate
for iatrogenic upper extremity venous thrombosis and/
or central vein stenosis as complications of prior central
venous catheterization. Placement of central venous cath-
eters (CVC) via the jugular or subclavian routes and PICC
lines are increasingly common procedures in modern
medicine to facilitate patient monitoring and administra-
tion of therapies and to provide parenteral nutrition.
Unfortunately, the use of CVCs can be complicated
by both venous thrombosis and stenosis which may be-
come symptomatic with either unilateral upper limb/neck
edema or even a full superior vena caval occlusion syn-
drome. Furthermore, occult central venous stenosis and/
or thrombosis may complicate what was initially thought
to be an uncomplicated catheter insertion in up to 65% of
patients, particularly where there has been prior interven-
Fig. 10.18. Breath-hold coronal contrast-enhanced MRA thin-slab tion [2534]. Hence visualization of the central thoracic
MIP image in equilibrium phase of potential living renal donor after veins prior to attempted CVC placement is playing an
administration of Vasovist. Image shows circumaortic left renal ve-
increasing role for patients with problematic vascular ac-
nous collar; part of the left renal venous drainage is complex, anas-
tomosing to lumbar veins and then passing retro-aortic and caudal
cess, with the aim of optimizing the lifespan, durability,
to join IVC. Note short length of right renal vein from union of two and effectiveness of central venous catheters. Ultrasono-
tributaries prior to cava graphy can assess the veins in the arms, however, the more
10.7 Upper Limbs and Thorax
127 10

central veins are inaccessible to direct ultrasonographic tous disorders may benefit from MRI, particularly as it
visualization while MRV provides excellent imaging. can be used for follow-up in assessment of response to
Visualization of the central veins of the thorax is also immunosuppressive therapies without concerns regard-
useful in assessment of superior caval obstructing diseases ing excessive radiation exposure. In this context the high
as, while CT will usually define simple SVC obstructions, relaxivity of a blood pool contrast agent is helpful for as-
from lung cancer, for example, more diffuse inflammatory sessing venous stent patency, increasing the conspicuity of
diseases such as fibrosing mediastinitis and granuloma- blood within the stent lumen ( Figs. 10.19, 10.20).



Fig. 10.19. A,B Breath-hold coronal contrast-enhanced MRA MIPs in high-resolution imaging of right arm shows sites of two of the previ-
first pass arterial and equilibrium phases of upper thoracic vessels in ous failed arteriovenous fistulae (arrows to anastomoses) and small-
patient with three previous bilateral failed arteriovenous fistulae. Note caliber basilic vein (asterisk smaller than and adjacent to brachial
occlusive damage to left axillary vein. C,D Subsequent steady state artery) unsuitable for basilic transposition operation
128 Chapter 10 Magnetic Resonance Venography

As in the lower limbs, dedicated small field-of-view,

high-resolution steady state imaging is of great value for
delineating specific lesions such as AVMs ( Fig. 10.21).
Indeed, such is the high spatial resolution achievable that
not only are arterial and venous structures readily distin-
guished but also excellent morphological musculoskeletal
information is obtained, such that these studies can obvi-
ate the standard T1w, T2w, and STIR imaging until now
usually employed for AVM assessment.

10.8 Conclusion

Venous evaluation using MRI with blood pool contrast

agents is a rich and relatively unexplored field of imaging
Fig. 10.20. Breath hold coronal Vasovist contrast-enhanced MRA base
image in equilibrium phase of upper thoracic vessels in patient with
where emerging applications bring real benefit through
overlapping nitinol stents in situ in left brachiocephalic vein through to enhancing patient diagnosis and hence treatment, whether
SVC note excellent conspicuity of blood pool within stent lumen. it be a dedicated venous study or the addition of venous



Fig. 10.21AC. High-resolution small field-

of-view imaging in steady state of right hand
after Vasovist administration in a patient with
extensive AVM of the little finger. Overview
MIP (A), example base image (B), and three
orthogonal plane MPRs (C) show excellent vas-
cular and musculoskeletal detail
129 10

imaging to an arteriographic examination such as in evaluation of venous limb disorders in those body areas
critical lower limb ischemia to add value. While stand- not subject to excessive respiratory movement. In future,
ard resolution CE-MRA techniques for MRV are readily vascular evaluation with blood pool contrast agents will
applicable for all body areas, the high spatial resolution likely routinely include high spatial resolution imaging to
imaging that is afforded by the prolonged steady state evaluate not only the arteries but also the veins as part of
phase of blood pool contrast agents is highly useful in the a comprehensive examination.

Take home messages

The use of Vasovist blood pool contrast agent High spatial resolution steady state imaging is not
simplifies the performance of contrast enhanced only improving diagnostic performance but can can
MRV. obviate the need for other time consuming imaging
Steady state imaging of Vasovist enhanced stud- sequences, for example in AVM assessment.
ies can be performed at very high spatial resolu- High spatial resolution steady state imaging is also
tion even on older generation MRI systems since opening up new diagnostic indications, this is par-
there is no requirement for very short TR/TE for ticularly so in relation to venous imaging and adds
these sequences. value to arteriographic MRA studies.

References 12. Froehlich JB, Prince MR, Greenfield LJ, Downing LJ, Shah NL,
Wakefield TW (1997) Bulls-eye sign on gadolinium-enhanced
1. Sidhu PS, Alikhan R, Ammar T, Quinlan DJ (2007) Lower limb magnetic resonance venography determines thrombus presence
contrast venography: a modified technique for use in thrombo- and age: a preliminary study. J Vasc Surg 26:809816
prophylaxis clinical trials for the accurate evaluation of deep vein 13. Wang MS, Haynor DR, Wilson GJ, Maki JH (2006) Intravascular
thrombosis. Br J Radiol 80:859865 hematocrit layering in equilibrium phase contrast-enhanced MR
2. Peterson DA, Kazerooni EA, Wakefield TW, Knipp BS, Forauer AR, angiography of the peripheral vasculature. J Magn Reson Imaging
Bailey BJ, Sullivan VV, Proctor MC, Henke PK, Greenfield LJ, Stanley 24:13931400
JC, Upchurch GR jr (2001) Computed tomographic venography is 14. Bastarrika G, Redondo P, Sierra A, Cano D, Martinez-Cuesta A,
specific but not sensitive for diagnosis of acute lower-extremity Lopez-Gutierrez JC, Cabrera J (2007) New techniques for the
deep venous thrombosis in patients with suspected pulmonary evaluation and therapeutic planning of patients with Klippel-
embolus. J Vasc Surg 34:798804 Trenaunay syndrome. J Am Acad Dermatol 56:242249
3. Prince MR, Sostman HD (2003) MR venography: unsung and un- 15. Minniti S, Visentini S, Procacci C (2002) Congenital anomalies
derutilized. Radiology 226: 630632 of the venae cavae: embryological origin, imaging features and
4. Prince MR, Grist TM, Debatin JG (2003) 3D Contrast MR angiogra- report of three new variants. Eur Radiol 12:20402055
phy, 3rd edn. Springer Verlag, Berlin Heidelberg New York 16. Raju S, Neglen P (2006) High prevalence of nonthrombotic iliac
5. Moody AR, Pollock JG, OConnor AR, Bagnall M (1998) Lower-limb vein lesions in chronic venous disease: a permissive role in patho-
deep venous thrombosis: direct MR imaging of the thrombus. genicity. J Vasc Surg 44:136143
Radiology 209:349355 17. Oguzkurt L, Tercan F, Pourbagher MA, Kizilkilic O, Turkoz R, Boyvat
6. Moody AR (2003) Magnetic resonance direct thrombus imaging. J F (2005) Computed tomography findings in 10 cases of iliac vein
Thromb Haemost 1:14031409 compression (May-Thurner) syndrome. Eur J Radiol 55:421425
7. Fraser DG, Moody AR, Davidson IR, Martel AL, Morgan PS (2003) 18. Nicholson T, Basile A (2006) Pelvic congestion syndrome, who
Deep venous thrombosis: diagnosis by using venous enhanced should we treat and how? Tech Vasc Intervent Radiol 9:1923
subtracted peak arterial MR venography versus conventional 19. Koc Z, Ulusan S, Tokmak N, Oguzkurt L, Yildirim T (2006) Double ret-
venography. Radiology 226:812820 roaortic left renal veins as a possible cause of pelvic congestion syn-
8. Lebowitz JA, Rofsky NM, Krinsky GA, Weinreb JC (1997) Gadolin- drome: imaging findings in two patients. Br J Radiol 79:152155
ium-enhanced body MR venography with subtraction technique. 20. Rozenblit AM, Ricci ZJ, Tuvia J, Amis ES jr (2001) Incompetent and
AJR Am J Roentgenol 169:755758 dilated ovarian veins: a common CT finding in asymptomatic pa-
9. Fraser DG, Moody AR, Davidson IR, et al (2003) Deep venous rous women. AJR Am J Roentgenol 176:119122
thrombosis: diagnosis by using venous enhanced subtracted peak 21. Belenky A, Bartal G, Atar E, Cohen M, Bachar GN (2002) Ovarian
arterial MR venography versus conventional venography. Radiol- varices in healthy female kidney donors: incidence, morbidity, and
ogy 226: 812820 clinical outcome. AJR Am J Roentgenol 179:625627
10. Demondion X, Bacqueville E, Paul C, Duquesnoy B, Hachulla E, Cot- 22. Nascimento AB, Mitchell DG, Holland G (2002) Ovarian veins: mag-
ten A (2003) Thoracic outlet: assessment with MR imaging in asymp- netic resonance imaging findings in an asymptomatic population.
tomatic and symptomatic populations. Radiology 227:461468 J Magn Reson Imaging 15:551556
11. Wang MS, Haynor DR, Wilson GJ, Leiner T, Maki JH (2007) Maxi- 23. Bell D, Kane PB, Liang S, Conway C, Tornos C (2007) Vulvar varices:
mizing contrast-to-noise ratio in ultra-high resolution peripheral an uncommon entity in surgical pathology. Int J Gynecol Pathol
MR angiography using a blood pool agent and parallel imaging. 26:99101
[Journal Article. Research Support, Non-U.S. Govt] J Magn Reson 24. Roditi GH, Buff BL, Esterbrook Longmaid H (1996) MR Venography
Imaging 26:580588 of left renal vein anomalies. Clin Radiol 51:861864
130 Chapter 10 Magnetic Resonance Venography

25. Taal MW, Chesterton LJ, McIntyre CW (2004) Venography at inser-

tion of tunnelled internal jugular vein dialysis catheters reveals
significant occult stenosis. Nephrol Dialysis Transplant 19:1542
26. Oguzkurt L, Tercan F, Yildirim S, et al (2005) Central venous ste-
nosis in haemodialysis patients without a previous history of
catheter placement. Eur J Radiol 55:237242
27. Murphy GJ, White SA, Nicholson ML (2000) Vascular access for
haemodialysis. Br J Surg 87:13001315
28. Kroencke TJ, Taupitz M, Arnold R, et al (2001) Three-dimensional
gadolinium-enhanced magnetic resonance venography in sus-
pected thrombo-occlusive disease of the central chest veins.
Chest 120:15701576
29. Shankar KR, Abernethy LJ, Das KS, et al (2002) Magnetic resonance
venography in assessing venous patency after multiple venous
catheters. Journal of Pediatric Surgery 37: 175-179
30. Rose SC, Gomes AS, Yoon HC (1996) MR angiography for mapping
potential central venous access sites in patients with advanced
venous occlusive disease. AJR Am J Roentgenol 166:11811187
31. Pagnan L, Tona G, Belgrano M, et al (2005) Direct contrast en-
hanced MR in the study of central venous accesses in children
receiving total parenteral nutrition. Radiol Med (Torino) 110: 241
32. Oxtoby JW, Widjaja E, Gibson KM, Uzoka K (2001) 3D gadolinium-
enhanced MRI venography: evaluation of central chest veins and
impact on patient management. Clin Radiol 56:887894
33. Finn JP, Zisk JH, Edelman RR, et al (1993) Central venous occlusion:
MR angiography. Radiology 187:245251
34. Planken RN, Tordoir JH, Duijm LE, de Haan MW, Leiner T (2007)
Current techniques for assessment of upper extremity vascula-
ture prior to hemodialysis vascular access creation. Eur Radiol
10 17:30013011

Whole-body MRA
Harald Kramer, Maximilian F. Reiser, and Konstantin Nikolaou

11.1 Introduction 132

11.2 Clinical Rationale 132

11.3 Whole-body MRA 134

11.4 Whole-body MRA Protocols 135

11.5 Whole-body MRA with Intravascular Contrast Agents 136

11.6 Clinical Applications 136

11.6.1 Whole-body MRA as a Replacement for Run-off Studies 137
11.6.2 Thromboembolic Diseases 137
11.6.3 Screening for Atherosclerotic Disease 137

References 138
132 Chapter 11 Whole-body MRA

11.1 Introduction

The morbidity and mortality statistics of developed coun-

tries still show vascular diseases in a leading position.
Probably the most common manifestation of the systemic
disease atherosclerosis is peripheral vascular occlusive
disease (PAOD), while the most threatening manifesta-
tions are carotid, renal, and coronary artery stenoses
[15]. All of these manifestations have in common that
they need adequate imaging for early detection and op-
timized treatment planning. There are a number of di-
agnostic tools for each of these vascular territories, but
none of them is regarded as the standard of reference
for the entire arterial vascular bed. Until a few years ago,
magnetic resonance angiography (MRA) was a preferred
diagnostic tool for the assessment of single anatomical
territories, but it was not possible to perform a whole-
body MRA. Today this drawback has been overcome, and
by combining parallel imaging strategies with whole-body
MR systems, the complete vasculature of the body can be
examined at once. However, the last persisting constraint
is the somewhat limited spatial resolution compared with
other imaging methods such as conventional digital sub-
traction angiography (DSA), duplex ultrasonography, or
computed tomography angiography (CTA) [610]. Due
to the absence of ionizing radiation and recent develop-
ments regarding hard- and software, as well as sequence
technology and new contrast agents (CA), magnetic reso- B
11 nance angiography (MRA) should be considered the best
Fig. 11.1A,B. Comparison of a DSA (A) and an MRA (B, colors in-
candidate and a comprehensive diagnostic tool to image
verted) dataset. Due to the higher spatial resolution, small muscle-
the complete body vasculature. arterial branches (arrows) as well as collateral vessels (arrowhead) are
better depicted

11.2 Clinical Rationale

With an increasing average age of the population in the in- the need for alternative, less harmful, and ideally non-
dustrialized world, the incidence and prevalence of athero- invasive imaging techniques. Imaging of the carotid arteries
sclerotic and other vascular diseases rise steadily. However, for example is among other techniques a domain of
treatment options for vascular-related diseases are improv- duplex ultrasound [14, 15] ( Fig. 11.2). With this method,
ing. In this respect, an early diagnosis of vascular disease, the superficially localized common carotid arteries as well
if possible in an asymptomatic stage, is of high interest. as the proximal parts of the internal carotid arteries are
There are several clinical and laboratory tests such as the excellently imaged; it is possible not only to measure the
ankle-brachial-index (ABI) for the diagnosis of PAOD or vessel lumen but also to assess the vessel wall. However, this
prognostically important risk factors for developing athero- method is very user dependent and cannot be used in every
sclerotic plaques such as the absolute levels and relations of anatomical region. The evaluation of the renal arteries, for
high- and low-density lipoproteins (HDL/LDL), but before example, is often challenging with ultrasound, because of
a treatment decision is made, detailed imaging of the dis- the location of these vessels deep in the body and the sur-
eased vascular region is required [11, 12]. As mentioned rounding structures such as the bowel, which can impair
above, there are several established vascular imaging mo- the accessibility and assessability. Because of its high spatial
dalities for different vascular regions. Digital subtraction resolution and easy application, CTA constitutes another
angiography (DSA) is still regarded as the standard of refer- promising and non-invasive method for imaging of nearly
ence in most vascular territories [13] ( Fig. 11.1). However, all vascular regions at a high spatial resolution [8, 16, 17]
the combination of ionizing radiation, invasiveness, and ( Fig. 11.3). On the other hand, ionizing radiation and the
potentially nephrotoxic contrast agents (CA) demonstrates need for iodinated CA constitute a definite drawback.
11.2 Clinical Rationale
133 11

Fig. 11.2A,B. Comparison of a duplex

ultrasound image (A) and a magnified MRA
image (B) showing an occlusion of the internal
carotid artery

Fig. 11.3AC. Examples of CTA datasets. (A) Anterior and (B) pos-
terior VRT view of a peripheral run-off CTA showing occlusion of the
arterial vessels of the left calf. C Coronal reconstruction of a pulmonary
CTA showing major central pulmonary embolism
134 Chapter 11 Whole-body MRA


Fig. 11.4AC. MRA examinations of different anatomical regions. tropic spatial resolution. C First pass carotid MRA with nice display of
A Run-off study acquired in three stations in arterial first pass without the carotid arteries
venous opacification. B The possibility of 3D reformatting due to iso-

Magnetic resonance angiography is the only non- most distal vessel station, i.e. the lower leg, due to venous
11 invasive imaging method which does not entail the dis- opacification [20] ( Fig. 11.5).
advantage of ionizing radiation and which provides good To overcome the limited range of table movement and
image quality at a high spatial resolution throughout the to avoid having to reposition the patient during whole-
entire body ( Fig. 11.4). Even the CA used are very well body MRA, crucial impetus was provided by independent
tolerated if renal function is not impaired [18]. research groups for technical implementations to achieve
Recent developments in MR system technology can whole-body MRA. Manually movable table platforms were
further expand the application of MRA from imaging developed, e.g., SKIP (Stepping Kinematic Imaging Plat-
dedicated anatomical regions to a real whole-body form, Magnetic Moments, Bloomfield, MI, USA) and An-
MRA [19, 20]. gioSURF (Angiographic System for Unlimited Rolling
Field-of-views, MR-Innovation GmbH, Essen, Germany),
which exceeded the range of movement of the motorized
11.3 Whole-body MRA patient tables used in MR scanners, thereby extending the
achievable FOV to approximately 200 cm and giving rise to
First attempts to image the entire arterial system from the real whole-body MRA [21, 22] ( Fig. 11.6). With these
skull base down to the feet were limited by MR system table platforms, a step-by-step whole-body MRA is possible
technology. The majority of MR systems have a restricted with five to six stations. The inadequate coverage of this
range of patient table movement of typically less than examination field with RF surface coils is circumvented by
150 cm. This means that whole-body imaging can be per- a stationary RF surface coil pair positioned in the isocenter
formed only with two contrast injections, requiring repo- of the magnet. A phased-array coil is system inbuilt in most
sitioning of the patient during the examination. This way, scanners in the patient table and delivers signals from the
in a first step only the thoracic and head and neck vessels posterior region of the patient. On the anterior side of the
can be displayed, and after repositioning of the patient the patient, a second phased-array surface coil is mounted on a
vessels of the lower body including the abdominal aorta coil glider, which is height-adjustable and held in position
and the iliac and peripheral arteries are imaged in a sec- by two arms at the isocenter of the magnet. The patient is
ond step. This second part covers a very large anatomical positioned on an MR-compatible table platform mounted
area which often leads to an impaired image quality in the on rollers. Using this technique, the patient can be manu-
11.4 Whole-body MRA Protocols
135 11

Fig. 11.5. Whole-body MRA dataset acquired

with a standard MR system. Note the gap between
the first MRA station (carotid arteries) and the sec-
ond part of the whole-body MRA after the patient
was repositioned

Fig. 11.6. Whole-body MRA dataset acquired

with a standard MR system in combination with
a rolling table platform. Whole-body-MRA is per-
formed in six steps from head to feet. Note the
beginning venous enhancement of the renal veins
(shorter arrow) and the portal vein (arrow)

Fig. 11.7. Whole-body MRA dataset acquired

with a dedicated whole-body MR system in four
steps from head to feet

Fig. 11.5. Fig. 11.6. Fig. 11.7.

ally pulled stepwise through the magnet passing between ered with dedicated surface coils before the examination.
the two surface coils. The RF surface coils provide the SNR During the examination, every anatomical region can be
required for high image quality without the need to com- moved to the isocenter of the magnet, and the activated
pletely enclose the patient in RF coils from head to toe. First coils or respective coil elements can be chosen as needed
results using this technique have been promising but come [19, 20, 23] ( Fig. 11.7).
along with a somewhat reduced spatial resolution, due to
the necessity of fast imaging and coverage of a large FOV.
In 2004, a new MR system generation with an exten- 11.4 Whole-body MRA Protocols
sive moving table in combination with a dedicated matrix
coil system for whole-body imaging was presented. With To achieve the best possible image quality, dedicated im-
this type of MR system, the patient does not need to be aging and/or optimized CA application protocols are rec-
repositioned, and no dedicated rolling table platform is ommended. When using standard extracellular CA, the
necessary. The systems patient table allows a movement of primary goal is to image every vessel region in a purely
more than 200 cm and the patient can be completely cov- arterial phase, without any venous contamination. When
136 Chapter 11 Whole-body MRA

imaging various vascular territories, this can be achieved mentioned dual injection CA application protocol [27].
by several techniques such as test-bolus or bolus-chase Before whole-body MRA with blood pool agents is done,
applications for a precise timing of the CA arrival or, po- some important questions have to be answered:
tentially, by additional venous compression to slow down 1. Do I need to image the entire vasculature in an arterial
venous enhancement, especially in the lower extremity. In first pass?
whole-body MRA, these techniques are of major impor- 2. If not, which vascular regions are most important and
tance, but they are often enough not sufficient to get opti- should be imaged during arterial first pass?
mal results. Here, acquisition time, spatial resolution, and
table movement have to be compromised [24]. It is cer- As mentioned above, it is not possible to image all vas-
tainly possible to image the first vascular region without cular territories at high spatial resolution without venous
venous contamination and at high spatial resolution, but in overlay with only one CA bolus. On the one hand, when
the following vascular regions image quality may decrease. using blood pool agents, spatial resolution can be compro-
For example, it is not possible to image the renal arteries mised during first pass of the contrast to get a whole-body
without venous overlay directly after imaging the carotid MRA dataset without venous overlay, because there is the
arteries. During data acquisition of the carotid arteries, opportunity to acquire high spatial resolution datasets
contrast runs down to the lower body part and leads to during the steady state or equilibrium phase [28]. On
venous contamination, for example, in the renal veins, and the other hand, a maximum of two vascular regions can
later on in the venous vessels of the lower leg. One solution be acquired at high spatial resolution in the arterial first
to restrict the amount of venous enhancement is to shorten pass, and every other region is acquired at high spatial
the acquisition time of every imaged station, which will resolution during steady state [2931] ( Fig. 11.8).
automatically result in a reduced spatial resolution and a In whole-body MRA with intravascular contrast dur-
decreased signal-to-noise ratio (SNR). ing steady state, there are significant inherent differences
The introduction of the aforementioned dedicated when examining various vascular beds. Thigh and calf
whole-body MR systems helped to overcome many of stations are not vulnerable to movement artifacts from
these limitations. The large range of table movement of- breathing or pulsation. Here, steady state imaging or
fers the flexibility to move to every vascular region in a respective acquisition time is more or less unlimited, and
very short time, and the combination with dedicated ma- acquisition times of more than 5 min can be tolerated,
trix surface coil systems allows for implementing parallel leading to an unprecedented spatial resolution. However,
11 imaging techniques for an enhanced data acquisition. For the upper body, with the aortic arch and neck and head
such a setup of hard- and software (i.e., MR sequences), vessels, as well as the abdomen, can be affected by breath-
several different imaging protocols for whole-body MRA ing artifacts and potentially by cardiac pulsation artifacts.
exist [20, 25, 26]. The most promising approach seems to This is mostly true for the thorax and abdomen, while the
be a protocol using two separate contrast injections, to internal carotid arteries are typically not affected by this.
avoid imaging of the entire arterial vasculature with only The most challenging vascular region to be acquired dur-
one contrast bolus. With such a dual-bolus technique, the ing first pass imaging of the intravascular contrast is the
arteries of the upper thorax, neck, and head are imaged abdominal station. Here, breathing and pulsation artifacts
first. Then the table moves the patient down to the calf occur in datasets with long acquisition times. Until there
station and during this table movement, the first contrast are respiratory-triggered acquisition strategies that can be
bolus is overtaken. This way, the most distal vascular used for imaging this region, steady state imaging will be
station, i.e., the calves, can be acquired without venous restricted to the duration of a single breath hold, and the
contamination. After a short break, the abdominal arter- full capacity of spatial resolution cannot be exploited.
ies are imaged with a second contrast bolus followed by
acquisition of the thigh arteries. With this protocol it is
possible to perform real whole-body MRA without dis- 11.6 Clinical Applications
turbing venous contamination at good arterial contrast.
Although most patients initially complain about problems
in a single vascular territory, vascular diseases typically
11.5 Whole-body MRA with Intravascular affect the entire arterial system, from the intracranial
Contrast Agents down to the pedal vessels. Therefore, whole-body vascular
imaging seems to be a fast and attractive diagnostic ap-
When using intravascular or blood pool contrast agents proach for different clinical scenarios. Here, application
such as Vasovist (Gadofosveset, Bayer Schering Pharma of an intravascular contrast agent such as Vasovist could
AG, Berlin, Germany) for whole-body MRA, it is certainly result in extended imaging time, potentially higher spatial
not possible (and also not necessary) to perform the afore- resolution, and larger anatomical coverage.
11.6 Clinical Applications
137 11


Fig. 11.8AC. Whole-body MRA dataset acquired in the steady state with Vasovist. (A) Arterial first pass im-
aging of every anatomical station without any venous enhancement is not possible, but every station can be
acquired in the steady state with high spatial resolution, limited only by motion due to breathing (B and C)

11.6.1 Whole-body MRA as a Replacement tic valve lesions) or the ascending aorta may affect the
for Run-off Studies entire arterial system with the potential fatal consequences
of stroke or limb ischemia. In these patients whole-body
Multistation MR angiography covering the abdominal is the perfect procedure for further diagnostic workup.
aorta down to the pedal arteries is a well-established Rather than looking only at the vascular area that becomes
technique for the evaluation of peripheral artery disease symptomatic at first, whole-body MRA can display the
(PAD). However, using the recent hard- and software embolic damage to the entire arterial system, discovering
developments described above, the examination can easily several other vascular regions that might become sympto-
be extended to whole-body MRA without increasing the matic in the future and can be treated in advance.
dose of contrast. The scan time, the in-room time, and the
time required for post-processing and reading of the ex-
amination increase to a certain degree; however, the high 11.6.3 Screening for Atherosclerotic Disease
prevalence of concomitant atherosclerotic findings in the
carotid arteries and thoracic aorta in PAD patients justify As the prevalence of atherosclerotic disease increases
these investments and improve patient care. and thus has a higher socioeconomic impact, screening
may appear reasonable, especially among patients at high
risk such as smokers or diabetics. Although most athero-
11.6.2 Thromboembolic Diseases sclerotic lesions are not completely reversible, lifestyle
changes, reduction of risk factors, and pharmaceutical
Arterial emboli originating from the heart (left atrial treatment can decelerate or even stop the progression of
thrombi in atrial fibrillation, left ventricular thrombi, aor- the disease.
138 Chapter 11 Whole-body MRA

Take home messages

Up to now, invasiveness, radiation exposure, con- Non-invasiveness, three-dimensionality, extended
trast dose limitations, and costs have been the coverage, and high contrast conspicuity are the
most critical hurdles preventing an all-encompass- characteristics of whole-body MRA.
ing approach to arterial imaging. Intravascular contrast agents such as Vasovist can
Whole-body MRA appears to be well-suited to visu- further increase image quality and by introducing
alize the systemic nature of atherosclerotic disease new aspects to MR imaging techniques such as the
and allows the depiction of relevant concomitant differentiation of first pass and steady state vascular
disease affecting other arterial territories than the imaging will most probably have a major impact
region that might become symptomatic first, with- on routine clinical imaging and image quality of
out potentially harming the patient by applying whole-body MR angiography.
ionizing radiation or high doses of iodinated CA.

References 17. Vasbinder GBC, De Haan MW, van Engelshoven JMA (2002) Ac-
curacy of CTA and 3D contrast-enhanced MRA as compared to
1. Adams MR, Celermajer DS (1999) Detection of presymptomatic intra-arterial digital subtraction angiography for assessment of
atherosclerosis: a current perspective. Clin Sci (Lond) 97:615624 the number of renal arteries in 356 subjects. Radiology 225 (Pro-
2. Cahan MA, et al (1999) The prevalence of carotid artery stenosis in ceedings): 400
patients undergoing aortic reconstruction. Am J Surg 178:194196 18. Michaely HJ, et al (2007) Nephrogenic systemic fibrosis (NSF)
3. Diehm C, Kareem S, Lawall H (2004) Epidemiology of peripheral implications for radiology [in German]. Radiologe 47:785793
arterial disease. Vasa 33:183189 19. Fenchel M, et al (2005) Whole-body MR angiography using a novel
4. Fowkes FG, et al (1991) Edinburgh Artery Study: prevalence of 32-receiving-channel MR system with surface coil technology:
asymptomatic and symptomatic peripheral arterial disease in the first clinical experience. J Magn Reson Imaging 21:596603
general population. Int J Epidemiol 20:384392 20. Kramer H, et al (2005) Cardiovascular screening with parallel imaging
5. Goyen M, et al (2003) Detection of atherosclerosis: systemic imag- techniques and a whole-body MR imager. Radiology 236:300310
ing for systemic disease with whole-body three-dimensional MR 21. Goyen M, et al (2002) Whole-body three-dimensional MR an-
angiography--initial experience. Radiology 227:277282 giography with a rolling table platform: initial clinical experience.
6. Green D, Parker D (2003) CTA and MRA: visualization without cath- Radiology 224:270277
eterization. Semin Ultrasound CT MR 24:185191 22. Shetty AN, et al (2002) Lower extremity MR angiography: universal
11 7. Leiner T, et al (2005) Peripheral arterial disease: comparison of retrofitting of high-field-strength systems with stepping kinematic
color duplex US and contrast-enhanced MR angiography for diag- imaging platforms initial experience. Radiology 222:284291
nosis. Radiology 235:699708 23. Nael K, et al (2007) High-spatial-resolution whole-body MR angi-
8. Prokop M (2000) Multislice CT angiography. Eur J Radiol 36:8696 ography with high-acceleration parallel acquisition and 32-chan-
9. Steffens JC, et al (2003) Bolus-chasing contrast-enhanced 3D MRA nel 3.0-T unit: initial experience. Radiology 242:865872
of the lower extremity. Comparison with intraarterial DSA. Acta 24. Ruehm SG, et al (2000) Whole-body MRA on a rolling table plat-
Radiol 44:185192 form (AngioSURF) [in German]. Rofo 172:670674
10. Koelemay MJ, et al (1996) Diagnosis of arterial disease of the lower 25. Nael K, et al (2007) Multistation whole-body high-spatial-resolu-
extremities with duplex ultrasonography. Br J Surg 83:404409 tion MR angiography using a 32-channel MR system. AJR Am J
11. Hirsch AT, et al (2006) ACC/AHA Guidelines for the Management Roentgenol 188:529539
of Patients with Peripheral Arterial Disease (lower extremity, renal, 26. Vogt FM, et al (2004) Venous compression at high-spatial-reso-
mesenteric, and abdominal aortic): a collaborative report from the lution three-dimensional MR angiography of peripheral arteries.
American Associations for Vascular Surgery/Society for Vascular Radiology 233:913920
Surgery, Society for Cardiovascular Angiography and Interven- 27. Nikolaou K, et al (2006) High-spatial-resolution multistation MR
tions, Society for Vascular Medicine and Biology, Society of In- angiography with parallel imaging and blood pool contrast agent:
terventional Radiology, and the ACC/AHA Task Force on Practice initial experience. Radiology 241:861872
Guidelines (writing committee to develop guidelines for the man- 28. van Bemmel CM, et al (2003) Blood pool contrast-enhanced MRA:
agement of patients with peripheral arterial disease) summary of improved arterial visualization in the steady state. IEEE Trans Med
recommendations. J Vasc Interv Radiol 17:13831397; quiz 1398 Imaging 22:645652
12. Norgren L, et al (2007) Inter-society consensus for the manage- 29. Klessen C, et al (2007) First pass whole-body magnetic resonance
ment of peripheral arterial disease (TASC II). J Vasc Surg 45 [1 angiography (MRA) using the blood-pool contrast medium gado-
Suppl]:S5S67 fosveset trisodium: comparison to gadopentetate dimeglumine.
13. Gregor M, et al (2002) Peripheral run-off CE-MRA with a 1.0 molar Invest Radiol 42:659664
gadolinium chelate (Gadovist) with intraarterial DSA comparison. 30. McGregor R, et al (2008) A multi-center, comparative, phase 3
Acad Radiol 9 [Suppl 2]:S398400 study to determine the efficacy of gadofosveset-enhanced mag-
14. Colquhoun I, et al (1992) The assessment of carotid and vertebral netic resonance angiography for evaluation of renal artery dis-
arteries: a comparison of CFM duplex ultrasound with intravenous ease. Eur J Radiol 65(2):316-25. Epub 2007 May 17
digital subtraction angiography. Br J Radiol 65:10691074 31. Rapp JH, et al (2005) Aortoiliac occlusive disease in patients with
15. Stanziale SF, et al (2005) Determining in-stent stenosis of carotid known or suspected peripheral vascular disease: safety and ef-
arteries by duplex ultrasound criteria. J Endovasc Ther 12:346353 ficacy of gadofosveset-enhanced MR angiography multicenter
16. Rankin SC (1999) CT angiography. Eur Radiol 9: 297310 comparative phase III study. Radiology 236:7178

Endoleak Imaging
Sandra A.P. Cornelissen, Mathias Prokop, Hence J.M. Verhagen, and Lambertus W. Bartels

12.1 Introduction 140

12.2 Non-shrinking Aneurysms 140

12.2.1 Endoleaks 140
12.2.2 The Endotension Problem 140

12.3 Challenges of Imaging Follow-up after EVAR 141

12.3.1 CT and CTA 141
12.3.2 MR and MRA 142

12.4 A new Approach to Endoleak Imaging: Vasovist-enhanced MRA 143

12.4.1 General principle 143
12.4.2 Examination Technique 143
12.4.3 Vasovist for Detection of Slow-flow Endoleak 143

12.5 Summary 146

References 146
140 Chapter 12 Endoleak Imaging

12.1 Introduction endoleak and leads to markedly increased intra-an-

eurysmal pressure (high-pressure endoleak). Type II
Until 15 years ago, abdominal aortic aneurysm treatment endoleaks originate from reversed flow in branch ves-
involved major abdominal surgery in which a prosthetic sels, such as lumbar arteries or the inferior mesenteric
graft was sewn into the aortic wall. As vascular surgeons artery, and lead to a pressure increase that depends on
were increasingly confronted with older patients with the amount of leakage. Type III endoleaks represent a
severe co-morbidity, attempts for less invasive aneurysm defect in the graft material or a modular disconnection
treatment were made. In 1991, Parodi et al. [1] were the and usually cause a marked increase in intra-aneurysmal
first to report successful endovascular aortic aneurysm pressure. Type IV endoleak denotes leakage via porous
repair (EVAR) in human patients. Their technique in- graft fabric.
volved cannulation of the common femoral artery and Accurate visualization and classification of endoleaks
endovascular placement of a Dacron tubular graft with is needed to evaluate whether treatment of an endoleak is
attached balloon expandable stents to anchor the graft necessary. Especially high-pressure endoleaks, i.e., type I
to the aortic wall. Since then, this technique has become and III endoleaks which originate directly from the arte-
widely available and has emerged as a frequently used rial circulation, seem to promote aneurysm growth and
alternative for open aneurysm treatment. Today, many increase rupture risk. Type I or III endoleaks have been
different types of endografts are available, and more var- associated with a significantly greater risk of rupture than
ied and also more challenging anatomies can be treated type II endoleaks [4]. This is why type I and III endoleaks
endovascularly. are more aggressively treated than type II endoleaks.
Due to the rapid development of endoprostheses, There is still controversy about whether treatment of type
limited experience about long-term durability is available II endoleaks is needed. In general, type II endoleaks are
for most grafts. At present, life-long imaging follow-up is treated in case of aneurysm growth, if possible by means
considered necessary in patients following EVAR. Imag- of endovascular techniques.
ing follow-up is needed to assess whether aneurysm size
regresses, whether blood leaks out of the endograft into
the aneurysm sac (endoleak), whether the endoprosthesis 12.2.2 The Endotension Problem
remains intact and at its original location, and whether
the endoprosthesis remains patent. Endotension is a term used for increased pressure within
Follow-up regimens differ worldwide but usually con- the aneurysm sac that leads to aneurysm growth without
sist of a computed tomography angiography (CTA) ex- evidence of endoleak (on CTA). Endotension has a preva-
amination shortly after the procedure, as well as repeated lence of 13%. Because the aneurysm grows in patients
12 CTA examinations each year for the rest of the patients with endotension, rupture risk is increased and treatment
life. Aneurysm size is an important parameter in this is indicated. However, because there is no visualized en-
follow-up. In case of aneurysm growth, the danger of doleak to direct treatment to, the main treatment option
aneurysm rupture is still present, and repeat intervention is currently conversion to surgery. For most patients, this
is indicated. In case of aneurysm shrinkage, aneurysm is not an attractive alternative because of their cardiovas-
treatment is considered successful. cular co-morbidity.
The etiology of endotension is unknown. It could well
be that the elevated intrasac pressure originates from an
12.2 Non-shrinking Aneurysms endoleak that was not diagnosed by CTA. This can be due
to the slow flow rate of the endoleak or the intermittent
Implantation of an endoprosthesis has the goal of ex- leakage of an endoleak. Most likely, slow flow leaks are
cluding the aneurysm sac from the circulation. This will due to type II or type IV leakage. Type I or III leakage
reduce pressure on the aneurysm wall and, through re- involves a direct communication with the arterial circula-
sorption of the ensuing thrombus, will lead to aneurysm tion and therefore has higher flow rates and should always
shrinkage. In case of a non-shrinking or growing aneu- be detectable by CTA.
rysm, the etiology is not always clear, but endoleaks are Alternatively, the elevated intrasac pressure can origi-
the most frequent cause. nate from inside the aneurysm sac. Very little is known
about the processes taking place inside the aneurysm sac
following EVAR. Possibly, the intra-aneurysmal throm-
12.2.1 Endoleaks bus is not a static entity; rather, repeated liquefaction
by fibrinolysis and clotting occurs. During fibrinolysis
Endoleaks are classified according to the source of leak- fluid accumulates inside the aneurysm sac, increasing the
age [2, 3]. Leakage via an attachment site is a type I intrasac pressure. Other inflammatory reactions or graft
12.3 Challenges of Imaging Follow-up after EVAR
141 12

infection can also cause fluid accumulation inside the 12.3.1 CT and CTA
aneurysm sac, resulting in intrasac pressure increase and
aneurysm growth. Pre-contrast combined with post-contrast CT imaging
Nevertheless, a substantial number of endotension at different time points after injection of contrast fluid is
cases could represent missed endoleak. For further in- currently the reference standard for endoleak detection.
vestigation, more sensitive techniques for visualizing en- The pre-contrast CT acquisition is used for differen-
doleaks are needed, such as MRI. Here we postulate tiating calcium from endoleak [5]. For measurement of
that the novel blood pool agent Vasovist (Gadofosveset, aneurysm diameter and for evaluating graft migration
Bayer Schering Pharma AG, Berlin, Germany) can play or kinking of the graft, the pre-contrast acquisition gives
an important role in endoleak visualization because it has enough information.
a high relaxivity and remains in the intravascular space Arterial-phase CTA is used for the detection of high-
for an extended period of time. This is interesting not flow endoleaks and gives the best anatomic information
only from a scientific point of view; accurate knowledge regarding in-stent thrombus, patency of side branches,
about endoleak in patients with endotension potentially and dissections, which sometimes occur as a complication
provides more patient-specific often endovascular of EVAR. Iezzi et al. demonstrated that the combination
treatment alternatives to prevent future aneurysm growth. of an initial pre-contrast and arterial acquisition 1 month
Adequate treatment of endotension potentially results in after EVAR and only an arterial acquisition at subsequent
shrinkage of the aneurysm sac. follow-up moments is sufficient for endoleak detection
This chapter describes current strategies and ques- [6]. In their opinion delayed-phase imaging to detect
tions in the follow-up after EVAR with special emphasis slow-flow endoleak is required only in case of an increase
on the visualization of endoleak. Advantages and disad- in aneurysm size.
vantages of visualizing endoleaks with CT and magnetic Venous-phase (60 s after injection of contrast me-
resonance imaging (MRI) are described. Furthermore, dia) or delayed-phase CT (100 s after injection) acquisi-
the application of Vasovist-enhanced MRI for endoleak tions are used to visualize slow flow endoleaks [5]. Intra-
visualization and MR protocols which can be used for vascular enhancement is far less compared with arterial
endoleak imaging are described. CTA because the scan is acquired after the first pass
of the injected bolus of contrast agent, so the contrast
agent has been diluted in the whole intravascular blood
12.3 Challenges of Imaging Follow-up volume and has leaked into the interstitium during its
after EVAR passage through capillaries. This is why the later this ac-
quisition is acquired after contrast media injection, the
A substantial number of endotension cases and non- lower the enhancement of the blood pool. Moreover, the
shrinking aneurysms probably represent endoleak not amount of contrast medium needed for such an acquisi-
visualized by conventional imaging techniques. The chal- tion should be adjusted to patient weight (as approxima-
lenge of radiologic follow-up after EVAR is to improve tion of the blood volume). Macari et al., however, argue
understanding of aneurysm sac behavior in such cases. As that pre-contrast and venous acquisitions are sufficient
said before, lack of aneurysm shrinkage can be caused by for endoleak detection; arterial phase imaging is, in their
slow flow endoleak or the intermittent leakage of an en- point of view, not necessary for the routine detection of
doleak. If an endoleak has a slow flow rate, a longer time endoleaks [7].
between injection and imaging is needed to allow for leak- The main problem with a delayed-phase scan is that
age of enhanced blood into the aneurysm sac (late-phase it is always a compromise: the later the acquisition with
imaging). Increasing this delay gives rise to some prob- regard to contrast media injection, the more contrast
lems. First of all, during late-phase imaging the images media is necessary or the lower the enhancement. If the
are not acquired during the first pass of a bolus of contrast blood is still sufficiently enhanced, the endoleak is bet-
material but at a later time, when contrast agent has mixed ter detected. If no endoleak is visualized on a late-phase
with the whole blood volume of a patient. Consequently, acquisition, it is still not clear whether there is a true ab-
the contrast agent concentration in the blood volume is sence of endoleak. Delayed acquisitions are needed for the
lower in late-phase imaging with respect to peak arterial- visualization of low-flow endoleaks. These acquisitions
phase imaging, which results in lower intravascular en- are less useful for imaging vascular anatomy. For diagnos-
hancement. Additionally, almost all of the contrast agents ing dissection or evaluating the patency of side branches
used nowadays distribute to the extracellular space, which arterial acquisitions are needed.
mean they leak into the interstitium during passage of tis- In summary, different protocols are used for endoleak
sue capillaries, which further decreases the intravascular detection. Obviously, in endoleak imaging with CT there
contrast agent concentration. is always a trade-off between the delay between injection
142 Chapter 12 Endoleak Imaging

and imaging and vascular enhancement. The delay be- MR compatible materials are MR-safe and result in diag-
tween injection and imaging should be as long as possible nostic images in 1.5 T systems. In addition, MR imaging
to give contrast agent enough time to leak to achieve the in the abdominal region is prone to artifacts. Care has to
accumulation of a detectable amount of contrast agent be taken to minimize ghosting artifacts originating from
in the aneurysm sac. However, it should not be too long, breathing. These artifacts can be prevented by placing a
because longer after injection the contrast agent already regional saturation slab ventrally on the subcutaneous fat
has disappeared from the blood, leaving only blood with- of the abdomen.
out contrast agent leaking into the aneurysm sac, which Different scan types can be used for endoleak visuali-
further deteriorates the visibility of the endoleak on CT zation. T1-weighted spin echo imaging and T1-weighted
images. spoiled gradient echo (MR angiographic) protocols can
be used. Spin echo imaging is preferred because the influ-
ence of the susceptibility artifacts caused by local field
12.3.2 MR and MRA inhomogeneities of the stent struts is less compared to
gradient echo imaging.
Recent studies have shown that magnetic resonance imag- A typical imaging protocol for endoleak imaging [11,
ing using Gd-DTPA is more sensitive for endoleak than 12] with Gd-DTPA contains a:
CTA [811], so it is logical to use MRI to further inves- 1. Pre-contrast transverse T1-weighted spin echo acqui-
tigate the role of endoleak in patients with endotension. sition
Additionally, with dynamic MRA more endoleaks can 2. Coronal dynamic 3D contrast enhanced (CE)-MRA
be classified into one of the categories described before during injection of Gd-DTPA (Magnevist, Bayer
compared to CTA [12]. Improved endoleak classifica- Schering Pharma AG, Berlin, Germany)
tion primarily results from dynamic MRA acquisitions in 3. First pass Coronal 3D CE-MRA
which the same volume is scanned multiple times result- 4. Post-contrast transverse T1-weighted spin echo
ing in visualization of the contrast bolus flowing through
the aortic trajectory in near real time. Lookstein et al. Voxels inside the aneurysm sac, outside the lumen of
demonstrated that endoleak classification by dynamic the endograft with a high signal intensity on the post-
MRA corresponded to the classification obtained by dig- contrast image and a low signal intensity on the pre-
ital subtraction angiography [13]. Accurate endoleak clas- contrast image represent endoleak. The appearance of
sification is important for endoleak treatment. In case of the aneurysm sac must be known before contrast admin-
a type II endoleak less aggressive treatment is indicated istration. For this reason T1-weighted spin echo acquisi-
than for a type I or III endoleak. tions are needed before and after contrast injection. In
12 Metal-related artifacts caused by the endoprosthe- most patients the intra-aneurysmal thrombus appears
sis can degrade the diagnostic value of MRI in patients dark before contrast injection on T1-weighted images,
after EVAR. Susceptibility artifacts, resulting from local but in some cases the thrombus inside the aneurysm
field inhomogeneity caused by metal in the stent struts, sac already has a high signal intensity before contrast
give rise to geometrical distortions and local signal loss injection. Presumably this is caused by the presence of
around the implants metallic parts. The severity of these methemoglobin in the aneurysm sac, which has a short
artifacts depends strongly on the susceptibility of materi- T1, which should not be interpreted as endoleak. In
als used in the endograft. Endografts with stent struts of Fig. 12.1 an example of an arterial CTA and delayed


Fig. 12.1AD. A Pre-contrast and B post-contrast T1-weighted spin- the endograft on B, which are not this hyperintense on A, and hyper-
echo images, C arterial CTA, and D delayed CT image of a patient with dense area on C represent endoleak
an endoleak. Hyperintense areas inside the aneurysm sac and outside
12.4 A new Approach to Endoleak Imaging: Vasovist-enhanced MRA
143 12

CT-image as well as the corresponding pre-contrast and protocol. Because Vasovist has a higher relaxivity than
post-contrast T1-weighted spin echo MR-images of a Gadolinium-DTPA, the injected dose is lower and it
patient with endoleak are shown.The post-contrast MR- should be injected at a lower speed. The approved dose
images were acquired approximately 3 min after injection of Vasovist, 0.12 ml/kg (0.03 mmol/kg) can be used and
of Gd-DTPA. The endoleak is hyperdense on CT-images injected at 1 ml/s followed by a saline chaser of 30 ml with
with respect to the rest of the aneurysm sac. On the MR the same injection speed.
images a hyperintense area is present inside the endoleak With this injection protocol the dynamic MRA proto-
on the post-contrast T1-weighted image which was less col as used with gadolinium-DTPA can be used unchanged
intense before contrast injection. with Vasovist. Subjectively, intravascular enhancement is
higher with Vasovist compared to gadolinium-DTPA.
Dynamic acquisitions acquired during injection of Vaso-
12.4 A new Approach to Endoleak Imaging: vist typically show rapidly decreasing intravascular en-
Vasovist-enhanced MRA hancement when the agent mixes with the whole blood
volume ( Fig. 12.2).
12.4.1 General principle After the dynamic series, late phase MRA images can
be acquired. Because Vasovist remains in the intravas-
Different classes of MR blood pool agents have been cular space for an extended period of time, MRA-images
developed; coated iron particles and Gadolinium-based can be acquired much longer after injection than with
protein binding agents (of which Vasovist is an ex- Gd-DTPA. For acquiring late phase MRA acquisitions
ample). Such agents remain in the intravascular space the flip angle of the steady state spoiled gradient echo
for a longer period compared to conventional contrast imaging should be adjusted, e.g. lowered, to the expected
agents. Coated iron particles remain in the intravas- T1 of blood. An example of two late phase MRA acquisi-
cular space because of their size and coating; Protein tions acquired at 6 and 15 min after injection is shown in
binding contrast agents remain intravascular by their Fig. 12.3.
reversible binding to human serum albumin [14]. In Additional postcontrast T1-weighted spin echo ac-
general, iron particles act as negative contrast agents, quisitions can be added up to 1 hour after injection to
resulting in signal voids in the MR image caused by fully utilize the longer intravascular residence time of
their local susceptibility effects which lead to dephasing Vasovist. Examples of T1-weighted spin echo images ac-
of neighbouring spins. This phenomenon makes them quired before and at different delays after administration
mostly suitable for T2*-weighted sequences [15]. How- of Vasovist are shown in Figs. 12.4 and 12.5.
ever, when short echo times are used, iron particles can The T1 of blood is lowered during injection of Vaso-
also be used as a positive contrast agent in T1-weighted vist. After injection, T1 rises again when the agent is
protocols [16]. diluted in the whole blood volume. The exact T1 of blood
Vasovist is a gadolinium-based protein binding con- with contrast agent at different times after injection is not
trast agent. For the visualization of endoleak, T1-weighted known. The development of blood T1 over time probably
images before and after injection of Vasovist can be also depends on patient-related factors, like the concen-
used. tration of albumin in the blood and renal function. Ide-
Vasovist bound to albumin remains in the intra- ally the TR and the flip angle of MRA sequences should
vascular space for an extended period of time. Conse- be adjusted to the T1 of blood expected at the specific
quently, the delay between injection and imaging can post-injection delay. Longer after injection a longer T1
be increased, allowing for the buildup of a higher con- is expected and the TR should be increased and the flip
centration of contrast agent in slow-flow endoleaks. angle should be adjusted to achieve maximal SNR. In the
Intermittent endoleak also has a higher chance of being abdomen this is a problem because the scan should pref-
visualized in this way. The longer the delay between erably fit in one breath hold.
injection and imaging the higher the chance that during
this time contrast agent leaks and accumulates in the
aneurysm sac. 12.4.3 Vasovist for Detection of Slow-flow

12.4.2 Examination Technique To investigate the possibility of visualizing slow flow en-
doleak, three patients were imaged with Vasovist who did
Slight changes need to be made to the imaging protocol not have evidence of endoleak on CTA and delayed CT.
in order to adapt it to the use of Vasovist instead of All patients had been treated with the original Excluder
Gadolinium-DTPA. The first adaptation is the injection endoprosthesis (W.L. Gore, Inc, Flagstaff, Arizona, USA).
144 Chapter 12 Endoleak Imaging



Fig. 12.2AI. Maximum inten-

sity projections of dynamic MRA
acquisitions acquired (A) before
and (B to I) following administra-
tion of Vasovist 0.12 ml/kg at
1 ml/s

This endograft was changed in July 2004 by the manu-
facturer because it was associated with lower aneurysm
shrinkage rates than other grafts. The cause was unclear
but the manufacturer decided to incorporate an addi-
tional layer of ePTFE (expanded polytetrafluoroethylene)
in the graft fabric because the low shrinkage rates were
assumed to result from graft porosity. However, in human
patients this had never been visualized.
Most probably, this graft porosity was not visualized A B
using both CT or MR techniques due to the low flow rate,
as explained above. In such cases it is useful to use Vaso- Fig. 12.3A,B. Maximum intensity projections of late-phase MRA ac-
vist. The scan protocol described above was used and quisitions. Both acquisitions are steady state spoiled gradient echo
acquisitions with a TR of 10 ms, TR of 2 ms and flip angle of 27 degrees.
late phase postcontrast T1-weighted spin echo acquisi-
A was acquired 6 min and B 15 min following injection of Vasovist in
tions were added more than 30 min after injection. Using the standard dosage of 0.12 ml/kg at 1 ml/s
this protocol, endoleak was indeed visualized on the late
phase images acquired more than 30 min after injection
( Fig. 12.6). The early postcontrast images acquired three this leakage most probably arises from graft porosity.
min after injection did not show endoleak. On the late These images illustrate it is indeed possible to visualize
phase T1-w spin echo images, leakage of contrast agent slow-flow endoleaks by increasing the time window be-
into the aneurysm sac in the direct vicinity of a leg of the tween injection and imaging when using the blood pool
endoprosthesis was visualized. Because of this location, agent Gadofosveset [17].
12.4 A new Approach to Endoleak Imaging: Vasovist-enhanced MRA
145 12

Fig. 12.4AC. A Pre-contrast,

B 3 min post-contrast, and
C 30-min post-contrast T1-
weighted spin-echo MR images
following administration of
Vasovist. This patient had an
Ancure endoprosthesis. In C
leakage of contrast agent is vis-
ible in the outer border of the
aneurysm sac

Fig. 12.5AC. A Pre-contrast,

B 3 min post-contrast, and
C 30-min post-contrast T1-
weighted spin-echo MR images
following administration of Vas-
ovist. This patient had a monoil-
iac endoprosthesis. No leakage
of contrast agent is visible in the
non-luminal aneurysm sac

Patient 1 Patient 2 Patient 3



early post-contrast


late-phase post-contrast


Fig. 12.6AI. Transverse T1-weighted spin-echo MR images. A,D,G the same patient and the same anatomical location. In the bottom row,
Pre-contrast, B,E,H early-phase post-contrast, C,F,I late-phase post- endoleak (arrowheads) arising from graft porosity is visible. (Reused
contrast T1-weighted images. Each column consists of images from with permission from [17])
146 Chapter 12 Endoleak Imaging

12.5 Summary endotension. It could well be that in a substantial number

of such cases endoleaks do play a role but are not visual-
In the follow-up of EVAR, visualization of endoleak plays ized by the currently used imaging strategies due to the
an important role, because endoleak hampers aneurysm slow flow rate or intermittent nature of the endoleak. By
shrinkage and promotes aneurysm growth, thus increas- using Vasovist it is possible to increase the time window
ing rupture risk. It is currently not understood why not all between contrast fluid injection and imaging to attain a
aneurysms without evidence of endoleak on CTA shrink. higher accumulation of contrast agent in the endoleak.
Furthermore, some aneurysms grow without evidence of This improves visualization of slow flow endoleaks and
endoleak on CTA and delayed CT which is referred to as increases the chance for visualizing intermittent endoleak.

Take home messages

Patients who have undergone EVAR need life-long With current CT and MRI imaging protocols endo-
imaging to rule out endoleak and further growth leaks are not visualized in a substantial number
of their aneurysm. of cases even when they are thought to be present.
After EVAR, some aneurysms grow without evi- Delayed-phase Vasovist-enhanced MRI is a prom-
dence of endoleak on CTA and delayed CT. This is ising technique to detect suspected endoleak in
referred to as endotension. cases where CT and convential MRI are negative.

References 10. Pitton MB, Schweitzer H, Herber S, Schmiedt W, Neufang A, Ka-

lden P, et al (2005) MRI versus helical CT for endoleak detec-
1. Parodi JC, Palmaz JC, Barone HD (1991) Transfemoral intraluminal tion after endovascular aneurysm repair. AJR Am J Roentgenol
graft implantation for abdominal aortic aneurysms. Ann Vasc Surg 185:12751281
5:491499 11. van der Laan MJ, Bartels LW, Viergever MA, Blankensteijn JD (2006)
2. Veith FJ, Baum RA, Ohki T, Amor M, Adiseshiah M, Blankensteijn Computed tomography versus magnetic resonance imaging of
JD, et al (2002) Nature and significance of endoleaks and endoten- endoleaks after EVAR. Eur J Vasc Endovasc Surg 32:361365
sion: summary of opinions expressed at an international confer- 12. van der Laan MJ, Bakker CJ, Blankensteijn JD, Bartels LW (2006)
ence. J Vasc Surg 35:10291035 Dynamic CE-MRA for endoleak classification after endovascular
12 3. White GH, May J, Waugh RC, Chaufour X, Yu W (1998) Type III and
type IV endoleak: toward a complete definition of blood flow in
aneurysm repair. Eur J Vasc Endovasc Surg 31:130135
13. Lookstein RA, Goldman J, Pukin L, Marin ML (2004) Time-resolved
the sac after endoluminal AAA repair. J Endovasc Surg 5:305309 magnetic resonance angiography as a noninvasive method to
4. van Marrewijk C, Buth J, Harris PL, Norgren L, Nevelsteen A, Wyatt characterize endoleaks: initial results compared with conventional
MG (2002) Significance of endoleaks after endovascular repair of angiography. J Vasc Surg 39:2733
abdominal aortic aneurysms: The EUROSTAR experience. J Vasc 14. Hartmann M, Wiethoff AJ, Hentrich HR, Rohrer M (2006) Initial
Surg 35:461473 imaging recommendations for Vasovist angiography. Eur Radiol
5. Rozenblit AM, Patlas M, Rosenbaum AT, Okhi T, Veith FJ, Laks MP, 16 [Suppl 2]:B15B23
et al (2003) Detection of endoleaks after endovascular repair of 15. Rohrer M, Bauer H, Mintorovitch J, Requardt M, Weinmann HJ
abdominal aortic aneurysm: value of unenhanced and delayed (2005) Comparison of magnetic properties of MRI contrast me-
helical CT acquisitions. Radiology 227:426433 dia solutions at different magnetic field strengths. Invest Radiol
6. Iezzi R, Cotroneo AR, Filippone A, Di Fabio F, Quinto F, Colosimo 40:715724
C, et al (2006) Multidetector CT in abdominal aortic aneurysm 16. Ersoy H, Jacobs P, Kent CK, Prince MR (2004) Blood pool MR an-
treated with endovascular repair: are unenhanced and delayed giography of aortic stent-graft endoleak. AJR Am J Roentgenol
phase enhanced images effective for endoleak detection? Radiol- 182:11811186
ogy 241:915921 17. Cornelissen SA, Verhagen HJ, Prokop M, Moll FL, Bartels LW (2008)
7. Macari M, Chandarana H, Schmidt B, Lee J, Lamparello P, Babb Visualizing type IV endoleak using magnetic resonance imaging
J (2006) Abdominal aortic aneurysm: can the arterial phase at with a blood pool contrast agent. J Vasc Surg 47:8614
CT evaluation after endovascular repair be eliminated to reduce
radiation dose? Radiology 241:908914
8. Cejna M, Loewe C, Schoder M, Dirisamer A, Hlzenbein T, Kret-
schmer G, et al (2002) MR angiography vs CT angiography in the
follow-up of nitinol stent grafts in endoluminally treated aortic
aneurysms. Eur Radiol 12:24432450
9. Haulon S, Lions C, McFadden EP, Koussa M, Gaxotte V, Halna P, et
al (2001) Prospective evaluation of magnetic resonance imaging
after endovascular treatment of infrarenal aortic aneurysms. Eur J
Vasc Endovasc Surg 22:6269

Gastrointestinal Bleeding
Joachim Lotz

13.1 Introduction 148

13.2 Diagnostic Challenge of GI Bleeding 148

13.3 Upper GI Tract 149

13.4 Lower GI Tract 149

13.5 Small Intestine 150

13.6 GI Bleeding and MRI 153

13.7 Blood pool Contrast Agents 153

13.8 Summary 154

References 155
148 Chapter 13 Gastrointestinal Bleeding

13.1 Introduction the United States in the period between 1996 and 2000
[2]. UGIB is about five times more frequent than LGIB
Gastrointestinal (GI) bleeding describes a large variety of [3, 4], with an incidence of approximately 50 per 100 000
clinical conditions and underlying pathology in any loca- population a year in a study from the Netherlands [5, 6].
tion of the GI tract. It is common practice to subdivide Bleeding from the small intestine makes up about 35%
GI bleeding disorders into upper und lower GI bleeding. of all GI bleeding sources [1].
Upper GI bleeding (UGIB) denotes bleeding sources in All variants of GI bleeding except those caused by
the esophagus and stomach, sometimes including the portal hypertension show a significant increase with age
duodenum up to the ligament of Treitz. Conversely, lower and tend to affect men more often than women [7, 8].
GI bleeding (LGIB) encompasses bleeding pathologies The underlying pathology of GI bleeding varies de-
that arise from the small intestine and the colon, up to and pending on its origin in the GI tract ( Table 13.1). Mu-
including the anal canal. cosal abnormalities and peptic ulcer disease are the lead-
Though the small intestine (duodenum-jejunum and ing cause of UGIB [9], followed by bleeding caused by
ileum) is regarded as part of the lower GI tract, underly- portal hypertension (variceal bleeding). Angiodysplasia,
ing pathology and its diagnostic workup are markedly lymphoma, and ulceration are the leading sources of
different from those of UGIB and bleeding of the colon. bleeding in the small intestine. Diverticulosis, colitis, and
Therefore, it has been proposed to be a separate entity [1] cancer constitute the majority of bleeding originating in
and will be dealt with accordingly in this text. the colon, sigmoid, and rectum [10].
GI bleeding is a major cause of morbidity in industrial
countries, resulting in over 300 000 admissions a year in
13.2 Diagnostic Challenge of GI Bleeding

Table 13.1. Most common causes of GI bleeding The diagnostic workup of GI bleeding depends on its
suspected site of origin, which can be inferred to some ex-
Location Cause Prevalence
tent from the clinical presentation, including the quantity
and appearance of the bleeding. Hematemesis (vomiting
Upper GI-tract Mucosal abnormalities 40 blood) points to a bleeding source in the upper GI tract.
[9, 19, 43]
Peptic ulcer disease 2574 Melena (black, tarry blood) indicates that blood has been
in the gastrointestinal tract for an extended period of
Varices 15
time. It is usually the result of UGIB, although its source
Esophageal inflammation 15 may also be the distal small bowel or even the ascending
colon. In the latter instance, the volume of bleeding is too
Mallory-Weiss tears 515
little to cause hematochezia (bright blood in stool) but
Small intestine Angiodysplasia
13 [1]
sufficiently large to provide hemoglobin for degradation.
Lymphoma Hematochezia originates in most cases from hemorrhoids
or pathologies of the distal colon. Approximately 10% of
Erosions / ulcers
all patients with rapid UGIB present with hematochezia
Lower GI tract Diverticulosis 555 [11]. In these patients, bleeding is typically massive. If the
[4446] bleeding is obvious but its source remains unknown even
Neoplasm 326
after initial diagnostics, it is termed obscure bleeding.
Colitis 329 Occult bleeding refers to bleeding that the patient is una-
Ischemia 618
ware of and may only be inferred from hemoccult tests or
changes in serological parameters.
Angiodysplasia 340 (incl. Most variants of GI bleeding are self-limited but tend
small bowel)
to recur. This is a particular challenge of the diagnostic
Anorectal 10 workup, as the source of bleeding is usually silent when
Obscure GI Tumor (small intestine),
diagnostic procedures are initiated. Only a minority of
bleeding [47], Meckels diverticulum, cases with massive active bleeding require emergency in-
age <40 years Dieulafoy lesion, heredi- terventions including hemodynamic stabilization, resus-
tary polyposis syndrome, citation, and surgical exploration. Nevertheless, mortality
M. Crohn
from GI bleeding increases with age and is estimated to be
Obscure GI Vascular lesion, ulcer as high as 1014% [7, 8, 12].
bleeding [47], (NSAID induced), Cameron Specific diagnostic approaches have been established
age >40 erosion
for the different parts of the GI tract. In general, endoscopy
13.4 Lower GI Tract
149 13

GI Bleeding
Upper GI Bleeding Obscure GI Bleeding Lower GI Bleeding

Hemodynamically Stable Hemodynamically Unstable Upper Endoscopy Hemodynamically Stable Hemodynamically Unstable

Routine Urgent Lower Endoscopy Flexible Sigmoidoscopy Upper Endoscopy

Ednoscopy Ednoscopy Colonoscopy

MSCT Hemodynamically unstable Hemodynamically Stable Colonoscopy

Angiography MSCT Capsule Endoscopy No Bleeding detected Mild to Severe Bleeding

MRT + Blood Pool CM

if negative: Angiography Double Balloon Endoscopy MSCT

Double balloon endoscopy or Scintigraphy
+/- capsule endoscopy

Double Balloon Endoscopy MSCT Angiography

MRT + Blood Pool CM


Intraoperative Endoscopy or
Exploratory Surgery

Fig. 13.1. Flow chart of diagnostic procedures for GI bleeding. (Based on Lee et al [19])

is the first choice in the workup of a GI bleeding. Cross- Other imaging modalities are employed for the stag-
sectional imaging has its role as a second-line diagnostic ing of local pathologies or the evaluation of lesions that
modality and adds valuable information about all sections extend beyond the wall of the upper GI tract. MSCT
of the GI tract. Especially in the emergency setting of and MRI are both accepted imaging modalities for this
massive active bleeding, multislice CT (MSCT) might be purpose, supplemented by ultrasound depending on the
the first-line diagnostic imaging approach to guide fur- hospitals policy. Angiography is used primarily for thera-
ther diagnostic and therapeutic steps. ( Fig. 13.1) peutic approaches in UGIB such as selective embolization
or in a venous approach transjugular intrahepatic
portosystemic shunts (TIPS). Both angiography and scin-
13.3 Upper GI Tract tigraphy can be used when endoscopy has not yielded a
definitive diagnosis or cannot be performed.
Endoscopy is the first choice when it comes to the evalu-
ation of UGIB. The main advantage of endoscopy is the
possibility to perform therapeutic interventions in the 13.4 Lower GI Tract
same session as the diagnostic procedure. Mucosal ab-
normalities as well as deep ulcerative lesions are readily As with UGIB, colonic and rectal pathologies leading to
detectable. Endoscopic sclerosis of varices is the standard LGIB are routinely evaluated with endoscopic techniques.
first-line therapy for bleeding from portal hypertension. In Colonoscopy has been reported to have a high diagnostic
the hand of the experienced clinician, a diagnostic upper yield in LGIB of the colon and rectum ranging from 45%
GI endoscopy takes about 20 min to perform, with extra to 95%, with an average in the high 80s [10]. About 10% of
time needed for interventions. Endoscopy-related morbid- suspected cases of LGIB have at least one bleeding source
ity occurs in fewer than 1% of cases [13]. More than 50% of in the upper GI tract. Therefore, gastroenteroscopy is
bleeding sources are detected with initial push endoscopy usually done following a negative colonoscopy for LGIB.
[14]. Recurrence of bleeding occurs in 1030%, especially Complications related to colonoscopy occur in about
in peptic ulcer disease but repeated gastroenteroscopy can 1.3% of cases [17]. Complete evaluation of the colon de-
reduce this rate significantly [15]. Up to 30% of patients pends strongly on an adequate bowel preparation. About
with active bleeding lesions require surgical intervention 20% of patients have a bleeding site that can be treated
for definitive therapy of the pathology [16]. with colonoscopy in the same session [18].
150 Chapter 13 Gastrointestinal Bleeding

Nuclear scans and angiography have their role in the

detection of occult and obscure bleeding as a second-line
modality. Usually this occurs in the setting of negative
endoscopic evaluation of the colon and upper GI tract.
With massive bleeding angiography may be of advan-
tage to rapidly diagnose and treat the origin of bleeding.
This is true especially in altered anatomy due to extensive
surgical procedures, where colonoscopy is difficult to per-
form. MSCT and MRI can be used for local and systemic * *
staging and as a preparation for angiographic or surgical
interventions [19].

13.5 Small Intestine

The diagnostic workup of bleeding originating from the

small intestine is limited by the length and tortuosity of Fig. 13.2. Image taken from a series of approximately 50 000 images
acquired during a capsule endoscopy for obscure LGIB. Small mucosal
this part of the GI tract. Prior to the advent of capsule en-
changes are seen (*) that were judged responsible for bleeding. No
doscopy, pathologies of the small intestine were the most further therapy was initiated
challenging of all parts of the GI tract, usually requiring
more than three different imaging modalities for suffi-
cient diagnosis [1]. detect the sources of occult or obscure bleeding have been
Conventional endoscopic equipment is limited to reported to be about 89% and 95%, respectively [22], with
the exploration of the upper GI tract, including in some a diagnostic yield of 92% in another study [23]. There
instances the duodenum up to the ligament of Treitz and have been initial reports of 20% failure to reach the colon
up to the ileojejunal junction in a retrograde approach. during reporting time [22]. Diagnostic accuracy is limited
Some technical variants of endoscopy have been intro- by food debris, fluid collections, and a limited field-of-
duced recently, the most promising being double-bal- view of 140 of the built-in optics. The most important
loon endoscopy [20]. With this technique it is possible to complication of this technique is capsule entrapment in
evaluate the whole small intestine, usually in a combined an unsuspected stenosis of the small bowl, in which case
antegrade (through the upper GI-tract) and retrograde the capsule has to be retrieved by endoscopy or surgery.
(through the colon) approach. Even if performed by an The incidence of this complication for the whole GI tract
experienced user, double-balloon endoscopy takes about is about 0.755% [24]. If stenotic lesions are anticipated,
13 1.52 h to complete and requires patient immobiliza- a radiopaque dummy capsule can be used in advance that
tion and thorough patient preparation. Double-balloon dissolves naturally within 2 days in the GI tract.
endoscopes have an additional working channel for in- Capsule endoscopy is the primary diagnostic modality
terventions. In a multimember approach, the diagnostic for obscure intestinal bleeding in the hemodynamically
yield of double-balloon enteroscopy in LGIB was 52%. stable patient following initial upper and lower endos-
Examination-related morbidity tends to be around 1%. copy. It is usually followed by an endoscopic examination
Failure to intubate the entire length of the small intes- if a target lesion has been identified and judged amend-
tine occurs in about 31% of cases if a rectal approach is able to endoscopic therapy. Capsule endoscopy followed
used [21]. by double-balloon endoscopy seems to be the most prom-
Capsule endoscopy is a recent development that al- ising diagnostic workup of obscure GI bleeding localized
lows an endoluminal diagnosis of pathologies of the small to the small intestine [25].
intestine. Capsule endoscopy is an ambulatory procedure. Scintigraphy was the mainstay of diagnosis for LGIB
Patients are asked to fast for 6 h prior to the examination. after negative upper and lower endoscopy until the intro-
The capsule is swallowed by the patient and travels down duction of capsule endoscopy. Active or intermittent ac-
the GI tract via natural peristalsis. During its passage tive bleeding is necessary for the scan to yield diagnostic
through the GI tract a built-in camera takes a picture information. A blood sample is taken from the patient,
every 2 s for approximately 8 h ( Fig. 13.2). The images the erythrocytes are labelled with 99m technetium, and
are transmitted wireless to a recorder attached to the pa- 20 ml is reinjected into a peripheral vein. An abdominal
tients waist. Battery life usually suffices for the coverage 2D plane scan is done. Usually images are taken every 3 s
of the small intestine. Most patients excrete the capsule for the first minute, every 5 min for the next 45 min, and
naturally within 2448 h. The sensitivity and specificity to then every 1560 min depending on the clinical setting.
13.5 Small Intestine
151 13

Fig. 13.3. Patient with obscure intermittent GI bleeding localized bleeding. Eighteen hours following injection, an activity pooling is
to the small intestine by upper and lower endoscopy. History of en- detectable in the upper right quadrant of the abdomen indicating
doscopic cholecystectomy 2 weeks prior to the examination. 99m Tc active bleeding (red arrow). Patient was transferred to angiography for
Scintigraphy with labeled erythrocytes. Early scans are negative for selective coil embolization

70% depending on the quantity and activity of bleeding

Table 13.2. Minimal bleeding rate detectable by different
diagnostic modalities. No data are published for MRI, but it is [26]. The yield seems to be higher when an immedi-
estimated to be within the same range as CT with intravenous ate pooling of activity is seen than when late pooling is
contrast. (rbc 99mTc: 99m Technetium-labeled erythrocytes) present [27]. However, when definitive lesions are verified
by other diagnostic modalities, the accuracy of a positive
Modality Minimal
bleeding rate
scan for lesion localization may be as low as 41% [28, 29].
Radionuclide scintigraphy still has its indication for ob-
Angiography intra-arterial unselective 6 ml/min
scure intermittent bleeding if endoscopy and other imag-
Angiography i.a. selective 0.5 ml/min ing modalities fail to localize a bleeding source.
Scintigraphy rbc 99m Tc <0.1 ml/min Digital subtraction angiography (DSA) has frequently
been used for obscure LGIB of the small intestine to de-
CT intra-arterial contrast 0.2 ml/min
tect and if found to embolize pathological vascular
CT intravenous contrast 6 ml/min
structures in the mesenteric territory. Angiography suc-
MRI n/a cessfully detects pathologies with a rate of bleeding above
0.5 ml/min. Selective probing of the celiac trunk and the
superior and inferior mesenteric artery is necessary. If
Activity accumulates in regions of intestinal bleeding, the bleeding cannot be visualized, super-selective angiog-
giving indirect clues about the site but not the cause raphy of the different vascular territories must be done,
of bleeding ( Fig. 13.3). Additional tomographic scans focusing the efforts on the clinically most suspected areas.
(SPECT) can further improve the localization of a bleed- If the bleeding is not detectable on a super-selective angi-
ing source. Scintigraphic scans are the most sensitive ogram, either the relevant vessel has not been catheterized
imaging technique for the detection of active bleeding or the rate of bleeding is too low to be detected. Sensitivity
( Table 13.2), with a diagnostic yield between 15% and and specificity for super-selective angiography have been
152 Chapter 13 Gastrointestinal Bleeding


Fig. 13.4AC. Digital angiography of the same patient as in Fig. 13.3. (arrow). B Super-selective catheterization and coil embolization of the
A Selective angiography of the superior mesenteric artery (SMA) aneurysm. C Control series with the catheter placed in the SMA with
depicts false aneurysm of a side branch of the middle colic artery successful embolization of the aneurysm

described as 89% and 100%, respectively, in the setting of detection of various pathologies in GI bleeding [31, 32,
13 occult bleeding [30]. Angiography remains important as 34, 35]. In 2003, Ernst et al. [36] published a series of 24
a second-line modality when endoscopy, cross-sectional patients with acute bleeding. In this series triphasic helical
imaging modalities, and scintigraphy have failed to local- CT was validated against colonoscopy, push enteroscopy,
ize the site of bleeding. It is the next therapeutic modality and surgery. CT accurately detected 15 (79%) of 19 bleed-
for lesions identified but not amendable by endoscopy ing sources. Filippone [35] recently published a review
( Fig. 13.4). Variations such as arterial CT angiography on a series of 32 patients with obscure GI bleeding using
have been used to improve sensitivity but have been re- 4-slice MSCT and nasoduodenal tube enteroclysis in com-
placed by and large by either capsule endoscopy or high- parison to capsule endoscopy. The combination of capsule
resolution MSCT or MRI. endoscopy and MSCT enteroclysis was able to detect a
Multiphase helical CT and MSCT have been described definitive cause of bleeding in 23 of the 32 patients, or an
as useful in the detection of bleeding sources in the small overall diagnostic yield of 72% as confirmed by surgery.
intestine with or without the use of small bowel ente- In 16 of 23 patients, findings in both imaging modalities
roclysis [31, 32]. The high resolution of up to isotropic were consistent. In five of 23 patients capsule endoscopy
0.6 mm and speedy data acquisition qualify MSCT as the depicted pathologies not seen on MSCT, whereas in two
modality of choice, especially in the setting of massive of 23 patients MSCT was able to detect neoplastic lesions
active bleeding prior to angiographic or surgical interven- not identified by capsule endoscopy. Pathologies not seen
tions ( Fig. 13.5). The minimal bleeding rate detectable in the MSCT enteroclysis were mostly superficial mucosal
by CT has been reported to be approximately 6 ml/min lesions. This review is remarkable, as it gives MSCT a
[33] though bleeding rates as low as 0.5 ml/min have higher priority in the diagnostic pathway of overt obscure
been detected in animal models [34]. Various small stud- bleeding than capsule endoscopy in the evaluation of ac-
ies, case reports, and reviews have described successful tive obscure GI bleeding.
13.7 Blood pool Contrast Agents
153 13

sequences similar to the late enhancement sequences

used for imaging of myocardial infarctions have been
proposed for the detection of subtle vascular changes or
subtle amounts of blood extravasation into the intestine.
Though promising larger validation studies using these
new imaging approaches with MRI have not yet been
published, initial reports are promising.
As with MSCT, the diagnostic capabilities of MRI
employing the latest imaging techniques have not yet
been adequately evaluated for their role in the diagnostic
workup of GI bleeding. Both modalities are underesti-
mated in their diagnostic yield in this specific area and
each institution will have to decide based on its local
setting whether MRI or MSCT will be its primary cross-
sectional modality for the diagnosis of GI bleeding.

13.7 Blood pool Contrast Agents

With the introduction of blood pool contrast agents, MRI

has the potential to combine the solely morphological
imaging of MSCT with the more functional imaging of
scintigraphy. The high relaxivity of blood pool contrast
Fig. 13.5. MSCT of the abdomen in the evaluation of an intermittent agents such as Vasovist (Gadofosveset, Bayer Schering
obscure LGIB. Maximum intensity projection in coronal plane depicts Pharma AG, Berlin, Germany))is the basis for a high-
active bleeding from angiodysplasia of the jejunum. Pathology was resolution dynamic imaging that provides detailed mor-
confirmed by endoscopy phological information about the vascular territory of
the abdomen. Pathological vascular structures and active
bleeding sites might be identified in these initial data
Based on only a small number of cases, this review sets, as has been shown with extracellular contrast agents
points to the potential of MSCT enteroclysis as an im- [38]. The extended intravascular retention time unique to
portant first-line modality for the evaluation of obscure blood pool contrast agents can then be used to perform
GI bleeding. repeated scans over as long as 24 h. Similar to scintig-
raphy, these repeated scans help to detect extravascular
contrast pooling as an indicator of extravasation of blood.
13.6 GI Bleeding and MRI This approach is most promising to localize the source
of intermittent active bleeding to a specific part of the
Magnetic resonance imaging (MRI) has not been a key intestine. Repetitive scans might also provide evidence
modality in the diagnosis of GI bleeding irrespective of that an abnormal structure is the source of bleeding. This
its source and location. There was a proof of concept for has been shown for intermittent bleeding from a gluteal
MRI in GI bleeding by Hilfiker in 1999 [37], and some aneurysm (Lotz, submitted).
case reports have been published on successful imaging of The need for repetitive scans is a drawback to this
duodenal angiodysplasia [38] and varices in the ascending approach, similar to scintigraphic scans. However, as the
colon [39]. New techniques of 3D GRE imaging (VIBE, intestinal blood deposits are propelled by natural peristal-
Thrive, LAVA) have been introduced since these reports sis, the exact timing of the scans is crucial to reliably de-
that allow for reliable and high-resolution T1 imaging fine the source of bleeding. After initial dynamic images
of the abdomen, usually within one breath hold. As has are acquired, we prefer to repeat coronal imaging every
been shown in MSCT imaging, MRI enteroclysis is suited 60 min for 6 h and additional scans 12 h and 1824 h after
for the detection of intestinal tumors and inflammatory initial contrast administration ( Table 13.3). Protocols
changes that might lead to GI bleeding, especially due to and image characteristics are the same for 1.5- and 3-T
its lack of radiation and superior extramucosal contrast as systems once sequence parameters have been adapted to
compared with MSCT [40, 41]. the longer T1 relaxation time in the 3-T environment.
Initial trials with high-resolution abdominal imaging Artifacts from air within the intestine tend to be more
using modified MRA sequences or inversion-recovery relevant in a 3-T environment [42]. A bowel preparation
154 Chapter 13 Gastrointestinal Bleeding

Table 13.3. Experimental protocol for diagnosis of intermit-

tent obscure GI bleeding with MRI and Vasovist. Whole ab-
domen is imaged with surface coils. Parallel imaging should
be used if available to minimize acquisition times. Sequences
are optimized for each system to be done in suspended res-

Time Sequences

12 h prior to scan Small and large bowel preparation,

same as for endoscopy: 4 l of fluid
within 12 h
Patients are to eat clear soup, no solid A
food for 12 h prior to and during the

0 Butylscopolamin 20 mg*
1. Transversal GRE T1
2. Transversal GRE T1; fat suppression
3. Transversal TSE T2; fat suppression
4. Dynamic 3D GRE T1 coronal, fat
a: w/o contrast
b: arterial phase
c: portalvenous phase (60 s p.i.)
d: venous phase (5 min p.i.)

20 min; 60 min; Butylscopolamin* 20 mg

every 1 h for 6 h; 1. Coronal 3D GRE T1; fat suppression
Late scan after 12 h 2. Transversal 3D GRE T1; fat sup-
Late scan after 1824 h pression

p.i. post injection of contrast medium

*Butylscopolamin: be aware of possible contraindications and maxi-
mal dosage within 24 h

is crucial to avoid overlap with T1 hyperintense intestinal

13 contents and to reduce the amount of air in the small in-
testine. We apply the same regimen for bowel preparation
as is done for small-bowel endoscopy. Enteroclysis might Fig. 13.6AC. MRI done in a patient with obscure massive inter-
be used in the initial scan, though its effect will not last for mittent LGIB caused by an aortoileal fistula. Contrast extravasation
the duration of the whole exam of 12 h and more. During to an ileal loop is seen 6 h after initial application of 10 ml Vasovist
the time of examination patients should avoid solid food (arrows). Origin of bleeding was defined to be at the upper end of a
as well as any food or liquid that might interfere with the stent placed at the aortic bifurcation into the right common iliac artery
T1 hyperintense signal of Vasovist. Intravenous butyls-
copolamin is mandatory to reduce artifacts from bowel
movement and to achieve higher image quality. Patients
are imaged in a supine position, although prone position-
ing might improve image quality if patients are able to co-
operate. Extravasation of blood is seen as a hyperintense 13.8 Summary
signal within the intestine not seen in the initial scans
( Fig. 13.6). The diagnostic approach to GI bleeding depends on its
There are no valid data on the sensitivity and spe- suspected location and the intensity of bleeding. The
cificity of this MRI approach to detect and localize GI ability to combine primary diagnosis with the option
bleeding, and no reliable information is available about of local treatment has made endoscopic techniques the
the minimal bleeding rate detectable with this technique. first-line diagnostic tool in the workup of GI bleeding.
These data will have to be derived from forthcoming These include classic push endoscopy as well as double-
studies. balloon endoscopy and colonoscopy. Capsule endoscopy
155 13

has emerged as a valuable modality for obscure bleeding In abdominal imaging, MRI has been inferior to
from the small intestine. MSCT in terms of speed of data acquisition and spa-
Cross-sectional imaging modalities are an important tial resolution. MSCT is the modality of choice for the
second option, if endoscopic techniques alone are not diagnostic evaluation of massive acute bleeding within
diagnostic or an extraintestinal pathology is suspected the small intestine in unstable patients. Recent 3D GRE
to be the primary cause of bleeding. Scintigraphy is well- sequences optimized for abdominal imaging have made
established for intermittent obscure bleeding, although MRI an alternative to MSCT in case of stable intermittent
additional cross-sectional imaging usually is employed bleeding not amendable by endoscopy. The combina-
to better characterize the causative lesion found on the tion of MRI with blood pool contrast agents promises
radionuclide scans. Angiography nowadays is usually per- to broaden the indications for MRI in GI bleeding.
formed with the aim of treatment. Diagnostic angiogra- Though it still has to be proven in adequate studies,
phy is the last resort if all other imaging modalities cannot MRI in combination with Vasovist has the potential to
identify the underlying pathology in serious GI bleeding. combine the high-resolution morphological imaging of
Intraoperative endoscopy has the highest diagnostic yield traditional MRI with the functional information usually
of all modalities but is restricted to patients with serious derived from scintigraphy about the origin and activity
blood loss from obscure bleeding sites. of bleeding.

Take home messages

Endoscopic techniques including capsule endos- and functional information usually derived from
copy are the primary imaging tools for GI bleeding. scintigraphy.
Cross-sectional imaging modalities are second-line The main benefit of MRI in GI bleeding, therefore,
options if endoscopy is inconclusive. might be in the workup of intermittent obscure
Up to now MRI has not been the preferred im- bleeding.
aging modality for the diagnostic workup of GI Further studies are needed to provide evidence of
bleeding. efficiency and reliability of MRI in the diagnosis of GI
MRI with blood pool contrast agents promises to bleeding.
close the gap between morphological imaging

8. Longstreth GF (1997) Epidemiology and outcome of patients

hospitalized with acute lower gastrointestinal hemorrhage: a
population-based study. Am J Gastroenterol 92:419424
1. Prakash C, Zuckerman GR (2003) Acute small bowel bleeding: a 9. Boonpongmanee S, Fleischer DE, Pezzullo JC, et al (2004) The
distinct entity with significantly different economic implications frequency of peptic ulcer as a cause of upper-GI bleeding is exag-
compared with GI bleeding from other locations. Gastrointest gerated. Gastrointest Endosc 59:788794
Endosc 58:330335 10. Strate L (2005) Lower GI bleeding: epidemiology and diagnosis.
2. Rubin TA, Murdoch M, Nelson DB (2003) Acute GI bleeding in Gastroenterol Clin North Am 34:643664
the setting of supratherapeutic international normalized ratio in 11. Segal WN, Cello JP (1997) Hemorrhage in the upper gastrointesti-
patients taking warfarin: endoscopic diagnosis, clinical manage- nal tract in the older patient. Am J Gastroenterol 92:4246
ment, and outcomes. Gastrointest Endosc 58:369373 12. Palmer K (2004) Management of haematemesis and melaena.
3. Peura DA, Lanza FL, Gostout CJ, Foutch PG (1997) The American Postgrad Med J 80:399404
College of Gastroenterology Bleeding Registry: preliminary find- 13. Landi B, Tkoub M, Gaudric M, et al (1998) Diagnostic yield of push-
ings. Am J Gastroenterol 92:924928 type enteroscopy in relation to indication. Gut 42:421425
4. Velayos FS, Williamson A, Sousa KH, et al (2004) Early predictors of 14. Lewis BS (1999) The history of enteroscopy. Gastrointest Endosc
severe lower gastrointestinal bleeding and adverse outcomes: a Clin N Am 9:111
prospective study. Clin Gastroenterol Hepatol 2:485490 15. Jensen DM, Kovacs TO, Jutabha R, et al (2002) Randomized trial
5. van Leerdam ME, Vreeburg EM, Rauws EA, et al (2003) Acute of medical or endoscopic therapy to prevent recurrent ulcer
upper GI bleeding: did anything change? Time trend analysis hemorrhage in patients with adherent clots. Gastroenterology
of incidence and outcome of acute upper GI bleeding between 123:407413
1993/1994 and 2000. Am J Gastroenterol 98:14941499 16. Laine L, Peterson WL (1994) Bleeding peptic ulcer. N Engl J Med
6. Vreeburg EM, Snel P, de Bruijne JW, Bartelsman JF, Rauws EA, 331:717727
Tytgat GN (1997) Acute upper gastrointestinal bleeding in the 17. Zuckerman GR, Prakash C (1998) Acute lower intestinal bleed-
Amsterdam area: incidence, diagnosis, and clinical outcome. Am J ing. I: Clinical presentation and diagnosis. Gastrointest Endosc
Gastroenterol 92:236243 48:606617
7. Longstreth GF (1995) Epidemiology of hospitalization for acute 18. Schuetz A, Jauch KW (2001) Lower gastrointestinal bleeding: ther-
upper gastrointestinal hemorrhage: a population-based study. apeutic strategies, surgical techniques and results. Langenbecks
Am J Gastroenterol 90:206210 Arch Surg 386:1725
156 Chapter 13 Gastrointestinal Bleeding

19. Lee EW, Laberge JM (2004) Differential diagnosis of gastrointesti- 40. Fidler J (2007) MR imaging of the small bowel. Radiol Clin North
nal bleeding. Tech Vasc Intervent Radiol 7:112122 Am 45:317331
20. Yamamoto H, Sugano K (2003) A new method of enteroscopy 41. Ryan ER, Heaslip IS (2008) Magnetic resonance enteroclysis com-
the double-balloon method. Can J Gastroenterol 17:273274 pared with conventional enteroclysis and computed tomogra-
21. Mehdizadeh S, Ross A, Gerson L, et al (2006) What is the learning phy enteroclysis: a critically appraised topic. Abdom Imaging
curve associated with double-balloon enteroscopy? Technical de- 33:3437
tails and early experience in 6 U.S. tertiary care centers. Gastroin- 42. Barth MM, Smith MP, Pedrosa I, Lenkinski RE, Rofsky NM (2007)
test Endosc 64:740750 Body MR imaging at 3.0 T: understanding the opportunities and
22. Pennazio M, Santucci R, Rondonotti E, et al (2004) Outcome challenges. Radiographics 27:14451462; discussion 14621444
of patients with obscure gastrointestinal bleeding after capsule 43. Esrailian E, Gralnek IM (2005) Nonvariceal upper gastrointestinal
endoscopy: report of 100 consecutive cases. Gastroenterology bleeding: epidemiology and diagnosis. Gastroenterol Clin North
126:643653 Am 34:589605
23. Apostolopoulos P, Liatsos C, Gralnek IM, et al (2007) Evaluation of 44. Schmulewitz N, Fisher DA, Rockey DC (2003) Early colonoscopy
capsule endoscopy in active, mild-to-moderate, overt, obscure GI for acute lower GI bleeding predicts shorter hospital stay: a ret-
bleeding. Gastrointest Endosc 66:11741181 rospective study of experience in a single center. Gastrointest
24. Pennazio M (2004) Small-bowel endoscopy. Endoscopy 36:3241 Endosc 58:841846
25. Rondonotti E, Villa F, Mulder CJ, Jacobs MA, de Franchis R (2007) 45. Strate LL, Syngal S (2003) Timing of colonoscopy: impact on
Small bowel capsule endoscopy in 2007: indications, risks and length of hospital stay in patients with acute lower intestinal
limitations. World J Gastroenterol 13:61406149 bleeding. Am J Gastroenterol 98:317322
26. Howarth DM, Tang K, Lees W (2002) The clinical utility of nuclear 46. Ohyama T, Sakurai Y, Ito M, Daito K, Sezai S, Sato Y (2000) Analysis
medicine imaging for the detection of occult gastrointestinal of urgent colonoscopy for lower gastrointestinal tract bleeding.
haemorrhage. Nucl Med Commun 23:591594 Digestion 61:189192
27. Ng DA, Opelka FG, Beck DE, et al (1997) Predictive value of tech- 47. Mujica VR, Barkin JS (1996) Occult gastrointestinal bleeding. Gen-
netium Tc 99m-labeled red blood cell scintigraphy for positive eral overview and approach. Gastrointest Endosc Clin North Am
angiogram in massive lower gastrointestinal hemorrhage. Dis 6:833845
Colon Rectum 40:471477
28. Hunter JM, Pezim ME (1990) Limited value of technetium 99m-la-
beled red cell scintigraphy in localization of lower gastrointestinal
bleeding. Am J Surg 159:504506
29. Chamberlain SA, Soybel DI (2000) Occult and obscure sources of
gastrointestinal bleeding. Curr Probl Surg 37:861916
30. Vernava AM, 3rd, Moore BA, Longo WE, Johnson FE (1997) Lower
gastrointestinal bleeding. Dis Colon Rectum 40:846858
31. Jain TP, Gulati MS, Makharia GK, Bandhu S, Garg PK (2007) CT
enteroclysis in the diagnosis of obscure gastrointestinal bleeding:
initial results. Clin Radiol 62:660667
32. Paulsen SR, Huprich JE, Hara AK (2007) CT enterography: nonin-
vasive evaluation of Crohns disease and obscure gastrointestinal
bleed. Radiol Clin North Am 45:303315
33. Ettorre GC, Francioso G, Garribba AP, Fracella MR, Greco A, Farchi
13 G (1997) Helical CT angiography in gastrointestinal bleeding of
obscure origin. AJR Am J Roentgenol 168:727731
34. Kuhle WG, Sheiman RG (2003) Detection of active colonic hemor-
rhage with use of helical CT: findings in a swine model. Radiology
35. Filippone A, Cianci R, Milano A, Valeriano S, Di Mizio V, Storto
ML (2007) Obscure gastrointestinal bleeding and small bowel
pathology: comparison between wireless capsule endoscopy and
multidetector-row CT enteroclysis. Abdom Imaging
36. Ernst O, Bulois P, Saint-Drenant S, Leroy C, Paris JC, Sergent G
(2003) Helical CT in acute lower gastrointestinal bleeding. Eur
Radiol 13:114117
37. Hilfiker PR, Weishaupt D, Kacl GM, et al (1999) Comparison of
three dimensional magnetic resonance imaging in conjunction
with a blood pool contrast agent and nuclear scintigraphy for the
detection of experimentally induced gastrointestinal bleeding.
Gut 45:581587
38. Erden A, Bozkaya H, Turkmen Soygur I, Bektas M, Erden I (2004)
Duodenal angiodysplasia: MR angiographic evaluation. Abdom
Imaging 29:1214
39. Chevallier P, Motamedi JP, Demuth N, Caroli-Bosc FX, Oddo F,
Padovani B (2000) Ascending colonic variceal bleeding: utility
of phase-contrast MR portography in diagnosis and follow-up
after treatment with TIPS and variceal embolization. Eur Radiol

Part IV New Horizons in Vascular


Chapter 14 Vasovist for Imaging Ischemic and Congenital Heart

Disease 159
Sebastian Kelle, Gerald Greil, Reza Razavi, and Eike Nagel

Chapter 15 Vasovist in Brain Tumor Imaging 169

Marco Essig and Clemens Fitzek

Chapter 16 Vasovist in Lymph Node Imaging: Present Status

and Future Development 181
Max J. Lahaye, R. Bert Jan de Bondt, Sanne M.E. Engelen,
Geerard L. Beets, and Regina G.H. Beets-Tan

Chapter 17 Vasovist for Breast Cancer Recognition 193

Joan C. Vilanova and Klaus Wasser

Chapter 18 Vasovist for Multiple Sclerosis Recognition 199

Michael Forsting

Chapter 19 Interventional MR Imaging 207

Michael Bock and Frank Wacker

Chapter 20 Vasovist in Lymphography 219

Christoph U. Herborn

Vasovist for Imaging Ischemic and

Congenital Heart Disease
Sebastian Kelle, Gerald Greil, Reza Razavi, and Eike Nagel

14.1 Introduction 160

14.2 Magnetic Resonance Coronary Angiography (MRCA) 160

14.3 Visualization of the Cardiac Venous System Using Vasovist 161

14.4 Detection of Myocardial Perfusion Defects with Vasovist 162

14.5 Late Gadolinium Enhancement 163

14.6 Use of Vasovist in Congenital Heart Disease 164

14.7 Conclusion 166

References 167
160 Chapter 14 Vasovist for Imaging Ischemic and Congenital Heart Disease

14.1 Introduction ings [39] only one of them has been approved by the
European Medicines Agency (EMEA): Vasovist for the
Cardiovascular MR (CMR) offers various options to de- peripheral vascular system [1013]. None of these con-
tect ischemia in patients with known or suspected coro- trast agents is approved for cardiac imaging [14].
nary artery disease (CAD). Over the past few years, the
evidence on many indications in the diagnosis of CAD
and heart failure with CMR has accumulated [1, 2]. CMR 14.2 Magnetic Resonance Coronary
provides combined diagnostic information, such as visu- Angiography (MRCA)
alization of the vascular tree, cardiac anatomy and func-
tion, as well as measurements of blood flow in the vessels, Despite the continuing interest of physicians, scientists,
ventricular wall motion, myocardial perfusion and vi- and patients in a non-invasive, non-radiation-based im-
ability. Most of the current applications for CMR imaging aging technique for the assessment of the coronary ar-
use contrast agents, either to improve image quality (e.g. teries, currently there is no approval for contrast agents
angiography) or to generate contrast (e.g. delayed gado- for MR coronary angiography (MRCA) [9]. For MRCA,
linium enhancement or perfusion). long scan times (several minutes) are needed to achieve
Blood pool contrast agents (BPCAs) such as Vasovist adequate spatial resolution and signal [15]. Thus, con-
(Gadofosveset, Bayer Schering Pharma AG, Berlin, Ger- ventional contrast agents are only marginally useful since
many) may have considerable advantages for cardiovas- their intravascular signal diminishes rapidly while back-
cular applications, since all their characteristics, includ- ground signal increases.
ing prolonged vascular retainment, longer half-life and In general, intravascular contrast agents have the ad-
stronger T1 relaxivity, can be used. The most promising vantage of providing adequate equilibrium-phase visu-
indications for CMR with Vasovist are angiography, per- alization of the coronary vasculature with high signal-to-
fusion imaging, and coronary artery imaging. BPCAs can noise ratio on 3D MR coronary angiography, due to their
be used to measure tissue blood volume and perfusion by greater T1 relaxivity and longer plasma half-life compared
using principles of indicator-dilution, and they may allow with extracellular agents, resulting in a greater blood sig-
for the evaluation of changes in capillary membrane in- nal but less myocardial enhancement. This allows the use
tegrity, which is advantageous for tissue characterization. of inversion-recovery sequences rather than T2-prepulse-
Even though a number of BPCAs have been developed based sequences, leading to a superior suppression of the
and tested for cardiovascular application in human be- myocardial signal ( Fig. 14.1).



Fig. 14.1A,B. Curved multiplanar reformatted images of a native after administration of Vasovist is demonstrated. Post-contrast, the
(A) and contrast-enhanced (B) image. A significant increase in visible myocardial signal is completely suppressed. (Reproduced by permis-
vessel length (arrow) and vessel sharpness of the right coronary artery sion from [9])
14.3 Visualization of the Cardiac Venous System Using Vasovist
161 14

In a preliminary study by Stuber et al., Vasovist was objective parameters such as visible vessel length and ves-
assessed in six patients, resulting in a substantial gain sel sharpness also improved significantly. The results of
in contrast-to-noise ratio (CNR) (improvement of 69% our study suggest that visualization of the distal parts of
compared with the non-enhanced images of the coronary the coronary arteries benefits from the administration of
arteries) [16]. These results of coronary artery imaging Vasovist more than imaging of the proximal segments. A
in a small group of patients demonstrated that Vasovist possible cause may be the reduction of saturation effects,
is well-suited for navigator-based techniques with high- which is more pronounced in the peripheral coronary
resolution imaging and an acquisition time of 14 5 min artery segments [9]. Consequently, the visible length of
for one three-dimensional scan (two scans were required the coronary arteries was increased in contrast-enhanced
to cover both the left and right coronary artery system) compared with non-enhanced images [8]. These results
using a targeted volume approach [9]. In a second report extend the use of Vasovist from targeted volumes to
by Nassenstein et al. [17] Vasovist was evaluated in 24 whole-heart imaging [9] ( Fig. 14.2).
volunteers, resulting in an overall improvement in image To date, it is not known how contrast-enhanced
quality and contrast-to-noise-ratio for MRCA after con- MRCA with Vasovist compares with state-of-the art
trast medium injection. Another study by Nassenstein et non-enhanced coronary MRA with navigator-corrected
al. investigated T1-values of blood and myocardium after SSFP-based sequences. In addition, the ideal time win-
administration of Vasovist for MRCA in volunteers. They dow for the MRCA data acquisition has not yet been
found that due to the long plasma half-life the T1-short- elucidated.
ening effect of Vasovist remains effective over 30 min,
which allows for multiple breath-hold or high-resolution
MRCA [18]. 14.3 Visualization of the Cardiac Venous
We compared 3D whole heart coronary artery imag- System Using Vasovist
ing pre and post administration of Vasovist and demon-
strated that the administration of Vasovist in a dosage A challenge of MRCA with BPCAs is the strong venous
of 0.05 mmol/kg body weight to healthy volunteers leads enhancement which may confound the delineation of
to an improvement of CNR and the blood-myocardial arteries. For peripheral vessels, this problem may be al-
contrast compared with non-enhanced images. Image leviated by performing venous subtraction [13].
quality, visible vessel length, and vessel sharpness in- The anatomic variability of the cardiac veins may
creased significantly on contrast-enhanced images. Other complicate LV lead implantation [19, 20]. In some pa-


Fig. 14.2A,B. Contrast-enhanced images demonstrating the left (A) vein, which covers the left circumflex artery; RCA right coronary artery;
and right coronary artery (B) using a 3D post-processing technique. LM D1 first diagonal branch; RVB right ventricular branch. (Reproduced by
left main stem; LAD left anterior descending artery; GCV great cardiac permission from [9])
162 Chapter 14 Vasovist for Imaging Ischemic and Congenital Heart Disease

cardiac vein for LV lead placement, thus helping to opti-

mize the implantation strategy [21].

14.4 Detection of Myocardial Perfusion

Defects with Vasovist

The wash-in kinetics and the local T1/T2 shortening in

the myocardial tissue depend on the concentration of
the contrast agent, the flow rate, diffusion of the contrast
agent into the interstitial tissue, relative tissue volume
fractions, contrast bolus duration, and recirculation ef-
fects [24]. For perfusion imaging, intravascular contrast
agents differ from conventional agents in several ways.
(a) The quantification of blood volume may be possible
from T1 values. (b) Due to the small distribution volume
(only 510% of the myocardial volume consists of blood)
[25] intravascular contrast agents result in a lower con-
trast effect, which may reduce their efficacy for perfusion
imaging. (c) The higher relaxivity of intravascular ver-
sus extracellular contrast agents may compensate for the
smaller distribution volume [26]. (d) In addition, there is
a potential for full quantification with minimal reliance
on complex models, since the diffusion component of the
Fig. 14.3. Vasovist enhanced 3D reconstruction of the heart. LA left contrast agent is smaller than that of conventional agents.
atrium; RA right atrium; LV left ventricle; RV right ventricle. The presence However, since a small fraction of Vasovist still diffuses
of the following veins at the crux cordis is demonstrated: CS coronary into the interstitium, quantification will be more complex
sinus; PIV posterior interventricular vein; GCV = great cardiac vein; PVLV =
than with truly intravascular macromolecules such as
posterior vein of the left ventricle. (Courtesy of Dr. Amedeo Chiribiri)
gadomer-17 or gadomelitol.
Several animal studies have shown that the detection
of perfusion defects during vasodilatation is feasible with
tients it can be difficult to access the cardiac veins with intravascular contrast agents [3, 2729]. The first study to
the guidewire. In these cases it helps to have detailed in- detect myocardial perfusion defects with Vasovist first
formation on the venous anatomy including side branches pass stress-induced (dipyridamol) perfusion was com-
available. pared with SPECT in an animal model of coronary artery
In a first study we demonstrated the capability of stenosis. The authors reported a prolonged hypointense
14 CMR to depict the complex anatomy of the venous sys- signal and lower peak signal intensity in ischemic myo-
tem of the heart in vivo through the combined use of cardium during stress. An MRI defect, classified as 75%
3D whole-heart imaging and the injection of blood pool reduction in peak myocardial signal intensity in the af-
contrast agents (16 volunteers and seven patients (Gado- fected territory, was detected in 92.3% of the animals. In
fosveset and Gadomer-17) [21]. We used the whole-heart the presence of mild coronary stenosis, there was uniform
approach to cover a three-dimensional volume. This im- enhancement with MRI and tracer uptake by SPECT.
aging technique was described formerly for the visualiza- The concordance of MRI and SPECT for detecting per-
tion of the whole coronary artery tree [22]. The strong fusion defects was 85%. In addition, three animals with
venous enhancement with BPCA allowed us to visualize myocardial infarction (i.e., triphenyltetrazolium chloride
the coronary veins from 210 min up to 2 h post injection [TTC] non-staining regions by postmortem analysis) had
( Fig. 14.3). hypoenhancing lesions on MR images. Only two of these
Thus, CMR may be proposed as the method of choice three animals with myocardial infarction also had per-
for the evaluation of patients who are candidates for car- fusion defects in SPECT images [30].
diac resynchronization therapy and LV lead positioning, In a different study by Jerosch-Herold et al. in pigs,
not only as a method to identify scar tissue, as proposed Vasovist was used to determine the relation between the
by van de Veire [23], but as a method capable of simulta- rate of myocardial signal enhancement during the first
neously evaluating the presence of scar, the LV function pass (upslope) and myocardial blood flow (MBF) and to
and viability in combination with identifying a suitable derive and validate a corrected perfusion reserve (PR)
14.5 Late Gadolinium Enhancement
163 14

Fig. 14.4. Perfusion MRI using MS-325: equatorial short-axis view. Fig. 14.5. Short-axis view post injection of MS-325 in a dosage of
There are clear perfusion defects in the anterior and septal segment 0.05 mmol/kg body weight in a patient with chronic myocardial infarc-
(arrows) tion demonstrated delayed enhancement in the anterior and septal
segment (arrows)

index from the upslope parameter. This was done with an nary stenosis, intravascular MR contrast agents provide
intravascular BPCA, because the characteristic time for longer delineation of the area at risk than extracellular
the vascular transit cannot be determined as easily for an agents [35]. In small-animal models, intravascular agents
extracellular tracer as for an intravascular tracer [31]. were shown to be more suitable than extracellular agents
First applications of Vasovist in patients with chronic for the prolonged delineation of no-reflow zones and for
myocardial infarctions have shown promising results defining small (few pixels) microvascular obstruction
for first pass perfusion imaging. The capability to de- areas. Thus, BPCAs may be useful for delineating micro-
tect chronic myocardial infarction with a combination embolization in human beings. Furthermore, these agents
of dynamic perfusion at rest and delayed enhancement have been used to demonstrate the progressive growth
was demonstrated ( Fig. 14.4). In a direct comparison, in the size of the no-reflow zone in mild, moderate, and
a larger number of infarcted segments showed perfusion severe myocardial injury [36] and the effect of the dura-
defects with Vasovist than with the extracellular contrast tion of reperfusion on the size of the no-reflow zone (up
agent [32]. to 24 h following reperfusion) [37].
In an animal study with pigs it was demonstrated that
Vasovist provided a robust visualization of acute myocar-
14.5 Late Gadolinium Enhancement dial infarction over a time-period of at least 2 h without
the need to change the inversion recovery time [38]. Our
There are only few reports about delayed enhancement group showed that Vasovist allows for a detection of
with BPCA. In an animal study in pigs using a combined chronic myocardial infarction in patients ( Fig. 14.5).
protocol for myocardial perfusion and viability with P792 However, in a direct comparison to extracellular con-
(Vistarem, Guerbet Group, Paris, France) in comparison trast agent, the transmurality of the scar and the number
to an extracellular contrast agent in a non-reperfused of scarred left ventricular segments were smaller with
infarction model, Peukert and co-workers reported that Vasovist [32]. This is most likely due to a slower diffu-
the BPCA allowed evaluation of myocardial viability [33]. sion of the contrast agent because of its albumin binding.
From a pig study, Saeed et al. reported a similar visualiza- However, the unbound fraction should behave similar to
tion of acute myocardial infarction with both intravascu- conventional contrast agents. Potentially, one could wait
lar and extracellular contrast. However, chronic infarc- much longer (on the order of hours) to allow sufficient
tions enhanced only with the extracellular but not the time for the contrast agent to diffuse. However, no data
intravascular agents [34]. In coronary occlusion and coro- on this question are currently available.
164 Chapter 14 Vasovist for Imaging Ischemic and Congenital Heart Disease

14.6 Use of Vasovist in Congenital Heart sential to delineate small intra- and extra cardiac struc-
Disease tures [41].
Patients younger than 10 years usually require general
Cardiac MRI gained significant importance for diagno- anesthesia or sedation during cardiac MRI examinations.
sis in infants, children and adults with acquired and con- Therefore, image data acquisition techniques using navi-
genital heart disease [1]. These imaging techniques need gator gating with prospective slice correction to compen-
to be applicable to patients of different age groups with a sate for respiratory motion in combination with ECG trig-
wide variety in size, weight, and physiologic parameters, gering are currently the preferred 3D imaging technique
such as different heart rates. For complete evaluation of in this group of patients [40, 41]. However, these imaging
the cardiovascular system in this group of patients the techniques are limited in their spatial resolution without
anatomy of the cardiovascular system and the coronary the use of contrast agents [41, 42].
arteries needs to be imaged ( Fig. 14.6). Additionally, Currently used first pass contrast-enhanced angiogra-
ventricular function and size as well as flow with phase- phy techniques are non-ECG-triggered techniques allow-
contrast imaging need to be assessed. ing data acquisition of the arterial and venous vascular
The complex anatomy of the main cardiac segments system within one breath hold [43, 44]. However, data
[39] and the great vessels require high-resolution 3D acquisition is limited to a period of one breath hold when
data acquisition techniques, particularly in infants and the contrast bolus passes the region of interest. Owing to
small children. Since a high percentage of these patients time constraints, image resolution is limited and intracar-
need follow-up examinations, highly standardized im- diac structures are blurred due to absent ECG triggering.
age data acquisition techniques are of great importance Imaging of complex vascular structures in patients
[40]. Optimal signal- (SNR) and contrast- (CNR)-to- with congenital heart disease requires special techniques.
noise ratios as well as high spatial resolution are es- For example, diagnosis of the complex pulmonary vas-



Fig. 14.6A,B. Reformatted left coronary artery system in a 22-year- tracking was used after i.v. infusion of Magnevist (Gd-DTPA; Bayer
old patient after the incompetent aortic valve was replaced with Schering Pharma AG, Berlin, Germany) (A). An additional inversion-
the pulmonary valve (Ross procedure). The pulmonary valve was recovery pulse was used for better suppression of the myocardium
replaced by a homograft valve. As the coronary arteries need to be re- after i.v. infusion of Vasovist (B). Better delineation of the left coronary
implanted in aortic position, stenosis may develop at the implantation artery system is demonstrated using Vasovist. LAD = left anterior de-
side. A free-breathing T2-prepared three-dimensional steady state free scending; LCx = left circumflex; Ao = aorta
precession (3D-SSFP) whole-heart approach with navigator gating and
14.6 Use of Vasovist in Congenital Heart Disease
165 14

cular blood supply in patients with pulmonary atresia is the catheterization laboratory. New time-resolved con-
very challenging [44]. Invasive cardiac catheterization trast-enhanced angiography techniques may provide this
is currently considered the gold standard for imaging information with first pass imaging using Vasovist [46].
the complex pulmonary vascular blood supply [44, 45]. Ventricular function has been assessed using two-di-
High-resolution images are required with high SNR and mensional (2D) cine magnetic resonance imaging (MRI)
CNR ratios [44, 45]. Optimal timing of the contrast for many years. K-space segmented cine gradient echo im-
bolus is necessary due to the delayed enhancement of aging was used previously and has recently been replaced
the pulmonary vascular tree in these patients. Contrast- by cine steady state free precession (SSFP) because of better
enhanced angiography is able to image the origin and SNR and CNR ratios [4749]. Currently, acquisition of 2D
proximal course of major aorto-pulmonary collateral cine SSFP images during multiple breath holds is consid-
arteries [44]. However, with these currently available ered the gold standard for assessing ventricular function,
techniques peripheral pulmonary artery vessel branching in particular the left ventricle [4852]. However, this may
and connection cannot be reliably imaged ( Fig. 14.7A). be difficult in seriously ill patients. Reduced temporal and
Multidetector computed tomography (MDCT) can pro- spatial resolution may reduce the reliability of the results
vide this information; however, the radiation dose and [49]. At 1.5 T, currently reduced data acquisition meth-
volume of nephrotoxic contrast agent are significant [45]. ods and new coil technology allow reduction of scanning
Using Vasovist with the previously described inversion- time using SSFP techniques for accurate assessment of
recovery technique, navigator gating, and prospective ventricular volumes [5356]. Because of B0 and B1 inho-
slice correction to compensate for respiratory motion in mogeneities 3T systems are not suitable using high-signal
combination with ECG triggering, results similar to those SSFP techniques [57]. The use of Vasovist would allow the
obtained with MDCT may be achieved without the use of application of high spatial and temporal resolution gradient
radiation ( Fig. 14.7B). echo sequences on 3-T systems due to increased blood pool
MDCT and the previously described inversion-recov- signal [57]. Increased SNR and CNR values may facilitate
ery 3D SSFP sequence do not provide time-resolved imag- highly automated segmentation algorithms for precise and
ing to describe the filling patterns of different lung areas. fast estimation of ventricular volumes [49, 58].
This is currently provided by selective injection of contrast Multidimensional flow is a very interesting technique,
agents in major aorto-pulmonary arteries (MAPCAs) in particularly in patients with congenital heart disease [59].


Fig. 14.7A,B. The reformatted left pulmonary artery system is shown of Magnevist (A). An additional inversion-recovery pulse was used for
in the same 22-year-old patient. A free-breathing T2-prepared three- better suppression of the lung tissue after i.v. infusion of Vasovist (B).
dimensional steady state free precession (3D-SSFP) whole-heart ap- Better delineation of the left pulmonary vascular bed is shown using
proach with navigator gating and tracking was used after i.v. infusion Vasovist. LPA left pulmonary artery
166 Chapter 14 Vasovist for Imaging Ischemic and Congenital Heart Disease


Fig. 14.8A,B. The magnitude images of in-plane phase-encoded are more homogeneous after administration of Vasovist compared
images of the previously shown 22-year-old patient following i.v. with Magnevist. Arch narrowed aortic arch after repair of coarctation
administration of Magnevist (A) and Vasovist (B). Signal and contrast of the aorta

First promising clinical applications in these patients 14.7 Conclusion

have been performed [60, 61]. Even though through-
plane phase-contrast imaging does not seem to benefit Vasovist provides advantages in enhanced depiction of
from contrast agents, our preliminary experience shows blood vessels as well as intracardiac structures, due to
increased in-plane signal in phase-contrast imaging greater relaxivity and shortening of the intravascular T1
( Fig. 14.8). than extracellular contrast agents. It offers the possibility
14 In summary, Vasovist would allow a standardized to minimize background signal, as the T1 changes remain
and comprehensive imaging protocol in patients with confined mainly to the intraluminal spaces. Vasovist
acquired and congenital heart disease. After a single in- seems to be effective in dynamic and steady state MR an-
jection of Vasovist a first pass high temporal resolution giography. Further future applications may be established
sequence would allow assessment of different enhance- with contrast-enhanced angiography in the 3-T environ-
ment kinetics of the arterial and venous vascular systems. ment using gradient echo sequences. Also in reduced data
An SSFP (1.5-T) or gradient echo (3-T) sequence with acquisition methods the increased intraluminal contrast
an inversion recovery prepulse and navigator gating for can be used for shortening of acquisition time for quanti-
prospective slice correction to compensate for respira- tative evaluation of cardiac volumes and function. Its use-
tory motion, in combination with ECG triggering, would fulness for assessing myocardial perfusion, microvascular
then be added. This would provide high spatial resolution permeability, and viability has been shown in preliminary
images with high SNR and CNR values of intra- and ext- small studies and animal experiments. However, its true
racardiac vascular structures. Qualitative and quantitative value remains to be assessed. Vasovist may further al-
functional information of the ventricles will be obtained low the design of highly standardized and reproducible
using 3D time-resolved high-resolution imaging of the MRI examination protocols in patients with acquired and
heart and vessels. Multidimensional phase-contrast imag- congenital heart disease. This is of particular importance,
ing will provide important qualitative and quantitative as these patients require regular follow-up examinations
flow information of the heart and vessels. throughout their life.
167 14

Take home messages

Vasovist allows enhanced depiction of blood ves- Its usefulness for assessing myocardial perfusion,
sels and intracardiac structures due to greater re- microvascular permeability, and viability has been
laxivity and shortening of the intravascular T1 than shown in preliminary small studies and animal
extracellular contrast agents. It offers the possibility experiments. Larger clinical trials are needed for
to minimize background signal, as the T1 changes further evaluation.
remain confined mainly to the intraluminal spaces. The previously shown advantages of Vasovist
Vasovist seems to be effective in dynamic and allow the design of highly standardized and re-
steady state MR angiography at 1.5 T. Further producible MRI examination protocols in patients
future applications may be established with con- with acquired and congenital heart disease. This
trast-enhanced angiography in the 3-T environ- is of particular importance, as these patients re-
ment using gradient echo sequences. quire regular follow-up examinations throughout
The increased intraluminal contrast can be used their life.
for reduced data acquisition methods to shorten
acquisition times at 1.5 T and 3 T.

References 10. Parmelee DJ, et al (1997) Preclinical evaluation of the pharmacoki-

netics, biodistribution, and elimination of MS-325, a blood pool
1. Pennell DJ, et al (2004) Clinical indications for cardiovascular agent for magnetic resonance imaging. Invest Radiol 32:741747
magnetic resonance (CMR): Consensus Panel report. Eur Heart J 11. Lauffer RB, et al (1998) MS-325: albumin-targeted contrast agent
25:19401965 for MR angiography. Radiology 207:529538
2. Hendel RC, et al (2006) ACCF/ACR/SCCT/SCMR/ASNC/NASCI/SCAI/ 12. Bluemke DA, et al (2001) Carotid MR angiography: phase II study
SIR 2006 appropriateness criteria for cardiac computed tomog- of safety and efficacy for MS-325. Radiology 219:114122
raphy and cardiac magnetic resonance imaging: a report of the 13. Grist TM, et al (1998) Steady state and dynamic MR angiog-
American College of Cardiology Foundation Quality Strategic raphy with MS-325: initial experience in humans. Radiology
Directions Committee Appropriateness Criteria Working Group, 207:539544
American College of Radiology, Society of Cardiovascular Com- 14. Mohs AM, Lu ZR (2007) Gadolinium (III)-based blood-pool contrast
puted Tomography, Society for Cardiovascular Magnetic Reso- agents for magnetic resonance imaging: status and clinical poten-
nance, American Society of Nuclear Cardiology, North American tial. Expert Opin Drug Deliv 4:149164
Society for Cardiac Imaging, Society for Cardiovascular Angiogra- 15. Jahnke C, et al (2004) Coronary MR angiography with steady state
phy and Interventions, and Society of Interventional Radiology. J free precession: individually adapted breath-hold technique ver-
Am Coll Cardiol 48:14751497 sus free-breathing technique. Radiology 232:669676
3. Kroft LJ, de Roos A (1999) Blood pool contrast agents for cardio- 16. Stuber M, et al (1999) Contrast agent-enhanced, free-breathing,
vascular MR imaging. J Magn Reson Imaging 10:395403 three-dimensional coronary magnetic resonance angiography. J
4. Herborn CU, et al (2003) Coronary arteries: contrast-enhanced MR Magn Reson Imaging 10:790799
imaging with SH L 643A experience in 12 volunteers. Radiology 17. Nassenstein K, et al (2006) MR coronary angiography with
229:217223 MS-325, a blood pool contrast agent: comparison of an inversion
5. Herborn CU, et al (2004) 2MR coronary angiography with SH L recovery steady state free precession with an inversion recovery
643 A: initial experience in patients with coronary artery disease. fast low angle shot sequence in volunteers [in German]. Rofo
Radiology 33:567573 178:508514
6. Paetsch I, et al (2004) Improved three-dimensional free-breathing 18. Nassenstein K, et al (2007) Magnetic resonance coronary angiog-
coronary magnetic resonance angiography using gadocoletic raphy with MS-325: in-vivo T1-measurements to improve image
acid (B-22956) for intravascular contrast enhancement. J Magn quality of navigator and breath-hold techniques. Eur Radiol (Epub
Reson Imaging 20:288293 ahead of print)
7. Paetsch I, et al (2006) Detection of coronary stenoses with con- 19. Abraham WT, Hayes DL (2003) Cardiac resynchronization therapy
trast enhanced, three-dimensional free breathing coronary MR for heart failure. Circulation 108:25962603
angiography using the gadolinium-based intravascular contrast 20. Puglisi A, et al (2004) Limited thoracotomy as a second choice
agent gadocoletic acid (B-22956). J Cardiovasc Magn Reson alternative to transvenous implant for cardiac resynchronisation
8:509516 therapy delivery. Eur Heart J 25:10631069
8. Klein C, et al (2003) Improvement of image quality of non-invasive 21. Chiribiri A, et al (2008) Visualization of the cardiac venous system
coronary artery imaging with magnetic resonance by the use of using cardiac magnetic resonance. Am J Cardiol 101:407412
the intravascular contrast agent Clariscan (NC100150 injection) 22. Jahnke C, et al (2005) Rapid and complete coronary arterial tree
in patients with coronary artery disease. J Magn Reson Imaging visualization with magnetic resonance imaging: feasibility and
17:656662 diagnostic performance. Eur Heart J 26:23132319
9. Kelle S, et al (2007) Whole-heart coronary magnetic resonance 23. Van de Veire NR, et al (2006) Non-invasive visualization of the
angiography with MS-325 (gadofosveset). Med Sci Monit cardiac venous system in coronary artery disease patients using
13:CR469474 64-slice computed tomography. J Am Coll Cardiol 48:18321838
168 Chapter 14 Vasovist for Imaging Ischemic and Congenital Heart Disease

24. Burstein D, Taratuta E, Manning WJ (1991) Factors in myocardial 44. Geva T, et al (2002) Gadolinium-enhanced 3-dimensional mag-
perfusion imaging with ultrafast MRI and Gd-DTPA administra- netic resonance angiography of pulmonary blood supply in pa-
tion. Magn Reson Med 20:299305 tients with complex pulmonary stenosis or atresia: comparison
25. Edelman RR (2004) Contrast-enhanced MR imaging of the heart: with x-ray angiography. Circulation 106:473478
overview of the literature. Radiology 232:653668 45. Greil GF, et al (2006) Imaging of aortopulmonary collateral arteries
26. Kelle S, Nagel E (2007) Cardiac magnetic resonance imaging with with high-resolution multidetector CT. Pediatr Radiol 36:502509
blood pool agents. European Cardiovascular Disease 1:7274 46. Du J, Bydder M (2007) High-resolution time-resolved contrast-
27. Zheng J, et al (2005) Single-session magnetic resonance coronary enhanced MR abdominal and pulmonary angiography using a
angiography and myocardial perfusion imaging using the new spiral-TRICKS sequence. Magn Reson Med 58:631635
blood pool compound B-22956 (gadocoletic acid): initial experi- 47. Bellenger NG, et al (2000) Reduction in sample size for studies of
ence in a porcine model of coronary artery disease. Invest Radiol remodeling in heart failure by the use of cardiovascular magnetic
40:604613 resonance. J Cardiovasc Magn Reson 2:271278
28. Gerber BL, et al (2002) Accuracy of contrast-enhanced magnetic 48. Bellenger N, Pennell D (2002) Assessment of cardiac function. In:
resonance imaging in predicting improvement of regional myo- Manning W, Pennell D (eds) Cardiovascular magnetic resonance
cardial function in patients after acute myocardial infarction. imaging. Churchill Livingstone, Philadelphia, p 103
Circulation. 2002 Aug 27;106(9):10839 49. Miller S, et al (2002) MR Imaging of the heart with cine true fast
29. Wilke N, et al (1995) Regional myocardial blood volume and flow: imaging with steady state precession: influence of spatial and
first pass MR imaging with polylysine-Gd-DTPA. J Magn Reson temporal resolutions on left ventricular functional parameters.
Imaging 5:227237 Radiology 223:263269
30. Kraitchman DL, et al (2002) MRI detection of myocardial perfusion 50. Moon JC, et al (2002) Breath-hold FLASH and FISP cardiovascular
defects due to coronary artery stenosis with MS-325. J Magn Re- MR imaging: left ventricular volume differences and reproducibil-
son Imaging 15:149158 ity. Radiology 223:789797
31. Jerosch-Herold M, et al (2003) Magnetic resonance imaging of 51. Kunz RP, et al (2005) Assessment of left ventricular function by
myocardial contrast enhancement with MS-325 and its relation to breath-hold cine MR imaging: comparison of different steady
myocardial blood flow and the perfusion reserve. J Magn Reson state free precession sequences. J Magn Reson Imaging 21:140
Imaging 18:544554 148
32. Thouet T, et al (2006) Detektion chronischer Infarktareale mit dem 52. Cottin Y, et al (1999) MR imaging of the heart in patients after
intravaskulaerem Kontrastmittel MS-325 in Patienten. Clin Res myocardial infarction: effect of increasing intersection gap on
Cardiol 95 [Suppl 5]:P142 measurements of left ventricular volume, ejection fraction, and
33. Peukert D, et al (2007) Improved evaluation of myocardial per- wall thickness. Radiology 213:513520
fusion and viability with the magnetic resonance blood pool 53. Kozerke S, Tsao J (2004) Reduced data acquisition methods in
contrast agent p792 in a nonreperfused porcine infarction model. cardiac imaging. Top Magn Reson Imaging 15:161168
Invest Radiol 42:248255 54. Tsao J, Boesiger P, Pruessmann KP (2003) k-t BLAST and k-t SENSE:
34. Saeed M, et al (2006) Discrimination of myocardial acute and dynamic MRI with high frame rate exploiting spatiotemporal cor-
chronic (scar) infarctions on delayed contrast enhanced magnetic relations. Magn Reson Med 50:10311042
resonance imaging with intravascular magnetic resonance con- 55. Jahnke C, et al (2007) Four-dimensional single breathhold mag-
trast media. J Am Coll Cardiol 48:19611968 netic resonance imaging using kt-BLAST enables reliable assess-
35. Roberts HC, et al (1999) MRI of acute myocardial ischemia: com- ment of left- and right-ventricular volumes and mass. J Magn
paring a new contrast agent, Gd-DTPA-24-cascade-polymer, with Reson Imaging 25:737742
Gd-DTPA. J Magn Reson Imaging 9:204208 56. Wintersperger BJ, et al (2006) Cardiac cine MR imaging with a 32-
36. Schwitter J, et al (1997) Influence of severity of myocardial injury channel cardiac coil and parallel imaging: impact of acceleration
on distribution of macromolecules: extravascular versus intravas- factors on image quality and volumetric accuracy. J Magn Reson
cular gadolinium-based magnetic resonance contrast agents. J Imaging 23:222227
Am Coll Cardiol 30:10861094 57. Tyler DJ, et al (2006) Cardiac cine MR-imaging at 3T: FLASH vs SSFP.
14 37. Bremerich J, et al (1998) Microvascular injury in reperfused in- J Cardiovasc Magn Reson 8:709715
farcted myocardium: noninvasive assessment with contrast- 58. Greil GF, et al (2007) Quantitative assessment of ventricular func-
enhanced echoplanar magnetic resonance imaging. J Am Coll tion using three-dimensional SSFP magnetic resonance angiogra-
Cardiol 32:787793 phy. J Magn Reson Imaging 26:288295
38. Krombach G, et al (2007) MS 325 zur Darstellung des ischmisch 59. Sorensen TS, et al (2005) Three-dimensional, isotropic MRI: a uni-
geschdigten Myokards. RFo [Suppl 1]: VO 205 fied approach to quantification and visualization in congenital
39. van Praagh R, van Praagh S (1992) Morphologic anatomy. In: Fyler heart disease. Int J Cardiovasc Imaging 21:283292
D (ed) Nadas pediatric cardiology. Hanley & Belfus, Philadelphia 60. Frydrychowicz A, et al (2007) Ascending-descending aortic bypass
pp 1726 surgery in aortic arch coarctation: four-dimensional magnetic
40. Sorensen TS, et al (2004) Operator-independent isotropic three- resonance flow analysis. J Thorac Cardiovasc Surg 133:260262
dimensional magnetic resonance imaging for morphology in con- 61. Beeri E, et al (1998) In vivo evaluation of Fontan pathway flow dy-
genital heart disease: a validation study. Circulation 110:163169 namics by multidimensional phase-velocity magnetic resonance
41. Fenchel M, et al (2006) Three-dimensional morphological mag- imaging. Circulation 98:28732882
netic resonance imaging in infants and children with congenital
heart disease. Pediatr Radiol 36:12651272
42. Martirosian P, et al (2007) Optimization of blood-myocardial con-
trast in 3D true FISP cardiac imaging at 1.5 T. Magn Reson Med
43. Greil GF, et al (2002) Gadolinium-enhanced three-dimensional
magnetic resonance angiography of pulmonary and systemic
venous anomalies. J Am Coll Cardiol 39:335341

Vasovist in Brain Tumor Imaging

Marco Essig and Clemens Fitzek

15.1 Introduction 170

15.2 First Experience with Vasovist in Human Brain Tumor Imaging 171

15.3 Practical Aspects 178

References 178
170 Chapter 15 Vasovist in Brain Tumor Imaging

15.1 Introduction be speculated that the complex is able to extravasate into

pathologic tissue if the gap between the vessel wall cells is
Vasovist (Gadofosveset, Bayer Schering Pharma AG, large enough.
Berlin, Germany) represents a third generation of contrast On the second effect, binding affinity to albumin in
agents. It is the first approved MR contrast medium with the extravascular space leads to a significant extension
a pronounced and reversible protein binding, resulting in of T1 and T2 relaxivities and hence to changes in MR
a long blood half-life. The agent is categorized as an intra- contrasting. The T1 relaxivity, a measure of the para-
vascular or blood pool agent and was developed for MR magnetic property, is about five times higher (at 1.5 T)
angiographic applications, allowing imaging in a submil- compared with the unbound state and with conventional,
limeter resolution range in the equilibrium phase. Gd-chelate-containing contrast agents with a direct effect
There are a growing number of reports about the role on the contrast properties. This effect is even higher at
of Vasovist in MR angiography [14]. Although several lower field strength [6]. This causes a rise in enhancement
potential alternative applications were investigated, the with a direct influence particularly on the vascular signal
extravascular behavior of the agent and its potential for in MRA but also on the contrast behavior in tissue.
imaging tissue pathologies have not yet been systemati- With respect to the brain the situation is even easier.
cally explored and reported. Because of the presence of the blood-brain barrier (BBB),
In this chapter we will present the first clinical experi- represented mainly by endothelial tight junctions, all cur-
ence with the extravascular contrast behavior of Vasovist rently available MR contrast media do not leak into the
in patients with brain tumors. All cases presented were brain tissue. The BBB blocks all molecules except those
seen within exploratory examinations on an off-label ba- that cross cell membranes by means of lipid solubility
sis with written informed consent of the patients. (such as oxygen, carbon dioxide, ethanol, and steroid
In order to explain the potential of Vasovist as an hormones) and those that are allowed in by specific
extravasal contrast agent, the specific pharmacodynam- transport systems (such as sugars and some amino acids).
ics and kinetics of the agent have to be evaluated. Quite Substances with a molecular weight higher than 180 Dal-
different from conventional extracellular MRI-CM, the tons (Da), which include all available imaging contrast
pharmacokinetics of Vasovist need to be understood. media, generally cannot cross the BBB [7, 8]. Therefore,
As described in detail in  Chap. 1, its pharmacokinetics the degree of enhancement in CNS neoplasm depends on
are characterized not only by a single, dominating elimi- several factors, including the leakage of the BBB, the intra-
nation half-life time t1/2, but rather by an initial (dis- vascular CM concentration and its magnetic properties,
tribution phase) half-life time t1/2, followed by a later the vascularity of the lesion, the blood flow in the vessels
(disposition phase) half-life time t1/2. The resulting supplying the tumor, and the degree of edema, as well
bi-exponential pharmacokinetics can be easily rational- as on technical and contrast media-specific parameters,
ized by a relatively high local concentration immediately mainly the relaxivity [8, 9].
post i.v. injection (first pass), during which the fraction In intra-axial primary tumors, mainly gliomas, the
of protein-bound CM molecules increases, depending on BBB can be compromised by neovascularization and di-
the relative local concentrations of human serum albu- rect tumor-related damage. Since non-neoplastic astro-
min and Vasovist molecules, respectively. During this cytes are required to induce BBB features of cerebral en-
initial phase, the larger fraction of unbound molecules dothelial cells, it is conceivable that malignant astrocytes
leads both to a more pronounced extravasation rate and have lost this ability due to dedifferentiate. Alternatively,
15 to a more pronounced glomerular filtration rate, com- glioma cells might actively degrade previously intact BBB
pared with the following equilibrium state. Hence, an tight junctions [10, 11]. In secondary tumors, mainly
averaged initial t1/2 is obtained in the order of 0.5 h, not metastases and extra-axial tumors, the lesion vessels are
significantly exceeding the t1/2 of conventional extracel- different from normal cerebral vasculature and have no,
lular MRI-CM [5]. or a strongly disturbed, BBB [12, 13].
However, after a few circulation cycles of the contrast In both primary and secondary neoplasms, a high al-
medium, the contrast agent molecules remaining in the bumin concentration has been described in the interstitial
vascular system can be regarded as homogeneously dis- space of the pathologic tissue [14, 15]. This theoretically
tributed in the entire blood volume and are found in a enables the extravasated Vasovist molecules to bind to
rather stable equilibrium state, characterized by approxi- albumin in the interstitial space, which substantially in-
mately 8590% of the molecules in the protein-bound creases the relaxivity and consequently the enhancement
state. Consequently, both rates of extravasation and renal characteristics of the agent.
elimination are substantially reduced, but the remaining In an early rat glioma model, Adzamali et al. [16] first
1015% of unbound molecules are still able to extravasate. described the use of two different intravascular contrast
But even if the agent is bound to serum albumin it might agents for imaging brain neoplasms. Although the bind-
15.2 First Experience with Vasovist in Human Brain Tumor Imaging
171 15


Fig. 15.1AC. Mechanism of parenchymal contrast enhancement

with Vasovist. Early after injection (A) the molecules are only partially
bound to albumin (approximately 5075%), which allows for an ex-
travasation of the unbound molecules (green dots) into the interstitial
space. The bound molecules are not able to pass the disrupted blood
brain barrier. In the second parenchymal phase (B) the Vasovist
molecules bind to albumin present in high concentration in tumors,
which keeps them from redistributing into the intravascular space.
Since there is a stable equilibrium between bound (85%) and un-
bound (15%) molecules, there is a trapping of contrast medium in the
interstitial space, causing an increase of enhancement over time. After
approximately 12 h the gradient changes toward the redistribution
from the interstitium into the vascular compartment (C); however, the
contrast enhancement can last for more than 24 h

ing capabilities of Vasovist are different in mice, the 15.2 First Experience with Vasovist
authors found a two- to threefold stronger enhancement in Human Brain Tumor Imaging
based on the applied contrast dose. They explained the
greater effectiveness of the albumin-bound agents by an To assess the early post-contrast imaging capabilities of
increased systemic persistence. While the standard agents Vasovist, ten consecutive patients with different types of
used peaked within the first 5 min after administration, intracerebral tumors were examined in a pilot study using
the intravascular agents showed a slower kinetics, requir- the standard dose of 0.03 mmol/kg body weight.
ing about 30 min to reach a maximum which then lasted All MR examinations were performed on standard
for about 3 h. Despite the species dependence of the bio- clinical 1.5-T systems (Magnetom Avanto and Magne-
physical properties of Vasovist, which indicates strongly tom Symphony, Siemens, Erlangen, Germany) using a
reduced protein-binding capabilities in rodents compared circular polarized head coil or an eight-channel head
with human beings [17], these findings in animal models coil. Eight patients were referred for follow-up MRI after
already suggest a very interesting and promising pharma- radiotherapy, two patients in the pretherapeutic workup.
cokinetic mode of action, which cannot be obtained with For all patients, data from previous MR examinations
conventional extracellular MRI-CM and may have a sig- with a standard extracellular contrast agent (1.5 years to
nificant additional diagnostic potential in cerebral tumor 2 weeks prior to the Vasovist examination) were available
perfusion MRI using protein-binding blood pool agents and served as a reference.
such as Vasovist [18]: The hypothesis includes a trap- The imaging protocol included a T1-SE, T2-FSE and
ping effect of contrast agent molecules after their pas- FLAIR prior contrast and contrast-enhanced T1-SE in
sage of the diseased BBB in the unbound state, followed axial and coronal orientation. In patients with skull-base
by protein-binding in the tumor compartment leading tumors, an additional fat-suppression T1-SE was per-
to both strongly reduced back-flow rates and increased formed. The time between contrast medium administra-
relaxivity ( Fig. 15.1). This mechanism is suggested to be tion and imaging was about the same for all patients at
responsible for the contrast enhancement as observed in 3 min after application.
our initial patient studies, which are described in the fol- Compared with the standard agents, the intensity of
lowing examples. the enhancement was rated equal for Vasovist despite
172 Chapter 15 Vasovist in Brain Tumor Imaging




Fig. 15.2AD. Vasovist-enhanced MRI study in a 55-year-old patient with meningioma. Intraindividual comparision with a conventional ECF
agent-enhanced study
15.2 First Experience with Vasovist in Human Brain Tumor Imaging
173 15

the substantially lower gadolinium concentration, which Depiction and better quantification of the BBB dis-
is less than 1/6 of the gadolinium dosage for standard ruption is essential for the grading of such lesions. This
extracellular agents used with 0.1 mmol/kg body wt further enables treatment decision and improved biopsy
dosage. The significantly higher relaxivity seems to com- planning in these heterogeneous lesions. Since the en-
pensate for the lower concentration. In the following we hancing tumor parts are the target of neurosurgical re-
present two examples of early post-contrast imaging with section, the information about the extension of a BBB
Vasovist. disruption substantially influences the neuronavigation-
In the first case, a 55-year-old female patient with assisted resection.
histologically confirmed meningioma of the left skull base Because of the very promising experience and based
was examined in an intraindividual comparative manner on the previous work in animals, a second study was
using a standard ECF agent for the first scan and for the started to assess the time dependence of the enhancement
second Vasovist. The tumor, presenting the typical MR properties of Vasovist. In a small series of only five pa-
imaging characteristics of a meningioma, was initially tients, the participants were asked to return for follow-up
diagnosed in 2000 and treated with stereotactic conformal examinations at 5, 12, and 24 h after a single injection
radiotherapy. The patient presented for a routine yearly (0.03 mmol/kg body weight) of Vasovist. The studies
follow-up without clinical symptoms. The initial exami- were performed with a 1.0 T MR system using a standard
nation showed an unchanged meningioma at the right circular head coil. In only one patient was a comparative
medial middle cerebral fossa ( Fig. 15.2A,B) presenting a study with standard extracellular agents obtained. Fol-
typical homogeneous enhancement with dural tail using a lowing Vasovist injection the enhancement gradually
standard contrast medium ( Fig. 15.2C). In the examina- increased over the first 12 h ( Fig. 15.4) and lasted more
tion with Vasovist ( Fig. 15.2D) the tumor is displayed in than 24 h. Follow-up studies on a larger series of patients
the same extension as in the previous examinination, with will systematically assess the potential of this new agent
visually identical tumor contrast. The dural tail and the and the exact time course of enhancement.
infiltration of the cavernous sinus are clearly visible. In the following we present two cases which show the
Meningiomas as primary extra-axial tumors present potential of Vasovist in late-enhanced imaging.
tumor vessels without a BBB, allowing the unbound and A 62-year-old patient suffering from renal cell carci-
part of the bound fraction of Vasovist to extravasate into noma with a fast progressive left-sided hemiparesis and
the extracellular space, which causes an enhancement epileptic seizures was referred for contrast enhanced MRI
comparable to that of standard MR contrast media 5 min to rule out cerebral metastasis. We found multiple intrac-
after contrast medium administration. erebral lesions of different sizes and with blurred borders.
The second example is of an intra-axial brain tu- The greatest metastasis left frontal was manifold lobulated
mor: a 67-year-old patient with malignant glioma and ( Fig. 15.4A).
clinical symptoms of seizures and minor neurologic defi- Following intravenous administration of Vasovist in
cits. The examination was performed with a standard a typical dosage there was a clear signal increase in
dose of Vasovist for diagnostic work and biopsy plan- T1-weighted images related to the BBB leakage. Even
ning. Unenhanced T1 ( Fig. 15.3A) and FLAIR imaging small lesions, e.g., a few millimeter-sized lesions left
( Fig. 15.3B) presents a large mass lesion in the right frontobasal, showed a comparable contrast uptake pattern
temporal lobe with strong enhancement ( Fig. 15.3C,D). ( Fig. 15.4B). Even 8 h after a single dosage (0.03 mmol/
The enhancement pattern and intensity was rated equal kg body wt) of Vasovist there was a persistent signal
to a previous exam only 1 week previously. Vasovist was increase within the lesions ( Fig. 15.4C,D). After 24 h this
in this case able to achieve a strong tumor enhancement signal increase was diminishing but remained evident
with clear delineation of the enhancing tumor parts, ( Fig. 15.4E,F).
which allows for biopsy and treatment planning both for In the second case, a 65-year-old patient suffered from
neurosurgery and for radiotherapy. Note the increasing a recurrent glioblastoma with solid and cystic compart-
enhancement from the axial scans to the coronal scans ments. The clinical course was slowly progressive after
acquired 3 min later. surgery. The initial images after intravenous administra-
Anaplastic and malignant gliomas typically present tion of Vasovist demonstrated a left temporal mass le-
with a disrupture of the integrity of the BBB, which sion with BBB leakage resulting in a strongly enhancing
causes an early enhancement as observed after Vasovist cystic tumor appearance ( Fig. 15.5A,B). Early imaging,
injection. The reason for the strong enhancement with acquired immediately after injection of the contrast me-
the blood pool agent might be twofold in this case: first, dium, revealed a multilocular tumor with a second, more
the extravasation of the unbound fraction and the signifi- solid and nodular emerging signal increase in projection
cantly higher relaxivity; second the extravastion of bound of the dorsal corpus callosum. Further images were ac-
molecules via the severely disrupted BBB. quired after 5 min ( Fig. 15.5C,D). Besides the lesion in
174 Chapter 15 Vasovist in Brain Tumor Imaging




Fig. 15.3AD.Vasovist-enhanced MRI study in a 67-year-old patient with glioblastoma

15.2 First Experience with Vasovist in Human Brain Tumor Imaging
175 15



Fig. 15.4AF. Time dependence of Vasovist enhancement in a pa-

tient with cerebral metastases of renal cell carcinoma
176 Chapter 15 Vasovist in Brain Tumor Imaging



Fig. 15.5AH. Time dependence of Vasovist enhancement in a 65-year-old patient with recurrent glioblastoma

the dorsal corpus callosum, a third small tumor nodule hour ( Fig. 15.5A,C,E). The enhancement was still intense
became visible which had not been detected on the initial even after 7 h ( Fig. 15.5G).
post-contrast scan (arrow in Fig. 15.5D). Additional scan- The lesions in projection of the corpus callosum were
ning was performed after 3 h ( Fig. 15.5E,F) and after best seen after 3 and 7 h. The size of the lesion increased
7 h ( Fig. 15.5G,H). No additional Vasovist was admin- substantially over time.
istered. As described above, the observed enhancement from
There was a continuous increase of enhancement Vasovist in the presented patient studies is based on the
within the larger right temporobasal lesion over the first extravasation of the complex into physiologic and patho-
15.2 First Experience with Vasovist in Human Brain Tumor Imaging
177 15



Fig. 15.5. Continued

logic tissues particularly in its unbound state, and hence proteins in the extracellular space. This increases the T1
the extravasation is correlated to the discussed specific relaxivity substantially with a direct effect on the contrast
pharmacokinetics. Because about 85% of the molecules properties. The strong increase in relaxivity seems to
are bound to albumin in the equilibrium state, the agent compensate for the substantially lower amount of gado-
acts preliminary as a blood pool agent. However, the non- linium concentration used at the approved dosage.
bound fraction, particularly during the initial distribution Even if the agent is bound to serum albumin it might
phase and approximately 15% of the molecules during be speculated that the complex is able to extravasate into
equilibrium phase, is able to extravasate and to bind to pathologic tissue if the gap between the endothelial cells
178 Chapter 15 Vasovist in Brain Tumor Imaging

of the vessel wall is large enough. This might be the case Initial clinical results suggest that Vasovist can de-
both in high-grade and malignant gliomas as well as me- tect a disturbed blood-brain-barrier in human patients
tastases. suffering from brain-tumors with a different signal in-
crease pattern over the time. With Vasovist, even 24 h
after injection pronounced contrast enhancement was
15.3 Practical Aspects still observed. Based on initial clinical experience, the
maximum signal increase and maximum distribution of
Vasovist can be used in exactly the same way conven- the contrast-enhancing tumor borders seem to range in
tioned as extracellular agents. The approved dosage is a time window between 3 and 12 h. Early imaging seems
0.03 mmol/kg body wt. With a 0.25-mol concentration to provide a contrast-to-noise ratio comparable to results
less than 1/3 of gadolinium is used, compared with a obtained with extracellular contrast agents.
0.1 mmol/kg body wt dosage. However, to obtain opti-
mal results with such a small amount of gadolinium, one
should take into account the advantages of the agent in References
respect of the percentage of unbound fraction, relaxivity,
and the length of stay in the vascular compartment. 1. Grist TM, Korosec FR, Peters DC, et al (1998) Steady state and dy-
namic MR angiography with MS-325: initial experience in humans.
To allow for a large number of unbound molecules at
Radiology 207:539544
the tissue site of interest the injection rate should be kept 2. Perreault P, Edelmann MA, Baum RA, et al (2003) MR angiography
high, e.g. 13 ml/s, followed by a saline flush of 30 ml at with gadofosveset trisodium for peripheral artery disease: phase
the same injection rate. Vasovist can be easily used in II trial. Radiology 229:811820
combination with a first pass or steady state MRA proto- 3. Goyen M, Edelmann M, Perreault P, et al (2005) MR angiography
of aortoiliac occlusive disease: a phase III study of the safety and
col. Due to the long-lasting residence time and the increas-
effectiveness of the blood-pool contrast agent MS-325. Radiology
ing relaxivity over the first 15 min [4], an equilibrium MR 236:825833
angiography is possible for at least 60 min post injection. 4. Nikolaou K, Kramer H, Grosse C, Clevert D, Dietrich O, Hartmann
The extended imaging window can be used to acquire im- M, Chamberlin P, Assmann S, Reiser MF, Schoenberg SO (2006)
ages with substantially higher spatial resolution. High-spatial-resolution multistation MR angiography with parallel
imaging and blood pool contrast agent: initial experience. Radiol-
With regard to extravascular imaging the protein
ogy 241:861872
binding causes a primarily unidirectional flow into the 5. Hartmann M, Wiethoff AJ, Hentrich HR, Rohrer M (2006) Initial
extracellular space, which results in an increase of the imaging recommendations for Vasovist angiography. Eur Radiol
enhancement and a longer imaging window. 16 [Suppl 2]:B15B23
Contrast enhancement in brain-tumor imaging is 6. Rohrer M, Bauer H, Mintorovitch J, Requardt M, Weinmann HJ (2005)
Comparison of magnetic properties of MRI contrast media solutions
time dependent. For common extracellular MR contrast
at different magnetic field strengths. Invest Radiol 40:715724
agents the BBB leakage was best seen 20 min after injec- 7. Bart J, Groen HJ, Hendrikse NH, et al (2000) The blood-brain barrier
tion compared with immediate post contrast imaging and and oncology: new insights into function and modulation. Cancer
with images which were acquired 10 min after injection Treat Rev 26:449462
[19]. Another group found the best scanning time point 8. Neuwelt EA (2004) Mechanisms of disease: the blood-brain bar-
rier. Neurosurgery 54:131140
between 8.5 and 38.5 min after injection [20]. Our own
9. Essig M, Weber MA, von Tengg-Kobligk H, Knopp MV, Yuh WT, Gie-
investigations found a decrease of the signal intensities al- sel FL (2006) Contrast-enhanced magnetic resonance imaging of
ready 20 min after injection [21] when using ECF agents, central nervous tumours: agents, mechanisms and applications.
15 even those with mild protein interaction. Top Magn Reson Imaging 17:89106

Take home messages

Vasovist as a strongly protein-binding contrast As expected from investigations in animal models, a
agent has the potential for high-quality contrast- relevant signal increase in T1-weighted images over
enhanced MR imaging of various cerebral tu- time can be clinically observed. Hours after injection of
mors. Vasovist the contrast-to-noise ratio and the enhanc-
Even at an early time point a robust enhance- ing area is significantly greater than in early imaging.
ment can be observed with a substantially Additional controlled studies are under way to
lower amount, compared with conventional assess the clinical potential of these agents, the
extracellular contrast agents. optimal imaging time, and the advantage over the
currently established MR contrast agents.
179 15

10. Kido G, Wright JL, Merchant RE (1991) Acute effects of human

recombinant tumor necrosis factor-alpha on the cerebral vascula-
ture of the rat in both normal brain and in an experimental glioma
model. J Neurooncol 10(2):95109
11. Schneider SW, Ludwig T, Tatenhorst L, et al (2004) Glioblastoma
cells release factors that disrupt blood-brain barrier features. Acta
Neuropathol (Berl) 107:272276
12. Fidler IJ, Yano S, Zhang RD et al (2002) The seed and soil hypoth-
esis: vascularisation and brain metastases. Lancet Oncol 3:5357
13. Groothuis DR (2000) The blood-brain and blood-tumor barriers:
a review of strategies for increasing drug delivery. Neurooncol
14. Seitz RJ, Wechsler W (1987) Immunohistochemical demonstration
of serum proteins in human cerebral gliomas. Acta Neuropathol
15. Nguyen TT, Pannu YS, Sung C, Dedrick RL, Walbridge S, Brechbiel
MW, Garmestani K, Beitzel M, Yordanov AT, Oldfield EH (2003)
Convective distribution of macromolecules in the primate brain
demonstrated using computerized tomography and magnetic
resonance imaging. J Neurosurg 98:584590
16. Adzamli K, Yablonski DA, Chicoine MR, et al (2003) Albumin-bind-
ing MR blood pool agents as MRI contrast agents in an intracranial
mouse glioma model. Magn Reson Med 49:586590
17. Eldredge HB, Spiller M, Chasse JM, Greeenwood MT, Caravan P
(2006) Species dependence on plasma protein binding and relax-
ivity of the gadolinium-based MRI contrast agent MS-325. Invest
Radiol 41: 229243
18. Zhou X, Caravan P, Clarkson RB, Westlund PO (2004) On the phi-
losophy of optimizing contrast agents. An analysis of 1H NMRD
profiles and ESR lineshapes of the Gd(III)complex MS-325+HAS. J
Magn Reson 167:147160
19. Yuh WT, Tali ET, Nguyen HD, Simonson TM, Mayr NA, Fisher DJ
(1995) The effect of contrast dose, imaging time, and lesion size in
the MR detection of intracerebral metastasis. AJNR 16:373380
20. Schrner W, Laniado M, Niendorf HP, Schubert C, Felix R (1986)
Time-dependent changes in image contrast in brain tumours after
gadolinium-DTPA. AJNR 7:10131020
21. Essig M, Tartaro A, Tartaglione T, Pirovano G, Kirchin MA, Spinazzi
A (2006) Enhancing lesions of the brain: intraindividual crossover
comparison of contrast enhancement after gadobenate dimeg-
lumine versus established gadolinium comparators. Acad Radiol

Vasovist in Lymph Node Imaging:

Present Status and Future Development
Max J. Lahaye, R. Bert Jan de Bondt, Sanne M.E. Engelen, Geerard L. Beets,
and Regina G.H. Beets-Tan

16.1 Introduction 182

16.2 Lymph Node Anatomy and Physiology 183

16.3 EUS and Ultrasound for Lymph Node Staging 183

16.3.1 Computed Tomography 184
16.3.2 FDG-PET 185

16.4 Magnetic Resonance Imaging 185

16.5 Lymph-node-specific MR Contrast Agents 186

16.5.1 Ultrasmall Superparamagnetic Iron Oxide 186
16.5.2 Gadolinium-based Lymph-node-specific MR Contrast Agents 188

16.6 Conclusion 188

References 190
182 Chapter 16 Vasovist in Lymph Node Imaging: Present Status and Future Development

16.1 Introduction ings. If the probability exceeds 20%, the neck should be
treated. Treatment consists either of a single modality of
Nodal tumor spread in cancer patients is associated with a therapy, such as a neck dissection or radiation therapy, or
poor prognosis [13]. Multimodality treatment in cancer a combination of both in which the decision will be driven
combining surgery with pre- and/or postoperative sys- by the treatment of the primary tumor [10]. Predictive
temic +/- radiation therapy aims at eradicating micro- models based on the histological and clinical behavior
scopic tumor deposits in the surgical bed and locoregional patterns of the tumor to assess which of the patients are at
nodes, as well as at sterilizing occult metastases in distant low and which are at high risk for local treatment failure
organs and distant lymph nodes. The effectiveness of mul- as with HNSCC, are presently being used mainly because
timodality treatment for the high-risk groups has been imaging tools to reliably identify pN0 patients are lacking.
shown in multiple cancer treatment trials [4]. Combined A substantial number of patients with HNSCC and an
preoperative modality treatment, however, carries a con- N0 neck as predicted by the presently available imaging
siderable perioperative morbidity and mortality [57] and methods are known to be understaged, because metas-
it is questioned whether, with multimodality treatment of tases in small cervical nodes remain beyond the detection
all cancer patients, the overall increased morbidity will levels of presently available imaging tools. As with rectal
outweigh the clinical effectiveness. The Dutch TME trial cancer patients, in HNSCC patients the development of
has shown in rectal cancer surgery that when rectal tumor accurate imaging methods to reliably distinguish between
nodes were involved, the local recurrence rate at 5 years the N0 and N+ necks could help clinicians to more confi-
was significantly lower for patients who received standard dently and accurately stratify their patients into a tailored
preoperative radiation therapy compared with patients individually based treatment.
who underwent immediate surgery. However, this was So far, no imaging tool has been reliable for detecting
at the expense of radiotherapy toxicity and a significant nodal metastases. A recent meta-analysis of all endolumi-
increase in perioperative morbidity in terms of perineal nal ultrasonography (EUS), computed tomography (CT),
wound leakage [8]. Furthermore, long-term complica- and magnetic resonance imaging (MRI) studies on nodal
tions such as bowel dysfunction were also described [9]. detection in rectal cancer patients has confirmed that
The trial also proved that subgroups of patients with ear- identifying nodal disease using the presently available
ly-stage rectal tumors without nodal involvement (stage noninvasive imaging techniques is not accurate enough
I), did not significantly benefit from radiotherapy in ad- for clinical decision-making [11]. Distinction between
dition to surgery, because these patients were already at malignant and benign nodes with noninvasive imaging
low risk for local failure. If it were possible to select these tools has been done only using morphological criteria
patients from the higher risk patients with bulky tumors such as size for CT and in addition to size, border contour,
and nodal involvement, then a more individually based and heterogeneity for EUS and MRI, consequently leaving
stratification of a multimodality treatment regime could small malignant nodes undetected. For example, in rectal
be introduced, restricting the more aggressive combined cancer treatment using the 9-mm cut-off to differentiate
preoperative treatment for the ugly cases and omit- benign from malignant nodes would lead to a consider-
ting preoperative irradiation in the good early tumors. able understaging of patients with nodal involvement be-
Likewise, in patients with a head and neck squamous cell cause more than half of the malignant rectal cancer nodes
carcinoma (HNSCC), the choice of workup and manage- are between 2 and 5 mm [12].
ment depends on the extent of lymph node metastases According to the literature for head and neck cancers,
from the primary tumor to the different levels in the neck, MRI is infrequently used for the assessment of lymph
as the presence of lymph node metastases (N+ neck) here node metastases, despite its superior contrast resolu-
is also an important prognostic factor. tion, and this again concerns the detection of metastases
16 As with rectal cancer, it is of clinical relevance for in very small lymph nodes [13]. The criterion used at
head and neck cancer whether any imaging tool can reli- present of an at least 10-mm cut-off short axial diameter
ably confirm the findings of a neck without lymph node for malignant nodes will result in misclassification of ma-
metastases at clinical examination (the clinically palpable lignant nodes as normal.
N0 neck). Weiss et al. performed a sensitivity analysis Consequently, if imaging is to develop into a more
to determine the optimal threshold for treatment of the powerful predictor for confident stratification of the
neck. Patients with a primary HNSCC with a clinically higher-risk cancer patients to a more aggressive treat-
N0 neck should be observed by a wait-and-see policy if ment and the lower-risk cancer patients to surgery only
the probability of occult cervical lymph node metastasis or even a wait-and-see policy, further research needs
is less than 20%. The calculated risk is based mainly on to be done to investigate advanced imaging techniques
the histopathological features, location, and extent of the to improve nodal staging. In this pursuit radiologists
primary tumor as assessed by clinical and imaging find- are aided by new developments in lymph-node-specific
16.3 EUS and Ultrasound for Lymph Node Staging
183 16

contrast agents. MR blood pool contrast agents, such as tive-tissue capsule ( Fig. 16.1). In these lymph nodules,
ultrasmall super paramagnetic iron oxide (USPIO), have dense masses of macrophages and lymphocytes are situated
shown better sensitivities and specificities for the detec- and separated by spaces called lymph sinuses. Through
tion of malignant nodes in various studies (concerning these lymph sinuses the lymph fluid passes, on its way
lymphoma, colon, rectum, lung, breast, head and neck, from the numerous afferent lymphatic vessels towards an
prostate, uterus, endometrial, ovarian, testicles, kidney, efferent lymphatic vessel. The efferent vessel leaves the
bladder, cervical, and vulval cancer). Another blood pool node at a concave area called the hilum. The lymph fluid
MR contrast agent approved for MR angiography is now consists of chyle, proteins, fat and white blood cells, pri-
on the market which is also promising for visualization marily lymphocytes. The lymph fluid flow to lymph nodes
of lymph nodes: Vasovist (Gadofosveset, Bayer Schering is anatomically defined by the lymphatic vessels and the
Pharma AG, Berlin, Germany) [14]. lymph node pattern. Metastasis can follow this structured
This chapter describes the present status of modern lymph flow to spread to the next lymph node. Cabanas et
imaging techniques for identifying nodal disease in high- al. demonstrated the existence of a specific lymph node
risk patients, focusing on colorectal and head and neck center, the so-called sentinel lymph node (SLN), which was
cancer, and the promising new techniques and novel con- the primary site of metastases in penile cancer [15]. When
trast agents that may help radiologists to improve the se- biopsy of the SLN is negative for metastatic disease, the
lection of oncology patients with lymph node metastases other lymph nodes downstream are likely to also be benign.
using non-invasive procedures. Thus, a specific node can be the doorway for cancer to
spread to other regional nodes. It is possible for a metastasis
in a sentinel node to grow and block the lymphatic flow,
16.2 Lymph Node Anatomy and Physiology thus redirecting lymph and tumor cells to other, possibly
not ordinarily sequential nodes. This hypothesis of a senti-
Knowledge of the anatomy and physiology of the lym- nel node has been firmly established in such malignancies
phatic system and lymph nodes is important in order as breast cancer, penile cancer, and melanoma, but not in
to have a basic understanding of lymph node imaging. several other malignancies such as thyroid, head and neck,
Lymph nodes are small, bean-shaped structures that are gastric, colorectal, cervical, and endometrial cancers.
usually less than 2.5 cm in length. Lymph nodes are
distributed largely in the drainage areas of body sites
most exposed to the external environment like limbs, 16.3 EUS and Ultrasound for Lymph Node
gastrointestinal tract, lungs, and the pharyngeal axis. Staging
The lymphatic system is an essential first line of defense
against pathogens. During the past decades radiologists have gained a great
The normal lymph node is separated into compart- deal of experience in lymph node imaging by means of
ments called lymph nodules and surrounded by a connec- EUS and/or ultrasound [16, 17]. Several morphological

Fig. 16.1. The anatomy of a normal lymph

184 Chapter 16 Vasovist in Lymph Node Imaging: Present Status and Future Development

characteristics are assessed to distinguish malignant from has been introduced but not widely adopted. First, beside
benign lymph nodes. Characteristics suggesting tumoral the fact that the technique is highly operator dependent
lymph node involvement are: enlargement, round shape, and invasive, it is limited in that many nodes outside the
irregular border, loss of the central hilum, and hypoecho- reach of the endosonography probe remain undetected.
geneity [18]. Second, if nodes adjacent to the tumor are suspected
However, none of these morphological characteris- based on their EUS appearance, cytological aspiration of
tics is accurate enough to identify a malignant node and these nodes, which can be done only by direct penetration
therefore (E)US alone, without fine-needle aspiration cy- of the needle through the tumoral wall, contaminates the
tology (FNAC), cannot be relied on for clinical decision- needle with tumor cells, leading to false-positive results
making in the staging of either head and neck cancer [31]. Finally, there is the high probability of sampling
or gastrointestinal tract cancer such as of the esophagus errors, especially when small nodes of less than 5 mm
and rectum. In rectal cancer EUS was slightly, although have to be aspirated. In rectal cancer these nodes have a
not significantly, better than CT or MRI in a large meta- 2050% chance of harboring tumor cells.
analysis evaluating the diagnostic performance for nodal New ultrasound techniques such as Doppler-ultra-
staging. However, large variations can be found in the ac- sound combined with micro-bubble contrast agents or 3D
curacy of EUS for nodal detection (6285%), illustrating EUS are being studied with promising results, although
its operator dependency [1922]. Controversial reports reports are limited mostly to experience at single centers,
on the performance of EUS-FNAC exist. Harewood et al. and the real question is whether advanced noninvasive
found no benefit from EUS-FNAC staging as compared EUS techniques in rectal cancer will gain widespread
with EUS staging alone [17], while Shami et al. showed acceptance if they turn out to be operator dependent in
significant improvement for EUS-FNAC compared with future multicenter studies [3234].
EUS alone, with a sensitivity of 93% [23].
Regarding lymph node staging in head and neck can-
cer, with the introduction of ultrasound-guided fine-nee- 16.3.1 Computed Tomography
dle aspiration cytology (USg-FNAC), sensitivity and spe-
cificity ranging from 6397% and 69100%, respectively Computed tomography has long been adopted as a tool
have been obtained [2426]. The associated diagnostic for identifying malignant retroperitoneal lymph nodes
odds ratio (DOR) for US alone was 40 whereas the DOR in several cancers. On CT imaging, mesorectal nodes
for USg-FNAC was significantly higher, up to 260, which measuring >5 mm can be identified as oval structures
means an enormous improvement in the performance of of soft tissue density within the fatty mesorectum [35].
preoperative cervical lymph node staging [13]. A draw- Conventional single and 4- to 16-slice CT machines
back of ultrasound in general is that the performance de- demonstrated accuracies varying between 22% and 77%
pends on the experience of the sonographer, and in rectal for nodal staging in rectal cancer [22, 3638]. In con-
cancer multicenter trials, EUS staging accuracy has been trast to EUS, CT has a much larger field of view and is
shown to drop by 30% when non-expert centers are doing less reader/operator dependent. However, due to its low
the examination [27]. The operator dependency has also contrast resolution, conventional computed tomography
been shown to impact cervical nodal US staging [28]. A techniques have relied primarily on size criteria. Size on
further drawback of USg-FNAC in head and neck cancer its own is insufficient for a reliable distinction between
staging is the poor accessibility of cervical nodes deeply malignant and benign lymph nodes, and the high fre-
located in anatomical regions such as the retropharynx, quency of malignant nodes between 2 and 5 mm in rectal
and for the evaluation of these nodes other cross-sectional cancer patients would only further limit CT in its ability
imaging such as CT or MRI is recommended nowadays. to identify rectal nodal disease.
16 Therefore, the recommendations in the guidelines for For the same reason, prediction of cervical lymph
diagnostic workup of head and neck cancer is to stage node metastases in head and neck cancer by means of
primarily with planar imaging techniques such as CT conventional single and 4- to 16-slice CT techniques
and limit US staging of head and neck cancer patients to remains low in specificity. The main criterion for the
experienced centers only. assessment of cervical lymph node involvement using
The drawbacks regarding operator dependency of the CT (and MRI) is the short axial diameter, and several
Usg-FNAC performance, combined with generally re- studies have been undertaken to determine the optimal
stricted availability, its inability to accurately access some cut-off size of the short axial diameter for discrimination
anatomical cervical regions at risk, and its inherently in- between metastatic and non-metastatic lymph nodes. In
vasive nature makes USg-FNAC a staging tool that is un- general, 10 mm is a commonly used cut-off size but a
likely to gain general acceptance in head and neck cancer range varying from 9 to 15 mm has been described [39,
diagnostic workup. Likewise, in rectal cancer EUS-FNAC 40]. Other criteria, such as the presence of necrosis or
16.4 Magnetic Resonance Imaging
185 16

extranodal tumor spread, are used but less valuable [26, However, only few reports are available on the use of
41]. Necrosis is seen mainly in lymph nodes exceeding FDG-PET for locoregional staging in primary head and
10 mm, and those lymph nodes will already be classified neck cancer. The low spatial resolution of presently avail-
as malignant in the first instance based on their large size. able PET machines may be the reason why interpretation
In the neck, with its densely packed structures of vessels of PET images in the complex anatomy of the head and
and muscles, it is very difficult to achieve an acceptable neck remains very difficult for staging of both the clini-
soft tissue contrast resolution with conventional CT tech- cally N+ neck as well as the N0 neck [40]. Despite these
niques, and it remains to be seen whether 64-slice CT shortcomings, Myers et al. reported a trend in increased
with improved contrast and spatial resolution will lead to accuracy for FDG-PET compared with CT for the detec-
better results in this regard. tion of cervical metastases in the N0 neck in HNSCC
Advances in MDCT techniques with 64-slice CT be- (PPV 100%, NPV 88%, accuracy 92% for PET versus
ing introduced in patient clinics could theoretically mean PPV 80%, NPV 75% and accuracy 76% for CT) [45]. This
an upgrading of CT for nodal disease detection, but so far study was performed in only a very small series of 14
no studies exist on 64-slice CT for evaluating rectal cancer patients with HNSCC in the oral cavity, oropharynx, and
nodes except for one, which indeed showed an improve- hypopharynx. Therefore, further and larger studies are
ment with an 85% accuracy [42]. needed to validate these results, also for other tumor sites
in the head and neck.
The introduction of hybrid techniques for whole-body
16.3.2 FDG-PET staging such as PET-CT, combining anatomical and func-
tional information, might lead to better results than with
18Fluorodeoxyglucose (FDG) positron emission tomog- PET or CT only as a single staging tool, but to date no
raphy (PET) has proven to be of additional value in the substantial evidence exists to support this assertion [47].
search for distant metastases in a wide variety of tumor In fact, to date a single study of 53 rectal cancer patients
types. Cancer cells have an increased glycolytic rate com- has been reported, showing a disappointing accuracy of
pared with normal cells, which 18FDG-PET is able to 72% for nodal staging using PET-CT [48].
detect. Although FDG-PET has proven its value for the
detection of distant extrahepatic disease in rectal cancer
patients, the results were unsatisfactory for locoregional 16.4 Magnetic Resonance Imaging
staging of the primary rectal tumor. An important limi-
tation of the currently available human PET scanners is In general, conventional MR techniques with 0.5- to
their low spatial resolution and consequent inaccuracy for 1.5-Tesla systems are not accurate for the detection of
the detection of low-bulk tumor, requiring at least 1 cm3 nodal metastases in cancer patients, despite the superior
tumor volume before it can be depicted by PET. Further- contrast resolution of MR imaging as compared with
more, the uptake of 18FDG within the primary tumor and all other existing conventional imaging techniques. Two
urinary bladder obscures the visualization of approximate meta-analyses in the literature report a similar ROC
mesorectal nodes in rectal cancer, because of the scatter curve for the detection of nodal metastases with conven-
artifacts. This may explain why in primary rectal cancer tional MRI, both showing insufficient performance for
staging with FDG-PET more than half of the malignant clinical decision-making [11, 22, 4955]. Most of these
nodes go undetected, resulting in an unacceptably low studies used size criteria to distinguish malignant from
sensitivity of 2129% [43, 44]. benign nodes. Only studies that investigated new criteria
For patients presenting with a head and neck tumor, it in addition to size, such as a border contour and signal
is of utmost importance to accurately predict lymph node homogeneity, demonstrated higher sensitivities and spe-
metastases in the lower cervical levels (level IV and VI), cificities for the prediction of lymph node involvement in
because lymph node involvement at these levels implies patients with rectal cancer, yet only large metastases, over
a significantly higher prevalence of distant metastases, 5 mm in size, could be accurately evaluated with regard
mainly to the lungs. In these patients a chest CT has so far to border and MR signal intensity [54, 55] ( Fig. 16.2).
been recommended in the workup as a first-line staging The introduction of higher field MR systems into clinical
tool [29]. practice could improve the disappointing performance
A recent study of patients presenting with cervical of conventional MRI. So far, two studies on rectal cancer
lymph node metastases, in which FDG-PET was em- staging with 3-Tesla MRI have reported promising accu-
ployed to search for the unknown primary tumor and dis- racies of 9195% for predicting nodal metastases in rectal
tant metastases, showed it to be superior to conventional cancer [56, 57], but larger studies are necessary before the
CT staging, and FDG-PET is now recommended for this role of high-field MR machines for staging rectal cancer
group of patients [30, 46]. can be established.
186 Chapter 16 Vasovist in Lymph Node Imaging: Present Status and Future Development


Fig. 16.2. A SPIR TSE T2-weighted image in the axial plane. On the 8 mm shows an indistinct border and heterogeneous signal intensity.
right side of the neck in level II an enlarged lymph node (arrowhead) Cytology confirmed the diagnosis of a metastasis. B At histological ex-
with a short axial diameter of 12 mm would be classified as malignant amination (H&E) of this latter lymph node, the region of heterogeneous
based on the size criterion; however, cytology revealed no metastasis. signal intensity corresponds to a metastasis (asterisk) with extranodal
Another, smaller lymph node (arrow) with a short axial diameter of only spread (arrow). Note disruption of the lymph node capsule (arrowhead)

In the staging of head and neck cancer MRI remains 16.5.1 Ultrasmall Superparamagnetic Iron Oxide
of limited use for malignant lymph node detection, as
shown in a recent meta-analysis in which MRI was the USPIO is a contrast agent that undergoes phagocytosis
least-performing noninvasive diagnostic modality [13]. by the reticuloendothelial system in the liver but also by
This is probably primarily due to the use of the short axis macrophages located in benign lymph nodes. Uptake of
diameter as a criterion whether a lymph node is benign USPIO within the node results in a decrease of signal
or malignant. In addition to the size, other morpho- intensity on T2*-weighted images due to increased suscep-
logical aspects, such as necrosis or enhancement pattern, tibility artifacts. This means that benign regions in a node
are used as criteria in daily practice, but the diagnos- appear hypointense on T2*-weighted images ( Fig. 16.3).
tic performances still have to be established in future In malignant nodes macrophages are displaced by tumor
studies. deposits. In these regions no uptake of USPIO will occur
16 ( Fig. 16.4) and they will be depicted as regions within the
node with high signal intensity (white regions). Weissleder
16.5 Lymph-node-specific MR Contrast et al. were the first to distinguish malignant from benign
Agents nodes in an animal model using USPIO [59]; they were
followed by Anzai et al., who demonstrated in a study
During the past decade, new MR contrast agents such as with healthy volunteers that the uptake of USPIO reached
ultrasmall super paramagnetic iron oxide (USPIO), Gad- its peak after 24 h [60]. Harisinghani et al. published the
ofluorine M, and Vasovist have been developed and have results of a study in 80 patients with prostate cancer and
proven in preclinical as well as clinical studies to be prom- showed that USPIO MRI distinguished malignant from
ising for oncology patients [14, 58, 59]. Of these, USPIO is benign nodes with 100% sensitivity and 95.7% specificity
at present the agent that has shown the largest evidence of [61]. A recent meta-analysis of 19 USPIO MR studies
being effective in the nodal staging of cancer patients. showed that USPIO MRI has a sensitivity and specificity of
16.5 Lymph-node-specific MR Contrast Agents
187 16


Fig. 16.3AC. A T2-weighted TSE image in the axial plane of the weighted image the same node in more detail showed general low
pelvis of a rectal cancer patient. On the conventional T2-weighted im- signal intensity, indicating a normal USPIO uptake pattern within the
age it is difficult to predict whether this 5-mm lymph node (arrow) is node. (C) At histological examination (H&E) this node showed a normal
malignant or benign. (B) On the corresponding USPIO-enhanced T2*- nodal architecture and no metastasis


Fig. 16.4AC. A T2-weighted TSE image in the axial plane of the indicating no USPIO uptake. At the border of the lymph node there
pelvis of a rectal cancer patient with a 7-mm mesorectal lymph node is still uptake of USPIO. C This correlated at histological examination
(arrow). B On the corresponding USPIO-enhanced T2*-weighted im- (H&E) with the centrally located tumor deposit (asterix) and the re-
age the same lymph node (arrow) showed central high signal intensity, maining surrounding normal lymphoid tissue (arrowheads)
188 Chapter 16 Vasovist in Lymph Node Imaging: Present Status and Future Development

88% and 96%, respectively